2024-03-29T08:48:48Zhttps:/www.ncbi.nlm.nih.gov/pmc/oai/oai.cgioai:pubmedcentral.nih.gov:321802001-06-04bmcphyspmc-openBMC PhysiolBMC Physiology1472-6793BioMed CentralPMC32180PMC3218032180112318871472-6793-1-111231887Research ArticleDifferential expression of Aquaporin 8 in human colonic epithelial cells and colorectal tumorsFischerHeléne1helene.fischer@cmm.ki.seStenlingRoger2Roger.Stenling.us@vll.seRubioCarlos3Carlos.Rubio@onkpat.ki.seLindblomAnnika1Annika.Lindblom@cmm.ki.seDept. Molecular Medicine, Karolinska Institute, Stockholm, Sweden.Dept. of Pathology, Umeå University Hospital, Umeå, Sweden.Dept. of Pathology, Karolinska Hospital, Stockholm, Sweden.2001231200111141220002312001Background
The gene expression pattern in tumor cells differs from that in corresponding normal cells. In order to identify differentially expressed genes in colorectal tumors and normal colorectal epithelium, a differential display experiment was used to compare RNA expression in normal and tumor tissue samples.
Results
One gene fragment was expressed only in normal tissue and not, or to a much lesser extent, in the adenomas, carcinomas and cancer cell lines. The isolated gene fragment was identical to Aquaporin 8 (AQP8), a water channel protein. In situ hybridization demonstrated that AQP8 was expressed in the cells facing the lumen in the normal colonic epithelium.
Conclusion
Our result suggests that the expression of AQP8 is a marker of normal proliferating colonic epithelial cells and suggest these cells to be involved in fluid transport in the colon.
Background
Many genes are known to be involved in colorectal carcinogenesis. Colorectal cancer is sometimes inherited in familial syndromes like Familial Adenomatous Polyposis (FAP) or Hereditary Nonpolyposis Colorectal Cancer (HNPCC) [1]. The identification of the predisposing genes in inherited tumors have meant great contribution to our understanding of colorectal carcinogenesis in these families but also of carcinogenesis in general [1]. The majority of colorectal tumors, however, are sporadic, and studies in sporadic tumors have revealed other genes of importance in colorectal tumor development, e.g. K-Ras [2], TGFβ [3], E-cadherin [4], SRC [5] and the PPARγ [6]. Genes may be altered in tumors compared to normal tissue without necessarily being causative. Tumor cells often de-differentiate or have their origin in immature cell types and loose expression of proteins associated with highly differentiated cells, and could in stead gain expression of embryonic proteins. These alterations can potentially be used as clinical markers to define a particular disease entity, useful for diagnosis, staging, population screening or even to detect the presence of occult metastatic disease, recurrent disease and response to treatment. The perhaps most known embryonic protein used as a tumor marker in malignancy is the CEA in colorectal cancer [7]. In an attempt to identify additional genes altered during the development of colorectal cancer, we compared the gene expression pattern between normal colon epithelium samples and colorectal tumors using the mRNA differential display PCR (mRNA DD-PCR) method. In this paper we report that the water channel protein AQP8 is expressed in the normal columnar epithelial cells of the colonic mucosa facing lumen and that adenomas and colorectal cancer do not seem to express this protein.
Results
One gene fragment of approximately 80 bp was expressed in the 4 normal tissue samples but not, or to a much lesser degree, in 24 colorectal tumors in the mRNA DD-PCR experiment. The band was cut and reamplified using the same primers, and cloned. Clones were then selected by size, and by using inverse northern hybridization it was found that all clones represented the same DNA sequence. Sequencing and subsequent search in the genome database showed that this differentially expressed gene fragment corresponded to the sequence of aquaporin 8 (AQP8).
The result was confirmed with RT-PCR on newly synthesized cDNA and northern blot hybridization (Fig. 1) on the original samples from the mRNA DD-PCR experiment. The number of samples used for northern blot was small depending on limited amounts of RNA from the samples. We extended the experiment and included besides the 28 samples used in the mRNA DD-PCR, another 7 sporadic colorectal cancers, 3 samples from normal colonic mucosa, 5 sporadic colonic adenomas and 3 HNPCC tumors from patients with known germline mutations in hMLH1. Using a semiquantitative RT-PCR it was found that samples from adenomas, rectal cancers, and colonic cancers, including those from HNPCC patients, expressed the AQP8 to various degrees, but generally to a much lesser extent than normal colonic tissue (Fig. 2). The expression of AQP8 in some tumors could be explained by a normal cell contamination in those samples.
Northern blot demonstrating AQP8 in normal colonic tissue, and not in adenomas, or carcinomas. Symbols first in sample identification relate to type; N, normal, H, HNPCC carcinoma, C, sporadic colonic carcinoma, R, sporadic rectal carcinoma, P, adenomatous polyp.
RT-PCR comparing expression of GADPH2 and AQP8 in normal colonic tissue (N), adenomas (P), colon (C) and rectal (R) carcinomas and carcinomas from HNPCC-patients (H). GADPH2, 25 cycles of PCR amplifications and AQP8, 35 cycles of amplification.
To study what cells expressed the AQP8 mRNA, in situ hybridization was used. In samples from 2 normal colonic mucosa there was a staining of mRNA AQP8 expression in the columnar surface epithelium only (Fig. 3a). No expression of AQP8 was detected in any of one sporadic rectal carcinoma, one sporadic colon carcinoma or one HNPCC carcinoma tested (the HNPCC carcinoma is shown)(Fig. 3c). The RNA in situ experiment suggested that only normal mucosa expressed AQP8.
In total 15 colon cancer cell lines, six microsatellite unstable (SW48, LOVO, HT116, RKO, LS174T, DDL1) and 9 microsatellite stable (SW480, SW620, HT29, CACO2, Crl7184, Crl7456, Crl2405, 238CCL, 233CCL), were tested for expression of AQP8 using RT-PCR. None of them demonstrated expression of the AQP8 transcript, supporting the hypothesis that tumor cells do not express AQP8 (Fig. 4).
Discussion
This study was designed to detect additional genes altered during colorectal carcinogenesis. One of the genes found, AQP8, was expressed in all normal colon samples but not, or to a less extent, in the colorectal tumors. The expression of AQP8 in tumor samples in Figure 2 can be explained in two ways. It is possible that adenomas and cancer cells do not express AQP8 in these samples and that the PCR products results from normal cell contamination. Alternatively it is possible that adenomas can develop in a cell already differentiated to express AQP8. The known function of this gene and the in situ hybridization result from this study suggest this protein to be a marker of fully differentiated epithelial cells. In the maturation process of cells from the colonic crypt, the stem cells differentiate to become either goblet cell in the crypt or surface epithelial cell [8]. This experiment suggests that the expression of AQP8 is limited to the differentiated mature columnar epithelial cells. None of the cell types in the crypt, including the stem cells seemed to express AQP8 in our in situ experiment. Tumors did not express AQP8. The reason for this can be that this gene has been downregulated in tumorigenes. Another, perhaps more likely, explanation could be that tumors arise in immature cells. There are other proteins whose expression is of late appearance in the life cycle of a colonic crypt cell, like the epthelial Na+ conductance channel ENaC [9] and maxi K+ channels [10], which could have been picked up by this experiment similar to AQP8. The reason why they were not picked could be related to the fact that using a limited number of primers not all differentially expressed genes will show up. Furthermore, since these genes are mostly expressed in distal colon cells, and not in proximal colon cells they would not have been identified by us as differentially expressed between normal and tumor tissue and therefore not selected for further studies.
Aquaporins (AQP) are members of the major intrinsic super family of integral membrane proteins, which act as specialized channels to facilitate the passage of water through the cell membrane in animals, plants, and bacteria. So far 10 AQP homologues named from 0 to 9 have been cloned in mammals. They are widely distributed and more than one AQP could be present in the same cell [11]. Several human AQP genes have been cloned and an increasing number of disturbances have been found associated to the abnormal function of these proteins [11]. The human AQP8 gene was cloned in 1998 and was shown to be important for cell fluid transport [12]. Expression of the gene was demonstrated by northern blot in human pancreatic and colonic tissue [12].
It is generally accepted that water transport in the gastro intestinal tract is secondary to osmotic driving forces created by active salt transport to hydrostatic pressure differences and absorption takes place despite high effective osmolarity of the faeces. Several studies have suggested the colonic crypts to be the major sites of both fluid absorption and agonist-stimulated fluid secretion [13]. Recently AQP8 was demonstrated to be expressed in the absorptive columnar cells of the jejunum and colon in rat by in situ hybridization and it was suggested that AQP8 had a role in water intake and/or outlet in the gastrointestinal tract [14]. In mice the AQP3, AQP4 and AQP8 are expressed in the colon surface epithelium, suggesting a combined role of these aquaporins in fluid transport [15, 16].
A) Antisens, RNA-in situ demonstrates AQP8 in normal colonic columnar epithelium from one patient. B) Sens, RNA-in situ demonstrate no signal. C) Antisens, HNPCC carcinoma demonstrates that there is no AQP8 expression in the tumor. Epithelial cells of the crypt formation on all photos give an autoflorescence seen as white borders in the pictures, the probe hybridization is seen as small white dots.
RT-PCR demonstrating that there is no expression of AQP8 in any of the tested colorectal cancer cell lines. Line one, normal mucosa demonstrate the band corresponding to AQP8 expression.
Conclusion
Our finding of AQP8 expression in human columnar surface epithelial cells, suggests these cells to be involved in water resorption in human colon. Since mRNA-in situ hybridization showed a positive signal for mRNA expression of AQP8 in the normal columnar surface epithelial cells and no signal from the cells in the crypt (Fig. 3), this indicates that the expression of AQP8 is a marker only of fully differentiated surface colonic epithelial cells.
Materials and MethodsTumors and cell lines
mRNA differential display RT-PCR (mRNA DD-PCR) was run on a panel of 24 fresh frozen colorectal tumors and 4 normal colon mucosa tissue (collected at the Umeå hospital 1987-1991). All tumors were at stage Dukes´ B. For the expression study with RT-PCR and Northern blot, we used the same tumors as above plus another 3 normal colonic tissues, 7 sporadic colorectal carcinomas, 5 sporadic adenomas (collected at Umeå Hospital 1987-1991) and 3 HNPCC tumors (collected at Karolinska Hospital, Stockholm). The RNA in situ study used samples from the above-described tissues. From all patients samples were obtained immediately after surgery and frozen at -70°C. The cell lines SW48, LOVO, HT116, RKO, LS174T, DDL1, SW480, SW620, HT29, CACO2, Crl7184, Crl7456, Crl2405, 238CCL, and 233CCL were cultured in Dulbecco´s modified Eagle´s medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum (Gibco), 100 U/ml penicillin (Gibco), and 100 μg/ml streptomycin (Gibco) in 90 mm tissue culture plates (Falcon).
RNA and DNA preparation
Total RNA was extracted from frozen tissue by homogenizing with a power homogenizer (IKA Labortechnik) in Trizol Reagent due to manufacturer's protocol (Life Technologies) and used for mRNA DD-PCR, RT-PCR, and northern blot. The RNA was treated with deoxyribonuclease I (Ambion Inc).
mRNA DD-PCR
mRNA Differential Display RT-PCR (mRNA DD-PCR) was performed essentially according to the method of Liang and Pardee, using the one base anchoring primer [17, 18]. Primers were synthesized by Genset Oligos. Anchoring primer for the AQP8 fragment was 5'-AAGCT11C-3' and the arbitrary primer was 5'-AAG CTT TTA CCG C-3'. Band was cut and reamplified with the same primers and then cloned without purification in a TA-vector (PCR II, Invitrogen). Clones were selected by size, followed by inverse hybridization [19]. DNA was sequenced with Thermo sequenase radiolabeled terminator cycle sequencing kit (Pharmacia & Amersham). Homology searches were performed using Blast search in GenBank databases.
RT-PCR
Primers designed for the isolated fragments were used on template from tumors and normal colonic tissue for comparison of expression. In a volume of 40 μl, 2 μg total RNA was reverse transcribed with the following conditions: 5 mM MgCl2, 50 mM KCL, 10 mM Tris-HCl, pH 8.3, 1 mM dNTP´s, 1 U RNase inhibitor (Perkin Elmer), 2.5 U MuLV Reverse transcriptase (Perkin Elmer).
Semiquantitative PCR
The PCR reaction volume was 25 μl containing 500 nM of each primer, 250 nM of DNTP´s, and 2.5 U of ampliTaq (Gibco, Life Technologies). cDNA was diluted 1:20 in water and 10 μl of the same dilution were used for all samples and controls. The primers used were as follows for AQP8 1200r 5'-AAA TGG GGT GCG GGA AAT GAG-3', aqp8f 800 5'-CCT GCT TG TTG GAC TGC TC-3', and aqp8f 150f 5'-TGG CGA GTG TCC TGG TAC-3'. Glyceraldehyde-3-phosphate dehydrogenase was used as internal control with the primers: F 5'-GAA GGC TGG GGC TCA TTT G-3' and R 5'-GAT GGC ATG GAC TGT GGT CA-3'. Thermocycler (PTC 225, Mj Research) was used with an initial denaturation step of 94°C for 2 min, followed by cycles of 94°C for 45 s, 59°C for 30 s, 72°C for 30 s, and a final elongation step of 72°C for 5 min. PCR products amplified after 25, 30, and 35 cycles were separated on a 1.5% agarose gel containing ethidium bromide for comparison between samples and controls.
Northern blot
10 μg total RNA from 3 normal colonic tissues, 2 HNPCC carcinomas, 3 sporadic colon carcinomas and 2 colon cancer cell lines, was separated by size using electrophoresis on 1.5% denaturing agarose gel. The RNA was transferred to a nylon plus membrane (Qiabrane, Qiagen). The membrane was UV-crosslinked and prehybridization and hybridization was as described [20]. A 600 bp gel-purified PCR fragment generated with primer 5'-TGT TTG CAC TGC TCT CTG G-3' and 5'-AAA TGG GGT GCG GGA AAT GT-3' was 32P-labeled with a random primer labelling kit (redi prime DNA labelling system; Pharmacia & Amersham) and used as a probe. The blot was reprobed with radiolabelled β-actin cDNA (Clontech) as loading control. Filters were analyzed with PhosphorImager (Fujix 1000)
RNA-in situ hybridization
Consecutive frozen tissue sections were taken from two normal colonic mucosa tissues, one sporadic colonic cancer, one sporadic rectal cancer and one HNPCC cancer. Fourteen μm sections were cut at -18°C in a cryostat (Jung CM 3000, Leica Instrument). One section was processed for RNA-in situ hybridization and the next was stained with Hematoxylin and Eosin (H&E). In situ RNA hybridization was performed as described previously [21]. About 40 bp oligonucleotides from base pairs 165-205 were synthesized as sense and antisense (GeneSet oligos) and used as probes.
Acknowledgements
The Swedish Cancer Foundation, The Cancer Foundation in Stockholm and The Lion's Cancer Research Foundation in Umeå supported this work. The authors would like to thank Susanne Petersson for valuable advice.
Maitotoxin, a potent cytolytic agent, causes an increase in cytosolic free Ca2+ concentration ([Ca2+]i) via activation of Ca2+-permeable, non-selective cation channels (CaNSC). Channel activation is followed by formation of large endogenous pores that allow ethidium and propidium-based vital dyes to enter the cell. Although activation of these cytolytic/oncotic pores, or COP, precedes release of lactate dehydrogenase, an indication of oncotic cell death, the relationship between CaNSC, COP, membrane lysis, and the associated changes in cell morphology has not been clearly defined. In the present study, the effect maitotoxin on [Ca2+]i, vital dye uptake, lactate dehydrogenase release, and membrane blebbing was examined in bovine aortic endothelial cells.
Results
Maitotoxin produced a concentration-dependent increase in [Ca2+]i followed by a biphasic uptake of ethidium. Comparison of ethidium (Mw 314 Da), YO-PRO-1 (Mw 375 Da), and POPO-3 (Mw 715 Da) showed that the rate of dye uptake during the first phase was inversely proportional to molecular weight, whereas the second phase appeared to be all-or-nothing. The second phase of dye uptake correlated in time with the release of lactate dehydrogenase. Uptake of vital dyes at the single cell level, determined by time-lapse videomicroscopy, was also biphasic. The first phase was associated with formation of small membrane blebs, whereas the second phase was associated with dramatic bleb dilation.
Conclusions
These results suggest that maitotoxin-induced Ca2+ influx in bovine aortic endothelial cells is followed by activation of COP. COP formation is associated with controlled membrane blebbing which ultimately gives rise to uncontrolled bleb dilation, lactate dehydrogenase release, and oncotic cell death.
Background
Maitotoxin (MTX), one of the most potent marine toxins known, is found in the "red-tide" dinoflagellate, Gambierdiscus toxicus, and is responsible in part for Ciguatera seafood poisoning. In all cells examined to date, MTX at subnanomolar concentrations causes a profound increase in cytosolic free Ca2+ concentration ([Ca2+]i) [1]. This occurs, not by release of Ca2+ from internal stores, but rather from activation of a ubiquitously-expressed, non-selective, Ca2+-permeable cation channel (CaNSC), present in the plasmalemma [2,3,4,5,6,7,8]. Channel activation is followed after a short lag by the activation or formation of large endogenous pores that allow organic molecules with molecular weights of <800 Da to cross the plasma membrane [8]. The activation of these pores can be determined experimentally by following the uptake of ethidium and propidium-based vital dyes. These dyes, of varying molecular weights, are normally excluded from the cytoplasm of intact viable cells, but gain access to the cell interior following pore formation where they bind to nucleic acids with a concomitant increase in dye fluorescence. The large pores activated by MTX have been referred to as cytolytic/oncotic pores, or COP, since their activation ultimately leads to the release of lactate dehydrogenase (LDH), an indication of necrotic or oncotic cell death [9].
The cell death cascade activated by MTX is not unique to this toxin. Stimulation of P2X purinergic receptors by ATP causes similar changes in cytosolic Ca2+ and vital dye uptake [10,11,12,13,14,15,16,17] suggesting that MTX activates a cell death cascade that is physiologically relevant, although the exact role of either the P2X- or MTX-induced cascade in normal cellular biology remains unknown. For the P2X receptor it has been suggested that the Ca2+-permeable channels grow in size to form the dye-permeable pores through either aggregation of channel subunits or through dilation of the existing channel pore structure [12,18,19,20,21]. In support of this model it has been found that the kinetics of ATP-induced pore formation in HEK cells, heterologously expressing the P2X7 receptor, appears to depend on molecular size of the permeating ionic species [19]. However, we recently showed that although MTX and ATP activate distinct channels, the characteristics of the ATP- and MTX-activated COP are indistinguishable [9]. These results suggest that the channel and COP are unique molecular structures. The aggregation or dilation model make specific predictions concerning the kinetics of dye uptake. In particular, this model predicts a delay between channel activation and dye uptake and this delay should be directly proportional to the molecular size of the permeating dye. Furthermore, if pores grow in size, dye uptake should be nonlinear with time. In contrast, if channel activation causes the formation or activation of a molecularly unique COP with fixed pore dimensions, the delay between channel activation and dye uptake should be independent of molecular size, but the subsequent rate of dye uptake should be linear and inversely proportional to molecular weight. To distinguish between these two models, the effect of MTX on [Ca2+]i, vital dye uptake and LDH release was examined in bovine aortic endothelial cells, a cell line particularly sensitive to the cytolytic effects of MTX. The results of the present study are consistent with the activation of an endogenous COP of fixed pore dimensions.
The opening of large pores in the plasmalemma is expected to cause a dramatic change in the ionic concentration gradients that normally exist between the extracellular and intracellular milieus, i.e., loss of K+, and gain of Na+, Ca2+, and Cl- by the cell. The concomitant flow of water into the cell as a result of this ionic redistribution, will drive the cell towards the Gibbs-Donnan equilibrium. The change in osmotic pressure will produce cell swelling and ultimately membrane rupture and release of large macromolecules from the cytoplasm. This final phase in the cell death cascade can be monitored experimentally by measuring the release of the ubiquitous cytosolic enzyme, LDH. The role of COP and the biophysical mechanism associated with this rather violent cellular event remains unknown. However, many cell types undergo membrane blebbing in response to changes in osmotic pressure. Such membrane blebbing, which may in a sense represent a cellular safety valve, has been reported during both ATP- and MTX-induced cell death [19,22], but it is unclear if blebbing occurs before, during, or after COP formation or LDH release. In the present study, the effect of MTX on vital dye uptake in BAECs was correlated with changes in cell morphology using single cell fluorescence videomicroscopy. The videos presented demonstrate that COP activation as indicated by vital dye uptake correlates in time with the formation of membrane blebs, and that membrane lysis, i.e., LDH release, is associated with dramatic bleb dilation.
Results and DiscussionMTX increases [Ca2+]i in BAECs
Fura-2-loaded BAECs were suspended in HBS, placed in a cuvette at 37°C, and the response to extracellular application of MTX was recorded as a function of time (Fig. 1). MTX produced an immediate, concentration-dependent increase in [Ca2+]i. At the highest concentration of MTX examined (3 nM), [Ca2+]i rapidly increased within 1-2 minutes from a mean ± SE resting level of 63 ± 4 nM to an apparent [Ca2+]i value close to fura-2 saturation. We previously showed that MTX-induced activation of COP allows leakage of fura-2 from human skin fibroblasts [8]. Thus, the apparent saturation of the fluorescence ratio observed in BAECs reflects, at least in part, a contribution from extracellular fura-2. This is particularly true at later times. Nonetheless, the major effect of increasing MTX concentration was on the rate of change of [Ca2+]i, and the apparent ED50 was approximately 0.3 nM MTX. MTX had no effect on BAEC [Ca2+]i in the absence of extracellular Ca2+ (data not shown), demonstrating that MTX activates Ca2+ influx, but does not cause the mobilization of Ca2+ from internal stores. These results are similar to those obtained in human skin fibroblasts, mouse THP-1 monocytes, and human embryonic kidney (HEK) cells [8,9] and supports the conclusion that the MTX-activated CaNSC is ubiquitously expressed.
Effect of maitotoxin (MTX) on fura-2 fluorescence in bovine aortic endothelial cells (BAECs). Fura-2-loaded BAECs were suspended in HBS and the fluorescence ratio recorded as a function of time as described in Materials and Methods. Six traces are shown superimposed. At time 100 sec, MTX was added during the individual recordings at the final concentration indicated to the right of each trace. Results shown are representative of 11 individual experiments.
MTX activates COP in BAECs
To evaluate MTX-induced activation of COP in BAECs, the uptake of ethidium (Mw 314 Da), YO-PRO-1 (Mw 375 Da) and POPO-3 (Mw 715 Da) was examined at various concentrations of MTX (Fig. 2). After a short lag, MTX produced a biphasic, concentration-dependent increase in ethidium, YO-PRO, and POPO-3 uptake into the cells. The lag-time for a given concentration of MTX was the same for each dye (Fig 2D and 2E; first dashed lined), but the rate of dye uptake during the initial phase was inversely proportional to molecular weight. Furthermore, at the highest concentration of MTX examined, the uptake of ethidium, YO-PRO, and POPO-3 during the initial phase was linear (Fig. 2D). These results suggest that the first phase reflects activation of a large pore (i.e., activation of COP) that exhibits finite permeability for each of the dyes tested and is distinct from the CaNSC activated by MTX. The molecular mechanisms responsible for the delay between MTX addition and COP formation (i.e., vital dye uptake) is unknown. The delay, which was also observed in fibroblasts, HEK cells, and THP-1 monocytes [8,9] suggests that there may be several biochemical steps between channel activation and COP formation or that the concentration of a critical second messenger must reach a threshold level before COP opens. This could explain the temperature sensitivity of vital dye uptake, previously noted for both MTX- [8,9] and purinergic receptor-induced [12,15] vital dye uptake. During the second phase, dye uptake was all-or-nothing and once initiated the rate of dye uptake was independent of dye molecular weight, i.e., ethidium, YO-PRO, and POPO-3 all enter the cell at the same time and at the same rate for each concentration of MTX examined. These results suggest that the second phase of dye uptake is related to cytolysis.
Effect of MTX on vital dye uptake in BAECs.Panel A. Ethidium bromide (EB) uptake in BAECs was determined from the increase in fluorescence as a function of time as described in Material and Methods. Five traces are shown superimposed. EB was added at 50 sec to BAECs suspended in HBS. At the time indicated by the arrow (2 min), MTX was added at the final concentration indicated to the right of each trace. Values shown are normalized to the maximum fluorescence obtained by addition of digitonin at the end of each trace. Panels B and C. Same as in panel A with YO-PRO-1 (blue) or POPO-3 (red) added at 50 sec. Panels D and E. Comparison of EB (black), YO-PRO (blue), and POPO-3 (red) uptake data for 1.0 and 0.3 nM MTX; note expanded fluorescence scale and time base. Results shown are representative of 5 individual experiments.
MTX induces LDH release
To test the hypothesis that the second phase of dye uptake reflects a large change in membrane permeability indicative of cytolysis, the uptake of ethidium and release of LDH (Mw 140,000 Da) were compared in parallel experiments performed on the same batch of cells (Fig 3). MTX again produced a biphasic increase in ethidium uptake into the cells. No LDH release was observed during the first phase of ethidium uptake consistent with the hypothesis that activation of COP occurs prior to release of LDH. Extracellular LDH was measurable 7 minutes after MTX addition and subsequently increased in parallel with the second phase of ethidium uptake. These results suggest that the second phase of dye uptake reflects a loss of membrane integrity and provides an explanation for the all-or-nothing uptake of ethidium, YO-PRO, and POPO-3, during this time frame. Note that more than 90% of the cell associated LDH is released within 15 minutes of MTX (0.3 nM) addition. For comparison, we previously reported that treatment of human skin fibroblasts for 10 minutes with 1 nM MTX caused release of 28% of cell-associated LDH [8]. Thus, BAECs are particularly sensitive to the cytolytic effects of MTX.
Comparison of EB uptake and LDH release in BAECs. EB uptake (red lines) was determined in the absence and presence of 0.3 nM MTX as indicated to the right of each trace. The solid lines are mean values from 3 experiments; for clarity ± SE values are only shown at selected time points, i.e. every 20 sec. In paired experiments, LDH release (black symbols) was monitored as described in Material and Methods. The values shown represent the mean ± SE LDH release determined in the presence (black circles) or absence (black diamonds) of 0.3 nM MTX. Note that LDH release in the absence of MTX was only determined at time zero and at time 17 min.
MTX produces similar changes in dye uptake in single BAECs
The ethidium uptake experiments shown in Figs. 2 and 3 reflect the average response of the entire population of BAECs within the cuvette. It is possible that the biphasic kinetics observed reflects, at least in part, heterogeneity within the cell population. To test this hypothesis, the effect of MTX was examined at the single cell level using fluorescence microscopy (Fig 4). BAECs were sparsely seeded on glass coverslips and mounted on the stage of an inverted fluorescence microscope. The epifluorescence in each cell was measured using a CCD camera and image analysis software. The uptake of ethidium by single BAECs, as indicated by the change in single cell fluorescence, was biphasic (Fig 4). The rate of ethidium uptake during the first phase was concentration-dependent, but the second phase appeared to be all-or-nothing as was seen in the cuvette experiments. Likewise, the delay between addition of MTX and the second phase of ethidium uptake, although variable from cell to cell, was concentration-dependent. Thus, the overall profile observed at the single cell level recapitulated the results obtained in the population experiments. One interesting difference between the cuvette and single cell experiments is the time at which the second phase begins. As seen in Fig 2A, the second phase of dye uptake started approximately 5 minutes after addition of 1 nM MTX. At the single cell level, the second phase is clearly delayed and on average begins approximately 10-12 minutes after 1 nM MTX addition (Fig 3D). Although we do not know the reason for this difference, the population studies were performed on dispersed cells in a cuvette whereas, the single cell studies were performed on BAECs attached to glass coverslips. Thus, adherent cells may be resistant to the gross changes in membrane structure that occurs during the second phase of the response to MTX.
Effect of MTX on EB uptake in BAECs at the single cell level.Panels A, B, and C. BAECs on glass cover-slips were perfused with HBS containing EB at 37°C and fluorescence was measured in individual cells as described in Material and Methods. At the time indicated by the arrow, the bath solution was exchanged for one containing 1 (black), 0.3 (red) or 0.1 (blue) nM MTX. Fluorescence recordings from 8 individual cells are shown in each panel. Panel D. The average single cell EB uptake as a function of time was determined from the data shown in panels A, B, and C. The concentration of MTX is indicated to the right of each trace. Results shown are representative of 3 individual experiments.
Previous studies suggested that ATP-induced pores in HEK cells heterologously expressing the P2X7 purinergic receptor, were formed either by dilation of the channel structure or by aggregation of channels subunits [19]. However, these studies relied on shifts in the reversal potential of whole-cell membrane currents. Although this is a sensitive technique for determining the time course of pore formation, these experiments do not eliminate the possibility that COP and the P2X channel are separate entities. In the present study, the effect of MTX on membrane permeability occurred in three distinct phases. The first phase reflects the activation of a CaNSC and a large and rapid increase in [Ca2+]i. After a short lag, the activation of COP allowed uptake of vital dyes into the cell and the lag was independent of dye molecular weight. At the highest concentration of MTX examined, dye uptake via COP was linear and inversely proportional to molecular weight. These results suggest that COP does not increase in size as a function of time, as would be predicted by the dilation or aggregation model. Furthermore, these results strongly suggest that the CaNSC and the COP are unique molecular structures.
MTX causes biphasic membrane blebbing
During the final phase of MTX-induced effects, there is a gross change in membrane integrity which allows rapid uptake of vital dyes, independent of dye molecular weight. This final phase is associated with release of LDH which is indicative of oncotic cell death. However, over the course of the single cell experiments we noticed that BAECs treated with MTX not only accumulate vital dyes, but also undergo an enormous reorganization of the plasma membrane to form giant membrane blebs. An example of MTX-induced membrane blebs is shown in the micrograph in Fig 5 which was obtained with Hoffman optics and simultaneous bright-field and fluorescence illumination. After 50 minutes of MTX treatment at room temperature, YO-PRO staining is prominent in the nucleus as expected, but is also visible in giant membrane blebs seen surrounding each cell in the field of view. At the present time, we do not know the composition of the membrane that comprise each bleb, or the structural features of the cell that would allow evagination or expansion of these membrane structures at specific sites on the cell surface. However, blebbing presumably requires a change in osmotic pressure within the cell and so we reasoned that the blebs may be associated with the second phase of vital dye uptake. To test this hypothesis, we compared the morphological changes at the cell surface with dye uptake at the single cell level using time-lapsed videomicroscopy. A montage showing bright-field and fluorescence image pairs is shown in Fig 6 and the complete time-lapsed video sequence is shown in Fig 1 of the Additional Data Files section (Montage.avi). As can be seen, the first phase of ethidium uptake is associated with the formation of membrane blebs (Fig 6, yellow arrows) with an average diameter of approximately 4 microns. Blebs begin to form within 3 minutes of MTX addition to the bath solution and remain at essentially a fixed size until approximately time 22 minutes. As seen in the graphic presentation of fluorescence in each cell (Fig 6, bottom right), bleb formation correlates with the initial phase of dye uptake. From time 28 to 38 minutes, the second phase of dye uptake is observed, and this phase is associated with large bleb dilation (Fig 6, red arrows). These results demonstrate that COP activity is associated with initial bleb formation, whereas LDH release (i.e., cytolysis) is associated with dramatic bleb dilation.
MTX-induced membrane blebbing in BAECs. This figure shows a photomicrograph of BAECs treated with 1.0 nM MTX in the presence of YO-PRO-1 for 50 min at room temperature. The image shown was obtained using Hoffman optics and dual bright-field/fluorescence illumination. Note the intense green fluorescence of the nucleus, indicative of YO-PRO uptake. Fluorescence is also detected in the cytoplasm and membrane blebs.
Comparison of EB uptake and membrane blebbing in BAECs. BAECs on glass cover-slips were perfused with HBS at 37°C. A phase (left) and fluorescence (right) image pair was obtained every 30 sec for 40 min. MTX (0.3 nM) was added to the bath solution at time 5 min. The montage shows image pairs at the times indicated above each pair. Bleb formation occurs between 8 min and 22 min as shown by the examples indicated by the yellow arrows. Bleb dilation occurs between 22 and 38 min as shown by the examples indicated by the red arrows. The complete time-lapse video is available in the Additional Data Files section (File 1). Ethidium uptake as a function of time was quantified as described in Materials and Methods in defined regions over each cell (colored squares indicated at time 38 min) and plotted as a function of time (lower right panel). Results shown are representative of 19 individual experiments using various concentrations of MTX and different vital dyes.
To better appreciate the dynamic nature of bleb formation and to clearly observe the rapid nuclear staining associated with bleb dilation, we created time-lapsed movies in which the bright-field and fluorescence images were merged into a single video (Fig 2, EB.avi and Fig 3, YO-PRO.avi; Additional Data Files section). By setting your video viewer to continuously loop (i.e., auto replay), and by focusing during each sequence on a single cell within the field of view, it can be seen that fluorescence increases slowly during bleb formation, but that intense rapid dye staining occurs during the bleb dilation phase. As seen in Fig 5, and in each of these videos, MTX-induced blebbing results in an enormous increase in membrane surface area which may represent stretching of the membrane or an evagination of existing caveolar structures associated with the plasmalemma of endothelial cells. Endothelial cells are also known to have an extensive system of intracellular vesicles which are used for transporting substances from the blood to the interstitial space, i.e., transcytotic vesicles. Thus, membrane blebbing could reflect a massive exocytotic event. Why this would occur at selective sites on the plasmalemma remains unknown. Interestingly, the videos show several examples of blebs forming on the surface of pre-existing blebs during the dilation phase. Thus, the structural element(s) that are required for localized evagination of membrane appears to be membrane-associated and/or "pulled" from the cytoplasm during the initial bleb formation. Irrespective of the exact mechanism, the experiments reported in the present study are the first to correlate MTX-induced vital dye uptake with alterations in cell morphology.
MTX-induced membrane blebbing requires Ca2+
Previous studies on human skin fibroblasts showed that MTX had no effect on [Ca2+]i, vital dye uptake, or fura-2 efflux when the cells were challenged in the absence of extracellular Ca2+ [8]. Although elevating [Ca2+]i with ionomycin had no effect on vital dye uptake, preliminary experiments in which fibroblasts were loaded intracellularly with the Ca2+ chelator BAPTA, showed that MTX-induced responses were significantly delayed, suggesting that a rise in [Ca2+]i is necessary, but not sufficient for MTX-induced COP formation. To determine if extracellular Ca2+ is also required for MTX-induced membrane blebbing, changes in cell morphology were correlated with ethidium uptake in the presence of Ca2+-free extracellular buffer (Fig. 7; the video showing the merged phase and fluorescence images for this experiment is given in the Additional Data Files section; see Ca-free.avi). MTX had no effect on BAECs incubated in Ca2+-free buffer. Importantly, no dye uptake or membrane blebbing was observed for 30 min after the addition of MTX to the bath solution. Subsequent readmission of Ca2+ to the bath produced an increase in ethidium uptake that again followed a biphasic time course, virtually identical to the profile observed when MTX was added to cells incubated with normal Ca2+-containing medium. As seen in Fig 7 and the video, dramatic membrane blebbing and ethidium uptake is observed 25 min after re-addition of Ca2+ to the bath. Control experiments in which Ca2+ was removed and added back to BAECs in the absence of MTX showed that these responses were dependent on the presence of MTX, i.e., no dye uptake or blebbing was detected in the absence of MTX (Fig 7). These results demonstrate that extracellular Ca2+ is necessary for MTX-induced COP formation and membrane blebbing, and suggest that both responses require Ca2+ influx and a concomitant rise in [Ca2+]i.
Effect of Ca2+ on EB uptake and membrane blebbing in BAECs. BAECs on glass cover-slips were perfused with Ca2+-free HBS at 37°C as indicated in the bottom panel. A phase and fluorescence image pair was obtained every 30 sec for 60 min. MTX (1 nM) was added to the bath solution at time 5 min. Selected phase images are shown at the indicated times (upper panels). The complete time-lapse video is available in the Additional Data Files section (File 4). Ethidium uptake as a function of time was quantified at the single cell level (for 11 cells) as described in the legend to Fig 6 and plotted as a function of time (lower panel). Results shown are representative of 3 individual experiments. For control, identical solution changes were made in the absence of MTX; the image shown (right middle panel) represents the 60 min time point, i.e., 25 min after re-addition of Ca2+ to the bath soluiton in the absence of MTX.
Conclusions
In conclusion, MTX treatment of BAECs causes a specific sequence of events (i.e., a cell death cascade) that is triggered by the activation of CaNSC and a rise in [Ca2+]i. This is followed by formation or activation of COP which is correlated with the formation of membrane blebs. COP appears to be a unique molecule associated with the plasma membrane and to have a fixed pore geometry and conductance for each vital dye examined. Furthermore, activation of COP provides the initial driving force for osmotic swelling and bleb formation. LDH release, indicative of the final phase of MTX-induced cell death, is associated with massive bleb dilation. Although, the molecular mechanisms associated with each step in the cell death cascade remain unknown, MTX may prove to be an important tool for understanding the biochemical and biophysical links between channel activation, COP formation, and membrane blebbing.
Materials and MethodsSolutions and reagents
Unless otherwise indicated, HEPES-buffered saline (HBS) contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 10 mM D-glucose, 1.8 mM CaCl2, 15 mM HEPES, 0.1% bovine serum albumin, pH adjusted to 7.40 at 37°C with NaOH. Ca2+-free HBS contained 0.3 mM EGTA and the same salts as HBS without added CaCl2. Fura-2 acetoxymethyl ester (fura-2/AM), ethidium bromide, YO-PRO-1 and POPO-3 were obtained from Molecular Probes (Eugene, OR, USA). Maitotoxin, obtained from LC Laboratories (Woburn, MA) or Wako Bioproducts (Richmond, VA), was stored at -20°C in ethanol. All other salts and chemicals were of reagent grade.
Cell Culture
Bovine aortic endothelial cells were cultured as previously described [23] using Dulbecco's modified Eagles medium (GIBCO) supplemented with 10% fetal bovine serum (Hyclone, Logan UT), 100 μg/ml streptomycin and 100 μg/ml penicillin (complete-DMEM). All cultures demonstrated contact-inhibited cobblestone appearance typical of endothelial cells.
Measurement of the apparent cytosolic free Ca2+ concentration
[Ca2+]i was measured using the fluorescent indicator, fura-2, as previously described [24]. Experiments were performed with cells in the twelfth to twentieth passage and 2-3 days post-confluency. Briefly, cells were harvested and re-suspended in HBS containing 20 μM fura-2/AM. Following 30 min incubation at 37°C, the cell suspension was diluted ∼ 10-fold with HBS, incubated for an additional 30 min, washed and resuspended in fresh HBS. Aliquots from this final suspension were subjected to centrifugation and washed twice immediately prior to fluorescence measurement. Fluorescence was recorded using an SLM 8100 spectrophotofluorometer; excitation wavelength alternated between 340 and 380 nm and fluorescence intensity was monitored at an emission wavelength of 510 nm. All measurements were performed at 37°C.
Measurement of vital dye uptake
An aliquot (2 ml) of dispersed cells suspended in HBS at 37°C was placed in a cuvette. Following addition of ethidium bromide (final concentrations of 5 μM), fluorescence was recorded as a function of time with excitation and emission wavelengths of 302/560 nm, respectively. All ethidium bromide fluorescence values are corrected for background (extracellular) dye fluorescence and expressed as a percentage relative to the value obtained following complete permeabilization of the cells with 50 μM digitonin. Uptake of POPO-3 and YO-PRO-1 was determined as described for ethidium with excitation/emission wavelengths of 530/565 and 468/510 nm, respectively.
For single cell measurement of vital dye uptake, BAECs in complete-DMEM were sparsely seeded on circular glass coverslips and used within 2-3 days of seeding. The coverslips were mounted in temperature-controlled perfusion chambers and placed on the stage of Nikon Diaphot inverted microscope. The cells were illuminated with light from a 75 watt xenon lamp using a O-5717 filter cube obtained from Molecular Probes. Epifluorescence was recorded using a Hamamatsu intensified CCD camera (model XC-77) and images were acquired and analyzed using Image-1 software (Universal Imaging, West Chester, PA). During each experiment, image pairs were collected at thirty second intervals. Images (8-bit gray scale) were stored as averages of sixteen video frames (phase images) or as accumulations of four video frames (fluorescent images) with shutter controllers switching between light and fluorescent illumination. The fluorescence images were used to determine dye uptake as a function of time. Using Image-1 software, regions were defined over single cells and the average fluorescence intensity of the region was quantified. The phase images were contrast enhanced using Debabelizer Pro software (Equilibrium, Sausalito, CA) and merged with the corresponding fluorescent images using Spot™ camera software (Diagnostic Instruments, Sterling Heights, MI). Time-lapse videos were created using Debabelizer Pro software with ethidium and YO-PRO epifluorescence displayed as red and green pseudocolor images, respectively.
Measurement of lactate dehydrogenase (LDH) release
Aliquots of dispersed cells (2 ml) were incubated at 37°C for various lengths of time in the presence and absence of MTX. The cells were pelleted by centrifugation for 15 sec at 12,000 rpm in an Eppendorf centrifuge (model 5415 C). The supernatants were removed, and placed on ice. Enzyme activity in aliquots (50 μl) of the supernatants was determined using the LD-L kit from Sigma. All values are expressed as percent LDH released relative to the value obtained following permeabilization of the cells with 50 μM digitonin.
Supplementary Material
File 1: Montage.avi
BAECs on glass cover-slips were perfused with HBS at 37°C. A phase (left) and fluorescence (right) image pair was obtained every 30 sec for 40 min. MTX (0.3 nM) was added to the bath solution at time 5 min. The time-lapse video was created from the captured images as described in Material and Methods, with a time compression of 3.5 minutes (i.e., 7 images) per second.
movie 1
File 2: EB.avi
BAECs on glass cover-slips were perfused with HBS at 37°C. A phase and fluorescence image pair was obtained every 30 sec for 40 min. MTX (0.3 nM) was added to the bath solution at time 5 min. The time-lapse video was created from the captured images as described in Material and Methods, with a time compression of 3.5 minutes (i.e., 7 images) per second. For this movie, the phase and EB fluorescence images were merged into a single video. EB fluorescence is given a red pseudocolor.
movie 2
File 3: YO-PRO.avi
BAECs on glass cover-slips were perfused with HBS at 37°C. A phase and fluorescence image pair was obtained every 30 sec for 40 min. MTX (0.3 nM) was added to the bath solution at time 5 min. The time-lapse video was created from the captured images as described in Material and Methods, with a time compression of 3.5 minutes (i.e., 7 images) per second. For this movie, the phase and YO-PRO fluorescence images were merged into a single video. YO-PRO fluorescence is given a green pseudocolor.
movie 3
File 4: Ca-free.avi
BAECs on glass cover-slips were perfused with Ca2+-free HBS at 37°C. A phase and fluorescence image pair was obtained every 30 sec for 60 min. MTX (0.3 nM) was added to the bath solution at time 5 min and Ca2+ was re-added to the bath solution at time 35 min. The time-lapse video was created from the captured images as described in Material and Methods, with a time compression of 3.5 minutes (i.e., 7 images) per second. For this movie, the phase and ethidium fluorescence images were merged into a single video. Ethidium fluorescence is given a red pseudocolor.
movie 4
Acknowledgements
This work was supported in part by NIH grant GM52019 and grant 9806267 from the American Heart Association. We gratefully acknowledge the technical assistance of Zack Novince and Justin Weinberg.
This study describes the functional interaction between the putative Ca2+ channel TRP4 and the cystic fibrosis transmembrane conductance regulator, CFTR, in mouse aorta endothelium (MAEC).
Results
MAEC cells express CFTR transcripts as shown by RT-PCR analysis. Application of a phosphorylating cocktail activated a Cl- current with characteristics similar to those of CFTR mediated currents in other cells types (slow activation by cAMP, absence of rectification, block by glibenclamide). The current is present in trp4 +/+ MAEC, but not in trp4 -/- cells, although the expression of CFTR seems unchanged in the trp4 deficient cells as judged from RT-PCR analysis.
Conclusions
It is concluded that TRP4 is necessary for CFTR activation in endothelium, possibly by providing a scaffold for the formation of functional CFTR channels.
Background
The cystic fibrosis transmembrane conductance regulator (CFTR) is well described as a low-conductance, cyclic nucleotide-regulated Cl- channel in epithelial cells [1]. Only recently, CFTR has also been detected in vascular endothelium [2]. Endothelial cells (EC) form an anticoagulative barrier but also control many other functions, such as regulation of the vascular tone by secretion of vasoactive compounds such as bradykinin, and autocoids, such as nitric oxide and prostacyclin [3]. These functions are modulated by a diversity of ion channels among which Cl- channels [4, 5]. Endothelial Cl- channels, the volume-regulated anion channel, VRAC, and Ca2+ activated Cl- channels, CaCC, have been shown to modulate EC electrogenesis, are possible mechano-sensors, serve as permeation pathways for amino acids and organic osmolytes and may be involved in regulation of the driving force for Ca2+ entry [for a review, see 6]. This list of Cl- channels has been extended with CFTR, which is functional in human umbilical vein endothelium and in human lung microvascular endothelial cells [1], but not in bovine pulmonary artery endothelial cells [6]. As we show in this work, it is also functional in mouse aorta endothelial cells.
MAEC express different types of putative ion channel transcripts which are encoded by genes of the trp family, trp1, 2, 3, 4, and 6 [7, 8]. TRP4 forms part of a store operated Ca2+ entry channel which is involved in the control of NO-dependent relaxation of the mouse aorta [8]. In addition, TRP4 has been shown to interact via a VTTRL motif in its C-terminal region with the first PDZ domain of the regulatory factor of the Na+- H+exchanger NHERF, which also interacts with PLCβ [9]. The two PDZ domain protein NHERF associates also with the actin cytoskeleton via members of the ezrin/radixin/moesin family [10, 11]. It is also well established that the C terminus of CFTR constitutes a PDZ-interacting domain (QDTRL for the last five C-terminal amino acids) that is required for CFTR polarization to the apical plasma membrane and interaction with the PDZ domain-containing protein NHERF [12]. Thus, both TRP4 and CFTR may bind to similar PDZ-domain proteins.
We have studied the functional expression of CFTR in both trp4 wild type and in trp4 deficient MAEC cells. We show here that CFTR is present in both cell types, but is not functional in trp4 deficient endothelial cells. These data may hint to a more general function of trp4 as regulator of other ion channels and to a novel regulatory mechanism for CFTR.
ResultsExpression of CFTR in mouse aorta endothelium
We have been unable to detect CFTR in bovine pulmonary endothelial cells [6], but its expression has recently been described in endothelium [1]. We have therefore assessed the expression of CFTR in mouse aorta EC (MAEC) by means of two sets of primers, the one detecting exon 5 through exon 9 of CFTR transcripts, and the other one detecting exons 23 and 24 of CFTR transcripts (figure 1A, B). The data show that CFTR is expressed in both wild-type and trp4 deficient MAEC cells, and are consistent with the recent detection of CFTR expression in human umbilical vein endothelium and human lung microvascular endothelial cells.
RT-PCR showing the expression of CFTR in mouse aorta endothelial cells A) cDNA from murine TRP4 +/+ and TRP4 -/- MAEC (lanes 1 and 2), human umbilical vein cells (HUVEC, lane 3) and human nasal epithelium cells (+, for a positive control) were amplified using primers P.CFTR-3249.5 and P.CFTR-3428.3, generating a 180 base pair fragment encoding partial sequences of exons 17a and 17b. Lane 5 is a negative control. B) cDNA from murine TRP4 +/+ and TRP4 -/- MAEC (lanes 1 and 2), from three different preparations of cDNA from HUVEC cells (lanes 3, 5 and 6) and human nasal epithelial cDNA as a positive control (lanes 4 and 7) were amplified using primers PF.CFTR-661.5 and P.CFTR-1360.3, generating a 700 bp fragment harboring exons 5 through 9. Lane 8 is the negative control of amplification.
Functional characterization of CFTR in MAEC
Subsequently, we have investigated the functional expression of CFTR in wild type MAEC. Application of the phosphorylating cocktail, containing10 μM forskolin and 100 μM IBMX, activated in these cells a current without any apparent rectification and voltage-independent kinetics, which reversed at -26 ± 6 mV (n = 6), i.e. close to the Cl-equilibrium potential, ECl (figure 2A). Its phenotype was completely different from that of Cl- currents activated by loading the cells with Ca2+ (figure 2B), which are outwardly rectifying, slowly activate at positive potentials and inactivate at potentials negative to ECl [13, 14]. Challenging the MAEC with a 25 % hypotonic extracellular solution activated another type of Cl- current that inactivates at positive potentials and shows a less pronounced outward rectification (figure 2C). The latter currents have been described in detail in other EC as VRAC, volume-regulated anion currents [6, 15, 16]. The cAMP-activated current reached a stationary value approximately 2 minutes after application of the phosphorylating cocktail (figure 3A), and disappeared after wash-out of the phosphorylation cocktail. The cAMP-activated current was observed in 16 out of 22 cells, but its density was rather small (4.9 ± 1.1 pA/pF at +50 mV, obtained from voltage ramps, n = 16) compared to that of the other Cl- currents. It was activated without any change in cell volume or elevation of intracellular Ca2+. Glibenclamide (50 μM) blocked the cAMP-activated current by 62 ± 4 % (n = 6) (figure 3D). Obviously, the profile of this current is similar to that of CFTR currents in other tissues, i.e. slow activation, linear I-V curve, time-independent kinetics and inhibition by glibenclamide, indicating that CFTR channels are also functionally expressed in MAEC, and coexist with at least two other types of Cl-channels.
At least three different Cl- channels exist in MAEC A. Current traces in a non-stimulated MAEC cell (upper traces) and in the same cell stimulated with the phosphorylation cocktail (lower traces) in response to voltage steps from +80 to -80 mV (decrement = 40 mV, VH = 0 mV). The bottom panel shows the corresponding I-V curves from the current amplitudes recorded at the end of each voltage step (open circles for the resting cell and open triangles for the stimulated cell). Note the voltage-independent kinetics of the current, the lack of rectification and the reversal potential close to ECl B. Current traces in an MAEC cell immediately after breaking into the cell and before it is loaded with Ca2+ (upper traces) and after equilibration with pipette Ca2+ (1 μM). Voltage steps form -100 to +100 mV (increment +20 mV), VH is -20 mV. Note the slow activation at positive potentials, and the inactivation at negative potentials, which are typical for CaCC currents. The corresponding I-V curves in the bottom panel illustrate the strong outward rectification of the CaCC current. C. Current traces from an MAEC cell before and during cell swelling induced by challenging the cell with a 25 % hypotonic solution. Same step protocol as in B. Note the inactivation of the current at positive potentials, which is a feature of volume-regulated anion channels (VRAC). The corresponding I-V curves at the bottom illustrate the weak outward rectification of the VRAC currents.
cAMP activated Cl- currents in MAEC A. Slow activation of the CFTR currents. Currents values were obtained form linear voltage ramps and were measured at +50 mV (holding potential, VH = 0 mV). B. Instantaneous current-voltage relationships measured from voltage ramps at the times indicated by solid circles (a and b) in panel A. C. Difference current, b-a, from the current measured after and before stimulation with the cAMP elevating cocktail. From such protocols current density was measured for the wild type and the trp4-deficient MAEC (see figure 4). D. Block of the cAMP activated current by 50 μM glibenclamide (measured from voltage ramps at +50 mV, VH = 0 mV)
Down-regulation of CFTR currents in trp4-deficient MAEC
The phosphorylating cocktail failed to activate a similar current in trp4-deficient MAEC. Figure 4 shows an example for stimulating wild-type and trp4- deficient MAEC. On the other hand, the CaCC current activated by loading MAEC with 1 μM [Ca2+]i was not significantly different from that in wild type MAEC, i.e. 16 ± 4 pA/pF in WT(n = 11) compared to 24 ± 6 pA/pF (n = 8) in trp4 -/- cells at +100 mV. Also VRAC was not significantly different in both cell types (peak currents at +100 mV: 58 ± 8 pA/pF, n = 7, in WT cells; 55 ± 11 pA/pF, n = 6 in trp4 -/- MAEC).
Failure of CFTR activation in trp4 deficient MAEC A. Current traces from a wild type MAEC obtained from steps protocols (voltage steps from +80 to -80 mV, decrement 40 mV, holding potential was VH = 0 mV). B. Same protocol as in panel A, but after application of the phosphorylating cocktail C. Same protocol as in A, but for a trp4 deficient cell. D. Same protocol as in C for the trp4 -/- cell after stimulation with the phosphorylating cocktail. E. Pooled current densities at +50 mV (obtained from voltage ramps) in wild type and trp4 deficient MAEC under resting conditions (cont) and after stimulation with the phosphorylating cocktail (cAMP). Note the lack of current activation by the phosphorylating cocktail in the trp4 deficient MAEC. Amplitude of the cAMP activated current in wild type (+/+) and trp4 deficient (-/-) endothelial cells, i.e. the difference between the current amplitudes in stimulated and unstimulated cells in panel E.
Discussion
We show that mouse aorta endothelial cells express functional CFTR channels, and that their activation is defective in mouse aorta endothelium of trp4-deficient mice, although these cells still expressed the CFTR transcripts.
CFTR expression has already been demonstrated in human umbilical vein (HUVEC) and lung microvascular endothelial cells (HLMVEC) [1], but was not detectable in bovine pulmonary artery endothelial cells [6]. We show here that CFTR is also expressed as a functional channel in mouse aorta endothelial cells. It has been suggested that the endothelial CFTR has 100% identity with the corresponding epithelial cDNA from exon 3 to exon 6, including exon 5 which is absent in cardiac CFTR [1]. CFTR is also present in corneal, non-vascular endothelium [17]. In MAEC cells, the first arterial cell type described to express CFTR, cAMP activated Cl- currents are smaller than Ca2+ activated and volume-regulated Cl- currents in the same cell type. Activation of CFTR in endothelial cells could be of functional interest in transendothelial transport, in pH regulation because of the CFTR permeability for bicarbonate, and intriguingly also as part of the NO signaling cascade because of their sensitivity to cGMP, which might be elevated during endothelial NO production.
Besides its role as a Cl- channel, CFTR also regulates the activity of various ion channels and transporters, mainly via direct protein-protein interactions [18, 19]. Some of these interactions might be mediated by the association of CFTR via its C-terminal PDZ-binding motif (D-T-R-L) with the PDZ-binding domains of other proteins, such as NHERF [11, 12, 20, 21]. CFTR also functions as regulator of ATP release [22, 23], which might be important for EC function because the released ATP could in turn bind to endothelial P2Y2 and P2X4 receptors [for a review, see 6].
The finding that CFTR, although detectable in RT-PCR, cannot be activated in trp4-defiecint mouse aorta endothelium might be of a special interest, since it shows that the functional activity of CFTR may be connected to the expression of trp4, a member of the trp-family that is involved in capacitative Ca2+ (CCE). CCE, mediated via store-operated channels (SOC), may affect CFTR activity indirectly via modulation of adenylyl or guanyl cyclases. It has indeed been shown that block of CCE by micromolar concentrations of La3+ ions inhibits cyclic AMP synthesis [24, 25]. In endothelial cells, CCE is the principle stimulus for sustained activation of eNOS [26], and the concomitant elevation of the cGMP concentration may in turn activate CFTR [27, 28].
Alternatively, TRP4 may interact more directly with CFTR by forming a signaling complex via binding of their PDZ-binding motifs to PDZ-domain proteins. The Na+/H+ exchanger regulatory factor, NHERF, has been identified as a protein that binds both TRP4 and CFTR [10]. Also light activated TRP channels in the Drosophila photoreceptor cluster with several other signaling proteins by associating with the PDZ-domain protein INAD. It is therefore not unlikely that TRP4 and CFTR associate in MAEC via similar PDZ-domain proteins, such as NHERF or the human INAD-Like (hINAD-L) protein [29].
Conclusions
This is the first report describing a functional interaction between a member of the TRP family and CFTR. It is therefore tempting to speculate that TRPs might be either regulators of CFTR or targets of CFTR regulating proteins.
Materials and MethodsCell isolation
Gene targeting for the construction of the trp4 deficient mouse has been described in detail elsewhere[8]. In short, part of the exon encoding transmembrane domain 4 and 5 and part of the linker between S5 and S6 was replaced by the neomycin-resistance gene and the deletion of the trp4 gene and lack of expression was confirmed [8].
The "primary explant technique" was used to study freshly isolated endothelial cells from mouse aorta of trp4 +/+ (wild type) and trp4 -/- mice [30]. In short, the aorta was removed form anesthetized and heparinized mice. Small pieces of the aorta were placed with their intimal side on matrigel-coated six well plates containing a very small volume of growth medium (mix for 100 ml: 80 ml DMEM (GIBCO BRL 41965) with in addition 10 ml FCS, 7.5 mg ECGF, 200 μl heparin (10 U/ml final), 2 ml penicillin/streptomycin (100 U/ml final, GIBCO BRL 15070), 1 ml L-glutamine (100x, GIBCO BRL 25030-024), and 1 ml Minimal Essential Amino acids so that the aortic pieces adhere to the substratum. Matrigel-coated plates contained TGFβ, FGF, and tPA. Additional endothelial cell growth factors were added (75 μg/ml, ECGS, endothelial cell growth factor supplement, E2759, Sigma). Addition of heparin induced an antiproliferative effect on smooth muscle cells and fibroblasts. After approximately 6 days, we removed the aortic pieces and let the endothelial cells that have migrated from the aortic segments grow to confluence. In this state, explanted endothelial cells can be studied on the matrigel support. For further investigations, cells were passaged by using dispase (Becton Dickinson, Two Oak Park, Bedford, MA, USA). Thereafter, growth medium was added to stop the dispase activity by dilution. Cells were then harvested, centrifuged and resuspended with DMEM.
Total RNA was isolated using TRIzol Reagent (Gibco BRL Life Technologies), according to the protocol provided by the manufacturer. After priming with oligo-dT oligonucleotides, cDNA was synthesized with MLV reverse-transcriptase (Life Technologies) according to the recommendations of the supplier. The cDNA was used as template in a polymerase chain reaction (PCR), using 5 μl of a cDNA synthesis reaction in a 50 μl amplification reaction. The amplification was performed in a DNA Thermal Cycler (Perkin Elmer). The temperature cycling conditions were: denaturation for 1 minute at 95°C, annealing for 1 minute at a temperature that is specific for each pair of primers (55°C for the primer couple P.CFTR-661.5 (5'-AAG TAT TGG ACA ACT TGT TAG TC-3' corresponding to nucleotides 661 to 683 of mouse CFTR cDNA, Accession number M69298) and P.CFTR-1360.3 (5'-TAA TTC CCC AAA TCC CTC CTC-3' 3' corresponding to nucleotides 1360 to 1340 of mouse CFTR cDNA, Accession number M69298), which amplify a 700 base pair fragment encompassing exon 5 (partial) through exon 9 (partial); and 60°C for the primer couple P.CFTR-3249.5 (5'-TGG AAT CTG AAG GCA GGA GTC-3' corresponding to nucleotides 3249 to 3269 of mouse CFTR cDNA, Accession number M69298) and P.CFTR-3428.3 (5'-TTC TCA TTT GGA ACC AGC GCA-3' corresponding to nucleotides 3428 to 3408 of mouse CFTR cDNA, Accession number M69298), which amplify a 180 base pair fragment excompassing exons 17a and 17b partially), extension at 72°C for 1 minute for fragments smaller than 1 kb for a total of 35-45 cycles, with a final extension step of 10 minutes to fully extend any remaining single stranded DNA. The first denaturation step was done for 6 minutes at 95°C.
Solutions and electrophysiology
For measurement of CFTR currents, we started the experiment by using a bath solution that contained (in mM): 150 NaCl, 6 KCl, 1 MgCl2, 1.5 CaCl2, 10 glucose, 10 HEPES, titrated with NaOH to pH 7.4. The Cl- equilibrium potential, ECl, is -36 mV. We then switched to a solution in which KCl had been substituted by CsCl. CFTR-channels were activated by a cocktail containing 100 μM IBMX (3-isobutyl-1-methylxanthine) and 10 μM forskolin (both from Sigma-Aldrich Chemie) dissolved in the bath solution. The pipette solution contained (in mM): 40 CsCl, 100 Cs-aspartate, 1 MgCl2, 0.1 EGTA, 4 Na2ATP, 10 HEPES, pH 7.2 with CsOH. Experiments were done at room temperature, ∼ 22°C.
Ca2+ activated Cl- currents were measured as described earlier [13, 14]. The bath solution contained (mM): 150 NMDG-chloride, 1 MgCl2, 1.5 CaCl2, 10 glucose, 50 mannitol, 50 nM charybdotoxin, 10 HEPES, titrated with NaOH to pH 7.4. Mannitol was used to suppress co-activation of volume-regulated anion channels (VRAC). Charybdotoxin (Sigma) was added to inhibit the big-conductance Ca2+ activated K+ channels, BKCa which is also present in MAEC cells [30]. The pipette solution contained (mM): 100 Cs-aspartate, 40 CsCl, 1 MgCl2, 4 Na2ATP, 1 μM Ca2+ buffered with 10 mM EGTA (CaBuf program, G. Droogmans, Leuven), 10 Hepes, pH 7.4 with CsOH.
Activation of VRAC has also been described in detail [16, 30]. In short, at the beginning of the patch-clamp recording, the Krebs solution was replaced by an isotonic Cs+ solution to suppress K+ currents, containing (in mM): 105 NaCl, 6 CsCl, 1 MgCl2, 1.5 CaCl2, 10 glucose, 90 mannitol, 10 HEPES, pH 7.4 with NaOH (320 ± 5 mOsm). Hypotonic solutions were obtained by omitting 90 mM mannitol from this solution (240 ± 5 mOsm). A pipette solution was used containing (in mM): 40 CsCl, 100 Cs-aspartate, 1 MgCl2, 1.93 CaCl2, 5 EGTA, 4 Na2ATP, 10 HEPES, pH 7.2 with CsOH (290 mOsm). This solution is slightly hypotonic compared to the Krebs' solution to avoid spontaneous activation of volume-sensitive Cl- currents. The concentration of free Ca2+ in this solution was buffered at 100 nM, which is below the threshold for activation of Ca2+ activated K+ and Cl- currents and also prevents activation of non-selective cation channels.
MAEC cells were patch clamped as described previously [8, 30]. Currents were monitored with an EPC-7 patch clamp amplifier (List Electronic, Germany). Patch electrodes had a resistance between 3 and 5 MΩ. An Ag-AgCl wire was used as reference electrode. VRAC and CaCC whole-cell membrane currents were measured in ruptured patches, sampled at 2 ms intervals (1024 points per record, filtered at 200 Hz), unless otherwise mentioned. All experiments were performed at room temperature (20-23°C).
The I-V curve of CFTR currents was derived from 400 ms linear voltage ramps from -100 mV to +100 mV or from 400 ms voltage steps to potentials ranging from +80 to -80 mV (in decrements of 40 mV) applied from a holding potential of 0 mV [31, 32]. Currents were sampled at 1 ms intervals (512 points) and filtered at 5 kHz.
Data analysis
Electrophysiological data were analyzed using the WinASCD software (G. Droogmans, Leuven). Pooled data are given as mean ± S.E.M. from n cells. Significance was tested using Student's paired t test (P < 0.05 are marked with an asterisk).
Acknowledgements
This work was supported by the Belgian Federal Government, the Flemish Government and the Onderzoeksraad KU Leuven (GOA 99/07, F.W.O. G.0237.95, F.W.O. G.0214.99, F.W.O. G. 0136.00; Interuniversity Poles of Attraction Program, Prime Ministers Office IUAP Nr.3P4/23), by "Levenslijn" (7.0021.99), a grant from the "Alphonse and Jean Forton Fonds - Koning Boudewijn Stichting" R7115 B0.
It has been proposed that GL15, a human cell line derived from glioblastoma multiforme, is a possible astroglial-like cell model, based on the presence of cytoplasmic glial fibrillary acidic protein.
Results
The aim of this work was to delineate the functional characteristics of GL15 cells using various experimental approaches, including the study of morphology, mechanism of induction of intracellular Ca2+ increase by different physiological agonists, and the presence and permeability of the gap-junction system during cell differentiation.
Immunostaining experiments showed the presence and localization of specific glial markers, such as glial fibrillary acidic protein and S100B, and the lack of the neuronal marker S100A. Notably, all the Ca2+ pathways present in astrocytes were detected in GL15 cells. In particular, oscillations in intracellular Ca2+ levels were recorded either spontaneously, or in the presence of ATP or glutamate (but not KCl).
Immunolabelling assays and confocal microscopy, substantiated by Western blot analyses, revealed the presence of connexin43, a subunit of astrocyte gap-junction channels. The protein is organised in characteristic spots on the plasma membrane at cell-cell contact regions, and its presence and distribution depends on the differentiative status of the cell. Finally, a microinjection/dye-transfer assay, employed to determine gap-junction functionality, clearly demonstrated that the cells were functionally coupled, albeit to varying degrees, in differentiated and undifferentiated phenotypes.
Conclusions
In conclusion, results from this study support the use of the GL15 cell line as a suitable in vitro astrocyte model, which provides a valuable guide for studying glial physiological features at various differentiation phases.
Background
Astrocytes are the most abundant cell type of the central nervous system, where they are closely involved in the modulation of the activity of neuronal components. Astrocytes play a pivotal role in several physio-pathological brain events that involve the synthesis and secretion of neurotrophic growth factors [1]. In addition, it has been shown that neurotrophin-mediated signalling may not be the only mechanism involved in astrocyte-neuron interactions. In fact, the presence of specific intercellular connections (gap junctions) between these two cell populations, which allow direct and selective cell-to-cell exchange of chemical signals (ions, small metabolites), may represent an additional, rapid and unique way for astrocytes to communicate with each other and to interact with adjacent neurons [2].
In mammalian astrocytes, extracellular physiological agonists are able to increase the concentration of intracellular Ca2+ ([Ca2+]i) via voltage-dependent channels or controlled release from internal stores (via inositol triphosphate receptors and/or ryanodine receptors) [3, 4, 5]. This is one of the most utilised mechanisms for modulating astrocyte functions. However, Ca2+ waves, which are transmitted from cell to cell via gap junctions (gjs), are thought to be important for co-ordination of astroglial function [6, 7, 8]. The genesis and propagation of Ca2+ waves were originally observed in brain-derived cell populations in culture and, more recently, this event has also been demonstrated in more integrated systems, such as brain slice preparations [4, 9] and living rat brain [10]. In spite of the large number of contributions published in the last decade, the mechanism(s) involved in the genesis and propagation of Ca2+ waves are not yet clear [11]. Moreover, there is insufficient data from in vivo experiments, especially those on human astrocytes.
About ten years ago, the GL15 cell line was established from human glioblastoma multiforme [12]. GL15 cells were characterised as an astroglial-like cell line by the study of the cell karyotype and immunohistochemical and cytogenetic demonstration of glial fibrillary acidic protein (GFAP) expression [12]. Moreover, other biochemical properties peculiar to astroglials were found in the GL15 cellular population that confirmed their astroglial origin; for example, expression of glutamine synthetase, taurine transport, transforming growth factor receptor expression and interleukin-induced apoptosis [13, 14, 15, 16]. Although the data derived from the previous studies support the presence of an astroglial phenotype, as yet no determination has been made concerning the main physiological characteristics of the GL15 cells in relation to their differentiation.
Therefore, we decided to focus our attention on one of the most important aspects of astrocyte physiology: the mechanism(s) of cell communication. Considering that in vivo astrocytes are capable of cellular communication both via membrane surface receptor-operated systems and/or gjs between two neighbouring cells, the investigation of the presence and activity of these mechanisms is fundamental in proposing GL15 cells to be an in vitro model of astrocytes.
For these reasons, we define the characteristics of this model by analysing some morphological aspects, the mechanism of [Ca2+]i increase induced by different extracellular physiological agonists and the expression and functional capacity of the gjs system in relation to the differentiative pathway.
ResultsMorphological analysis
Undifferentiated GL15 cells have a heterogeneous morphology. Figure 1 shows confocal microscopy images of the different typologies of these cells. Fluorescent markers were used to make evident the structural characteristics of these cells, namely, the microtubule array of the cytoskeleton (Paclitaxel-Bodipy, green fluorescence) (Fig. 1A), the nuclear structure and the mitochondria network (Propidium Iodide, red fluorescence, and Mitotracker-Green488, green fluorescence, respectively) (Fig. 1B), and the F-actin organization (Phalloidin-Alexa594, red fluorescence) (Fig. 1C). These data made it clear that cytoskeleton organization and lobate nuclei were characteristic of each cellular element, confirming the population heterogeneity.
Confocal images of different GL15 cell typology. Panel A: an overlay of single-section images of cells, stained with Paclitaxel-Bodipy (green fluorescence), shows tubulin organization in the cytoskeleton; panel B: an overlay of single-section images of cells, double-stained with Mitotracker-Green488 and Propidium Iodide, highlights mitochondrial network (green fluorescence) and lobate nuclei (red fluorescence); panel C: single median-section image of two cells, stained with Phalloidin-Alexa594 (red fluorescence), shows actin organization in the cytoskeleton. Bar = 25 μm.
Replacement of the 10% FCS in the culture medium with 2% HS, in order to induce cell differentiation, caused the GL15 cell growth rate to slow and apoptosis to be triggered. Therefore, the differentiation phase was completed within 9-12 days. The percentage of apoptotic and dividing nuclei (assayed by staining with the fluorescent nuclear probe DAPI) present during the differentiation period (0-12 days) are reported in Table 1. The percentage of mitotic cells fell sharply from 9-12% (0-3 days) to 1-2% after 12 days, whereas the presence of apoptotic nuclei increased during the first 3 days (from 5% to 16%) and regained its initial value at day 12 of differentiation.
GL15 nucleus morphology analysis during differentiation phases
Differentiation days
Nucleus typology
0
3
6
9
12
normal
86 ± 10.2
72 ± 8.3
91 ± 7.5
87 ± 9.4
94 ± 11.3
mitotic
9 ± 2.3
12 ± 2.1
2 ± 0.8
2 ± 1.5
1 ± 0.9
apoptotic
5 ± 1.5
16 ± 3.1
7 ± 2.6
8 ± 2.3
5 ± 1.9
Data derived from DAPI-stained nuclei (see Materials and Methods) observed using a fluorescence microscope. Each value represents the mean ± s.d. of the percentage of nuclei with different morphology: regular lobate nuclei (normal), two distinct nuclear lobes (mitotic) and enriched pluri-lobate nuclei (apoptotic). During selected differentiation times (0, 3, 6, 9 and 12 days), the percentages were calculated from values derived from five randomly selected fields each containing about 50-60 cells on slices, in duplicate for each sample.
The characteristics of the proliferation rate and apoptosis process in GL15 cells can also be correlated with the morphological analysis carried out over the same time interval. In particular, in 10% FCS-supplemented DMEM or after only 1-2 days in 2% HS-supplemented medium, the cells, characterised by a high proliferation rate (see the mitotic nuclei percentage in Table 1), showed an extreme morphological heterogeneity with elongated, fibroblast-like cells in a range of sizes together with larger, flat and more polygonal-shaped elements (Fig. 2A). In contrast, the 10-day confluent cells cultured in the presence of 2% HS showed apparent homogeneous morphology. Most of the cells were of small size with a regular, round shape (Fig. 2B); the extremely rare mitoses and numerous thin cell elongations, which form a sort of web within the culture, were strongly suggestive of resting or differentiated cells (Fig. 2B).
GL15 phenotypes. Panels A and B represent, respectively, phase-contrast images of sub-confluent undifferentiated cells, grown in D-MEM containing 10% FCS, and differentiated cells incubated for 10 days in D-MEM supplemented with 2% HS. Panels C and D show, respectively, confocal GFAP localization in undifferentiated and differentiated GL15 cells immunostained with Cy3-conjugated anti-GFAP antibody (red fluorescence). The images (single focal plane at intermediate cell section) show no detectable difference in GFAP distribution between the two phenotypes. Panels E and F visualize, in undifferentiated and differentiated GL15 cells respectively, single focal plane at intermediate cell section images showing S100 expression in cells immunostained with OregonGreen-conjugated anti-S100B antibody (green fluorescence). The fluorescence signal indicates that S100B is localized in the perinuclear area. This finding is more evident in the differentiated phenotype. The same samples were also stained with TexasRed-conjugated anti-S100A, but no red fluorescent emission was detected. Bar = 50 μm.
To better define the glial features of GL15 cells, we also tested for the presence of specific astrocytic markers (GFAP and S100B). Immunolabelling for GFAP and the use of confocal microscopy highlighted, in both undifferentiated and differentiated phenotypes, a diffuse expression of the protein in the cytoplasmic compartment (Figs 2C and 2D). Under the same conditions, we also tested the expression of S100A and S100B, two Ca2+-binding proteins characteristically distributed in the central nervous system, but with different modalities: S100A is predominantly expressed in neurons, whereas S100B is normally expressed in and secreted by astroglial cells [17]. Both undifferentiated and differentiated GL15 cells were S100B-positive (Figs 2E and 2F, respectively). A marked distribution of S100B was found in the perinuclear area, above all in the differentiated phenotype (Fig. 2F), and the undifferentiated phenotype also showed a diffuse cytoplasmic localization (Fig. 2E). In contrast, S100A was not detected (data not shown). This result is particularly significant, because it indicates a specific astroglial feature of the GL15 line.
[Ca2+]i levels
We also studied the membrane activating systems of GL15 cells by analysing, in single cells, the [Ca2+]i variations triggered by extracellularly applied stimuli for which the concentration and effect on astroglial cells are well known. Figures 3, 4, 5 and 6 show the temporal analysis of single cell [Ca2+]i variation expressed as normalized fluorescence values (see Materials and Methods).
KCl-evoked Ca2+ spikes in GL15 cells. The graph shows the trace of intracellular Ca2+ variations in a single cell responding to extracellular 50 mM KCl pulses (arrows). The solid trace represents the [Ca2+]i variations found in 100% (n = 60) of the differentiated tested cells. The trace in the box represents the [Ca2+]i variations found in 70% (n = 47) of the undifferentiated cells. Time (s=seconds) is indicated on the abscissa; the ordinate gives the normalized fluorescence value (f/f0).
Temporal analysis of [Ca2+]i variations induced by L-glutamic acid in GL15 phenotypes. Of the undifferentiated cells (n = 60), 83% were responsive to addition of 300 μM L-glutamic acid (L-glu) (arrow) with the same kinetics although with different amplitudes (A). In panel B the traces represent the three equally probable [Ca2+]i kinetics (fast increase and fast decay; fast increase and slow decay; Ca2+ waves) recorded in 100% (n = 70) of differentiated cells tested. Time (s=seconds) is indicated on the abscissa; the ordinate gives the normalized fluorescence value (f/f0).
ATP-induced [Ca2+]i variations in differentiated GL15 cells. Panel A shows a single Ca2+ spike induced by 200 μM ATP (arrow) in 32% of differentiated cells (n = 69); addition of 2 μM thapsigargin (Tg) (arrow) evoked a further increase in [Ca2+]i. In 18% (12 out of 69) of differentiated cells, 200 μM ATP (arrow) induced Ca2+ oscillations, partially inhibiting the thapsigargin-induced Ca2+ increase (B). Time (s=seconds) is indicated on the abscissa; the ordinate gives the normalized fluorescence value (f/f0).
Spontaneous [Ca2+]i variations in both undifferentiated and differentiated GL15 cells. Panels A, B and C show, respectively, a single [Ca2+]i spike, low frequency [Ca2+]i oscillations and high frequency [Ca2+]i waves. 50% (n>200) of the cell population (differentiated or not) seemed to prime spontaneous intracellular Ca2+ movements. Time (s=seconds) is indicated on the abscissa; the ordinate gives the normalized fluorescence value (f/f0).
The main graph in Fig. 3 shows the repetitive Ca2+ variations induced by 50 mM KCl that were found in 100% of the differentiated cells (Δ of [Ca2+]i increase = 7 ± 2; n = 60). In the undifferentiated GL15 cells, only 70% of the cell population responded to KCl, with [Ca2+]i variations showing a kinetic shape similar to that observed for the differentiated ones, but with a lower amplitude (Δ of [Ca2+]i increase = 1 ± 0.5; n = 47; trace in the box). We also tested the cell response to glutamic acid (L-glu) and ATP which, like other physiological extracellular signals, induce [Ca2+]i variation in astroglial cells via membrane-receptor systems. Starting from a concentration of 300 μM, L-glu induced in 83% of undifferentiated cells a rapid increase in [Ca2+]i that regained the basal value within 4 minutes (Δ of [Ca2+]i increase from 1 to 3.5; n = 60) (Fig. 4A). In contrast, 100% of differentiated cells were responsive to 300 μM L-glu (n = 70), but with different kinetics and time-course. In fact, as shown in Fig. 4B, three different Ca2+ pathways, with the same probability percentage, were recorded in this phenotype: i) a single transient [Ca2+]i spike, ii) a [Ca2+]i variation with a shape similar to that observed in the undifferentiated phenotype, and iii) [Ca2+]i oscillations. Another physiological stimulus able to trigger an intracellular Ca2+ response in GL15 cells was the presence of 200 μM ATP in the medium. While this purinergic agonist did not induce [Ca2+]i variation in the undifferentiated population (0 responsive cells out of 53 tested cells; data not shown), in the 10-day low serum-treated cells 200 μM ATP caused single Ca2+ spikes or oscillations in 50% (n = 69) of the stimulated cells, with a ratio of 2:1 between the two possibilities (Figs 5A and 5B). Another important observation is that the different types of [Ca2+]i variation are derived from different involvements of intracellular Ca2+-stores. In fact, in 18% of the differentiated cells that showed Ca2+ oscillations, 200 μM ATP almost completely emptied the intracellular Ca2+-stores. This was confirmed by the further addition of thapsigargin, a well-known blocker of internal store Ca2+-pumps [18], which caused a slight increase in [Ca2+]i (Fig. 5B). In contrast, if the same thapsigargin concentration was added to the cell that responded to the presence of 200 μM ATP with a single spike, the alkaloid induced a large and sustained increase in Ca2+ (Fig. 5A).
Another distinctive aspect of this cell line was the presence of spontaneous Ca2+ variations, a phenomenon frequently observed in cells of astroglial origin. GL15 cells showed at least three different kinetics of Ca2+ oscillations: i) single transient variation (Fig. 6A) ii) low frequency oscillations (Fig. 6B) and iii) high frequency oscillations (Fig. 6C) (see also additional data: Movie 1 for the original data used to perform this analysis). Spontaneous Ca2+ variations were observed in both GL15 phenotypes. The phenomenon seemed to be independent of the presence of external Ca2+, because spontaneous [Ca2+]i variations were detectable both in the presence of 1.8 mM extracellular Ca2+ and in Ca2+-free external medium containing 5 mM EGTA (data not shown). To quantify this phenomenon, we tested more than 200 undifferentiated and differentiated cells; among these, 50% of each cell population seemed to be able to prime spontaneous intracellular Ca2+ movements. However, considering the variability of the genesis of spontaneous [Ca2+]i oscillations, this cell percentage could be even higher.
Gap-junction analysis
The complex process of astrocytic Ca2+ signalling involves not only the inner cellular pathway (information flow throughout the sub-cellular compartments) and the extracellular, receptor-mediated, chemical signal transduction (i.e. neurotransmitters - or stimuli triggered by growth factors), but also the gap junction intercellular communication (GJIC).
In order to assess the astrocytic properties that are shared by the GL15 cell line, it was important to investigate their capacity to establish functional GJIC. To achieve this aim, functional, immunocytochemical and molecular analyses were performed on GL15 phenotypes.
i) Functional analysis of GL15 junctional coupling. A microinjection/dye-transfer assay was employed to determine GJIC strength in differentiated and undifferentiated GL15 monolayers, analysed at different culture densities. Cells from undifferentiated, proliferating (sub-confluent) monolayers were clearly shown to be junctionally coupled. When the saturation density (confluence) of the cultures was reached, a significant reduction of the cell coupling was observed (48% of the value found in the proliferating counterpart). In the case of differentiated, resting GL15 cells (confluent cultures), dye-transfer was almost completely restricted to the cells initially loaded with the dye (Table 2). The GJIC capacity of GL15 cultures is shown in Fig. 7.
Pattern of dye coupling in GL15 cell cultures Sub-confluent monolayers of undifferentiated GL15 cells (A and B) show the presence of dye-permeant junctional channels. When the monolayers reached confluence (C and D), the dye-spreading capacity of the cells was reduced. Almost no dye spreading is observed in differentiated confluent GL15 cultures (E and F). Fluorescence (B, D and F) and the corresponding phase contrast (A, C and E) photographs are taken on formaldehyde-fixed cells, 15 min after dye injection. The star symbol indicates the microinjected cell. Bar = 50 μm.
Quantitative analysis of dye-coupling in GL15 cell cultures
n° of dye-coupled cells/
GL15 cell cultures
injection ± s.e.m.
% of communicating cells
(n°)
undifferentiated
8.13 ± 0.74
100
sub-confluent
(63)
undifferentiated
3.91 ± 0.42
48.09
confluent
(64)
differentiated
1.59 ± 0.25
19.55
confluent
(59)
The communication capacity of GL15 cells was quantified by counting the number of fluorescent cells surrounding the microinjected ones (n° of dye-coupled cells/injection) ± s.e.m.; n° = number of independent microinjections.
ii) Immunocytochemical analysis of connexin 43 (cx43) in GL15 cells. It is well known that cx43 is the main gap-junction protein expressed by astrocytes (both in vivo and in vitro); hence, its expression could have been responsible for the GJIC observed in GL15. The immunocytochemical localization of cx43 protein was performed by confocal microscopy on GL15 cultures kept in differentiating and proliferating conditions identical to those described above. The results showed the presence of the cx43 antigen in all the GL15 populations tested; however, the antigen distribution differed between populations, and its quantity was strictly related to the GJIC extent specific to each culture condition. In the highly communicating, sub-confluent, undifferentiated cells the punctate immunopositive reaction, typical of cx43 gjs aggregates, was localized close to the plasma membrane at cell-cell contact areas (Figs 8A and 8B; see also additional data: Movie 2 for the original data used to perform this analysis). The quantity gradually decreased when the cells reached the confluence status (Fig. 8C), and almost completely disappeared in the non-communicating, confluent differentiated GL15 cells (Fig. 8D). In all cases, randomly distributed cx43-positive staining was also found in the cytoplasmic compartment of GL15 cells. This could correspond to the unphosphorylated cx43 isoform that is usually defined as non-functional. Also in this case, an inverse relation was found between the GJIC capacity of the culture and the quantity of cx43 antigen located in the cytoplasm.
Expression of cx43 in GL15 cells. Confocal microscopy image acquisitions of cells stained with anti-cx43 antibody revealed by OregonGreen-conjugated anti-IgG. In panels A sub-confluent undifferentiated GL15 cells are also stained with Propidium Iodide. The image shown in A is an X-Y projection of a tri-dimensional reconstruction of 12 sections. Confluent undifferentiated GL15 cells (B) appear to poorly express cx43 in respect to sub-confluent undifferentiated cells (A). This feature was also observed in the confluent differentiated phenotype (C). Panels B and C represent a single focal plane at intermediate cell section. Bar = 25 μm.
iii) Immunoblot analysis of cx43 in GL15 cells. The results of the immunoblot analysis of cx43 expression in the various GL15 cell populations (shown in Fig. 9), confirmed the results reported above. The anti-cx43 antiserum recognised, in control samples (rat heart and IAR 203 cell line), both the functional phosphorylated (44 and 47 kDa) and the unphosphorylated (42 kDa) isoforms of the connexin, whereas in GL15 cells only a small quantity of the 42 kDa antigen was detected. This observation may explain the very low communication capacity of this cell line. The quantity of the functional phosphorylated cx43 isoform could have been, in this case, under the limit of immunoblot sensitivity. Only in the samples of GL15 cells was an upper signal sized around 50 kDa detectable. This signal is probably due to a non-specific reaction of the anti-rabbit secondary antiserum with a protein only expressed by human cells. The possibility that it was caused by a slowly migrating, altered (and non-functional) isoform of cx43, typical of this cell line, is very unlikely. If this were the case, a very high total quantity of the cx43 antigen/single GL15 cell would be expected (much more than in the case of the IAR 203 cells, used as a positive control because of their extremely high cx43 expression level). However, the data obtained by the immunolocalization of the protein showed the quantity of cx43 antigen/single cell to be very low.
Immunoblot analysis of cx43 protein expression in GL15 cell line 40 μg of total protein extracts from undifferentiated and differentiated GL15 monolayers are electrophoresed and blotted according to Material and Methods. Normal rat heart and IAR 203 rat liver epithelial cells are included as positive controls for cx43 expression. Molecular-mass markers are shown in kDa. All the GL15 cultures express cx43. Note that, in both the rat heart and the IAR 203 samples, all the non-phosphorylated (42 kDa) and the phosphorylated (44 and 47 kDa) isoforms of cx43 are detected by the anti-serum. The protein extracts used for each lane are, (A) rat heart, (B) undifferentiated sub-confluent GL15 cells, (C) undifferentiated confluent GL15 cells, (D) differentiated sub-confluent GL15 cells, (E) differentiated confluent GL15 cells, and (F) IAR 203 cells. Results are representative of three immunoblots.
Discussion
The name glia - derived from the Greek word for glue - per se indicates why this cell compartment was considered the "less interesting element" of the nervous system until only a few years ago. Today the conception that researchers have of this cell population has completely changed, and, in general, theories concerning the role of astrocytes have been radically modified [19]. A recent, novel hypothesis put forward the possibility that this cell type is directly involved in the activity of brain tissue, and not just as a growth-factor producer. This hypothesis is strongly supported by the observation that astrocytes display rapid electrical responses to neuronal activity and/or modifications of transductive systems [2, 4], like those of extracellular purines, which are known to be active in primary cultures of astrocytes [3].
Our experiments demonstrate that GL15 cells possess characteristics typical of astrocytes. The glial nature of this cell line was demonstrated by a number of experimental approaches, which collectively show that they have several astrocyte-like physiological properties. Immunostaining with anti-astrocyte protein-specific antibodies against GFAP and S100 showed the presence of specific glial markers GFAP and S100B, and the lack of the neuronal marker S100A, a particular isoform of S100 [17]. S100B and GFAP proteins were present in GL15 cultures, and the distribution pattern of S100B was relative to the differentiation status of the examined cells.
The presence of voltage-gated channels, correlated directly or indirectly with the possible generation of different transmembrane potentials, was indicated by data obtained from incubation of GL15 mature phenotype with 50 mM KCl, which induces membrane depolarisation. Under these conditions, a transient increase of [Ca2+]i was observed, similar to that obtained in other excitable cells such as neurons, muscle cells or mature astrocytes [20]. By employing [Ca2+]i variation as an indicator for the presence of Ca2+-related receptors in these cells, it was further possible to demonstrate that the mature phenotype of GL15 cells is associated with different agonist-receptor systems. The presence of appropriate glutamate or ATP concentrations in the experimental medium induced significant [Ca2+]i variation with specifc and agonist-related characteristics. However, upon testing with undifferentiated GL15 cells, these effects were either not present (i.e. ATP-induced [Ca2+]i variation), or less marked (i.e. KCl-induced), or evident in a different way (i.e. L-glu-induced).
Our decision to examine [Ca2+]i as probe for definitive astrocytic features of GL15 cells was based on the fact that the specific role played by astrocytes in many brain functions is achieved by Ca2+ signalling mechanism(s). Astrocytes express many different pathways, through which they react to external stimuli by variation of [Ca2+]i. In fact, these cells contain different forms of inositol triphosphate-coupled receptors that increase Ca2+ signalling, for example via the glutamate pathway and purine-activated systems. Ionotropic receptors, which open Ca2+ channels, are also found in this cell type. Furthermore, the presence of voltage-activated Ca2+ channels that permit Ca2+ fluxes from outside was demonstrated in both primary cultures and brain slices [4, 11].
Therefore, it is well elucidated in the literature that the mechanisms involved in intracellular Ca2+ regulation (release from and uptake to Ca2+-stores, capacitative and inductive Ca2+ currents, Ca2+ oscillations, and Ca2+ waves, etc) are present in astrocytes. If one considers that in these cells changes in [Ca2+]i underlie a reciprocal communication system between neurons and astrocytes, assessing the presence and definition of the Ca2+-system signalling in the GL15 cell line is fundamental in examining the reliability of this cell line as an astrocyte model.
All the Ca2+ pathways previously described for astrocytes are present in GL15 cells, although they are differentially regulated according to the differentiation status of the cell. It is possible to record oscillations in intracellular Ca2+ levels, generated either spontaneously or in the presence of ATP or Glutamate. On the other hand, depolarising agents like 50 mM KCl are unable to induce this phenomenon, which may therefore be considered independent from the voltage-operated Ca2+ channel status. This interpretation is credible, especially considering oscillating responses were also observed when extracellular Ca2+ was absent in the experimental medium. We hypothesise from the data that in the GL15 population, Ca2+ oscillations originate during Ca2+-release from internal stores. The presence of a complete fan of receptor-operated responses in GL15 cells is correlated with the mature phenotype, whereas intercellular communication of astrocytes is associated with the immature status of GL15 cells. In vivo, astrocytes have the possibility to form 'functional syncytia' by establishing cytoplasmic connections through specific intercellular channels (gap junctions), which in turn provide pathways for the direct exchange of ions, small metabolites and water. Cultured astrocytes, like their in vivo counter-parts, are extensively coupled by gjs (cx43 being the predominant junctional protein expressed by these cells) [21].
Cx43 is expressed in specific brain regions by different glial populations like the other connexins (e.g. cx30), depending on the developmental stage of the tissue, different physio-pathological conditions and/or growth-factor influence [22, 23, 24].
Our results demonstrate that GL15 cells express the cx43 protein and form junctional channels in which permeability is directly related to the proliferation rate, and decreases when the differentiative status is reached.
The reduction of cell coupling when confluence is achieved supports the proposed role of GJIC in regulating astrocyte migration. In the differentiated phenotype, the very low levels of GJIC compared to that found in non-proliferating, undifferentiated confluent monolayers, demonstrated that even in a situation of 'mitotic arrest', factors other than those linked with cell-cycle progression (although probably dependent on the developmental features of the cells) may be involved in the regulation of GL15 cell coupling. These data confirm that physiologically, GL15 behave like glial cells with respect to both cell-cycle-dependent and differentiation-related regulation, and astrocytic junctional coupling (cx43 expression) [25]. A further point to note is that in the in vitro conditions cx43 expression is limited to type I astrocytes [26].
It was observed that even when subjected to culture conditions allowing maximal junctional communication (sub-confluent, undifferentiated monolayer), the extent of GL15 dye-coupling was low, especially when compared to results obtained from other cultured cells of astrocytic origin [27]. This functional evidence closely reflects the data from immunocytochemical and immunoblot analyses, which indicate that in GL15 cells the junctional protein cx43 is expressed at low levels and its localization is mainly cytoplasmic, probably in the nonfunctional, unphosphorylated isoform. The presence of low levels of cx43 and limited intercellular coupling of GL15 cells is likely to be due to the neoplastic origin of this cell line, which is also observed in many other tumour-derived cell types such as C6 glioma cells [28]. Another possible explanation for the low communication capacity of GL15 cells might be the culture conditions. It has been demonstrated that the efficiency of junctional coupling of cultured astrocytes is positively influenced by interactions with other cell types (e.g. neurons and meningeal cells) [29, 30]. Moreover, in astrocytes (both in situ and in vitro), the different levels of dye-coupling (and cx43 expression) depend on factors such as the specific regional origin of these cells in the CNS or degree of maturation, thus suggesting that gjs formation in this cell type undergoes very complex environmental regulation [24].
Another important feature observed in our results was the correspondence between the modality of Ca2+ wave propagation in GL15 cells, and their GJIC capacity. Although several factors influence the extent of Ca2+ wave propagation, gjs permeability is a major determinant [11]. Heterogeneity in GJIC extent may, therefore, partly explain the heterogeneity of Ca2+ wave modulation and propagation in the GL15 population.
Conclusions
In conclusion, the data reported in this paper support the reliability of the GL15 cell line as a suitable in vitro model for astrocytes, which should aid in the investigation of their distinctive physiological properties, and subsequently contribute to clarifying the complex role of this cell type in the brain. It is important to remember that, by simply utilising the differentiated or undifferentiated phenotype of this cell line, it is possible to study the modality by which the cells communicate with each other, either via gjs and/or membrane receptors. The proposed model becomes even more fascinating when the human origin of this cell line is considered. This new astrocyte model provides a stepping-stone in the efficient analysis and interpretation of problems regarding the role of astrocytes during modulation and remodelling of the nervous system, their contribution to the electro-physiological activity of neurons and other relevant mechanisms.
Materials and MethodsCell culture
GL15 cells were routinely cultured in growth medium (GM): D-MEM (Dulbecco's Modified Eagle's Medium) supplemented with 10% fetal calf serum (FCS), 100 IU/ml penicillin-100 μg/ml streptomycin and 2 mM L-glutamine. The cells were maintained at 37°C in a 5% CO2 humidified atmosphere. The medium was changed twice weekly and the confluent cell monolayers were regularly sub-cultured, washing with phosphate-buffered solution (PBS) and treating with 0.5% trypsin-0.2% EDTA at 37°C for 5 minutes. Long-term differentiating cultures were obtained after a 10-day incubation of sub-confluent GL15 cultures in a differentiation medium (DM): D-MEM supplemented with 2% horse serum (HS), 100 IU/ml penicillin-100 μg/ml streptomycin and 2 mM L-glutamine. Undifferentiated cells were tested after a 1-2 day incubation in GM, while the differentiated phenotype was observed after culturing for 10 days in DM. All the experimental procedures were performed on cells of 60°-70° passage. All media, sera, antibiotics and culture solutions were purchased from Life Technologies Italia srl (S. Giuliano Milanese, Italy). Sterile culture plastics were purchased from Falcon (Plymouth, UK). All other reagents were of analytical grade purity.
Nuclear morphology
The cells were plated on 12 mm-glass coverslips (BDH Italia, Milano, Italy) at a density of approximately 2 × 104 cells/well and incubated in DM for 0, 3, 6, 9 and 12 days. At the selected times the cells were rapidly washed in PBS and fixed by 3.7% paraformaldehyde (Sigma, St. Louis, Missouri, USA) at room temperature (r.t.) for 15 minutes. The cells were then permeabilized with absolute methanol (Sigma) at r.t. for 5 minutes. After a single PBS washing, the samples were stained with 1 μg/ml 4',6-diamidino-2-phenylindole (DAPI; Molecular Probes, Eugene, Oregon USA) at r.t. for 5 minutes. After a double PBS washing, the coverslips were dried and mounted using a commercial mounting medium (Molecular Probes) and the DAPI fluorescence observed using an inverted Olympus IX 50 microscope (Olympus, Hamburg, Germany) equipped with an oil objective lens (Uapo/340 40X/1.35 oil Iris, Olympus). Regularly lobate nuclei were considered to belong to viable cells in cycle (indicated as normal); nuclei presenting two distinct lobes were considered to be in mitosis (indicated as mitotic), while enriched pluri-lobate nuclei were considered to be apoptotic. The quantification of each nuclear typology observed was represented by the percentage mean ± s.d.. Each nucleus typology percentage was calculated from values derived from five randomly selected fields each containing about 50-60 cells on slices, in duplicate for each sample.
Fluorescence labelling and confocal microscopy
GL15 cells were grown on 12 mm-glass coverslips (plating density: 5 × 104 cells/well) as undifferentiated or differentiated cells. At the selected times the cells were rapidly washed with PBS and then fixed with 3.7% paraformaldehyde (Sigma) at r.t. for 5 minutes. The cells were then permeabilized with 0.1% Triton X-100 (Sigma) at r.t. for 5 minutes. These samples were used for fluorescent dye- or immuno-staining protocols. The dyes: Paclitaxel-Bodipy, Mitotracker-Green488, Propidium Iodide, Phalloidin-Alexa594 (all obtained from Molecular Probes) were used according to manufacturer's instructions. The immunolabelling procedure was preceded by 1 hour incubation of the cultures in 10% bovine serum albumin (BSA) at r.t.. Direct immunostaining detected the presence of GFAP, S100A and S100B proteins after 1 hour incubation (at 37°C) with the following antibodies, respectively: Cy3-conjugated mouse monoclonal anti-GFAP IgGs, Texas Red-conjugated mouse monoclonal anti-S100A IgGs, and Oregon Green-conjugated mouse monoclonal anti-S100B IgGs (all of these antibodies were obtained from Sigma and used at 1:100 dilution). Anti-S100A and anti-S100B were conjugated with their respective fluorochrome using FluoReporter Protein Labelling Kit (Molecular Probes). The presence of cx43 antigen was revealed by immunofluorescence: after 1 hour incubation at 37°C of the cells with primary mouse monoclonal anti-cx43 antibody (diluted 1:60) (Chemicon International Inc., Temecula, CA), followed by 1 hour incubation at 37°C with secondary Oregon Green-conjugated anti-mouse IgGs (diluted 1:100; Molecular Probes). The cells were washed 3 times for 5 minutes at r.t. with 0.1% Tween 20 (Sigma) in PBS, dried and then observed.
Fluorescence images were obtained by using a Bio-Rad MRC-1000 confocal system (BioRad Laboratories, USA) with an Axiovert 100 microscope equipped with a 63X/1.25 PLAN NEOFLUAR oil immersion objective lens (Zeiss, Jena, Germany). The Kr/Ar laser potency, photomultiplier and pin-hole size were kept constant for the entire experimental procedure. Images were acquired using CoMOS/MS-DOS software and then processed using LaserSharp/OS2 software (BioRad).
[Ca2+]i measurements
GL15 cells were plated on 25 mm-glass coverslips at a density of 9 × 104 cells/well and tested as undifferentiated or differentiated cells. At the beginning of each experiment, the cells were washed with the normal external solution (NES), a buffered solution containing (in mM) 140 NaCl, 1.8 CaCl2, 2.8 KCl, 2 MgCl2, 10 Glucose and 10 HEPES/NaOH, at pH 7.4. The cells were then incubated for 30 minutes at 25°C with 3 μM Fluo3 acetoximethylester (Fluo3/AM, Molecular Probes) dissolved in NES supplemented with 10 mg/ml BSA. The loaded cells were rinsed and maintained for an additional 15 minutes at 25°C in NES to allow the complete de-esterification of the dye. In these experimental conditions each cell sample showed good preservation of the intracellular Ca2+ dye fluorescence emission. The coverslips were then transferred into an Attofluor chamber (Molecular Probes). Stimulating agents were added in less than 1 second to the cells kept at r.t.. A high-speed wavelength switcher Polychrome II (Till Photonics, Germany) equipped with a 75 W stabilised Xenon lamp (Ushio Inc., Japan) provided the excitation beam. The Polychrome II was connected to an Olympus IX 50 microscope equipped with an oil objective lens (Uapo/340 40X/1.35 oil Iris). The fluorescence emission was acquired by C6790 Hamamatsu camera and analysed using the Argus Hisca 1.7 software (Hamamatsu, Hamamatsu, Japan). The traces in Figs 3,4,5,6 were ratios (1 ratio/second) calculated off-line as f/f0, where f is the fluorescence emission of a single FLUO3-loaded cell at time range from 1 to x seconds and f0 is the fluorescence emission of the same cell at time 0 [31].
Dye-Transfer Assay
Undifferentiated (sub-confluent and confluent) and differentiated (confluent) GL15 cells, cultured onto 60-mm Petri dishes, were tested for their capacity to establish functional gjs as described by Mazzoleni et al. [32]. Briefly, single cells within the monolayers of three separate dishes were microinjected with a 10% (wt/vol) solution of the gap-junction-permeant fluorescent tracer Lucifer Yellow CH (Sigma) in 0.33 M LiCl. Microinjections were performed using glass capillary needles (Clark Electromedical Instruments, Edenbridge, UK), prepared with an automatic puller (Narishige, Tokyo, Japan) and driven by a Narishige micromanipulator (SYF II) linked to an Olympus IMT2 microscope. The fluorescent dye was injected under nitrogen pressure using an Eppendorf microinjector (Hamburg, Germany). Five minutes after the last injection, the cells were fixed with 4% paraformaldehyde in PBS and the dye-transfer capacity of the cells was observed using the IMT2 epifluorescence system. The extent of gap junction intercellular communication (GJIC) was then quantified by counting the number of fluorescent cells surrounding the microinjected ones (n° of dye-coupled cells/injection). At least 25 independent microinjection trials/dish were taken into account for the precise quantification of the GJIC competence of the culture. Data were expressed as mean ± s.e.m.. Bright-field and fluorescence images were taken on Kodak T-MAX 400 (400 ASA) films.
Immunoblot analysis
Cells from six 100-mm Petri dishes were pooled for each previously described culture condition and the whole-cell extracts (40 μg protein of total lysate/lane) were first resolved by electrophoresis on 10% sodium dodecyl sulphate-polyacrylamide gel [33] and then transferred onto a nitrocellulose membrane (Schlesher and Shuell, Keene, NH). Protein of total lysate from rat heart and IAR 203 rat liver epithelial cells [34] were included as positive controls for cx43 expression. Total protein concentration was determined using a Bio-Rad DC Protein Assay kit (Bio-Rad, Segrate, Italy); equal sample protein-loading was verified by Comassie Blue staining of identical gels run in parallel, using a 0.25% Comassie Blue solution (R 250/G 250 1:1) (Bio-Rad). Membranes were hybridized with polyclonal rabbit anti-cx43 antibody (1:1000) (Chemicon International Inc.), followed by reaction with peroxidase-conjugated anti-rabbit IgGs (1:2000) (Amersham Pharmacia Biotech Italia, Cologno Monzese, Italy). Immunopositive reaction was detected by the enhanced chemiluminescence method (ECL, Amersham Pharmacia Biotech Italia) and revealed using autoradiography films (Hyperfilm-ECL, Amersham Pharmacia Biotech Italia).
Supplementary Material
Structural and functional GL15 characteristics
To better visualise the structural and functional GL15 characteristics, we also include two movies in which it is possible to observe:
Movie 1: A representative experiment showing spontaneous intracellular Ca2+ oscillations. Each frame was acquired with a ratio of 1 frame/second. The total number of frames was 120 with a total real time of 120 seconds compressed in the 30 seconds of the movie. Size bar = 50 μm.
Movie 2: Tri-dimensional reconstruction of cx43-immunostained undifferentiated GL15 cells. Size bar = 25 μm.
Movie 1
Movie 2
Acknowledgements
We wish to thank Dr Francesca Rovetta for helpful discussions and Peter A. Mattei for his assistance in preparing the manuscript. This work was supported by research grants from MURST-Italy (Ministero dell'Università e della Ricerca Scientifica e Tecnologica) to G.F. and from the University of Brescia to G.M.
Although various endothelium-dependent relaxing factors (endothelial autacoids) are released upon the elevation of endothelial cytosolic free Ca2+ concentration (EC [Ca2+]i), the quantitative relationship between EC [Ca2+]i and vascular tone remains to be established. Moreover, whether the basal release of endothelial autacoids is modulated by basal EC [Ca2+]i is still unclear. We assessed these issues by using a novel method that allows simultaneous recording of EC [Ca2+]i and vascular displacement in dissected rat aortic segments.
Results
Receptor-dependent (acetylcholine) or independent (ionomycin) agonists caused immediate EC [Ca2+]i elevation followed by vasorelaxation in preparations pre-contracted with phenylephrine. Low doses of agonists induced small EC [Ca2+]i elevations (about 100 nmol/L) and concomitant half-maximal vasorelaxation. At high doses, agonists elevated EC [Ca2+]i to μmol/L range with little additional vasodilatation. When EC [Ca2+]i was plotted against the vasorelaxation, the curves were almost identical for both acetylcholine and ionomycin treatments, in the presence or absence of various endothelial autacoid inhibitors. Calcium-free solution reduced basal EC [Ca2+]i and induced a drastic vasoconstriction. Endothelial autacoid inhibitors reduced EC [Ca2+]i changes and abolished both agonist-induced vasodilatation and calcium-free solution-induced vessel contraction. When the EC [Ca2+]i was completely chelated by 40 μmol/L BAPTA, the acetylcholine-evoked vasorelaxation could be abolished as well. However, when the EC [Ca2+]i was partially chelated by 20 μmol/L BAPTA, the acetylcholine-evoked vasorelaxation was almost unaffected.
Conclusions
These results indicate that vascular tone is modulated by subtle changes of EC [Ca2+]i level, which seems to serve as an integrating signal in both basal and stimulated states.
Background
Vascular endothelium plays an important role in controlling vascular tone by secreting a variety of endothelium-derived relaxing factors (endothelial autacoids), namely NO, prostacyclin (PGI2), and L-NNA/indomethacin-insensitive relaxing factor [1, 2, 3]. In response to various chemical and physical stimuli, an elevation of endothelial cytosolic free Ca2+ concentration (EC [Ca2+]i) followed by the activation of calcium-dependent enzymes/channels and the consequent production of endothelial autacoids [4]. Although EC [Ca2+]i appears to mediate the release of endothelial autacoids, the direct relationship between EC [Ca2+]i and vascular contractility in intact vessels remains to be established.
Among all three calcium-dependent endothelial autacoids, NO is the primary regulator responsible for the modulation of vascular tone. The administration of NO synthase (NOS) inhibitors increases blood pressure [5], indicating an important role of NO on the regulation of basal tone. The basal EC [Ca2+]i level in intact rat aorta [6, 7, 8] or in culture (ref. [9] as an example) is less than 100 nmol/L. Studies using purified endothelial NOS or EC membrane fraction indicate that EC NOS may exhibit a basal enzyme activity under an extremely low calcium level [10, 11], suggesting that EC [Ca2+]i under basal conditions may be too low to regulate the minimal activity of EC NOS. Moreover, it has been reported that isometric contraction induces a calcium-independent activation of endothelial NOS [12]. Nevertheless, the EC [Ca2+]i level was not monitored in these studies. Therefore, the role of EC [Ca2+]i level under basal conditions remains ambiguous and still needs to be validated.
Here we reported the development of a novel method that allowed simultaneous measurement of both EC [Ca2+]i and vascular displacement in an opened vascular preparation. The goal of this study is to establish a firm relationship between rat aortic EC [Ca2+]i elevation and vascular contractility by using a receptor-mediated agonist (acetylcholine, ACh) and a receptor-independent agonist (ionomycin). Furthermore, similar approaches were carried out under the conditions that either the EC [Ca2+]i was reduced by exposing the specimen to a calcium-free solution, or its elevation was prevented by a calcium chelator pretreatment.
Results
In our system using rat aortic en face preparations, neither EC [Ca2+]i elevation nor vasodilatation was observed during or at the onset of flow. When exposed to phenylephrine (PE), the specimen contracted immediately and reached maximal within a few min without alteration of EC [Ca2+]i (Fig. 1c, initial contraction phase of Fig. 2). When various concentrations of ACh or ionomycin were added to the preparation, we observed EC [Ca2+]i elevations that were followed by vascular dilatation (Fig. 2). Both the elevation of EC [Ca2+]i and the subsequent vasodilation occurred in a dose-dependent manner. While vasorelaxation was very sensitive to low concentrations of agonists, it became saturated at high concentrations. As a comparison, the EC [Ca2+]i level progressively increases in response to increasing concentrations of either ACh or ionomycin.
Mounting of the aortic en face preparation for simultaneous measurements of EC [Ca2+]i and vascular displacement (a). Movement of EC images in response to PE (b). The arrows point to an arbitrarily selected endothelial calcium image that shifts upward during vascular contraction in response to cumulative concentrations of PE. Bar equals 100 μm. Effects of PE on the EC [Ca2+]i and contraction of aortic en face preparation (c). PE causes vasoconstriction (○) without alteration of EC [Ca2+]i (●) (n = 5)
Agonist-induced EC [Ca2+]i elevation and vasorelaxation. A typical example of endothelial [Ca2+]i and corresponding vascular response to cumulative concentrations of ACh (a) or ionomycin (b) in a PE pre-contracted aortic en face preparation. The dash and solid arrows indicate the time of application of PE and agonists, respectively. Both agonist-induced EC [Ca2+]i elevation (●) and corresponding vasorelaxation (○) increase in a dose-dependent manner.
Summarized results of EC [Ca2+]i and concomitant vessel dilatation in the presence of cumulatively increasing of agonists are shown at Fig 3. To examine the relative importance of various endothelial autacoids in our system, the preparations were incubated in the presence or absence of a combination of endothelial autacoid inhibitors. The relative contribution of a specific endothelial autacoid was estimated by performing experiments in the presence of the inhibitors of another two endothelial autacoids. The most potent agonist-induced endothelial autacoid in rat aortae appeared to be NO, which by itself was able to contribute agonist-induced vessel dilatation completely. As a comparison, the L-NNA/indomethacin-insensitive factor(s) alone resulted in 40% vasodilatation at most, whereas PGI2 played insignificant role in this system.
Agonist-evoked EC [Ca2+]i elevation and vascular relaxation. Various combinations of endothelial autacoid inhibitors (10-4 mol/L L-NNA, 3 × 10-5 mol/L barium sulfate and 10-3 mol/L ouabain, and 10-5 mol/L indomethacin) were included in either ACh (a) or ionomycin (b) experiments to block various endothelial autacoids. (●): control in the absence of inhibitors; (○): contribution of NO in the presence of barium sulfate/ouabain and indomethacin; (Δ): contribution of L-NNA/indomethacin-insensitive dilatation in the presence of L-NNA and indomethacin; (*): contribution of PGI2 in the presence of L-NNA and barium sulfate/ouabain.
When the ACh-evoked increase of EC [Ca2+]i was plotted against the extent of corresponding vascular relaxation, the curve was log-linear and it was almost identical to that from ionomycin experiments (Fig. 4). No apparent threshold value was observed in either agonist treatment, indicating a drastic increase in vasorelaxation occurred at a minimal increase of EC [Ca2+]i. The half-maximal relaxation induced by agonist application was accompanied with less than 100 nmol/L increase of EC [Ca2+]i. High doses of ionomycin evoked EC [Ca2+]i elevation to the micro-molar range without further dilating the vessel preparations. Moreover, the EC [Ca2+]i-vasodilatation curves from ACh (Fig. 5a) or ionomycin (Fig. 5b) treatments were similar to each other under any inhibitor combination of endothelial autacoids.
Relationship between EC [Ca2+]i elevation (abscissa) and vascular relaxation (ordinate). Experiments were carried out in the absence of endothelial autacoid inhibitors. Results were taken from the control experiments in Fig. 3. The lines in (a) and (b) were the fitting curves of ACh and ionomycin groups, respectively.
Relative contribution of various endothelial autacoids in agonist-evoked EC [Ca2+]i elevation (abscissa) and vascular relaxation (ordinate). Results were taken from inhibitor experiments in Fig. 3. (●): contribution of NO; (○): contribution of L-NNA/indomethacin- insensitive dilatation; (*): contribution of PGI2.
In order to examine the role of basal EC [Ca2+]i level in the regulation of vascular tone, PE-precontracted vascular preparations were subsequently exposed to the calcium-free solution followed by ACh (Fig. 6). As mentioned earlier, when the vascular preparation was exposed to 50 nmol/L PE, it contracted without concomitant endothelial [Ca2+]i elevation. When the preparation encountered the calcium-free solution, it contracted further with diminution of the basal EC [Ca2+]i level. The application of 10-7 mol/L ACh, which would almost fully dilate the vascular preparation under normal conditions, induced marginal EC [Ca2+]i elevation and vasorelaxation (Fig. 6a). Results from this part of study are summarized in Table 1. The PE-precontracted aortic en face preparations contracted further (almost 2-fold) by the exposure to a calcium-free solution, whereas the average basal EC [Ca2+]i decrease was about 35 nmol/L. The role of endothelial autacoids in these calcium-free solution effects was further examined by performing similar experiments in the presence of all endothelial autacoid inhibitors (Fig. 6b). Pretreatment of endothelial autacoid inhibitors prevented further vasoconstriction induced by the calcium-free solution without affecting EC [Ca2+]i depression (table 1, n = 5), indicating that the calcium-free solution-induced vasoconstriction was attributed to the reduction of endothelial autacoids. When endothelium-denuded preparations were used, the vascular tone did not increase by the exposure to calcium-free solution. In denuded preparations, many vessel segments gradually dilated when exposed to the calcium-free solution, possibly due to the delayed reduction of [Ca2+]i level in smooth muscle cells (data not shown).
Effects of phenylephrine and calcium-free Krebs-Ringer buffer on EC [Ca2+]i and corresponding vasoconstriction.
PE (5 × 10-8 mol/L)
Basal
Normal
Ca2+-free
EC [Ca2+]i (nmol/L)
- inhibitors
92.4+10.4
85.6+15.5
51.4+5.1*
+ inhibitors
80.0+15.2
83.8+9.9
53.5+10.3*
Contraction (%)
- inhibitors
0
100
211.0+29.9*
+ inhibitors
0
100
103.3+1.6*#
Results were taken from experiments similar to the one shown in Fig 4. Data are expressed as mean ± SE (n = 5). - inhibitors: without pretreatment of endothelial autacoid inhibitors + inhibitors: pretreatment with all endothelial autacoid inhibitors * P < 0.05 (Ca2+-free vs. normal buffer) # P < 0.05 (with inhibitors vs. no inhibitors)
Effects of PE and calcium-free Krebs-Ringer buffer on EC [Ca2+]i (●) and corresponding vasoconstriction (○). The aortic en face preparations were precontracted by 5 × 10-8 mol/L PE and subsequently exposed to calcium-free Krebs-Ringer buffer and 10-7 mol/L ACh (a). Experiments were carried out using intact preparation in the absence (a) or presence (b) of all endothelial autacoid inhibitors.
In order to investigate whether the EC [Ca2+]i elevation was necessary for agonist-induced vasorelaxation, the vessel preparation was pretreated with either 20 μmol/L or 40 μmol/L of dimethyl 1,2-bis (2-aminophenoxy) ethane-N, N, N', N'-tetraacetic acid-acetoxymethyl ester (BAPTA-AM), a calcium chelator, for 45 minutes before the exposure to either ACh or ionomycin. When pretreated with 20 μmol/L of BAPTA-AM, the dose-response curves of ACh-induced [Ca2+]i elevation were shifted to the right and the maximal value of EC [Ca2+]i elevation was less than 200 nmol/L (Fig. 7a). Nevertheless, maximal vasorelaxation could be reached. These results provide the evidence that only small EC [Ca2+]i elevations were needed to induce a large vasodilatation. Both ACh-evoked EC [Ca2+]i elevation and vasodilatation were completely blocked when the preparations were pretreated with 40 μmol/L of BAPTA-AM (Fig. 7a). In the absence of inhibitors, ionomycin at 10-8 mol/L was capable of inducing 60 nmol/L EC [Ca2+]i elevation and 50% vasorelaxation under normal conditions (Fig. 3b). However, at the same concentration it failed to evoke any response in the preparations pretreated with 20 μmol/L of BAPTA-AM (Fig. 7b). When exposed to 5 × 10-8 mol/L ionomycin, this preparation remained unresponsive for the first couple minutes then it responded by a gradual increase of EC [Ca2+]i with concomitant vasorelaxation to almost maximal extent. The preparation pretreated with 40 μmol/L of BAPTA-AM was rather unstable when ionomycin was added.
The dose-response curves of (a) ACh and (b) ionomycin on EC [Ca2+]i (●) elevation and corresponding vasodilatation (○) of PE pre-contracted aortic en face preparation pre-treated with 2 × 10-5 mol/L (solid lines) or 4 × 10-5 mol/L (dash lines) BAPTA-AM. (n = 5)
Discussion
We have successfully developed a technique which allows simultaneous measurements of both EC [Ca2+]i and vascular displacement in a tissue flow chamber system. Perhaps due to technical difficulties, studies on EC [Ca2+]i and simultaneous measurements of vascular contractility have not been very popular. Although a few studies have reported the measurement of EC [Ca2+]i level on intact vessels [13, 14, 15], results from these studies may be confounded by the fluorescent signal from smooth muscle cells or by the relatively poor signal-to-noise ratio due to light transmission through layers of smooth muscle cells. Consequently, the absolute values of EC [Ca2+]i in vascular tissues are not reported. As a comparison, the fluorescence signal from our optical system only comes from endothelial cells [7]. Moreover, since the fluorescence signal remained strong for hours, this made the characterization of both agonist-induced EC [Ca2+]i elevation and downstream vascular response possible.
Some previous studies have reported that flow induces endothelial NO synthase phosphorylation or Ca2+ entry, which in turn leads to the formation of NO and vasorelaxation in rat arterioles and rabbit aortae and coronary arteries [13, 14, 16, 17]. However, endothelial responses to flow may be variable in different vessels or species. Other studies have shown flow-induced vasoconstriction in rat femoral arteries and rabbit resistant arteries [18, 19]. In this study using isolated rat aortae, neither EC [Ca2+]i elevation nor vasodilatation was observed during or at the onset of flow. Thus, the constant flow applied throughout our experiments was not likely to induce a Ca2+-indpendent endothelial NO synthase activation.
We hypothesize that EC [Ca2+]i may play an integration role in the regulation of vascular tone. A schematic conclusion of present study (Fig. 8) is based on the following convergent observations. Regardless receptor dependent or not, the agonist-induced vasorelaxation was quantitatively correlated with EC [Ca2+]i level in rat aorta, suggesting a decisive role of EC [Ca2+]i signaling on vasorelaxation. Besides, when the EC [Ca2+]i elevation was completely suppressed by pretreatment of 40 μmol/L BAPTA-AM, the ACh-induced vasorelaxation was inhibited as well (Fig. 7a). This result indicated that ACh- induced vasorelaxation required an increase of EC [Ca2+]i level. When the rat aortic tissue was exposed to a calcium free solution, a drastic vasoconstriction was observed along with a small reduction of basal EC [Ca2+]i level, implying that the basal secretion of endothelial autacoids was also modulated by EC [Ca2+]i level. Consistent with our finding, the calcium-free solution also induced vasoconstriction in rabbit carotid artery [16]. Since it has been established that endothelial NO is constitutively released even in the absence of agonist [20], our results further suggested that EC basal [Ca2+]i level may play an important role in the regulation of basal release of NO.
Schematic representation of various EC [Ca2+]i levels and corresponding vessel tone. See text for detail.
Although there might be multiple explanations for the vasoconstriction induced by a calcium-free solution, it is most likely due to suppressed EC [Ca2+]i -dependent release of endothelial autacoids under unstimulated conditions. First, this vasoconstriction is accompanied with a concomitant EC [Ca2+]i reduction (Fig. 6). Second, it could be completely blocked by the presence of endothelial autacoids inhibitors (Fig. 6). Third, if the neurotransmitter release from nerve terminals was blocked by exposure to a calcium free solution, the vessel segments should dilate instead. Similarly, if the calcium-free solution affected the smooth muscle cells, the vessel segments should dilate as well. Vessel dilation actually happened when exposed to calcium-free solution for prolonged time periods (data not shown). Therefore, the basal release of endothelial autacoids appeared to be modulated by the basal EC [Ca2+]i level.
Like in many non-excitable cells, the store-operated capacitative calcium entry in ECs is the main mechanism responsible for the agonist-induced calcium entry [21, 22]. Since the ACh-induced vasorelaxation and EC [Ca2+]i elevation were relatively small under calcium-free conditions (Fig. 6a), calcium influx appeared to be the major source of EC [Ca2+]i elevation and played a major role in modulating the endothelium-dependent vasorelaxation. It is interesting to find that the agonist-induced EC [Ca2+]i elevation, most likely the capacitative calcium entry, was suppressed in the presence of autacoids inhibitors such as L-NNA, indomethacin, or ouabain and barium (Fig 3). Various factors, including NO, cytochrome P450, membrane potential, and protein tyrosine kinase inhibitors, have been reported to affect the capacitative calcium entry [22]. Our findings support the notion that the effect of endothelial autacoid inhibitors on vasorelaxation might be partially due to the modulation of capacitative calcium entry into endothelial cells.
It has been proposed that NO synthesis can be activated through calcium independent manner during mechanical stretch of cultured human ECs [23] or during isometric contraction of rabbit aortic rings [12]. Although a sustained increase of NO secretion has been observed after mechanical stimulation of cultured ECs [23], it may be actually mediated by small increases in EC [Ca2+]i level. Moreover, since L-NNA could potentially suppress EC [Ca2+]i level, the interpretation of L-NNA induced vasoconstriction as the calcium-independent activation of endothelial NOS [12] might not be entirely appropriate. In our opinion, the role of basal EC [Ca2+]i in regulating NO production could be overlooked in studies without directly monitoring the EC [Ca2+]i level.
We used L-NNA and indomethacin as inhibitors of NO, and PGI2, respectively. Although others and we have shown that the combination of L-NNA and indomethacin almost completely abolish the ACh-induced vasorelaxation [24, 25], a 40% of L-NNA/indomethacin-insensitive dilatation was observed in this study. Despite the uncertain identity of the 'L-NNA/indomethacin-insensitive factor' here, the ouabain/barium combination could completely block its dilation effect (Figs. 3 and, 5). Thus the fact that L-NNA/indomethacin-insensitive dilation contributed to a significant portion of agonist-induced vasorelaxation in the current setup cannot be explained by our choice of inhibitors used. We believe that this discrepancy is largely due to the higher sensitivity of vascular contractility measurement in our current open preparation than in the conventional vessel ring preparation. In the absence of any inhibitors, while ACh at 10-8 mol/L induced about 35% displacement of opened vessel segment (Fig. 3), only about 10% tension reduction was observed in vessel rings [24].
In conclusion, agonist-induced vasorelaxation depends on subtle changes of EC [Ca2+]i level and basal EC [Ca2+]i level modulates basal autacoids secretion. The EC [Ca2+]i varies from a basal level of less than 100 nmol/L up to 1000 nmol/L upon full stimulation by ACh or ionomycin and down to 50 nmol/L in calcium-free solution. Major changes of aortic contractility occurs when the EC [Ca2+]i level varies between 50 nmol/L and 200 nmol/L.
Materials and MethodsVessel Preparation and Fura-2 Loading
This study was conducted in conformity with the policies and procedures detailed in the "Guide for Animal Care and Use of Laboratory Animals". Male Wistar rats (6-8 weeks old) were anesthetized by ether inhalation, and their thoracic aortae were immediately obtained. Aortic rings (5 mm long) were immersed in an organ chamber containing Krebs-Ringer solution (118.0 NaCl, 4.8 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 24 NaHCO3, 0.03 Na2-EDTA, and 11.0 glucose in mmol/L, pH 7.4) and gassed with 95% O2-5% CO2. The rings were labeled for 1 hr at room temperature with 10 μmol/L fura-2 AM in the presence of 0.025% pluronic F-127 [8]. In some experiments, the labeling procedure was carried out in the presence of 20 or 40 μmol/L BAPTA-AM for 45 minutes. The excess chemicals were washed away with Krebs-Ringer solution.
Experimental Setup
The basic experimental setup for endothelial [Ca2+]i measurement was previously described [7], with some modification during vessel mounting in order to allow simultaneous measurement of vascular endothelial [Ca2+]i and vascular smooth muscle contraction. One side of the longitudinally opened vessel segment was fixed in the direction of blood flow with insect pins. The corners on the opposite side were passively stretched to an optimal tension (4 g) and pinned onto the base plate (Fig. 1a). Thus, this arrangement allows free vessel movement in response to the addition of vasoactive chemicals. After vessel mounting, the flow chamber was placed on the microscope stage and perfused with Krebs-Ringer buffer at 30°C under a constant flow rate of 700 μl/min (estimated shear stress 3 dyne/cm2) for an equilibration period of 1 hr. The agonist-induced EC [Ca2+]i elevation and vascular activity were determined in the presence of 50 nmol/L PE, followed by subsequent exposure to cumulative ACh (10-9 - 10-5mol/L) or ionomycin (10 -9 - 3 × 10-6 mol/L). The relative movement of endothelial cells was used as an index of vascular tone, whereas fluorescence ratio images from fura-2 labeled endothelial cells provided quantitative information about EC [Ca2+]i.
Determination of In Situ Endothelial [Ca2+]i and Vascular Tone
The calibration of endothelial [Ca2+]i in each preparation was described previously [7]. Briefly, at the end of each experiment, the EC [Ca2+]i was calibrated by applying ionomycin (5 μmol/L) in the presence of 5 mmol/L EGTA, followed by 10 mmol/L CaCl2. Finally, manganese (5 mmol/L) was added in order to quench cytosol fura-2 fluorescence. Endothelial [Ca2+]i was estimated after subtracting the background and autofluorescence according to the established formula [26]. Axon image workbench software (Axon Instruments, Foster City, CA, USA) was used for digital image acquisition and EC [Ca2+]i analysis.
To calculate the vascular displacement, MetaMorph software (Universal Imaging Corporation, PA, USA) was used to trace and analyze a series of images. The fluorescence image was focused sharply, and a particular endothelial cell in the initial image was selected as a location marker. In the presence of a specific concentration of PE, this image was forced to shift inward during experiments. After the marker had reached its final equilibrium location, the total displacement represented the maximal extent of vascular contraction induced by PE. When the preparation was subsequently exposed to either ACh or ionomycin, the marker moved back towards its initial location, indicating vasorelaxation. The extent of vasodilatation was calculated as the percentage of pre-contraction value.
To validate the experimental procedure for the measurement of vessel contractility, we calculated results acquired from different EC markers located within the same calcium image. Although the displacement value of each marker may vary from point to point, the relative movement at each point was identical when normalized by the maximal pre-contraction displacement at that point. Moreover, if we on purposely focused at the free-moving edge where some smooth muscle cells were exposed, the relative movements were identical when either ECs or smooth muscle cells were used as the marker. Therefore, there was no detectable relative movement between ECs and the smooth muscle cells underneath, indicating that the ECs in our preparation were firmly attached to basement membrane.
Basic Experimental Protocol
An initial series of experiments were performed on aortic en face preparations for the simultaneous measurement of endothelial [Ca2+]i and vasodilatation in the presence of PE. The endothelial calcium image shifted in a dose-dependent manner to the cumulatively increasing concentration of PE (Fig. 1b). The changes in vascular tone were expressed as a percentage of maximal displacement, which was achieved at a PE concentration of 10 μmol/L. Moreover, the endothelial [Ca2+] i was unaltered during the experimentation (Fig. 1c). The EC50 of PE was approximately 50 nmol/L; this concentration was thereafter used for the following experiments to test the vascular tone responses to different vasodilators. EC [Ca2+]i and vascular displacement were determined in PE pre-contracted preparation in the presence of cumulative supplement of ACh (a receptor-dependent agonist) or ionomycin (a receptor-independent agonist). In some experiments, various endothelial autacoid inhibitors were applied to evaluate the relative importance of NO, L-NNA/indomethacin-insensitive dilatation, and PGI2 in our system [27]. To investigate the role of basal EC [Ca2+]i level in regulating the vascular tone, PE-precontracted preparations were subsequently exposed to calcium-free buffer in the presence of PE.
Reagents
Reagents for preparing Krebs-Ringer solution were purchased from Merck (Darstadt, Germany). Other reagents were obtained from Sigma (St. Louis, MO, USA).
Statistical analysis
Results were expressed as mean + SEM. Sample sizes were indicated by n. Dose responses of agonist-induced [Ca2+]i elevation were analyzed by ANOVA with a repeated measures design. Differences between two groups were compared by using unpaired Student's t-test with P < 0.05 considered as statistically significant.
Acknowledgments
This work was supported by grants from the National Science Council and National Health Research Institute in Taiwan, R.O.C. (Grant numbers NSC89-2320-B006-034, 070, 125 and NHRI-GT-EX89S834L).
A family of aspartate-specific cysteinyl proteases, named caspases, mediates programmed cell death, apoptosis. In this function, caspases are important for physiological processes such as development and maintenance of organ homeostasis. Caspases are, however, also engaged in aging and disease development. The factors inducing age-related caspase activation are not known. Xanthurenic acid, a product of tryptophan degradation, is present in blood and urine, and accumulates in organs with aging.
Results
Here, we report triggering of apoptotic key events by xanthurenic acid in vascular smooth muscle and retinal pigment epithelium cells. Upon exposure of these cells to xanthurenic acid a degradation of ICAD/DFF45, poly(ADP-ribose) polymerase, and gelsolin was observed, giving a pattern of protein cleavage characteristic for caspase-3 activity. Active caspase-3, -8 and caspase-9 were detected by Western blot analysis and immunofluorescence. In the presence of xanthurenic acid the amino-terminal fragment of gelsolin bound to the cytoskeleton, but did not lead to the usually observed cytoskeleton breakdown. Xanthurenic acid also caused mitochondrial migration, cytochrome C release, and destruction of mitochondria and nuclei.
Conclusions
These results indicate that xanthurenic acid is a previously not recognized endogenous cell death factor. Its accumulation in cells may lead to accelerated caspase activation related to aging and disease development.
Background
Xanthurenic acid is formed upon tryptophan degradation by indoleamine-2,3 dioxygenase (IDO). The end products of this degradation pathway are, alternatively, nicotinate and xanthurenic acid. IDO activity is stimulated by superoxide radicals, liposaccharides and interferon-γ [1]. Kynurenine aminotransferase (KAT), the enzyme directly responsible for xanthurenic acid formation from 3-hydroxykynurenine, is found in the cytoplasm and mitochondria, and is highly expressed in the retina [2, 3]. Xanthurenic acid is present in blood and urine at concentrations of 0.7 and 5-10 μM, respectively [4, 5]. A several - fold increase is observed in vitamin B6 deficiency and some diseases such as tuberculosis [6, 7]. Xanthurenic acid's presence in the blood is linked to malaria development, and in the lenses to senile cataract formation [8, 9, 10]. Xanthurenic acid binds covalently to proteins, leads to their unfolding, and to cell death [11]. Here, we report that xanthurenic acid induces cell death associated with caspase-3, -8, and -9 activation, nuclear DNA cleavage, and cytochrome C release. However, cell death is not associated with cytoskeleton breakdown, usually observed because of actin depolymerization by caspase-3-cleaved gelsolin [12].
Results and DiscussionXanthurenic acid leads to caspase-3 activation in smooth muscle cells
Xanthurenic acid induced cell death in a concentration-dependent manner. Death was associated with caspase-3 activation (Fig. 1). Caspase-3 can be induced by two primary pathways: by activation of cell surface receptors (FAS, TRAIL, TNF), or by activation of a stress response pathway leading to cytochrome C release from mitochondria and caspase-9 activation [13, 14, 15]. Procaspase-3 (CPP32) is cleaved at ASP 175, leading to an autocatalytic process which liberates the active p17 fragment [16]. In this study, xanthurenic acid-dependent cleavage of CPP32 was observed by Western blotting (Fig. 1e), and formation of caspase-3 p17 was detected by immunostaining (Fig. 1f, Fig. 3a).
Caspase-3 activation and apoptosis upon xanthurenic acid (XAN) accumulation in vascular smooth muscle cells (VSMC). a, VSMC after one week of growth in MEM without XAN (left panel), and in the presence of XAN at a concentration of 5 μg/ml (right panel). The nuclear DNA fragmentation at 40 μg/ml, see Fig. 2b. b, Concentration of xanthurenic acid in the cell extract used for Western blot analysis. XAN was determined in the cell extracts as described previously [11]. c, Apoptotic cell death measured as a ratio of the cells with condensed or fragmented nuclei to total nuclei. d, Caspase-3 activity measured by caspase-3 substrate Ac-DEVD-pNa cleavage, e, Western blot of active caspase-3 in VSMC after exposure to XAN. Procaspase-3 (CPP32) was cleaved in the presence of XAN with the formation of pl7 and pl2. f, Immunofluorescence staining for caspase 3 pl7 in control cells (left panel) and in VSMC exposed to 40 μg/ml of XAN for one week (right panel).
Cleavage of caspase-3 substrates PARP, DFF45/ICAD, and gelsolin in VSMC exposed to xanthurenic acid (XAN) for one week. DFF45 and PARP were slightly upregulated in the presence of XAN. a, Western blot of DFF45/ICAD. b, Nuclear fragmentation in control cells (left panel) and VSMC grown in the presence of XAN at 40 μg/ml for one week (right panel). c, Western blot of PARP. d, Western blot of gelsolin.
Detection of apoptotic proteins by immunofluorescence in human retinal pigment epithelium cells after exposure to 20 μM xanthurenic acid for one week (magnification 200-fold). Panels a-c from left to right represent immunodetection of apoptotic protein with respective antibody, staining of nucleus with Hoechst, and merge of both. No staining for these enzymes was observed in control cells. a, Detection of active caspase-3 pl7. b, Detection of PARP p85. c, Detection of active caspase-9. Panel d, Detection of N-half of gelsolin bound to the cytoskeleton. From left to right: without xanthurenic acid, with 10 μM xanthurenic acid, with 20 μM xanthurenic acid. e, Detection of cleaved gelsolin containing polyphosphoinositide binding domain by antibody directed against GPIP1 in apoptotic cells. From left to right: first two panels, control cells; Hoechst staining, no staining with antibody directed against GPIP1. Next two panels: cells incubated for one week with 20 μM xanthurenic acid; Hoechst staining, staining with antibody against GPIP1.
Xanthurenic acid provokes degradation of caspase-3 substrates DFF-45, PARP, and gelsolin
Caspase 3 is required for the degradation of DFF45/ICAD with formation of the carboxy-terminal fragment p11, which is necessary for DNA cleavage [17, 18]. In cells exposed to xanthurenic acid, DFF45 was cleaved with generation of the p11 fragment, recognized with an antibody directed against full-length DFF45 (Fig. 2a). Processed DFF45 leads to internucleosomal cleavage, and indeed the DNA of cells exposed to xanthurenic acid was fragmented as shown by Hoechst 33342 staining and fluorescence microscopy (Fig. 2b). The amount of PARP protein was increased in xanthurenic acid-exposed cells, and PARP was degraded to the apoptotic p85 fragment (Fig. 2c, 3b), which was reported to be formed upon caspase-3 cleavage [19]. Also gelsolin was cleaved in cells in a xanthurenic acid concentration-dependent manner with formation of p41, the amino terminal part of gelsolin, called "N-half" (Fig. 2d. Fig. 3d,e). The latter is a product of caspase-3 activity [12, 20]. The p41 fragment was further degraded, indicating that xanthurenic acid activated additional enzyme(s) involved in gelsolin processing. It was reported that N-half leads to depolymerization of actin filaments [20].
Apoptosis induced by xanthurenic acid did not cause cytoskeleton depolymerization
In the presence of xanthurenic acid the caspase-3 cleaved gelsolin did not cause cytoskeleton breakdown (Fig. 3d). To the contrary, the elongated cytoskeleton was strongly stained for N-half of gelsolin, and the staining increased with xanthurenic acid concentration (Fig. 3d), the condition which led to activation of caspase-3 (Fig. 3a) and caspase-9 (Fig. 3c). N-half contains two polyphosphoinositide (PIP) binding domains. Phosphatidylinositol 4,5-bisphosphate and phosphatidylinositol 3,4-bisphosphate form a stable complex with gelsolin, which prevents caspase-3 and -9 activation [21]. We prepared a part of N-half containing amino acid residues 162-187 (human sequence of gelsolin), which contains a PIP binding domain in position 162-169, and raised a polyclonal antibody against the peptide, called GPIP1. This antibody stained the cytoskeleton of xanthurenic acid-exposed cells but not of the control cells (Fig. 3e). This indicates that after xanthurenic acid-dependent gelsolin cleavage the sequence containing GPIP1 binds to the cytoskeleton. The cleavage of gelsolin in the presence of xanthurenic acid did not lead to breakdown of the cytoskeleton (Fig. 3e), in contrast to experiments where gelsolin was overexpressed [12].
Xanthurenic acid induces mitochondrial damage
We observed activation of caspase-9 in cells after exposure to xanthurenic acid, indicating the triggering of a mitochondrial pathway (Fig. 3c) and release of cytochrome C in the cells exposed to xanthurenic acid (Fig. 4a,b,c). We used Mitotracker CMXRos to detect mitochondria and a possible loss of the mitochondrial membrane potential, previously seen when gelsolin had been cleaved [22]. Mitotracker CMXRos dramatically sensitized mitochondria to xanthurenic acid. Whereas in the experiments reported above cells were exposed for one week to xanthurenic acid, 3 hours of exposure to xanthurenic acid together with Mitotracker CMXRos sufficed to cause relocalisation of mitochondria from the nuclear region (Fig. 4d) to the cytoskeleton (Fig. 4e). At higher xanthurenic acid concentration mitochondria lost their membrane potential (Fig. 4f,g,h,i,j), released cytochrome C (Fig. 4f) and strong staining for N-half of gelsolin was observed (Fig. 4f). Under these conditions cells died within 24 hours.
Analysis of mitochondria in human pigment retinal epithelial cells exposed to xanthurenic acid. Panels a-c (magnification 200-fold), Cytochrome C release in the presence of 20 μM xanthurenic acid for one week. a, Staining for cytochrome C. b, Hoechst staining. c, Merge of both. Panels d-f (magnification 800-fold), Mitochondria were stained with Mitotracker CMXRos, and cells were stained with antibody against gelsolin (N-half). d, Control cells, e, Cells exposed to 5 μM xanthurenic acid (see Methods), f, Cells exposed to 20 μM xanthurenic acid. Lack of staining with Mitotracker CMXRos, co-staining for cytochrome C (orange), and N-half of gelsolin (green). Panels g-h (magnification 4000-fold), g, Single control mitochondrion stained for cytochrome C. h, Single mitochondrion stained with Mitotracker CMXRos. i, Same as in j, but staining with Mitotracker CMXRos. j, Staining for cytochrome C in cells exposed to 20 μM xanthurenic acid.
Effector caspases in demolition phase of xanthurenic acid-induced apoptosis
In the presence of xanthurenic acid the cell death was associated with DFF p11 formation, which is characteristic for caspase-3 activity. Recently, it was reported that executioner caspases-3, 6, and 7 play non-redundant roles during the demolition phase of apoptosis. Caspase-6 and caspase-7 are not involved in DNA degradation but in lamin A degradation, and caspase-7 activation provokes PARP cleavage [23]. Our Western blot analysis showed that in the presence of xanthurenic acid lamin A cleavage does not occur in VSMC suggesting that these caspases are not activated in the presence of this compound. This suggests that the cleavage of PARP occurred due to caspase-3 activation. Plectin, a cytolinker responsible for the mechanical stability of the cytoskeleton, is cleaved by caspase-8 at ASP 2395. The cytoskeleton lost intermediary fibers due to plectin cleavage, and cells lost their integrity. Full-length caspase-8 co-localises with mitochondria and active caspase-8 is translocated to plectin [24].
In the presence of xanthurenic acid caspase-8 was activated in VSMC (Fig. 5a, b). In the control cells caspase-8 was observed in the perinuclear region (Fig. 5a, left panel), and in the presence of the xanthurenic acid the antibody directed against caspase-8 stained the whole cytoskeleton (Fig. 5b, right panel). In the presence of xanthurenic acid plectin was translocated from the mitochondrial region to the cytoskeleton (Fig. 5c, left and right panel, respectively). Western blot analysis using an antibody specific for the amino terminus of plectin showed the formation of the fragment of about 130 kD characteristic for caspase-8 cleavage. However, the p130 fragment was further degraded, indicating that xanthurenic acid activated additional enzyme(s) involved in plectin processing (not shown). Caspase-8 activation in the presence of xanthurenic acid is an upstream event for the cell pathology in the presence of xanthurenic acid. Downstream caspase-3 can be directly activated by caspase-8 or via the mitochondrial pathway consisting of the translocation to mitochondria of truncated BID (t-BID) [25, 26]. In our study, t-BID was translocated in the presence of xanthurenic acid from the cytoplasm to mitochondria (Fig. 5d, left and right panel respectively) indicating that caspase-8 activation is an upstream event leading to the observed cytochrome C release, which in turn can activate caspase-9 by apoptosome formation [27]. Caspase-8 can be activated by death receptors, or by an amplification loop with caspase-9 [28, 29, 30]. The pathways of caspase-8 activation and cytochrome C release in the presence of xanthurenic acid are currently under study in our laboratory.
Analysis of caspase-8 and caspase-8 substrate-proteins, plectin and BID, in pig VSMC. a, Western blot analysis of caspase-8 in the cells after exposure to xanthurenic acid (XAN) for one week. Procaspase-8 was cleaved in the presence of XAN with the formation of pl8. Panels b-d, (magnification 800-fold) immunodetection of caspase-8 and plectin and BID in control cells (left panels), and after exposure to 40 μM of xanthurenic acid for 96 hours (right panels), b, detection of caspase-8. c, detection of plectin. d, detection of BID.
Conclusions
Our results indicate that an accumulation of the tryptophan metabolite, xanthurenic acid, leads to cleavage of caspases substrates and apoptosis. Unexpectedly, xanthurenic acid-induced apoptosis is associated with an abnormal function of cleaved gelsolin. The results indicate that xanthurenic acid is an important factor involved in aging and disease development.
Materials and MethodsReagents
We used the following polyclonal antibodies from Santa Cruz Biotechnology Inc. CA, USA: antibodies against full length caspase-3, caspase-8, DFF 45/ICAD, and PARP, amino terminus of gelsolin, plectin, cytochrome C, and carboxy terminus of BID. Immunocytochemistry was performed using primary antibodies against active caspase-3 pl7 (BD PharMingen, San Diego, CA, USA, and Promega, Madison, USA), active caspase-9 (BioLabs Inc., New England, UK), and anti-PARP p85 (Promega). Secondary IgG-Texas Red, fluoresceine (FITC)-conjugated antibodies and Mitotracker CMXRos were from Molecular Probes, Leiden, The Netherlands. Other reagents were from Sigma if not specified.
Preparation of polyclonal antibody directed against GPIP1
GPIP1, a peptide comprising 25 aminoacids corresponding to residue 162-187 (NH2-KSGLKYKKGGVA-SGFKHVVPNEVVV-COOH) in human gelsolin sequence (Swiss-Prot P06396) was synthetized (95% purity) by MWG AG Biotechnology, Ebersberg, Germany. 200 μg of the peptide were injected 3 times into rabbits in 3 weeks intervals. Sera were used for Western blots and immunoflourescence in a 1:100 dilution.
Cell culture
Primary vascular smooth muscle cells (VSMCs) were prepared from porcine aorta. Retinal pigment epithelium (RPE) cells obtained from a 59 years old eye donor were provided by Dr. M. Boenke, Department of Ophthalmology, University Hospital, Bern. The cells were cultivated in Minimal Essential Medium (MEM) with Earle's salts (Life Sciences, Basel, Switzerland). Cells were grown under a humidified atmosphere of 5% CO2 in air at 37°C in MEM supplemented with 10% fetal bovine serum, penicillin (10 U/ml), streptomycin (10 μg/ml) and fungizone (250 ng/ml). When confluent, they were incubated for one week in MEM or MEM supplemented with xanthurenic acid. A 20 mM stock solution of xanthurenic acid was prepared in 0.5 M NaHCO3, and diluted in 0.05 M NaHCO3.
Cytotoxicity and apoptosis assay
Cells were observed with differential interference contrast optics on a Zeiss Avionert 405 M inverted microscope. Cell viability was determined by staining the cells with Hoechst 33342 and propidium iodide (PI) (Juro, Switzerland) using 50 μg/ml of each dye. Fragmented, apoptotic, nuclei were observed with excitation at 350 nm, and necrotic nuclei at 530 nm. The extent of apoptosis was estimated by examination of 300 nuclei from each sample.
Cell lysis and immunobloting
Cell were washed twice with cold 0.01 M phosphate buffer, pH 7.4. For Western blotting, cells were lysed in buffer containing 50 mM Tris (pH 8.0), 150 mM NaCl, 1 %Triton X-100, and the following protease inhibitors: 1 mM phenyl-methylsulfonyl fluoride, and leupeptin, aprotinin, and pepstatin, each at 1 μg/ml. The concentration of proteins was calculated from the absorption maximum at 280 nm, as described previously [11], and concentration of xanthurenic acid from its absorptiom maximum at 342 nm (εM 6500). The lysate was centrifuged for 10 min at 14 000 g, and the supernatant was boiled in loading-buffer for 5 min. Proteins (50 μg per lane) were separated by SDS-PAGE containing 10 or 12.5% acrylamide. After transfer to Hybond ECL membrane (Amersham Pharmacia Biotech AB, Uppsala, Sweden) the proteins were probed with the appropriate antibodies. Chemilunimescence ECL system (Amersham Pharmacia Biotech AB, Uppsala, Sweden) was used for the detection of peroxidase-conjugated secondary antibody.
Caspase-3 activity
Caspase 3 cleavage activity is based on the spectrophotometric detection of the chromophore p-nitroaniline at 405 nm after cleavage from the substrate Ac-Asp-Glu-Val-Asp-p-nitroaniline (Ac-DEVD-pNa) (Bachem, Basel, Switzerland). Caspase activity was measured after 1 hour of incubation of 200 μM Ac-DEVD-pNa at 37°C with the cell extract containing 25 mM HEPES (pH 7.5), 300 mM NaCI, 10 mM KCl, 1.5 mM MgCl2, 10% glycerol, 0.1 mM DTT, 1 mM phenylmethylsulfonylfluoride, and aprotinin, leupeptin, and pepstatin, each at 1 μg/ml.
Immunofluorescence
Cells grown on glass coverslips were fixed for 10 min at room temperature in 4% paraformaldehyde in 0.1 M PIPES, pH 6.8, washed in PBS and permeabilized for 5 min in PIPES containing 0.05% saponin (65 μl per coverslip), washed in PBS, incubated with cold aceton for additional fixing and permeabilisation, and again washed in PBS. Cells were incubated for 1.5 hour with the first antibody diluted in PBS containing 1% bovine serum albumine, and after washing incubated for 1.5 hour with the secondary antibody. The coverslips were then washed in PBS and incubated for 10 min with 65 μl of 4% paraformaldehyde solution containing 1 μl of Hoechst 33342 dye (1 mg/ml), washed in PBS, and incubated with Antifade Kits (Molecular Probes, Leiden, The Netherlands) according to the supplier's instruction. Staining of mitochondria was performed using Mitotracker CMXRos, as follows: confluent cells culture were pre-incubated without or with xanthurenic acid in MEM medium for 3 hours. The medium was removed and replace with medium containing 100 nM Mitotracker CMXRos. After an incubation for 45 min Mitotracker CMXRos was removed and replaced by MEM medium. The cell were cultivated for the next 20 hours, stained additionally with antibody against gelsolin (as above) and then observed by fluorescence microscopy.
Acknowledgements
This work was supported by grant awarded to H. Z. M. by the Swiss National Foundation (32-59183.99). We thank Drs. E. Berrou and M. Boenke for helpful discussions and Mrs D. Zuercher for retinal cell culture preparation.
The rapid increase in the development of mouse models is resulting in a growing demand for non-invasive physiological monitoring of large quantities of mice. Accordingly, we developed a new system for recording electrocardiograms (ECGs) in conscious mice without anesthesia or implants, and created Internet-accessible software for analyzing murine ECG signals. The system includes paw-sized conductive electrodes embedded in a platform configured to record ECGs when 3 single electrodes contact 3 paws.
Results
With this technique we demonstrated significantly reduced heart rate variability in neonates compared to adult mice. We also demonstrated that female mice exhibit significant ECG differences in comparison to age-matched males, both at baseline and in response to β-adrenergic stimulation.
Conclusions
The technology we developed enables non-invasive screening of large numbers of mice for ECG changes resulting from genetic, pharmacological, or pathophysiological alterations. Data we obtained non-invasively are not only consistent with what have been reported using invasive and expensive methods, but also demonstrate new findings regarding gender-dependent and age-dependent variations in ECGs in mice.
Background
Although electrocardiograms (ECGs) have been obtained in conscious mice, currently reported techniques require restraint [1] or anesthesia and surgical implantation of telemetry devices [2, 3, 4]. Anesthesia, however, may depress cardiovascular function, and adequate recovery after transmitter implantation in mice is nearly 3 weeks [3, 5]. Accordingly, we developed a non-invasive technique for obtaining ECGs in conscious mice by placing the animal on a platform embedded with paw-sized ECG electrodes connected to an amplifier. This method is much less traumatic, requires no anesthesia or surgery, and promotes rapid screening of large quantities of mice. ECG data we obtained non-invasively in conscious mice are comparable to those recently published using surgically implanted telemetry devices [2, 4]. To test the efficacy of our system, we evaluated ECGs in mice of either sex, of several strains, and of different ages. Moreover, we tested whether our system could detect ECG alterations in response to pharmacological challenge by isoproterenol. The baseline heart rate data and responses to the β-adrenergic agonist isoproterenol we recorded non-invasively in mice are comparable to data published using invasive methods[6]. We developed an ECG signal processing, analyzing, and database Web portal, which we named e-MOUSE™, accessible to the biotechnology community. The advantages of the ECG recording and analyses paradigm we developed are clear, given the high cost of breeding, housing, and transporting mice, and the call for comprehensive yet widely available phenotyping tests [7].
Results
We recorded ECGs in 10 males and 10 females each of C57BL/6, 129/Sv, and FVB/N mice, three inbred strains commonly used to model human diseases. A representative ECG recording from an adult male C57BL/6 mouse is shown in Fig. 1. Since the mice are conscious, baseline artifact and noise are apparent in the unfiltered signals. Yet, the P-waves and T-waves are discernible by eye and interpretable by the software algorithmic processing of the signal digitized at 2500 samples per second. Analyses of the digitized signals via e-MOUSE™ demonstrated significantly faster heart rates in C57BL/6 female mice than in males (741 ± 2 bpm vs. 692 ± 5 bpm, P < 0.05, n = 10) and shorter QRS duration (7.5 ± 0.2 ms vs. 8.1 ± 0.2 ms, P < 0.05). Table 1 summarizes results in 129/Sv mice (10 males and 10 females), again demonstrating faster heart rate and shorter QRS duration and QT interval in female mice than in age-matched males. Rate corrected QT (QTc) in male 129/Sv mice was significantly longer than in females (66.7 ± 1.9 ms vs. 58.4 ± 1.6 ms, P < 0.05). Gender differences in heart rate were also apparent in the FVB/N strain [736 ± 5 bpm in males (n = 10) vs. 706 ± 7 bpm in females (n = 10), P < 0.05].
Electrocardiogram from a conscious adult male C57BL/6 mouse at baseline, with indication of ECG parameters.
Electrocardiographic parameters in male and female 129/Sv, C57BL/6, and FVB conscious mice.
129/Sv
C57BL/6
FVB
Males
Females
Males
Females
Males
Females
(n = 10)
(n = 10)
(n = 10)
(n = 10)
(n = 10)
(n = 10)
Heart Rate (bpm)
571 ± 13
689 ± 12 *
692 ± 5
741 ± 2 *
736 ± 5
706 ± 7 *
HR var (bpm)
15 ± 4
20 ± 4
12 ± 3
15 ± 3
33 ± 2
16 ± 2 *
PR (ms)
31.9 ± 1.3
28.3 ± 0.6*
26.0 ± 0.8
24.4 ± 0.7
25.4 ± 0.6
27.7 ± 0.6 *
QRS (ms)
9.6 ± 0.3
8.5 ± 0.1*
8.1 ± 0.2
7.5 ± 0.2*
7.4 ± 0.1
7.7 ± 0.1
QT (ms)
70.2 ± 1.8
54.7 ± 2.0*
55.4 ± 0.8
53.8 ± 0.6
53.9 ± 1.2
55.7 ± 1.4
QTc (ms)
66.7 ± 1.9
58.4 ± 1.6*
59.0 ± 0.8
58.7 ± 0.3
59.3 ± 1.2
59.9 ± 1.6
Values are mean ± SE. *, P < 0.05 vs. males within strain. HR var, heart rate variability.
Reducing the spacing of the electrodes allowed us to record ECGs in 6-day and again in 12-day old nursling C57BL/6 mice (8 males and 5 females) and return them to their mother. The ECG from a 6-day-old C57BL/6 nursling female is shown in Fig. 2. Heart rate variability in neonates was significantly less than in adults (2.5 ± 0.4 bpm vs. 21 ± 2 bpm, P < 0.05). Heart rate in nursling females was significantly slower than in adult females (655 ± 6 bpm vs. 741 ± 2 bpm, P < 0.05). Interestingly, the gender differences in heart rate we observed in adult mice were absent in 6-day old and 12-day old neonates. Upon weaning (21 days old), however, gender differences in heart rate became apparent [697 ± 14 bpm in males (n = 5) vs. 750 ± 8 bpm in females (n = 3), P < 0.05].
Electrocardiogram from a conscious neonate (6 days old) female C57BL/6 mouse. Heart rate variability in neonates was significantly less than in adult mice.
The acute increase in heart rate within the first minutes following one intraperitoneal injection of the β-adrenergic agonist isoproterenol (2.5 μg/g) was significantly less in female compared to male C57BL/6 adult mice (+5 ± 2% females vs. +12 ± 2% males, P < 0.05, n = 5 for each group). Saline injection had no effect on heart rate. After 3 days of repeated isoproterenol injection (2.5 μg/g, once at 9AM and once at 9PM), heart rate in males prior to the 6th injection was significantly reduced compared to control heart rate in males prior to the 1st injection, whereas the heart rate in females did not change. Both male and female mice exhibited significant acute increases in heart rate (+27 ± 6% males and +22 ± 3% females) after the 6th injection of the drug on day 3 (P < 0.05 compared to % increases following 1st injection on day 1). Fig. 3 illustrates significant electrocardiographic alterations immediately after the 1st administration of isoproterenol on day 1 in a C57BL/6 male mouse. ST segment depression, suggestive of acute subendocardial ischemia, was consistently observed in C57BL/6 male mice but not in all female mice immediately following administration of isoproterenol. Heart weight in males and females treated with isoproterenol for 3 days was significantly increased (16%) compared to hearts from mice treated with saline for 3 days (P < 0.05).
Electrocardiogram from the C57BL/6 male mouse of Fig. 1 recorded immediately after an intraperitoneal injection of 2.5 μg/g isoproterenol on day 1, showing ST segment depression suggestive of acute subendocardial ischemia.
Discussion
This report describes the development of a system for non-invasively recording ECGs in conscious mice; neither attachment of wires, nor anesthesia, nor surgical implantation of devices is required. Our heart rate data in conscious mice are comparable to those obtained using sutured steel wire [8] and chronic indwelling catheters [6]. Our ECG measurements are comparable to those obtained using implantable telemetry devices [2, 4]. This non-invasive method, standardized protocol, and Internet-accessible mouse ECG analyses software (e-MOUSE™) may reduce the heterogeneity in data collected from different laboratories [7].
Strain & Gender Differences in ECGs in Adult Mice
To our knowledge, this is the first study to describe gender differences in heart rate in genetically homogenous strains of conscious mice. Studies in humans [9, 10] and in rats [11, 12] have shown females to have faster heart rates than males, differences that disappear with age [10, 13]. In C57BL/6 mice and in 129/Sv mice, we found that conscious females exhibited significantly faster heart rates than male mice. In FVB/N mice, however, we noted faster heart rates and briefer ECG time intervals in male compared to female mice. Mitchell et al., using implanted transmitters, reported slightly but insignificantly faster heart rates in male compared to female FVB mice [4]. Gender and strain differences in heart rate may reflect differences in hormone activity, which are known to affect cardiovascular autonomic regulation differentially in males and females [9] and vary among mouse strains [14]. Our strain-dependent observations agree with results of invasive studies [8, 6] and support the importance of genetic factors in influencing heart rate.
Developmental Changes in ECGs in Conscious Neonatal Mice
We found that the gender differences apparent in C57BL/6 adults were absent in nursling mice. Reduced activity in newborns could contribute to reduced heart rate and heart rate variability. However, observations in quietly resting adult mice (n = 3), monitored 30 min after placement on the recording platform, of reduced heart rate (471 ± 20 bpm) and increased heart rate variability (53 ± 6 bpm) suggest that the results in neonates may be attributable to reduced sympathetic and parasympathetic signaling [15]. The time-domain heart rate variability data we obtained non-invasively are comparable to those obtained using implanted telemetry devices [2, 16]. The developmental changes in heart rate we measured in mice parallel those in human neonates [13, 17], lambs [18], and newborn rats [19]. Moreover, our observations in neonatal mice are in agreement with Wang et al. who reported progressive and significant shortening of R-R intervals during neonatal development [1]. Accordingly, our approach, adaptable to the study of small mice that might otherwise die from anesthesia or surgery, might be valuable in examining developmental changes and abnormalities [13, 17].
Effects of Sympathetic Stimulation in Conscious Mice
The blunted heart rate increase in response to β-adrenergic stimulation in conscious C57BL/6 female mice is consistent with gender related differences in baroreflex control of heart rate in healthy humans [20]. Estrogen has been shown to enhance baroreflex sensitivity [21] and attenuate the response of heart rate to administration of isoproterenol in rats [22]. After repeated injections of isoproterenol, heart rate decreased and the QRS duration increased significantly in male mice, but did not change in female mice. Yet, both males and females demonstrated increased sensitivity to acute isoproterenol injection after repeated administration that resulted in significant cardiac hypertrophy [23, 24]. We observed a 16% increase in heart mass in both sexes after 3-days of isoproterenol treatment compared to mice treated with equivolume of saline. Although β-adrenoceptor desensitization [23] or reduced ATPase activity [25] may account for the decrease in basal heart rate in the hypertrophied male hearts, the increased sensitivity to isoproterenol may support the hypotheses of ischemia-induced activation of β-adrenoceptor kinase [26] or modulation of G-protein signaling to preserve adenylate cyclase activity [27]. To our knowledge, this is the first report that describes electrocardiographic evidence of myocardial ischemia in conscious mice, although ST segment changes have characterized isoproterenol-induced ischemia in rat [28] and man [29]. Why males consistently developed ST segment depression following isoproterenol administration and females did not remains unanswered. However, significant gender-related differences in the expression of glutathione S-transferases (GST) have been observed in mice, with female mice expressing significantly more of this antioxidant than male mice [30]. In mice, isoproterenol-induced oxidative stress may be attenuated in females due to higher levels of GST [31].
Study Limitations
Mice are sensitive to even modest handling [3, 8] and transport [3, 32]. The ECG indices we measured may reflect physiologic responses to the experimental environment relative to its home cage. We encourage a 10 min acclimation period prior to recording data to attenuate effects consequent to handling and transport. Usually mice establish contact between 3 electrodes and 3 limbs within 5 minutes after acclimation to record a continuous ECG for approximately 2 seconds. The duration of the ECG recording should be considered in the interpretation of heart rate variability [33]. The age, gender and strain variations in heart rate may reflect age, gender and strain variations in physiologic responses to transport [32], handling [3], and adaptation to repeated measurements [8]. Yet, the inter-strain differences in heart rate we non-invasively obtained after ECG recordings of short duration are in agreement with invasive experimental techniques intended to monitor heart rate in mice as caged [6, 8]. Our measurements were made in daytime hours, disrupting the less active phases of the mouse circadian cycle. Future innovation might incorporate an array of conductive electrodes into the animals' cages to eliminate the effects of handling and transport, and perhaps enable timed recordings.
Materials and MethodsMice
Young adult mice (9 ± 1 weeks old; 21 ± 1 g) from 3 commonly used inbred strains, C57BL/6, 129/Sv, and FVB/N (The Jackson Laboratory, Bar Harbor, ME), were housed in standard conditions within the Animal Resource Facility at the Beth Israel Deaconess Medical Center, Boston MA. Nursling C57BL/6 mice of either sex were examined at post-natal day 6 and 12 and returned to their mothers. The same mice, weaned, were examined at 3 weeks of age.
ECG recording
Mice were gently removed from their cages and positioned on the ECG recording platform. An array of gel-coated ECG electrodes (Red Dot; 3 M, St. Paul, MN) were embedded in the floor of the platform and spaced to provide contact between the electrodes and animals' paws. For adults, the spacing between electrodes was 3 cm, and for nurslings, the spacing was reduced to 2.5 cm. Filter paper, with openings for the electrodes, prevented mouse urine from short-circuiting the signals. The electrodes were connected to an amplifier (HP78901A, Hewlett-Packard, Andover, MA) by a shielded 3-electrode lead set (M1605A Snap, Hewlett-Packard, Andover, MA). Since even modest handling of mice may induce alterations in heart rate [6], each mouse was permitted to acclimatize for 10 min prior to collection of baseline data. The signals were digitized with 16-bit precision (DI-220, DATAQ Instruments, Inc., Akron, OH) at a sampling rate of 2500 samples/second. When mice were sitting or otherwise positioned such that the paws were not in contact with three electrodes, the output from the amplifier was discarded. Only data from continuous recordings of 20-30 ECG signals were used in the analyses. Data were transmitted to the mousespecifics.com Internet site (Mouse Specifics, Inc., Boston, MA) using standard file-transfer protocols for ECG signal analyses by e-MOUSE™.
ECG analyses
Each signal was analyzed using e-MOUSE™, an Internet-based physiologic waveform analysis portal. e-MOUSE™ incorporates Fourier analyses and linear time-invariant digital filtering of frequencies below 2 Hz and above 100 Hz to minimize environmental signal disturbances. The software uses a peak detection algorithm to find the peak of the R-waves and to calculate heart rate. Heart rate variability was calculated as the mean of the differences between sequential heart rates for the complete set of ECG signals. Subsequently, determination of 1st and 2nd derivatives and algebraic "if-thens" search the ECG signals for probable P-wave peaks and onset and termination of QRS complexes. Since the T-wave is not separate from the QRS in rodent ECGs [28, 34], there have been discrepancies in the definition of the QT interval and reported values [4]. In accord with Mitchell et al. [4], we routinely included the inverted and/or biphasic portions of the T-wave in our calculations of the QT interval. We defined the end of the T-wave of each signal as the point where the signal returned to the isoelectric line [1, 35] [the mean voltage between the preceding P-wave and QRS interval]. The QT intervals were rate corrected (QTc) by application of the equation recommended by Mitchell et al. [4] for use in mice. The software plots its interpretation of P,Q,R,S, and T for each beat so that spurious data resulting from unfiltered noise or motion artifacts may be rejected. e-MOUSE™ then calculates the mean of the ECG time intervals for each set of waveforms.
β-adrenoceptor stimulation
To test the sensitivity of our system for describing ECG alterations in response to a drug, C57BL/6 mice were given either an intraperitoneal injection of 2.5 μg/g (-)-isoproterenol (Sigma, St. Louis, MO) (5 males, 5 females) dissolved in 0.1 ml saline or an equal volume of saline (5 males, 5 females), twice daily for 3-days. ECGs were recorded within 5 min prior to each injection and between 1 and 2 min after each injection to capture the peak of the response to the drug [6]. After the last measurement, mice were euthanized by intraperitoneal administration of pentobarbital (150 mg/Kg), consistent with the American Veterinary Medical Association Panel on Euthanasia guidelines. Excised hearts, excluding atria and blotted dry, were then weighed.
Statistics
Data are presented as the means ± SE. Comparisons between genders among strains were performed using Student's t-test for unpaired observations. Effects of isoproterenol or saline injections within groups were performed using Student's t-test for paired observations, and between group comparisons using Student's t-test for unpaired observations. Differences were considered significant with P < 0.05.
Conclusions
We developed a non-invasive technique for obtaining ECGs in conscious mice. We developed an Internet-accessible portal for analyses of mouse electrocardiograms. Using this system, we demonstrated significant strain, gender and age-dependent differences in electrocardiograms in mice. Moreover, we demonstrated significant gender-dependent differences in the cardiovascular response to β-adrenoceptor stimulation. Our results may suggest that the stimulatory effects of genes and drugs on cardiac function may be more profound in male or masked in female mice. This non-invasive and rapid ECG phenotyping technique may improve the quality and increase the quantity of data collected from mouse models.
Acknowledgements
Mr. Diego Yepes is gratefully acknowledged for his innovative design approaches to the ECG acquisition system. Ms. Jeanne Smith provided excellent administrative support for this project and manuscript submission. I. Amende received generous support from Förderkreis zur Verbesserung des Gesundheitswesen e.V.
The multisubunit (α1S,α2-δ, β1a and γ1) skeletal muscle dihydropyridine receptor (DHPR) transduces membrane depolarization into release of Ca2+ from the sarcoplasmic reticulum (SR) and also acts as an L-type Ca2+ channel. To more fully investigate the function of the γ1 subunit in these two processes, we produced mice lacking this subunit by gene targeting.
Results
Mice lacking the DHPR γ1 subunit (γ1 null) survive to adulthood, are fertile and have no obvious gross phenotypic abnormalities. The γ1 subunit is expressed at approximately half the normal level in heterozygous mice (γ1 het). The density of the L-type Ca2+ current in γ1 null and γ1 het myotubes was higher than in controls. Inactivation of the Ca2+ current produced by a long depolarization was slower and incomplete in γ1 null and γ1 het myotubes, and was shifted to a more positive potential than in controls. However, the half-activation potential of intramembrane charge movements was not shifted, and the maximum density of the total charge was unchanged. Also, no shift was observed in the voltage-dependence of Ca2+ transients. γ1 null and γ1 het myotubes had the same peak Ca2+ amplitude vs. voltage relationship as control myotubes.
Conclusions
The L-type Ca2+ channel function, but not the SR Ca2+ release triggering function of the skeletal muscle dihydropyridine receptor, is modulated by the γ1 subunit.
Background
In skeletal muscle, the dihydropyridine receptor (DHPR) consists of α1S, α2-δ, β1a and γ1 subunits [1]. This complex is responsible for the L-type Ca2+ current and serves as the voltage sensor for excitation-contraction (EC) coupling. In the latter process, the movement of electrical charges in the α1S subunit promotes a conformational change that opens the ryanodine receptor type-1 (RyR1) in the sarcoplasmic reticulum membrane (SR) leading to an increase in cytosolic Ca2+ [2, 3]. The functional interactions between the DHPR subunits necessary for opening the Ca2+ channel are only partially known. Further, the interactions between DHPR subunits and RyR1 also are incompletely understood. The α1 subunit is a large protein that contains the basic functional elements of the L-type Ca2+ channel, including the Ca2+ selectivity, voltage-dependent gating, and sensitivity to dihydropyridines. The cytoplasmic loop between repeats II and III of the α1S subunit interacts closely with RyR1 and is an important determinant of skeletal type EC coupling [4, 5]. A region in the cytoplasmic loop between repeats I and II of the α1 subunit, referred to as the AID [6], binds tightly with a 30 amino acid region on the β1 subunit, referred to as the BID [7]. β subunits are ∼ 55 to 65 kDa proteins essential for channel assembly and/or membrane targeting, as well as for modulation of channel kinetics [8]. The α2-δ subunit is a highly glycosylated ∼ 175 kDa protein formed by two disulfide-linked peptides encoded by the same gene. Transmembrane topology and functional analyses suggest the α2-δ subunit is composed of a single transmembrane domain and a short cytoplasmic tail of only five residues [9, 10]. Given this topology, the α2-δ subunit is most likely to interact with the α1 and/or the γ1 subunits. The γ1 subunit is a ∼ 32 kDa skeletal muscle-specific protein with four presumptive transmembrane domains [11, 12, 13]. The transmembrane topology of the γ1 subunit and the critical binding domains are unknown at this time.
EC coupling is initiated by voltage-dependent charge movements in the S4 segments of the DHPR α1S subunit, whose expression is dependent on the presence of the β1a subunit [14]. The C-terminus of the β1a subunit has also been shown to be important in EC coupling, presumably by interaction with RyR1 [15]. The role of the α2-δ subunit on skeletal EC coupling is unknown, but in heterologous expression systems it has been demonstrated to increase the amount of charge movement [16]. The role of the γ1 subunit in Ca2+ channel function is of particular interest given the discovery of a second γ subunit, γ2 or stargazin, which is expressed in neurons and is responsible for the stargazer mutation in mice [17]. Subsequently, several other γ subunit genes have been identified and shown to be expressed in brain and peripheral tissues [18, 19]. In the present study we examined the role of the γ1 subunit in L-type Ca2+ current and EC coupling in skeletal myotubes. Absence of the γ1 subunit slows inactivation and produces a depolarizing shift in the Ca2+ current inactivation vs. voltage curve, in agreement with results from an independently produced γ1 knockout mouse [20]. However, absence of γ1 does not affect the voltage dependence or the magnitude of charge movements and Ca2+ transients. Overall, the γ1 subunit appears to promote inhibition of the Ca2+ channel function of the skeletal DHPR. While this subunit is clearly non-essential for activation of the L-type Ca2+ channel and for triggering skeletal-type EC coupling, γ1 appears to specifically modulate the Ca2+ channel function of the skeletal DHPR.
Results and DiscussionInactivation of the DHPR γ1 subunit by gene targeting
Gene targeting was used to replace the entire coding region of the γ1 gene with the neo gene, as described in Materials and Methods (Fig. 1). Targeted ES cell clones were injected into C57Bl/6J blastocysts and chimeras obtained. After additional breeding, a line containing the γ1 targeted allele was obtained. Fig. 1A shows Southern blots from DNA of control (+/+), γ1 het (+/-) and γ1 null (-/-) mice, after digestion with EcoRI and hybridization to probe 1. The γ1 het and γ1 null mice show the predicted 12.5 kb restriction fragment indicative of the targeted allele (Fig. 1B). Similarly for the Southern blot of DNA digested with HindIII and hybridized to probe 2, the γ1 het and γ1 null mice show the predicted 3.5 kb restriction fragment indicative of the targeted allele (Fig. 1C). The γ1 het and γ1 null mice have no visible abnormalities and are fertile. Western blots showing the expression of the DHPR subunits are shown in Fig 1D. The amount of the γ1 subunit present in control, γ1 het and γ1 null samples as a function of total membrane protein was determined in two independent membrane preparations. These data indicate that the amount of γ1 expression in membranes from γ1 null muscle was not detectable above background (no primary antibody) as expected from the fact that the entire gene was deleted. In membranes isolated from the γ1 het mice, the level of expression of the γ1 subunit was 64% of control values for each of the two independent determinations, consistent with what might be expected from the loss of one allele. Whether there is compensation by increased expression of other γ subunits such as the γ6 and γ7 subunits, which were shown recently to be present in skeletal muscle by RT-PCR [19], or alteration of the amount of the other subunits remains to be determined in the future.
Inactivation of the γ1 subunit by gene targeting. A) i. The normal γ1 gene, which contains 4 exons (represented by black boxes). ii. The targeting vector contains a fragment with 4.4 kb of homology to a region 5' of exon 1 and a 2.6 kb fragment with homology to a region 3' of exon 4 (thick horizontal lines). These fragments are separated by the neo gene, which is used to select the ES cells in culture. iii. The modified γ1 locus. When recombination occurs in the regions indicated by the X's, all 4 exons of the γ1 gene are replaced by the neo cassette. B, C) Southern blots of DNA, either digested with EcoRI and hybridized to probe 1 (B), or DNA digested with HindIII and hybridized to probe 2 (C), from control (+/+), γ1 het (+/-) and γ1 null (-/-) mice. These data show the predicted size bands for the wild type and targeted alleles. D) Western blots for each of the four skeletal DHPR subunits from control (+/+), γ1 het (+/-) and γ1 null (-/-) mice. The γ1 subunit is absent from the γ1 null mice and at approximately half the normal level in the γ1 het mice.
Increase in Ca2+ current density and decrease in inactivation in γ1 null and γ1 het myotubes
We examined the impact of the γ1 null mutation on DHPR properties using electrophysiological techniques in primary cultures of skeletal myotubes derived from mouse embryos. Fig. 2A shows whole-cell Ca2+ currents in control, γ1 het and γ1 null myotubes in response to 500-ms depolarizing voltage steps from a holding potential of -40 mV. Current traces for only three voltage steps are shown for clarity. All three cell types displayed the slow, high-voltage activated L-type Ca2+ current typical of skeletal myotubes. When the holding potential was -80 mV, we also observed a T-type Ca2+ current, which was seen in approximately half of the cells tested (data not shown). Ca2+ conductance vs. voltage curves were computed in Fig. 2B for the three cell types. The sigmoidal lines correspond to a Boltzmann fit to the population mean while parameters fitted to each cell are shown in Table 1. These data show that the maximum conductance, Gmax, for both the γ1 het and γ1 null cells are significantly larger (p<0.05) than the values for control cells. However, the half-activation potential (V1/2) and the steepness of the curve (factor k) were not significantly different (see Table 1). Thus, full or partial elimination of the γ1 subunit produced a significant increase in the Ca2+ current density without a change in the voltage-dependence of the Ca2+ conductance.
L-type Ca2+ conductance in γ1 knockout myotubes. A) Whole-cell L-type Ca2+ current when control (+/+), γ1 het (+/-), or γ1 null (-/-) myotubes are depolarized to either -10 mV, +30 mV or +60 mV from a holding potential of -40 mV. The pulse duration was 500 ms. The cell capacitance was 262 pF, 221 pF, and 286 pF, for the control, γ1 het and γ1 null cells, respectively. B) Voltage dependence of the Ca2+ conductance for 8 control, 8 γ1 het and 12 γ1 null cells. The curves correspond to a Boltzmann fit of the population mean with the following parameters. Gmax = 132 pS/pF, V1/2 = 13.4 mV and k = 5.3 mV for control cells; Gmax = 160 pS/pF, V1/2 = 13 mV and k = 5.8 mV for γ1 het cells; and Gmax = 167 pS/pF, V1/2 = 9.9 mV and k = 4.6 mV for γ1 null cells.
Boltzmann parameters of Ca2+ conductance, charge movements and Ca2+ transients of control, γ1 het and γ1 null myotubes.
CELL TYPE
G-V
Q-V
Gmax (pS/pF)
V1/2 (mV)
k (mV)
Qmax (fC/pF)
V1/2 (mV)
k (mV)
Control, +/+
131 ± 6.2
14.7 ± 2.1
4.2 ± 0.4
5.3 ± 0.4
11.4 ± 2.5
15.4 ± 1.1
(n)
(8)
(8)
γ1 het, +/-
161 ± 11.5a
14.3 ± 2.8
5 ± 0.8
5.3 ± 0.3
14.4 ± 0.6
14.7 ± 2.0
(n)
(8)
(4)
γ1 null, -/-
170 ± 9.6b
10.9 ± 0.9
4.5 ± 0.4
5.3 ± 0.9
17.8 ± 1.9
15.8 ± 2.2
(n)
(12)
(8)
ΔF/Fo-V
Inactivation
ΔF/Fo(max)
V1/2 (mV)
k (mV)
V1/2 (mV)
k (mV)
Control, +/+
3.9 ± 0.3
4.6 ± 1.2
10.2 ± 1.3
-4.6 ± 3.6
7.3 ± 0.2
(n)
(6)
(6)
γ1 het, +/-
3.3 ± 0.5
8.2 ± 3.0
10.1 ± 2.0
8.1 ± 2.3c
8.0 ± 0.6
(n)
(4)
(8)
γ1 null, -/-
3.7 ± 0.4
7.3 ± 1.6
9.0 ± 1.1
8.9 ± 0.8d
8.1 ± 0.8
(n)
(6)
(8)
Entries (mean ± SEM) correspond to Boltzmann parameters fitted to each cell (number of cells n, in parenthesis). The only statistically significant differences (p<0.05) were for either γ1 het or γ1 null cells compared to control cells as indicated. The t-test p values were 0.004, 0.009, 0.011, 0.009 for a-d, respectively.
At intermediate potentials such as +30 mV (Fig. 2A), the Ca2+ current of γ1 het and γ1 null cells inactivated more slowly than that of control cells. To examine changes in inactivation, we used the paired-pulse protocol shown in Fig. 3A, which measured the Ca2+ current not inactivated by the pre-pulse. It has been shown that inactivation of skeletal L-type Ca2+ channels approaches steady-state extremely slowly, on the time scale of tens of seconds [21]. However, for comparative purposes and to be consistent with previous measurements in Ca2+ channels lacking γ1 [20, 22], we chose a much shorter protocol. Cells were held at -80 mV, then stepped to a series of depolarizing (pre-pulse) voltages for 2.5 s to promote inactivation, then stepped to -50 mV for 100 ms to close non-inactivated channels. The non-inactivated Ca2+ current was measured during a 200-ms test pulse to +20 mV. In control cells, there was almost complete inactivation of the Ca2+ current when the pre-pulse potential was in the positive range. This can be seen by comparing the amplitude of the Ca2+ current during the test pulse in traces 1, 2, and 3, corresponding to pre-pulse potentials to -70 mV, +30 mV, and +60 mV, respectively. In contrast, γ1 het and γ1 null cells showed less inactivation during the pre-pulse, and thus a significantly larger Ca2+ current during the test pulse. Furthermore, pre-pulse potentials larger than +30 mV did not promote additional inactivation, which can be seen by comparing traces 2 and 3 in the γ1 het and γ1 null cells. Fig. 3B shows the maximum Ca2+ current during the test pulse plotted as a function of the pre-pulse potential. In control cells, inactivation was complete when the pre-pulse potential was above +30 mV (-6.9 ± 0.8 pA/pF at -70 mV vs. -0.2 ± 0.2 pA/pF at +70 mV, n = 6 cells). In γ1 null cells, inactivation reached a plateau when the pre-pulse potential was more positive than +20 mV (-8.5 ± 0.9 pA/pF at -70 mV vs. -3.2 ± 0.7 pA/pF at +70 mV, n = 8 cells). Incomplete inactivation was also seen in γ1 het cells, although this was less pronounced than in γ1 null cells. Since the pre-pulse was not sufficiently long for inactivation to reach steady-state, we are not certain whether the observed incomplete inactivation reflects a component of the γ1 null Ca2+ current that inactivates extremely slowly or a "truly" non-inactivating component. To compare the voltage-dependence of the inactivation produced by the 2.5-second pre-pulse, the Ca2+ current during the test pulse was normalized and fit to the Boltzmann equation (Fig. 3C). The curves for γ1 het and γ1 null cells showed a significant (p<0.05) shift to more positive potentials compared to control cells. The same conclusion was reached by comparing the Boltzmann parameters of individual cells in Table 1. Since control cells had a more negative half inactivation potential, the higher Ca2+ conductance of γ1 het and γ1 null cells reported in Fig. 2 could be explained by a larger steady-state inactivation of the control Ca2+ current at the -40 mV holding potential. To determine if this was the case, we measured the Ca2+ conductance during the pre-pulse for the same cells analyzed in Fig 3B,C. The pre-pulse was delivered from a holding potential of -80 mV, and therefore the pre-pulse current should be less contaminated by steady-state inactivation. Gmax at the peak of the Ca2+ current during the pre-pulse was 146 ± 12 pS/pF for controls, 172 ± 18 pS/pF γ1 het and 186 ± 13 pS/pF for γ1 null cells with t-test statistical significance for controls vs. γ1 null cells of p = 0.034 and for controls vs. γ1 het cells of p = 0.064. Thus the higher Ca2+ conductance of γ1 het and γ1 null cells persisted at the more negative holding potential. These Ca2+ conductances were approximately 10% higher than those at the -40 mV holding potential reported in Table 1. The additional current could have been contributed by T-type Ca2+ channels (not shown), which are available at the -80 mV holding potential but are inactivated at the more positive holding potential. Also, since the T-type Ca2+ current density was highly variable (see above), this could explain the higher p values compared to those reported in Table 1. In summary, decreasing the amount of the γ1 subunit protein or eliminating it completely resulted in myotubes with a higher Ca2+ current density, a similar voltage-dependence of activation, and a reduced rate of inactivation. These results, with the exception of the change in Ca2+ current density, are in remarkable agreement with those obtained when L-type Ca2+ channel subunits are coexpressed without the γ1 subunit in heterologous cells (see below). They also are similar to results obtained elsewhere in γ1 null myotubes using a 5-second pre-pulse [20]. Furthermore, the incomplete inactivation of the γ1 null Ca2+ current resembles that observed in ts A201 cells expressing DHPRs specifically lacking γ1 and using a 3-second inactivation protocol [22].
Slow inactivation of the Ca2+ current in γ1 knockout myotubes. A) Diagram to scale of the two-pulse protocol used to inactivate the L-type Ca2+ current at each of 15 potentials and then the test potential (+20 mV) to measure the remaining non-inactivated current. Ca2+ currents during the pre-pulse and test-pulse phases of the protocol are shown for a pre-pulse depolarization to -70 mV (trace 1), +30 mV (trace 2) and +60 mV (trace 3) in a control, a γ1 het, and a γ1 null myotube. The cell capacitance was 317 pF for the control, 175 pF for the γ1 het and 348 pF for the γ1 null cell. B) The maximum Ca2+ current during the test pulse is plotted as a function of the pre-pulse potential for 6 control, 8 γ1 het and 8 γ1 null cells. C) The non-activating component was subtracted and the curves were normalized to show the voltage-dependence of the inactivating component. I is the maximum test current, Imin is the test current at +50 mV and Imax is the test current at -70 mV. The curves correspond to a Boltzmann fit of the population mean with the following parameters. [(I-Imin)/Imax]max = 1, V1/2 = -3.8 mV and k = 8.4 mV for control cells. [(I-Imin)/Imax]max = 1, V1/2 = +15.6 mV and k = 8.1 mV for γ1 het cells. [(I-Imin)/Imax]max = 1, V1/2 = +9.7 mV and k = 7.9 mV for γ1 null cells.
The effect of the γ1 subunit on Ca2+ current kinetics has been somewhat confusing until now, because it has not been possible to examine its role using homologous subunits and a homologous expression system. In amphibian oocytes, γ1 was assayed in a complex with α1C, β1a and α2-δ [23, 24]. These investigators found γ1 produced a slight decrease in Ca2+ current density, no change in activation kinetics and a positive shift in the peak Ca2+ current vs. voltage curve. In HEK293 cells using the same subunit composition, the γ1 subunit did not alter the kinetics or density of Ca2+ currents [25]. However, in both oocytes and HEK293 cells, the γ1 subunit increased the rate of Ca2+ channel inactivation and produced a large hyperpolarizing shift in steady-state inactivation [23, 25, 26]. The results obtained here in skeletal muscle demonstrate that absence of the γ1 subunit increases Ca2+ current density, decreases the inactivation rate, and shifts the voltage-dependence of inactivation in the depolarizing direction. Freise et al. [20] showed that the increase in Ca2+ current was not a result of changes in single channel conductance and suggested the effects were due to changes in channel open probability. However, this remains to be tested directly. Expression of the γ2 subunit in BHK cells expressing α1A, α2-δ and β1a also shifts the voltage-dependence of inactivation ∼ 7 mV in the hyperpolarizing direction [17]. In summary, the two γ subunits that have been characterized to date, namely γ1 in the homologous skeletal myotube system described here and γ2 described above in heterologous BHK cells [17], appear to be inhibitory. In each case, the presence of γ results in fewer Ca2+ channels available for activation.
Absence of changes in the EC coupling voltage sensor of γ1 nullmyotubes
The impact of the γ1 null mutation on myotube EC coupling was inferred from intramembrane charge movements, the bulk of which have been shown to originate from the DHPR voltage sensor [27, 28]. Additionally, we measured Ca2+ transients evoked by voltage (i.e., by the movement of the voltage sensor) and compared these two sets of data. To measure DHPR charge movements and to separate these charge movements from those produced by other voltage-gated channels, a) ionic currents were blocked; b) the linear component of the cell capacity was subtracted using a P/4 procedure and by analog compensation; and c) we used a pulse protocol that eliminated the immobilization-sensitive charge movements from voltage-gated Na+ channels and presumably also from T-type Ca2+ channels. Previous studies in dysgenic (α1S null) myotubes transfected with α1S showed that the pulse protocol and the internal and external solutions used here (see Materials and Methods) detected charge movements of a magnitude 5 to 8 fold larger than those present in non-transfected dysgenic myotubes [28]. Thus we are confident that the technique adequately measures intramembrane charge movements in DHPR voltage sensors present in the plasma membrane. Fig. 4A shows "gating-type" currents produced by charge movements in response to 3 of 14 step voltages delivered to each cell (-40, +10, and +70 mV). Charge movements did not occur at -40 mV in any of the cell types. Traces at +10 mV and +70 mV show a transient current at the onset of the voltage step, the ON charge, and an equal inverted current at the end of the voltage step, the OFF charge. The voltage-dependence of the OFF charge estimated by integration is shown in Fig. 4B. The OFF transient was usually less contaminated by ionic current than the ON transient, and for that reason it provides a better estimate of the total DHPR charge. The main contaminant of the ON transient was a non-linear outward current, presumably a K+ channel that was not always completely blocked by the pipette solution. The OFF charge increased in a sigmoidal fashion and saturated at potentials more positive than +60 mV in all cases. These data were adequately fit by a Boltzmann equation shown by the lines. We found that the fitted maximum charge (Qmax) expressed by the 3 cell types was ∼ 5 fC/pF, in agreement with previous determinations in normal myotubes [29]. Furthermore, Boltzmann parameters determined for each cell and then averaged were not significantly different in any of two-by-two comparisons between control, γ1 het or γ1 null cells (Table 1). In summary, the magnitude of intramembrane charge movements and its voltage dependence were unaltered by partial or complete removal of the γ1 subunit. Furthermore, in agreement with previous data [29], no additional charges moved when pulses were delivered from a holding potential of -120 mV (data not shown). Thus, failure to detect changes in the charge vs. voltage characteristics of γ1 null myotubes compared to controls cannot be explained by charge movements in the γ1 null myotubes occurring at potentials more negative than the chosen -80 mV holding potential.
Voltage dependence of intramembrane charge movements in γ1 knockout myotubes. A) Whole-cell current produced by charge movements in response to 3 of 14 25-ms voltage steps delivered to the same cell. The voltage steps shown are to -40 mV, +10 mV, and +70 mV. The cell capacitance was 241 pF for the control, 292 pF for the γ1 het and 253 pF for the γ1 null cell. B) Voltage dependence of charge movement obtained by integration of the OFF transient current for 8 control, 4 γ1 het, and 8 γ1 null cells. The curves correspond to a Boltzmann fit of the population mean with the following parameters. Qmax = 5.2 fC/pF, V1/2 = 13.2 mV and k = 9 mV for control cells. Qmax = 4.9 fC/pF; V1/2 = 15 mV and k = 14.5 mV for γ1 het cells. Qmax = 5.4 fC/pF, V1/2 = 19.6 mV and k = 16 mV for γ1 null cells.
Absence of alterations in the voltage dependence of charge movement in the γ1 null cells should also be reflected in the voltage-dependence of Ca2+ transients. Changes in intracellular Ca2+ were measured using confocal line-scan imaging of fluo-4 fluorescence. Fig. 5A shows a line-scan image of a Ca2+ transient stimulated by a 50 ms depolarization to +90 mV from a holding potential of -40 mV in each cell type. The scan direction was perpendicular to the long axis of the myotube in a region selected on the basis of a low resting fluorescence and a high stimulated fluorescence (red/yellow color). The two traces on top of each image correspond to the time course of the fluo-4 fluorescence averaged across the entire line-scan at -10 mV, which is close to the threshold, and at +90 mV, which produced the maximum Ca2+ transient. Increase in cytosolic Ca2+ started at the onset of the depolarization and peaked ∼ 100 ms later in the three cell types. For all cells, the peak fluorescence change for the depolarization to +90 mV was >3 ΔF/Fo units, which is equivalent to >4 times the resting fluorescence. The relatively long recovery time following the depolarization is due to the low Ca2+ buffering power of the internal recording solution and is in agreement with previous studies [27, 30]. The voltage dependence of ΔF/Fo measured at the peak of the transient is shown in Fig. 5B. All curves had the same profile which consisted of a threshold for peak Ca2+ release at -10 mV, an increase in peak Ca2+ with pulse potential from -10 mV to +30 mV and a plateau at more positive potentials. Furthermore, two-by-two comparisons of Boltzmann parameters fitted to each cell did not reveal statistically significant differences (see Table 1). In summary, the voltage-dependence of Ca2+ transients and the maximum cytosolic Ca2+ concentration during the transient were not significantly affected by partial or total absence of the γ1 subunit. However, we cannot rule out changes in the rate of SR Ca2+ release since this measurement requires an entirely different internal solution [31]. A recent report indicates that the SR Ca2+ release rate is slightly higher in γ1 null myotubes [32].
Voltage-dependence of Ca2+ transients in γ1 knockout and normal myotubes. A) Confocal line-scan images of fluo-4 fluorescence in response to a 50 ms step to +90 mV from a holding potential of -40 mV. Hot colors represent high fluorescence (yellow>red). The pulse was delivered 100 ms after the start of the line-scan as indicated at the bottom of the figure. Images have a horizontal dimension of 2.05 seconds in all cases. The vertical dimension was 15, 28, and 24 microns for the control, γ1 het, and γ1 null cells, respectively. The two curves on top of the image show the time course of the fluorescence intensity at -10 mV and +90 mV. 1 ΔF/Fo unit corresponds to a doubling of the cell resting fluorescence. B) Voltage-dependence of the peak ΔF/Fo for 6 control, 4 γ1 het, and 6 γ1 null cells. The curves correspond to a Boltzmann fit of the population mean with the following parameters. ΔF/Fomax = 3.8, V1/2 = 10 mV, k = 9.4 mV for control cells. ΔF/Fomax = 3.6, V1/2 = 10.8 mV; k = 10.9 mV for γ1 het cells. ΔF/Fomax = 3.7, V1/2 = 4.5 mV; k = 10 mV for γ1 null cells.
Involvement of DHPR subunits in voltage sensing and EC coupling
The α1S pore subunit of the skeletal DHPR is the main determinant of charge movements and EC coupling [4, 27, 28]. However, other subunits of the DHPR are critical modulators of both functions. In heterologous expression systems, the α2-δ subunit was shown to affect the rate of Ca2+ current inactivation and charge movement immobilization [33]. Studies investigating the role of the β subunit on charge movement vary with the expression system used. In amphibian oocytes, β subunits increase the coupling between charge movements and channel opening [34, 35]. We have shown in skeletal muscle cells lacking the β1 subunit that reintroduction of variants of β1 has little effect on either charge movement or the kinetics of activation of the expressed Ca2+ current [15]. However, the carboxyl terminus of β1a is an important determinant of the efficiency of the coupling between charge movement and Ca2+ release [15]. In contrast to these results, earlier studies in which similar β subunits were co-expressed with α1 subunits in oocytes indicated β subunits had major effects on Ca2+ current density and gating kinetics [36, 37]. While the reason for these discrepancies is unknown, they highlight the fact that in skeletal muscle, the gating characteristics of the L-type Ca2+ channel might be controlled by interactions beyond those taking place among subunits of the DHPR. For example, a region of RyR1 has been shown to be essential for L-type Ca2+ current expression [38]. Such interaction is critical for skeletal DHPR function and is unlikely to occur in heterologous expression systems.
Conclusions
The ability of the γ1 subunit to selectively modulate the pore function of the DHPR without modulation of charge movements or the voltage dependence of Ca2+ transients is unique, especially since other DHPR subunits participate in both functions. In all likelihood, the charge movement protocol failed to detect gating currents responsible for opening the Ca2+ channel, which are quite small and are only resolved for depolarizations >200 ms [39] compared to the 25 ms used here (Fig. 4). A possible shift in the voltage-dependence of these charges recruited by long depolarizations would be consistent with the shift in Ca2+ current inactivation and remains to be resolved in γ1 null myotubes. However, the protocol accurately measures the immobilization-resistant charge movements that are known to be required for skeletal-type EC coupling [31, 40]. Therefore, the γ1 subunit is unlikely to play a critical role in the activation of SR Ca2+ release, in agreement with a recent report [32]. However, shifts in voltage dependence below the limit of resolution (see Materials and Methods) and effects on charge movement and Ca2+ release inactivation cannot be completely ruled out. The γ1 knockout mice provide a unique resource to understand the function of this protein in myotubes in molecular detail.
Materials and MethodsGene targeting
Bacteriophage clones containing portions of the γ1 gene were isolated from a mouse 129Sv genomic library (Stratagene, La Jolla, CA #946305). To inactivate the γ1 gene, a 4.4 kb BamHI/XhoI fragment upstream of exon 1 was subcloned into a multiple cloning site 5' of the neo and Herpes virus TK gene that was in pBluescript (Stratagene). A second 2.6 kb EcoRI/BamHI fragment 3' of exon 4 of the γ1 gene was cloned into a site between the neo and TK genes (Fig. 1Aii). Thus, the targeting vector contains 4.4 kb of identity with the native γ1 gene 5', and 2.6 kb of identity 3' of the neo cassette, respectively. The targeting vector also contains the HSV TK gene to allow for negative selection of non-recombinants. A unique BamHI restriction site was used to linearize the plasmid prior to its introduction into mouse ES cells by electroporation. The γ1 gene contains 4 exons, and correct targeting results in the deletion of all 4 exons, which will inactivate the gene (Fig. 1Aiii). Probe 2 was used for identification of targeted clones [41]. 5 μg of the targeting vector was electroporated into 5 × 106 AB1 ES cells. G418 and FIAU (1-(2'-deoxy-2'-fluoro-γ-D-arabinofuranosyl)-5-iodouracil) resistant clones were analyzed by Southern blotting after digestion with either EcoRI and hybridization to probe 1 or digestion with HindIII and hybridization to probe 2 (Fig. 1). Of three clones targeted on the 5' end, two also were targeted on the 3' end. These were expanded and used to produce germline chimeras. Mice homozygous for the targeted allele were viable, with no apparent abnormalities. Mice heterozygous and homozygous for the targeted mutation are referred to as γ1 het and γ1 null, respectively.
Western blots
Total microsomes were prepared from the forelimb and hind limb muscles of control, γ1 het, and γ1 null adult mice as previously described [42]. Total microsomes were washed twice with 0.6 M KCl buffer and 100 μg of total membrane protein was applied to a 5-15% SDS polyacrylamide gel. Proteins were transferred to PVDF membranes and analyzed with either anti-α1S mAb IIC12 antibody [43], or anti-α2-δ protein G-purified guinea pig antibody #1 [13], or anti-β1a affinity-purified sheep #6 antibody [44], or with anti-γ1 affinity-purified guinea pig #16 antibody and the appropriate secondary antibodies [11]. The subunits were visualized using SuperSignal ECL reagent (Pierce, Rockford, IL). The images were captured on a Chemi-Imager (Alpha Innotech, San Leandro, CA) set to a level just below saturation.
Primary cultures of mouse myotubes
Primary cultures were prepared from hind limbs of day 18 embryos (E18) as described previously [45]. Dissected muscles were incubated for 9 min at 37°C in Ca2+/Mg2+-free Hanks balanced salt solution (in mM: 136.9 NaCl, 3 KCl, 0.44 KH2PO4, 0.34 NaHPO4, 4.2 NaHCO3, 5.5 glucose, pH 7.2) containing 0.25% (w/v) trypsin and 0.05 % (w/v) pancreatin (Sigma, St. Louis, MO). Mononucleated cells were resuspended in plating medium containing 78% Dulbecco's modified Eagle's medium (DMEM) with low glucose (Life Technologies, Rockville, MD), 10 % horse serum (HS, Sigma, St. Louis, MO), 10 % fetal bovine serum (FBS, Sigma, St. Louis, MO), 2% chicken embryo extract (CEE, Life Technologies, Rockville, MD) and plated on plastic culture dishes coated with gelatin at a density of ∼ 1 × 104 cells per dish. Cultures were grown at 37°C in 8 % CO2. After the fusion of myoblasts (∼ 7 days), the medium was replaced with a FBS-free medium (88.75 % DMEM, 10 % HS, 1.25% CEE) and cells were incubated in 5% CO2. All media contained 0.1 % v/v penicillin and streptomycin (Sigma, St. Louis, MO). All measurements were made when the cells had been in culture for 7-10 days. During this time the maximal current density did not change [45].
Electrophysiological measurements
Whole-cell recordings were performed as described previously [29] using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Linear capacity and leak currents were compensated with the circuit provided by the manufacturer. Series resistance, Rs, was compensated up to the point of amplifier oscillation with the Axopatch circuit. Considering all cells in the present study, the linear capacitance (mean ± SE) of control, γ1 null, and γ1 het myotubes was 256 ± 25 pF (38 cells), 246 ± 16 pF (34 cells) and 223 ± 24 (23 cells), respectively. Rs of the same cells before analog compensation was 7 ± 0.5 MΩ, 7.2 ± 0.4 MΩ, and 6.3 ± 0.5 MΩ, respectively. Following compensation, the effective Rs was <2 MΩ in all cases measured, although this was not measured systematically. Considering a mean cell capacitance of 240 pF, a typical maximum current density of 5 to 10 pA/pF (see Fig. 3B), and Rs (effective)<2 MΩ, the voltage error was <2.4 to <4.8 mV. All experiments were performed at room temperature. Patch pipettes had an open tip resistance of 1-2 MΩ when filled with the pipette solution. Ca2+ currents were measured from a holding potential of -40 mV. Test pulses of 500 ms in 5 mV increments ranged from -40 mV to +60 mV. Ca2+ conductance vs. voltage curves were obtained by extrapolation of the Ca2+ current to the reversal potential of the cell. To measure inactivation, cells were held at -80 mV, then stepped to a series of depolarizing (pre-pulse) voltages, from -70 mV to +70 mV in 10 mV increments, for 2.5 seconds to promote inactivation, then stepped to -50 mV for 100 ms to close non-inactivated channels, then stepped for 200 ms to the test potential of +20 mV to measure Ca2+ currents, then stepped to the -80 mV holding potential for 5 seconds to permit recovery from inactivation. Charge movement was measured using a protocol that eliminated immobilization-sensitive components [15, 28, 29]. Voltage was stepped from a holding potential of -80 mV to -35 mV for 750 ms, then to -50 mV for 5 ms, then to test potential for 25 ms, then to -50 mV for 30 ms and finally to the -80 mV holding potential. Subtraction of the linear component was assisted by a P/4 procedure following the pulse paradigm listed above. P/4 pulses were separated by 500 ms and had a duration of 25 ms. Previous studies showed this choice of pulse protocol and internal and external solutions resulted in nifedipine-sensitive charge movements of a magnitude ∼ 5 fold larger than those detected in dysgenic myotubes lacking α1S [29].
Solutions
The external solution in all cases was (in mM) 130 TEA methanesulfonate, 10 CaCl2, 1 MgCl2, 10 HEPES titrated with TEA(OH) to pH 7.4. For Ca2+ transients and Ca2+ currents, the pipette solution was (in mM) 140 Ca2+-Aspartate, 5 MgCl2, 0.1 EGTA (Ca2+ transients) or 5 mM EGTA (Ca2+ currents), and 10 MOPS-CsOH pH 7.2. For charge movements, the pipette solution was (in mM) 120 NMG (N-methyl glucamine)-Glutamate, 10 HEPES-NMG, 10 EGTA-NMG pH 7.3. For charge movements, the external solution was supplemented with 0.5 mM CdCl2 and 0.5 mM LaCl3 to block the Ca2+ current and 0.05 mM TTX to block residual Na+ current.
Ca2+ transient measurements
Ca2+ transients were measured using a confocal microscope in line-scan mode as described previously [46, 47, 48]. Cells were loaded with 4 mM fluo-4 (fluo-4 acetoxymethyl (AM) ester, Molecular Probes, Eugene, OR) for 20 to 40 min at room temperature. Stocks of fluo-4 (1 mg/ml) were made in DMSO and stored frozen. All experiments were performed at room temperature. Cells were viewed with an inverted Olympus microscope with a 20X objective (N.A. = 0.4) and a Fluoview confocal attachment (Olympus, Melville, NY). The 488 nm spectrum line necessary for fluo-4 excitation was provided by a 5 mW argon laser attenuated to 20% with neutral density filters. The fluorescence intensity, F, was calculated by densitometric scanning of line-scan images and was averaged over the entire width of the cell. The background fluorescence intensity (Fo) was averaged in the same manner from areas of the same image prior to the voltage pulse. The fluorescence unit ΔF/Fo corresponds to (F-Fo)/Fo. The limit of fluorescence detection, based on microscope settings and the average resting fluo-4 fluorescence, was ∼ 0.1 ΔF/Fo units. That is, we could detect a ∼ 10% change above the cell resting fluorescence. Using a pseudo-ratiometric equation for estimating the cytosolic free Ca2+ [48] and assuming a resting free Ca2+ of 100 nM [48], the nominal limit of resolution of free Ca2+ change was ∼ 200 nM.
Data and error analysis
For each cell, or for the population average, the voltage dependence of charge movements (Q), Ca2+ conductance (G), and peak intracellular Ca2+ (ΔF/Fo) were fitted according to a Boltzmann equation A = Amax/(1+exp(-(V-V1/2)/k)) where Amax was either Qmax, Gmax or ΔF/Fomax; V1/2 is the potential at which A = Amax/2; and k is the slope factor. For a fit of steady-state inactivation, the term (V-V1/2) in the equation above was replaced by (V1/2-V). Ca2+ conductance for each cell was computed from the extrapolated reversal potential and the maximal Ca2+ current at each voltage. In all figures, the symbols and error bars correspond to the population mean ± 1 SEM. The curves correspond to a Boltzmann fit of the mean. In addition to Student t-tests reported in Table 1, one-way ANOVA tests were conducted among the 11 sets of Boltzmann parameters of control, γ1 het and γ1 null cells of Table 1. The null hypothesis (equality of means) was not rejected at a level of significance p<0.05 except in the two cases also identified by t-tests, namely Gmax and V1/2 inactivation. An F-test, which formally tests for the difference between the variances of independent populations, was used to determine which of the fitted Boltzmann parameters (Amax, V1/2, k) had the largest variance, and thus the largest experimental error. The ranking order in variance error was V1/2>Amax>k. Furthermore, the error in V1/2 was the largest for the fluorescence measurement and the smallest for the conductance measurement. This result was expected since the size of the voltage step used in the V1/2 determination was 5 mV for conductance, 10 mV for charge movements, and 20 mV for fluorescence. The number of voltage steps in fluorescence measurements was kept to a minimum to avoid rundown of the Ca2+ transient due to photobleaching. The limit of resolution of V1/2, the parameter with the largest error, was estimated with a z-test which formally tests for the difference between the mean of a population and a hypothetical mean. We determined the smallest difference between a experimental and a hypothesized V1/2 resolvable at a level of statistical confidence p<0.05. In the calculations, we used the data with the largest V1/2 variance for each measurement (conductance, charge movement and fluorescence). Also, we assumed a hypothetical S.D. equal to the experimental S.D. The estimated limits of statistical resolution were ∼ 3.9 mV for conductance, ∼ 6.5 mV for charge movements, and ∼ 8.5 mV for fluorescence measurements. Thus, a difference in population mean V1/2 of less than the indicated limit would not be statistically significant, given the variance and the number of determinations of the present study. Statistical analyses were performed with Analyse-it software (Leeds, UK)
Supported by National Institutes of Health Grant HL-47053 (R.C, P.A.P., and R.G.G.), AR46448 (R.C.), National Science Foundation IBN-9319540 (R.G.G. and P.A.P) and by a predoctoral fellowship from Wisconsin Heart Association (C.A.A). K.P.C. is an investigator of the Howard Hughes Medical Institute.
Dopamine was shown to stimulate the perivitelline fluid secretion by the albumen gland. Even though the albumen gland has been shown to contain catecholaminergic fibers and its innervation has been studied, the type of catecholamines, distribution of fibers and the precise source of this neural innervation has not yet been deduced. This study was designed to address these issues and examine the correlation between dopamine concentration and the sexual status of snails.
Results
Dopaminergic neurons were found in all ganglia except the pleural and right parietal, and their axons in all ganglia and major nerves of the brain. In the albumen gland dopaminergic axons formed a nerve tract in the central region, and a uniform net in other areas. Neuronal cell bodies were present in the vicinity of the axons. Dopamine was a major catecholamine in the brain and the albumen gland. No significant difference in dopamine quantity was found when the brain and the albumen gland of randomly mating, virgin and first time mated snails were compared.
Conclusions
Our results represent the first detailed studies regarding the catecholamine innervation and quantitation of neurotransmitters in the albumen gland. In this study we localized catecholaminergic neurons and axons in the albumen gland and the brain, identified these neurons and axons as dopaminergic, reported monoamines present in the albumen gland and the brain, and compared the dopamine content in the brain and the albumen gland of randomly mating, virgin and first time mated snails.
Background
Dopamine is commonly found in the molluscan central nervous system (CNS). In some gastropods, dopamine has been implicated in the regulation of many physiological activities such as feeding in Helisoma trivolvis[1,2], Limax maximus[3], Aplysia[4,5] and Lymnaea stagnalis[6], respiration in L. stagnalis[7,8], gill movement in Aplysia californica[9,10], and egg laying behaviour in L. stagnalis[11]. Saleuddin et al. [12] have shown that dopamine stimulates protein secretion from Helisoma duryi albumen gland.
The albumen gland in pulmonate snails is an accessory gland of the female reproductive tract (Fig 1). It synthesizes and secretes perivitelline fluid (PVF), which is composed mainly of proteins and polysaccharides [13]. Mature oocytes are released by the ovotestis and travel via the hermaphroditic duct into the carrefour, where the albumen gland duct empties. In the carrefour the eggs are fertilized and then are coated with the PVF. The importance of the PVF lies in the fact that it is a major nutrients source for the developing embryo since the oocytes themselves contain very little vitellogenic protein. [14,15]. The secretion of the PVF and the arrival of oocytes at the carrefour must be synchronized, suggesting a precise control of the PVF release [16]. de Jong-Brink and Goldschmeding [17] identified a neuronal plexus in the duct of the albumen gland and the carrefour, which suggested that a nervous mechanism may be involved in the control of the PVF release. Furthermore, catecholamine-containing axons were identified in the albumen gland, carrefour and some other reproductive organs of L. stagnalis and other species of pulmonate snails [18,19]. It was also shown that the PVF secretion by the albumen gland could be stimulated by forskolin, cAMP, brain extract [16] and dopamine [12].
A diagram of the reproductive system of H. duryi
Dopaminergic neurons have been localized in the CNS of some snails such as L. stagnalis[20-23], Helix pomatia[24,25], Aplysia californica[5,26] and Planorbis corneus[27], and they were mapped in the buccal ganglia of Helisoma trivolvis[2,28] but not other ganglia. In H. trivolvis, Trimble et al. [28] showed that 3H-dopamine accumulates specifically in the buccal, cerebral, pedal, left parietal and visceral ganglia, and the left pedal ganglion contains a greater amount of dopamine than the right. Furthermore, using glyoxylic acid Harris and Cottrell [29] and Syed et al. [30] identified a giant dopaminergic neuron in the left pedal ganglion in the CNS of H. trivolvis.
The purpose of our study was to establish possible sources of dopamine that may be regulating the secretion of the PVF from the albumen gland. Although Brisson and Collin [18] showed the presence of catecholamine-containing neurons in the albumen gland and carrefour they neither specified the type of catecholamines nor their distribution within the albumen gland and the carrefour. Furthermore, since the albumen gland is known to be innervated by the CNS [17] and the localization of dopaminergic neurons in the CNS of Helisoma has not been studied, we focused on these investigations. In this study we describe distribution of catecholaminergic neurons and their axons in the H. duryi CNS and albumen gland utilizing well accepted methods that employ anti-tyrosine hydroxylase (TH) IgG and glyoxylic acid [3,20-22,24,25,31,32]. Tyrosine hydroxylase is an enzyme in the pathway of catecholamine synthesis; it converts tyrosine into DOPA and can be used to localize neurons producing catecholamines. The application of glyoxylic acid converts dopamine and other catecholamines into intensely fluorescent 2-carboxymethyl-dihydroisoquinoline derivatives [33].
Using high performance liquid chromatography with electrochemical detection (HPLC-ED) we report monoamines present in the albumen gland and the CNS, identify catecholaminergic neurons found in the albumen gland and CNS as dopaminergic, report the amount of dopamine present in these organs in randomly mating snails and compare it to that of virgin and first time mated snails to determine whether the PVF secretion caused any changes in the amount of dopamine present in these organs. We compared dopamine quantity in snails of different sexual status because research in this lab identified differences between virgin and randomly mating snails, such as differences in egg mass production and synthetic activity of the albumen gland [34]. Furthermore, known centers involved in regulating reproduction (endocrine dorsal bodies and neurosecretory caudodorsal cells) also show changes after mating [35,36]. Following from the above data we formulated a hypothesis: after mating eggs are fertilized and coated with the PVF and since dopamine stimulates the PVF release it might be spent in animals that have mated and formed an egg mass. If this hypothesis is true a difference in dopamine quantity in the albumen gland would be observed between virgin and first time mated animals. Furthermore, since the albumen gland is innervated by the CNS, dopamine is either synthesized centrally in neuronal cell bodies in the CNS and then transported along axons to the albumen gland or locally in axons in the albumen gland. Our experiments were designed to test which mechanism is valid. If first time mated animals have lower levels of dopamine in the CNS the first mechanism applies. Randomly mating animals were treated as control.
Results
In H. duryi both the albumen gland (Figs. 2a,2b,2c, 3a,3b,3c) and the CNS (Figs. 4a,4b,4c, 5a,5b,5c, 6a,6b,6c, 7a,7b) contained dopaminergic neurons and fibers. In the CNS their number, size and location are recorded in Table 1. Unless otherwise stated, the mapping illustrates neurons that were found deep in the ganglia. Since HPLC analysis revealed that dopamine is the only catecholamine present in the albumen gland and there is very little norepinephrine in the brain compared with dopamine, the structures that were positive when probed with glyoxylic acid and anti-TH IgG are presumed to contain dopamine.
Confocal images of the anti-TH IgG treated tissues. TH-IR neurons are indicated with arrowheads.a: a part of the albumen gland (AG) with attached carrefour (cf). Lumen (In) is clearly visible within the carrefour. TH-IR nerve endings form a tract (arrow) that originates in the carrefour. TH-IR neuronal cell bodies are seen within the tract b: a part of the albumen gland with attached carrefour. Closely packed together TH-IR cell bodies form walls of the carrefour. c: a part of the albumen gland with TH-IR axons forming the nerve tract with TH-IR neuronal cell bodies visible within the tract. Scale bar = 100 μm.
Fluorescence micrograph of glyoxylic acid treated tissues. a: the albumen gland. Dopaminergic nerve endings form a tract. Neuronal cell bodies are seen within the tract (arrowhead). b: the albumen gland. Dopaminergic nerve endings form varicosities (arrowheads). c: a part of the albumen gland (AG) with attached carrefour (cf). Dopaminergic nerve endings form a tract (arrowhead) that originates in the carrefour. Scale bar = 100 μm.
Confocal images of the anti-TH IgG treated CNS. TH-IR neurons are indicated with arrowheads. a: the CNS without cerebral ganglia. The giant dopaminergic neuron (LPeD1) is in the left pedal ganglion. TH-IR axons are visible in the ganglia and some nerves. A thick LPeD1 axon is well defined. It passes through the left pleuro-pedal and pleuro-parietal connectives giving off branches and leaving the left parietal ganglion in left parietal nerve (1 pm). b: the CNS with a part of the cerebral ganglia A TH-IR neuron is present in the left parietal ganglion. TH-IR axons are clearly visible in the ganglia and some nerves. c: the CNS without cerebral ganglia. The giant dopaminergic neuron (LPeD1) is in the left pedal ganglion. TH-IR axons are clearly visible in the ganglia and some nerves. A symmetrical group of three TH-IR neurons is clearly visible in the right pedal ganglion. A single TH-IR neuron is present in the visceral ganglion. The structures shown are CG-cerebral ganglia, Pd-pedal ganglia (L-left, R-right), Pl-pleural ganglia (L-left, R-right), Pr-parietal ganglia (L-left, R-right), V-visceral ganglion, cc-cerebral connective, cpc-cerebropedal connective, in-intestinal nerve, pn – pedal nerve. Scale bar = 100 μm.
Confocal images of the anti-TH IgG treated CNS. TH-IR neurons are indicated with arrowheads. a, b and c are a series of confocal sections through the cerebral ganglia with a being the first section close to the dorsal surface of the ganglia and c being close to the ventral surface. TH-IR axons run through the cerebral commissure (cc) and exit the cerebral ganglia in tentacular (tn), median lip (mm) penis (pen) and frontolateral (fn) nerves. TH-IR neurons are located in the vicinity of axons. Neither TH-IR cells nor axons were found in dorsal bodies (DB). Scale bar = 100 μm.
Fluorescence micrograph of glyoxylic acid treated tissues. Blue fluorescing cells are indicated with arrowheads. a: The left parietal (LPr) and visceral (V) ganglia. A single blue-green fluorescing neuron is visible in the visceral ganglion. Blue fluorescing fibers are visible in the intestinal (in) and left parietal (lprn) nerves, viscero-parietal and pleuroparietal connectives. b: the pedal ganglia. The LPeD1 is clearly visible in the left pedal ganglion. Groups of small dopaminergic neurons are located on the periphery. Blue fluorescing fibers are visible in the pedal nerves (pn) and cerebro-pedal connectives (cpc). c: the cerebral ganglia. Cerebral commissure (cc), cpc, median lip (mm), frontolateral (fn) and tentacular (tn) nerves contain blue fluorescing fibers. The structures shown are Pd-pedal ganglia (L-left, R-right), Pl-pleural ganglia (L-left, R-right), Pr-parietal ganglia (L-left, R-right). Scale bar = 100 μm.
Fluorescence micrograph of glyoxylic acid treated buccal ganglia. a: dorsal surface b: ventral surface. The structures shown are B20 – neuron that is similar to B20 neuron in Aplysia[4], CBC – cerebro-buccal connective, ET – esophageal trunks, HBN-heterobuccal nerve, N1a – neuron that is involved in the control of feeding in Helisoma[2], PBN – posterior buccal nerve, VBN – ventral buccal nerve. Scale bar = 100 μm.
The distribution of dopaminergic neurons and axons in the CNS of H. duryi. Filled circles indicate neurons that showed green fluorescence after glyoxylic acid treatment but were not immunostained after anti-TH IgG treatment. Gray circles were both green fluorescent after glyoxylic acid treatment and TH-IR. The axons shown are those probed with anti-TH IgG. a: dorsal surface, b: ventral surface. Anterior is up. The structures shown are BG-buccal ganglia, CG-cerebral ganglia (L-left, R-right), DB-dorsal bodies, Pd-pedal ganglia (L-left, R-right), Pl-pleural ganglia (L-left, R-right), Pr-parietal ganglia (L-left, R-right), V-visceral ganglion. ao – aorta, an – anal nerve, fn – frontolateral nerve, in – intestinal nerve, lprn – left parietal nerve, mln – median lip nerve, pen – penis nerve, rprn – right parietal nerve, tn – tentacular nerve.
Comparison of the distribution and size of tyrosine hydroxylase immunolabeled and glyoxylic acid- induced green fluorescent neurons in ganglia of the H. duryi CNS.
Ganglia
Tyrosine hydroxylase-imniunoreactive
Green fluorescing
Size (μm)
Number(Left+Right)
Size (μm)
Number min-max (Left+Right)
Buccal
Not treated with anti-TH IgG
5–15
4
15–25
76–86
Cerebral
25–50
16
25–50
24–28
Pedal
-
-
5–15
10–12
15–25
6
15–25
6
75+
1
75+
1
Parietal-pleura
1–25–50
2
25–50
2
visceral gangli
a
Total
15–75+
25
5–75+
123–139
The fixation preserved the tissues probed with antibodies well. Distribution of dopaminergic neurons in H. duryi CNS is diagrammed in Figures 8 and 9, based on examination of 30 brains stained with glyoxylic acid and 20 brains probed with anti-TH IgG. Although some variation was found in brains stained with glyoxylic acid in the number and position of neurons, the distribution of neurons was consistent in brains probed with anti-TH IgG. The TH-immunolabeling was intense and contrasted well with the clear background causing neuronal cell bodies and axons to appear well defined. Neurons stained with glyoxylic acid were of intense fluorescence and generally contrasted well with background except for neurons in the cerebral ganglia where the presence of blue-green fluorescing axons found in the nerves leaving the cerebral ganglia interfered with the identification of the dopaminergic neurons (Fig. 6c). Blue-green fluorescing axons in the brain were of low contrast and staining intensity. The buccal ganglia were stained with glyoxylic acid only.
The distribution of dopaminergic neurons and axons in the buccal ganglia of H. duryi.a: dorsal surface, b: ventral surface. B20 – neuron that is similar to B20 neuron in Aplysia,[4], CBC – cerebro-buccal connective, ET – esophageal trunks, HBN-heterobuccal nerve, N1a – neuron that is involved in the control of feeding in Helisoma[2], PBN – posterior buccal nerve, VBN – ventral buccal nerve.
The distribution of dopaminergic fibers in H. duryi albumen gland is diagrammed in Figure 10, based on examination of 20 albumen glands stained with glyoxylic acid and 15 albumen glands probed with anti-TH IgG. The blue-green fluorescence was intense and contrasted well with the clear background whereas the intensity of TH-immunoreactivity was weaker. The number of TH-immunoreactive (TH-IR) neurons and axons found in the albumen gland was less compared to the number of blue-green fluorescing neurons and axons.
The distribution of dopaminergic neurons and axons in the albumen gland of H. duryi (not to scale). Dopaminergic nerve endings form a uniform network that is consistent across all parts of the albumen gland (AG) except for the central region, where the axons are parallel forming a nerve tract. Neuronal cell bodies (open circles) are visible in the vicinity of the tract. The nerve tract originates from the carrefour (cf). A great number of closely packed together dopaminergic cells (open circles) make up walls of the carrefour.
Mapping of tyrosine hydroxylase-immunoreactive neurons and comparison with glyoxylic acid-induced blue-green fluorescenceBuccal ganglia
Altogether 48–54 blue-green fluorescent neurons on the dorsal side (Fig. 7a) and 32–36 on the ventral side were observed (Figs. 7b). All the neurons occurred in bilaterally symmetrical groups. A single 15 μm diameter neuron was found off the center towards the buccal commissure and could be seen from both ventral and dorsal surfaces. Its location suggests its similarity to the B20 neuron in Aplysia, which is involved in feeding behaviour [4]. Other neurons were smaller (5–10 μm). On the dorsal surface a single neuron was located near the posterior buccal nerve and another single neuron was at the root of the ventral buccal nerve. The location of the latter neuron suggests that it is the same neuron as the N1a neuron in H. trivolvis[2], and the B65 neuron in Aplysia californica[5], which evoke the buccal motor program. Other neurons on both dorsal and ventral surfaces were located in groups composed of three to seven neurons (Fig. 7a,7b). The buccal commissure, cerebro-buccal connective, heterobuccal nerve, ventral buccal nerve, posterior buccal nerve and esophageal trunks contained blue-green fibers (Fig. 7a,7b). The number of fibers was high in all nerves except the posterior buccal nerve, which contained only a few axons. The buccal ganglia were stained with glyoxylic acid only due to the difficulty of manipulating them for the antiserum treatment.
Cerebral ganglia contained symmetrically distributed neurons, 20–30 μm in diameter. Anti-tyrosine hydroxylase antiserum revealed 16 TH-IR neurons in both left and right ganglia (Fig. 5a,5b,5c). They made up three groups: a pair of neurons was located at the root of the tentacular nerve, another pair was located at the root of the frontolateral nerve (these neurons were located close to the dorsal surface of the ganglia), and a group of 4 neurons was located in the center of the ganglia. The number of blue-green fluorescing neurons varied from 24 to 28. They were distributed in similar manner as TH-IR neurons (Fig. 6c). The cerebral commissure contained a great number of TH-IR fibers that passed through the length of the ganglia and exited in the frontolateral, median lip, penis and tentacular nerves (Figs. 5a,5b,5c). All the neurons were located in the vicinity of the TH-IR fibers. Glyoxylic acid revealed dopaminergic fibers in the frontolateral nerve, median lip nerve, tentacular nerve and cerebral commissure but no fibers were observed within the cerebral ganglia (Fig. 6c). Neither dopaminergic neurons nor axons were found in the dorsal bodies.
Pedal ganglia
The left and right pedal ganglia contained a total of 7 dopaminergic neurons that were both TH-IR and blue fluorescent (Figs. 4a,4b,4c, 6b). Six dopaminergic neurons (10–20 μm in diameter) made up a bilaterally symmetrical group located near the roots of nerves leaving the pedal ganglia. A single giant dopaminergic neuron (LPeD1) was seen in the left pedal ganglion at the pleural side of the ganglion located close to the pedal commissure (Figs. 4a,4c, 6b). The pedal commissure contained a few TH-IR fibers that passed through the length of each ganglion and exited with the nerves leaving the pedal ganglia. Neurons located in the bilaterally symmetrical group were in the vicinity of the nerve fibers. A single thick, intensely stained TH-IR LPeD1 axon passed through the left pleuro-pedal and pleuro-parietal connectives giving off branches and leaving the left parietal ganglion in the left parietal nerve (Figs. 4a,4c). Several TH-IR fibers exited the right pedal ganglion and passed through the right pleuro-pedal, pleuro-parietal, and viscero-parietal connectives giving off branches, and connecting with the fibers coming from the left parietal ganglion before exiting via the anal and intestinal nerves. The cerebro-pedal connective also contained TH-IR fibers. In the glyoxylic acid treated brain an additional group composed of 5 small (5–10 μm) neurons was located near the pleuro-pedal connective (Fig. 6b). Glyoxylic acid revealed dopaminergic fibers in the pedal commissure, pleuro-pedal, cerebro-pedal connectives and the nerves leaving the pedal ganglia (Fig. 6b).
Parietal-pleural-visceral ganglia complex
Although a great number of TH-IR fibers were found throughout the parietal-pleural-visceral ganglia complex (see above) no dopaminergic neurons were seen in either right or left pleural ganglia or the right parietal ganglion (Figs. 4a,4b,4c, 6a,6b). The left parietal ganglion contained a single dopaminergic neuron 25–30 μm in diameter located in the path of the TH-IR LPeD1 axon that passed through the complex (Fig. 4b). The visceral ganglion contained a single similar size neuron, also located on the path of dopaminergic fibers going through the complex (Fig. 4c, 6a). With glyoxylic acid treatment no fibers were found within the ganglia, only in the pleuro-pedal, pleuro-parietal, viscero-parietal connectives, the left and right parietal, anal and intestinal nerves.
Albumen gland
Dopaminergic nerve endings formed a uniform network that was consistent across all parts of the gland except for the central region, where the axons were parallel forming a nerve tract (Figs. 2a, 2c, 3a,3b,3c). Neuronal cell bodies were present in the vicinity of the tract. The albumen glands of virgin and mated animals showed a similar distribution of dopaminergic neurons and fibers (not shown). The origin of the nerve tract was traced to the carrefour, which was positive when probed with glyoxylic acid and anti-TH IgG (Figs. 2a,2b, 3c). A great number of dopaminergic cells that were closely packed together made up the walls of the carrefour and dopaminergic axons made up the nerve tract that went into the albumen gland.
HPLC
Among monoamines serotonin and dopamine were present in great amounts whereas norepinephrine was present in insignificant amounts in the CNS of H. duryi (Fig. 11). In the albumen gland dopamine was the only monoamine. The content of dopamine in the CNS and the albumen gland of randomly mating, virgin and first time mated snails is summarised in Figure 12. No significant difference was found when the concentrations of dopamine in the CNS and the albumen gland were compared between the three experimental groups (P>0.05).
Chromatograms showing an example of high performance liquid chromatography with electrochemical detection for the CNS and the albumen gland of randomly mating H. duryi (DA – dopamine, 5-HT – serotonin, NE – norepinephrine, LD – L-Dopa). NE eluted prior LD. a: elution profile of the CNS. Note the NE peak is higher than the LD peak, and the DA and 5-HT peaks show very clearly; b: elution profile of the albumen gland. Note NE and LD peaks are barely visible, 5-HT peak is absent, DA peak is very clear. A distinct peak that elutes prior to serotonin was not identified. Scale bar = 4.8 minutes.
Bar graph showing dopamine content in the central nervous system (CNS) and the albumen gland (AG) of H. duryi measured by high performance liquid chromatography with electrochemical detection and expressed in nmol of dopamine per mg of protein (mean ± SEM), sample size 6. Abbreviations: rm – randomly mating, v – virgin, ftm – first time mated snails.
Discussion
Our work reported here followed the studies of a number of researchers [12,16-19]. We demonstrated the presence of the neuronal cell bodies and their axons in the H. duryi albumen gland and identified them as dopaminergic, described their distribution and possible origin. The analysis of biogenic amines by HPLC demonstrated that dopamine is a major catecholamine in the CNS and the only monoamine in the albumen gland. We measured the amount of dopamine in the albumen gland and the CNS of randomly mating, virgin and first time mated animals. In addition, we mapped dopaminergic neurons in the H. duryi CNS.
Comparison of the distribution of the dopaminergic neurons in the CNS of Helisoma duryi to that in other gastropods
The distribution of dopaminergic neurons and their axons in the CNS of H. duryi is illustrated in Figs. 8 and 9. Since some norepinephrine was detected by HPLC it is possible that some of the mapped neurons contain norepinephrine and not dopamine. However, the amount of norepinephrine detected by HPLC was small compared with dopamine (Fig. 11), and its presence could not have interfered with our data to any great extent. Our conclusion coincides with that of other researchers who detected a significantly greater amount of dopamine compared with norepinephrine in the CNS of some other gastropods [3,24,28].
Dopaminergic neurons have been mapped in the brains of other snails [5,20-27], and they were mapped in the buccal ganglia of Helisoma trivolvis[2,28]. The map of the dopaminergic neurons in the buccal ganglia done by Quinlan et al. [2] is similar to the map obtained in this experiment except for a few additional neurons found in the present study. The locations of some of the neurons identified in the buccal ganglia suggest their similarity to B20, N1a and B65, well studied neurons in other snails [2,5]. Trimble et al. [28] showed that 3H-dopamine accumulates only in the buccal, cerebral, pedal, left parietal and visceral ganglia of the H. trivolvis brain. Their results were consistent with the findings in our study: dopaminergic cells were found only in the buccal, cerebral, pedal, left parietal and visceral ganglia. The left pedal ganglion contained the LPeD1 neuron, which explains why Trimble, et al. [28] found the amount of dopamine in the left pedal ganglion to be greater than in the right pedal ganglion, and confirms findings of other researchers, who identified a giant neuron in the left pedal ganglion of H. trivolvis[29,30].
Neither dopaminergic neurons nor their fibers were found in the dorsal bodies even though the dorsal bodies are located in close proximity to the cerebral ganglia and were shown to be involved in the regulation of the albumen gland activity [37,38]. These findings are contrary to the findings of Elekes et. al. [22] who observed small dopaminergic cells and Werkman et. al. [23] who found dopaminergic axons within the dorsal bodies of L. stagnalis.
The number of dopaminergic neurons in the CNS of L. stagnalis and Helix pomatia was much greater than that found in H. duryi; however, the majority of neurons were located in the same ganglia. Distribution of dopaminergic neurons in the CNS of H. duryi that we mapped was almost identical to that of Planorbis corneus in the number of neurons, their size and position in the ganglia. The giant dopamine cell was found in the right pedal ganglion in L. stagnalis (but not in Helix pomatia or Aplysia californica) whereas in H. duryi and Planorbis corneus it was found in the left pedal ganglion. This difference was described in detail by other researchers [39-42] and can be explained by the fact that L. stagnalis is dextrally spiraled whereas H. duryi and Planorbis corneus are sinistrally spiraled leading to the restriction of some organs to one side of the snail's body. Dopaminergic fibers were found in abundance throughout the brain of H. duryi running in similar fashion as in L. stagnalis. Helix pomatia and Planorbis corneus.
Comparison of the distribution of the TH-IR neurons to the distribution of glyoxylic acid-induced blue-green fluorescing neurons
In the cerebral and pedal ganglia a few blue-green fluorescing cells did not correspond to TH-IR neurons. A plausible explanation for this discrepancy can be the overestimation of the number of blue-green cells in the cerebral ganglia due to the presence of blue-green nerves leaving the cerebral ganglia that may have interfered with identification of the dopaminergic neurons (Fig. 6c). In the pedal ganglia the neurons that TH antiserum failed to recognize were small and possibly contained insufficient amount of the enzyme for the antibodies to bind to. Nevertheless, most of the identified neurons were consistent in their location and size after the use of both techniques. Both techniques showed that all the nerves leaving the brain contained some dopaminergic fibers. More nerve fibers were localized with TH antiserum than with glyoxylic acid and they were of better contrast.
Comparison of the distribution of the dopaminergic axons in the albumengland of mated and virgin Helisoma duryi
Brisson and Collin [18] and Brisson [19] demonstrated the presence of catecholamine-containing axons in the albumen gland, carrefour and some other reproductive organs in L. stagnalis and other species of pulmonate snails and concluded that the catecholaminergic system might be important to the carrefour region for transport and orientation of genital fluxes. In our study using HPLC-ED we provide proof that neurons and axons present in the albumen gland and the carrefour are dopaminergic and we describe their distribution in these organs. We observed a tract of dopaminergic axons that runs through the central region of the albumen gland and originates from the carrefour (Figs. 2a,2c, 3a,3c, 10). We also detected the nerve cell bodies in the vicinity of the tract. In other parts of the gland dopaminergic axons extended from the central tract and formed varicosities. de Jong-Brink and Goldschmeding [17] showed that in L. stagnalis the gonadal branch of the intestinal nerve, which originates from the visceral ganglion, innervates the gonad, carrefour, albumen gland and its secretory duct. The albumen gland and its secretory duct also receive intrinsic nervous input derived from neurons in the carrefour region. Therefore, we suggest that the tract and the cell bodies observed within the tract could constitute the intrinsic innervation of the albumen gland that originates from the carrefour. Furthermore, a dopaminergic neuron was found in the visceral ganglion of the CNS of H. duryi, and since the albumen gland is innervated by the intestinal nerve that originates from the visceral ganglion this could be the neuron that controls the albumen gland activity. However, in order to establish this premise conclusively a thorough tracing of the axon of the dopaminergic neuron found in the visceral ganglion is required.
HPLC analysis revealed that in the CNS of H. duryi serotonin and dopamine were major monoamines whereas norepinephrine was present in insignificant amounts (Fig. 11). These data are in accordance with findings of other researchers [3,24,28]. In the albumen gland dopamine was the only monoamine, which is in agreement with research conducted in our laboratory by Dr. S. Mukai (unpublished results) who demonstrated that of all neurotransmitters tested only dopamine caused a significant increase in protein secretion by the albumen gland.
No marked difference was noticed when the distribution and the number of dopaminergic nerve endings in the albumen glands of randomly mating and virgin animals were compared. In addition, the dopamine concentration in both the CNS and the albumen gland of randomly mating, virgin and first time mated animals (Fig. 12) did not differ significantly between experimental groups. Since the axons of the dopaminergic neurons stained positively with anti-TH IgG and TH is an enzyme in the dopamine synthetic pathway, it is reasonable to assume that dopamine is synthesized in the axons of these cells and therefore is not transported from the cell bodies in the CNS. As a result it is perhaps not surprising that dopamine levels in the CNS did not change much following egglaying. A possible explanation for the lack of difference in the amount of dopamine present in the albumen gland is that HPLC measured the dopamine content only and as soon as dopamine is used to signal the PVF release it is either resynthesized or taken back into nerve terminals by a reuptake mechanism, which commonly occurs in molluscs [43]
Conclusions
Our results represent the first detailed studies regarding the catecholamine innervation and quantitation of neurotransmitters in the albumen gland. Earlier research in our laboratory established the importance of dopamine in the snail's reproductive system, especially in the regulation of protein secretion by the albumen gland. Our data confirmed and extended these findings by localizing the catecholaminergic neurons and axons in the albumen gland and the CNS of H. duryi using glyoxylic acid and anti-TH IgG. Using HPLC-ED we identified these neurons and axons as dopaminergic, reported monoamines present in the albumen gland and the CNS, and compared the dopamine content in the CNS and the albumen gland of randomly mating, virgin and first time mated snails.
The distribution of dopaminergic neurons in the brain of H. duryi was similar to that of other gastropod snails. Nerve fibers found in the albumen gland formed a nerve tract with neuronal cell bodies present in the vicinity of the tract in the central region, and a uniform network in other areas of the gland. The nerve tract originated from the carrefour. The visceral ganglion contains a dopaminergic neuron and its axon runs within the intestinal nerve, suggesting that this neuron could control the albumen gland activity. No difference in dopamine quantity in the CNS and the albumen gland of snails of different mating times was observed, which can be explained by the local dopamine production within the albumen gland axons or by a reuptake mechanism.
Materials and MethodsAnimals
Randomly mating H. duryi were taken from the stock population, which was raised in 6 L tanks containing snails of different maturity, from recently hatched youngsters to sexually mature adults. Virgin snails were isolated 2–3 weeks after hatching and placed in separate plastic containers. All animals were kept at 22°C and photoperiod of 16L/8D. Snails were maintained in dechlorinated water, which was changed once a week, and fed boiled lettuce three times per week, occasionally supplemented with fish food pellets. Virgin animals were allowed to mate and within 24 hours after mating were dissected. This group of animals was termed first time mated snails.
Immunocytochemistry
The brain and the albumen gland of mated H. duryi were dissected out in HEPES-buffered Helisoma saline (51.3 mM NaCl, 1.7 mM KCl, 4.1 mM CaCl2, 1.5 mM MgCl2, 5.0 mM HEPES, ph 7.4, 120 mOsm/L) and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer pH 7 at 4°C for 4 hours. Prior to fixation the albumen gland was subjected to 0.5% protease treatment in phosphate buffer. After fixation tissues were washed for 12 hours in phosphate buffered saline pH 7.2 (PBS) containing 4% Triton X-100 and then processed for immunocytochemistry. The tissues were incubated with a monoclonal mouse anti-TH IgG (Incstar, Stillwater, MN) diluted 1:1000 in PBS with 5% normal goat serum and 4% Triton X-100 for 72 hours at 4°C. Prior to incubation with anti-TH IgG the albumen gland was treated with 0.2% trypsin in PBS and 0.1% CaCl2 for 20 minutes. Then the tissues were incubated in goat anti-mouse IgG (1:100) conjugated to rhodamine (Sigma-Aldrich, Oakville, On) in PBS with 5% normal goat serum and 4% Triton X-100 for 12 hours at 4°C in the dark. Between each step tissues were washed with PBS several times. After several more rinses the tissues were mounted between two coverslips (rectangular 60 × 22 mm and circular 18 mm) in a 3:1 mixture of glycerol and PBS. Prior to mounting a circle was made on each rectangular coverslip with melted sealing wax (50% Vaseline, 50% dental wax) to prevent tissue distortions. Then the coverslips were affixed with adhesive tape to an aluminium slide (80 × 36 × 1 mm) with a 25 mm circular window so that the circular coverslip faced down within the window. In such orientation the preparation could be viewed from both sides. Specimens were viewed and photographed with a Biorad MRC 600 confocal microscope with a Krypton/Argon laser, YHS filterblock, single channel, excitation filter 568 DF10 nm, dichroic reflector 585 DRLP and an emission filter 585 EFLP nm. The location of dopaminergic neurons was drawn onto standardized maps. No staining was observed in the control experiments where the same procedure was followed except for the absence of primary antiserum.
Glyoxylic acid histochemistry
The dopaminergic neurons and their axons in the albumen gland and CNS were visualized with the glyoxylic acid-induced histofluorescence technique. The brain and the albumen gland of mated and virgin H. duryi were dissected out in HEPES-buffered Helisoma saline and pinned in a Sylgard-lined dish containing 220 mM glyoxylic acid and 40 mM HEPES, pH 7.0, and incubated for 30 minutes at 4°C. Then tissues were unpinned and arranged on glass coverslip, dried at room temperature for 30 minutes, heated at 100°C for 4 minutes and mounted in mineral oil between two coverslips (rectangular 60 × 22 mm and circular 18 mm). Prior to mounting the same procedure was followed as described above for the immunohistochemistry. The specimens were viewed and photographed with Zeiss optics (Carl Zeiss Canada, Toronto, Canada) that consisted of a mercury lamp HBO 100 W/2 with modulator, vertical illumination by the III RS condenser containing excitation filters BP 405/14, chromatic splitter FT 420, and a barrier filter LP 418. Emission spectra were measured using a Zeiss continuous interference monochromator with a range of 400–700 nm. With these filters two distinct colours were observed: yellow, indicating serotonin-containing neurons and blue-green, indicating catecholamine-containing neurons [33]. Location of dopaminergic neurons was drawn onto standardized maps. Distribution of serotonin-containing neurons in the Helisoma brain was described elsewhere [31,44,45]. No staining was observed in the control experiments where the same procedure was followed except for the absence of glyoxylic acid.
HPLC-ED
Dopamine was assayed by using HPLC-ED. The CNS and the albumen gland of randomly mating, virgin and first time mated snails were dissected in HEPES-buffered Helisoma saline then placed in 50 μl HPLC buffer (see later) and kept on ice. The mixture was sonicated, centrifuged at 16,000 rpm for 10 minutes at 4°C and then filtered through a 0.22-μm nylon filter. Ten microliters were injected into a Brownlee RP-18 Spheri-5 HPLC column (4.6 mm × 22 cm). The mobile phase, pumped at 0.7 ml/minute, contained 75 mm NaH2PO4 (pH 2.75), 0.3 mM sodium octyl sulfate, 0.05 mM EDTA, 3.5% acetonitrile and 5% methanol. Detection was achieved electrochemically using an ESA model 5100 A detection system coupled to an ESA model 5010 dual coulometric detector (ESA, Inc., Bedford, MA). The first detector was set at 0.025 V to act as a screen and dopamine was detected using the second detector set at 0.2 V at a sensitivity 20 nA. A guard cell inserted before the injector valve was set at 0.5 V to preoxidize possible contaminants in the mobile phase. The output of the second detector was recorded on a Spectra Physics 4270 integrator (Spectra Physics, San Jose, CA), and dopamine was quantified using external standard method. Samples were spiked with dopamine to confirm the identity of the oxidizable substance. In addition, the I-V curve for dopamine was found to be identical for the peak eluting with authentic dopamine. Samples were also spiked with norepinephrine, L-dopa and serotonin to confirm the identity of other peaks. A Bradford protein assay (Bio-Rad, Hercules, CA) was done using human gamma globulin as a standard to determine the amount of protein in each tissue sample. The concentration of dopamine was expressed in nmol of dopamine per mg of protein per tissue.
Abbreviations in the text
CNS – central nervous system
HPLC-ED – high performance liquid chromatography with electrochemical detection
PBS – phosphate buffered saline
PVF – perivitelline fluid
TH – tyrosine hydroxylase
TH-IR – tyrosine hydroxylase-immunoreactive
Abbreviations in the figures
AG – albumen gland
an – anal nerve
bc – buccal connective
B20 – neuron that is similar to B20 neuron in Aplysia
BG – buccal ganglia
CBC- cerebro-buccal connective
cc – cerebral commissure
cf – carrefour
CG – cerebral ganglia
cpc – cerebro-pedal connective
DA – dopamine
DB – dorsal bodies
ET – esophageal trunks
fn – frontolateral nerve
HBN – heterobuccal nerve
5-HT – serotonin
in – intestinal nerve
L or 1 – left
LD – L-Dopa
ln – lumen
LPeD1 – giant dopaminergic neuron present in left pedal ganglion
mln – median lip nerve
N1a – neuron that is involved in the control of feeding in Helisoma[2]
NE – norepinephrine
PBN – posterior buccal nerve
Pd – pedal ganglion
PI – pleural ganglion
plprc – pleuroparietal connective
pn – pedal nerves
Pr – parietal ganglion
pm – parietal nerve
R or r – right
V – visceral ganglion
VBN – ventral buccal nerve
vprc – visceroparietal connective
TH-IR – tyrosine hydroxylase-immunoreactive
tn – tentacular nerve
Acknowledgements
This work was supported by grants from the NSERC of Canada. We wish to thank Dr. B.G. Loughton, Dr. H.R. Khan, Dr. S.T. Mukai and Dr. R.P. Croll for help in preparing the manuscript.
Cyclic nucleotides can relax vascular smooth muscle by mechanisms distal to myosin regulatory light chain (MRLC) phosphorylation. This mechanism, termed relaxation without MRLC dephosphorylation, may be regulated by ser16 phosphorylation of heat shock protein 20 (HSP20).
Results
Confocal imaging of HSP20 in smooth muscle tissues revealed that HSP20 was present throughout the cytoplasm, although some focal regions of the cytoplasm were found to contain more HSP20 than the remaining cytoplasm. The distribution of HSP20 within the cytoplasm was not altered by histamine, forskolin, or nitroglycerin.
Conclusion
Cytoplasmic localization of HSP20 is consistent with a potential function of HSP20 as a regulator of smooth muscle contractile force.
Introduction
In general, contractile stimuli induce smooth muscle contraction by increased myoplasmic [Ca2+], activation of myosin light chain kinase (MLCK), and phosphorylation of myosin regulatory light chains (MRLC) [1]. Phosphorylation of MRLC on ser19 allows the muscle filaments actin and myosin to interact and form cross-bridges, thus generating contraction [2]. In most cases, smooth muscle relaxation proceeds via a reversal of this contraction process: withdrawal of myoplasmic [Ca2+], inactivation of MLCK, and MRLC dephosphorylation [3].
Cyclic nucleotide induced smooth muscle relaxation appears to be more complex. When submaximally stimulated swine carotid artery was treated with nitroglycerin, the relaxation was associated with reductions in myoplasmic [Ca2+] and MRLC phosphorylation [4]. Cyclic nucleotides are known to reduce myoplasmic [Ca2+] by multiple mechanisms (reviewed in [5]).
However, when maximally stimulated swine carotid artery was treated with nitroglycerin, stress decreased significantly, but myoplasmic [Ca2+] and MRLC phosphorylation only transiently decreased so that sustained values did not significantly differ from levels observed in maximally contracted tissues [4]. This phenomenon is termed relaxation without MRLC dephosphorylation, and has been observed with activators of guanylyl cyclase, such as NO, and with phosphodiesterase inhibitors that increase intracellular [cGMP] [6].
Recently, cAMP- and cGMP-dependent relaxation was found to associate with phosphorylation of heat shock protein 20 (HSP20) on ser16[7-9]. We found that a peptide from HSP20 had a sequence homology with troponin I, and that this peptide bound to thin filaments and relaxed skinned smooth muscle [8]. We hypothesized that binding of ser16 phosphorylated HSP20 to the thin filament was responsible for relaxation without MRLC dephosphorylation. If HSP20 regulates contraction by binding to thin filaments, then HSP20, at least during cAMP or cGMP induced relaxation, should colocalize with thin filaments. We therefore evaluated the intracellular localization of HSP20.
Results
If HSP20 regulates contraction by binding to thin filaments, then HSP20, at least during cAMP or cGMP induced relaxation, should colocalize with thin filaments. In 10 μM histamine and 10 μM nitroglycerin treated swine carotid artery, confocal HSP20 immunostaining was present throughout the cell, however, there were some regions with more intense staining (Fig. 1).
Intracellular localization of HSP20 in swine carotid arteries. Representative cross-sectional (top) and longitudinal (bottom) confocal micrographs showing the distribution of HSP20 immunostaining in 10 μM histamine and 10 μM nitroglycerin treated swine carotid artery. Orientation refers to the long axis of the cells. The image is 160 microns wide. The micrographs show that HSP20 immunostaining was present throughout the cell, however, there were regions with more intense staining.
Colocalization analysis revealed that HSP20 staining did not colocalize with the nuclear stain SYT013 (Fig. 2, bottom). These data suggest that HSP20 is primarily a cytosolic protein. Colocalization analysis comparing actin and HSP20 was more complex (Fig. 2, top). Most pixels that stained for actin also stained for varying amounts of HSP20. There were only a few pixels that stained only for HSP20. These data suggest that HSP20 is partially colocalized with actin in the cytoplasm of swine carotid artery.
Co-localization of HSP20 and actin but no co-localization of HSP20 and STY013, a nuclear stain, in swine carotid arteries.A (top left). Representative micrograph shows partial co-localization of actin (green) and HSP20 (red). B (top right). An colocalization plot of staining intensities is shown at right and some regions containing HSP20 (horizontal axis) also contained actin (vertical axis). The presence of pixels in the center and upper right comer indicates that some of the HSP20 and actin colocalized. C (bottom left). Representative micrograph shows no co-localization of STY013, a nuclear stain (green) and HSP20 (red). D (bottom right). A colocalization plot of staining intensities is shown at right and demonstrates that regions containing HSP20 (horizontal axis) staining did not stain for the nucleus (vertical axis). E. Reference for colocalization plots; color indicates number of colocalized pixels. Since these studies were performed on intact smooth muscle tissues with high concentrations of HSP20 and actin, imaging would not be expected to detect individual filaments. Each image is 190 microns wide.
Activation of some regulatory proteins (e.g. protein kinase C and rhoA) induces their redistribution in the cytoplasm [12-14]. Fig. 3 shows confocal HSP20 immunostaining in smooth muscle tissues. There was no apparent intracellular redistribution of HSP20 induced by stimulation with histamine or relaxation of histamine stimulated tissues by nitroglycerin or forskolin.
Lack of stimulus dependence of HSP20 localization in swine carotid arteries. Cross-sectional representative confocal micrographs showing the distribution of HSP20 immunostaining in swine carotid artery that was A) unstimulated (top), B) treated with 10 μM histamine for 30 min (second from top), C) treated with 10 μM histamine for 30 min with 10 μM nitroglycerin added for the last 20 min (third from top), and D) treated with 10 μM histamine for 30 min with 10 μM forskolin present for the last 20 min (bottom). The image is 140 microns wide. The micrographs show that the distribution of HSP20 did not appreciably change during histamine stimulation or relaxation induced by nitroglycerin or forskolin.
Discussion
Overall, these data suggest that HSP20 is a cytosolic protein that does not significantly translocate during cyclic nucleotide-induced relaxation. Localization studies cannot prove that a protein is physiologically relevant. However, cytosolic localization HSP20 would be required if HSP20 were to regulate smooth muscle contractile force by binding to thin filaments. Further studies are required to prove that HSP20 phosphorylation is the mediator of relaxation without MRLC dephosphorylation.
Material and methods
Swine common carotid arteries were obtained from a slaughterhouse and transported at 0°C in physiological salt solution (PSS). PSS contained (mM): NaCl, 140; KCl, 4.7; 3-[N-morpholino] propane sulfonic acid (MOPS) 5; Na2HPO4, 1.2; CaCl2, 1.6; MgSO4, 1.2; D-glucose, 5.6; pH adjusted to 7.4 at 37°C. Dissection of medial strips, mounting and determination of the optimum length for stress development at 37°C was performed as described [10].
Recombinant HSP20 was made in BL21 bacteria from a human HSP20 clone (B53814) obtained from the EST collaboration. Gel purified HSP20 was provided to a commercial vendor who injected a rabbit to make anti HSP20. Sequence of HSP20 was confirmed by mass spectroscopy. After confirmation of an antigenic response, serum was collected and frozen for further use. Prior studies showed that preincubation of the HSP20 antibody with recombinant HSP20 abolished immunstaining of a blot containing swine carotid HSP20 [11].
After appropriate pharmacological treatment, swine carotid artery tissues were fixed in 4% para-formaldehyde in PBS (pH 7.0) at 4°C for 2 hours. Tissues were washed three times in PBS for 30 minutes each, dehydrated in an ascending series of ethanol (30%, 50%, 70%, 85%, 95%, and 100%), infiltrated into Technovit 7100 (Heraeus Kulzer GmbH, Wehrheim, Germany) over night, and embedded in Technovit 7100 at room temperature. Semi-thin sections (2–3 μm) were prepared using a LKB Rotary retracting microtome with glass knives. The slide was blocked with 1% BSA in PBS for one 1 hr and then incubated with anti-HSP20 diluted at 1:125 in the blocking solution over night at 4*C, washed with PBS three times of 5 minute each, incubated with CY3-conjugated anti-rabbit IgG (Sigma) diluted at 1:250 for 2–4 hours at room temperature, washed with PBS three times of 5 min each and mounted in Gel-Mount (Fisher) medium. Photographs were taken with an Olympus Flowview confocal microscope.
Acknowledgements
The authors would like to thank Marcia Ripley for technical support. Smithfield of Gwaltney, Smithfield, VA donated the swine carotid arteries. Grants from the Mid Atlantic American Heart Association and the Jeffress Trust supported this research.
The active hormonal form of vitamin D (1,25-dihydroxyvitamin D) is the primary regulator of intestinal calcium absorption efficiency. In vitamin D deficiency, intestinal calcium absorption is low leading to an increased risk of developing negative calcium balance and bone loss. 1,25-dihydroxyvitamin D has been shown to stimulate calcium absorption in experimental animals and in human subjects. However, the molecular details of calcium transport across the enterocyte are not fully defined. Recently, two novel epithelial calcium channels (CaT1/ECaC2 and ECaC1/CaT2) have been cloned and suggested to be important in regulating intestinal calcium absorption. However, to date neither gene has been shown to be regulated by vitamin D status. We have previously shown that 1,25-dihydroxyvitamin stimulates transcellular calcium transport in Caco-2 cells, a human intestinal cell line.
Results
In the current study, we have demonstrated that Caco-2 cells express low but detectable levels of CaT1 mRNA in the absence of 1,25-dihydroxyvitamin D treatment. CaT1 mRNA expression is rapidly up regulated (4-fold increase at 4 h and 10-fold at 24 h) by treatment with 1,25-dihydroxyvitamin D (10-7 moles/L). Moreover, the increase in CaT1 mRNA expression preceded by several hours the vitamin D induction of calbindin D9K, a putative cytosolic calcium transport protein.
Conclusion
These observations are the first to demonstrate regulation of CaT1 expression by vitamin D and are consistent with a new model of intestinal calcium absorption wherein vitamin D-mediated changes in brush border membrane CaT1 levels could be the primary gatekeeper regulating homeostatic modulation of intestinal calcium absorption efficiency.
Background
Low intestinal calcium absorption efficiency in humans is common and contributes to negative calcium balance and bone loss [1]. Recent epidemiological evidence suggests that low fractional calcium absorption is a significant risk factor for future osteoporotic hip fracture [2]. The primary homeostatic mechanism for altering calcium absorption in response to low dietary calcium intake and increased calcium requirements during growth is an increase in plasma 1,25-dihydroxyvitamin D, the active hormonal form of vitamin D produced in the kidney [3]. 1,25-dihydroxyvitamin D is a seco-steroid that acts on gene transcription through a nuclear steroid receptor protein and also may have nongenomic actions mediated by a putative plasma membrane vitamin D receptor [4]. Changes in vitamin D status primarily alter the rate of Ca absorption in the proximal small intestine. During transcellular calcium transport across the absorptive enterocyte, calcium must pass across an apical brush border membrane, then transverse the interior of the cell, and exit against an electrochemical gradient across the basolateral membrane. Changes in the rate of vectorial Ca transport in the intestine however must be accomplished without a major disruption in the free intracellular Ca concentration, which is used as an important intracellular signal in the enterocyte [5] as in other cells.
It has been known for over 30 years that vitamin D status influences the level of calbindin D, a cytosolic calcium-binding protein [6] believed to be the rate-limiting step in calcium absorption [7] by acting as an intracellular calcium ferry or chaperone. Calcium extrusion at the basolateral membrane is an active process involving a Ca-ATPase that may also be affected by vitamin D status [8]. However, although it has been known for more than 20 years that the rate of Ca transport across the intestinal brush border membrane can be increased by 1,25-dihydroxyvitamin D [9], the identity of the calcium transport protein involved in Ca influx has not been ascertained. Recently, two novel, closely related, epithelial calcium channels, CaT1 (also known as ECaC2) and ECaC1 (also known as CaT2), have been cloned from rat [10] and rabbit [11] respectively and shown to mediate calcium influx when expressed in Xenopus oocytes. CaT1 and ECaC1 are found in several tissues, including 1,25-dihydroxyvitamin D-responsive epithelia in the intestine and kidney. These epithelial calcium channel genes are products of evolutionary local gene duplication found on chromosome 7 and represent a new family of Ca-selective ion channels involved in active Ca (re)absorption [12]. Furthermore, it has been suggested that the control of the number of active channels at the cell surface is critical for regulation of active calcium transport and that the epithelial calcium channel could confer the rate-limiting step in transcellular calcium transport [13].
Peptide hormones and other cell surface-acting signal molecules trigger increases in the concentration of calcium ions inside the cell by release of calcium from stores within the cell, or activation of Ca++ ion channels in the plasma membrane allowing calcium to enter from outside the cell. One of the most common mechanisms that regulate entry of calcium is by capacitative or store-operated calcium entry [14]. The epithelial Ca-release activated channel, ICRAC, is a highly calcium-selective plasma membrane ion channel that is activated on lowering of either intracellular calcium concentration or intracellular calcium stores. The CaT1 channel has been recently expressed in mammalian CHO cells and patch clamp studies have shown it to have the characteristics of ICRAC[15]. Nearly ubiquitous, store-operated calcium entry is essential in a wide variety of cellular processes including muscle contraction, protein secretion, metabolism, cell differentiation, cell division and apoptosis [14].
Intestinal calcium absorption is increased after feeding a low calcium diet or under conditions of increased calcium needs, such as during growth, via a parathyroid hormone (PTH)-vitamin D mediated process [3]. A lowering of serum calcium levels stimulates the secretion of PTH that then activates the renal conversion of inactive 25-hydroxyvitamin D to the active hormonal metabolite 1,25-dihydroxyvitamin D. At the intestine, 1,25-dihydroxyvitamin D, acting through a nuclear receptor, stimulates transcription of vitamin D-mediated genes, including calbindin D, involved in calcium transport thereby acting as the primary hormonal regulator of active intestinal calcium transport [3]. It is thus likely that any putative key molecular player in calcium absorption, such as CaT1 or ECaC1, should be a prime target for hormonal regulation by 1,25-dihydroxyvitamin D. However, to date, neither CaT1 nor ECaC1 expression has been shown to be related to serum 1,25-dihydroxyvitamin D or vitamin D status [10,16].
We have developed [17-22] a unique human intestinal cell culture model of vitamin D-mediated transepithelial Ca transport using Caco-2 cells, a colon adenocarcinoma cell line. This cell line spontaneously differentiates in culture and expresses biochemical and morphological characteristics of well-differentiated small intestine-like enterocytes [17,23]. We have previously shown that Caco-2 cells possess a classical nuclear vitamin D receptor [18] that directly mediates vitamin D-dependent gene expression [22], and exhibits a high degree of specificity for various vitamin D metabolites and analogs [20,21]. Most importantly, in response to short-term (<24 h) 1,25-dihydroxyvitamin D treatment, Caco-2 cells increase transepithelial calcium transport [17] and calbindin D9K expression [19], a cytosolic calcium binding protein. Kinetic analysis of Ca transport across Caco-2 monolayers grown on permeable filter supports indicated that 1,25-dihydroxyvitamin D treatment increased the Vmax of Ca transport [17], which would be consistent with a vitamin D-dependent increase in the number of calcium transporters.
In the current paper, we have used the Caco-2 cell culture model to demonstrate that CaT1 mRNA is preferentially expressed in the human enterocyte compared to ECaC1, and for the first time demonstrate that 1,25-dihydroxyvitamin D treatment can cause a rapid and marked increase in CaT1 mRNA expression in the enterocyte. These observations suggest a likely important role for this calcium transporter in vitamin D-dependent calcium absorption.
Results and DiscussionCaT1 is expressed in Caco-2 cells
We initially addressed the question of whether Caco-2 cells express ECaC1 or the closely related CaT1 gene. This is important because it has been suggested that because of the high degree of sequence similarity (97%) between the CaT1 and ECaC1 genes care should be taken in interpreting data obtained by hybridization, PCR or other methodologies based on nucleotide sequences [24]. First, the PCR product derived from the putative ECaC1 or CaT1 gene was purified and sequenced. Based on new sequence information (GenBank accession number AF304464 (ECaC1) and AF365927 (CaT1), [25]) the nucleotides determined following the sequencing of our PCR product from Caco-2 cells was an identical match for CaT1 rather than ECaC1 (data not shown). Second, the sequences of the coding region of the CaT1 and ECaC1 genes (GenBank AF365927 and AF304464) indicate that an additional identification step based on restriction fragment length polymorphism (RFLP) analysis could also be used to confirm the identity of the CaT1 mRNA in Caco-2 cells. As described in the Method section, a 259 bp PCR product was generated with a second set of PCR primers that would specifically encompass a differential Bgl1 restriction enzyme cut site present in CaT1 but not ECaC1. Prior to electrophoresis on a 2% agarose gel, the amplified PCR product was either left untreated, or treated with Bgl1 or BssHII restriction enzymes. There is no predicted BssHII cut site in our amplified region of either CaT1 or ECaC1 based on the recently published sequence of these two genes (GenBank AF365927 and AF304464). Thus, BssHII should serve as negative control for the restriction enzyme digestion step. In contrast, the location of the single BglI cut site in CaT1 should result in the generation of two smaller bands (≅ 180 bp and ≅ 80 bp) if the amplified PCR product was CaT1. Since ECaC1 does not have a predicted BglI site in the amplified PCR product, no smaller bands should result if ECaC1 was being amplified in the PCR reaction. As shown in Figure 1, we found that there were two smaller bands of the predicted size and no larger 259 bp band remaining after Bgl1 digestion, confirming the identity of CaT1 as the product being amplified in Caco-2 cell. In addition, BssHII restriction enzyme digestion did not result in smaller bands, consistent with a likely specific action of Bgl1 digestion of CaT1 under the digestion conditions used. In addition, further attempts at PCR amplification of Caco-2 cell RNA with other PCR primers designed to specifically amplify ECaC1 did not result in an amplified product under our conditions (data not shown). Thus, based on this collective information, we conclude that CaT1, but not ECaC1, is expressed in Caco-2 cells. It needs to be pointed out, however, that since Caco-2 cells are not derived from normal human intestinal tissue, it is possible that ECaC1 could be expressed in the human intestine. However, our observation in Caco-2 cells is consistent with the recent report by Barley and colleagues [16] indicating a high expression level of CaT1/ECaC2 in human intestine based on Northern blotting, and similar findings by Peng et al. [25] using quantitative real-time PCR for CaT1, who also noted the apparent absence of ECaC1/CaT2 in human intestinal tissue.
Restriction Enzyme Digest of PCR Product From Caco-2 Cells. The amplified PCR product for the apical calcium transporter from Caco-2 cells (see text) was digested with Bgl1 or BssHII restriction enzymes. Digestion with Bgl1 resulted in the generation of two smaller appropriately sized products (180 bp and 80 bp) as would be expected only if the PCR product was CaT1 due to the presence of this restriction enzyme cut site only in CaT1 but not ECaC1 in the region amplified (GenBank acc. no. AF365927 nucleotide location 545–803). BssHII digestion did not result in any smaller bands being observed. Thus was expected since this enzyme was not predicted to cut either CaT1 or ECaC1 (GenBank acc. no. AF304464) in the region of interest and thereby could serve as a negative control to exclude the possibility that nonspecific fragmentation of under the restriction digest conditions.
CaT1 mRNA is regulated by 1,25-dihydroxyvitamin D
Having established the identity of CaT1 in Caco-2 cells, we then studied the effect of 1,25-dihydroxyvitamin D treatment on CaT1 mRNA expression. If CaT1 is an important regulator of intestinal calcium absorption, it is likely that it would be regulated by 1,25-dihydroxyvitamin D. As shown in Figure 2, treatment of cells for 24 h with 10-7 mol/L 1,25-dihydroxyvitamin D caused a 10-fold increase (p < 0.001, ANOVA, n= 10 studies) in CaT1 mRNA expression. To our knowledge, this is the first demonstration in either experimental animals or human tissue that intestinal CaT1 expression is up regulated by 1,25-dihydroxyvitamin D. This finding may provide an important molecular key to understand the mechanism of vitamin D-mediated intestinal calcium absorption. Moreover, given the potentially central role of this epithelial calcium channel as an important modulator of cellular calcium influx in a variety of cell types [15], it will be of interest to investigate whether the vitamin D effect on CaT1 expression extends beyond the enterocyte. Vitamin D status is known to influence a wide range of physiological systems including cells involved in the immune system, endocrine function, prostate, skin and muscle. Vitamin D effects on CaT1 expression and cellular calcium influx could thereby theoretically modulate intracellular calcium and cell signaling in a variety of different tissues providing insight into understanding the physiological effects of vitamin D throughout the body. Peng and colleagues [26] have reported that the steroid hormone androgen influences CaT1 mRNA expression in LNCaP prostate cells. We have not yet evaluated the effects of androgens on CaT1 expression in Caco-2 cells. However, we have found (S. Taparia, unpublished studies) that all steroid hormones do not share this effect on CaT1 expression because 17-β estradiol (10 -7 mol/L for 24 h) had no effect on CaT1 mRNA expression in Caco-2 cells. In any case, the demonstrated effect of 1,25-dihydroxyvitamin D on CaT1 mRNA shown here in Caco-2 cells is consistent with the idea that the active metabolite of vitamin D could increase the number of CaT1 calcium transporters in the plasma membrane and could therefore be responsible for the higher rates of calcium influx seen across the brush border membrane in vitamin D-replete animals [9,27] thereby playing a key gatekeeper role in increasing the efficiency of intestinal calcium absorption and maintaining calcium homeostasis.
1,25-Dihydroxyvitamin D Up Regulates The Expression Of CaT1 mRNA In Caco-2 Cells. Cells were treated with ethanol vehicle or 1,25-dihydroxyvitamin D (100 nM, 10 -7 mol/L) for 24 h. Total RNA was harvested and CaT1 mRNA (28 cycles) and GAPDH mRNA (21 cycles) were determined by RT-PCR. The PCR products were electrophoresed on a 2% agarose gel and stained with ethidium bromide for visualization and digital capture. The upper picture in the illustration shows the CaT1 PCR products from 6 individual wells of a single experiment (3 control and 3 + 1,25-dihydroxyvitamin D). CaT1 expression in the gel has been juxtaposed on top of the corresponding GAPDH expression. GAPDH is a housekeeping gene used to standardize CaT1 mRNA expression. Low levels of CaT1 mRNA are discernable in the absence of vitamin D treatment and the up regulating effect of 1,25-dihydroxyvitamin D by 24 h of treatment is clearly evident. The bar graph shows the combined mean (SEM) effect of 1,25-dihydroxyvitamin D on CaT1 expression normalized for GAPDH from 10 independent studies. 1,25-dihydroxyvitamin D treatment caused a significant (p < 0.001) 10-fold increase in CaT1 mRNA expression compared to vehicle treated control cells.
1,25-dihydroxyvitamin D causes a rapid and marked up regulation of CaT1 expression
In another series of studies, we addressed the temporal and dose-response effect of 1,25-dihydroxyvitamin D on CaT1 mRNA expression in Caco-2 cells. As shown in Figure 3, the response of CaT1 mRNA to 1,25-dihydroxyvitamin D is rapid (n = 6 studies). Treatment with 10-7 mol/L 1,25-dihydroxyvitamin D increased (~2-fold) CaT1 mRNA by 2 h, the earliest time point investigated, and clearly elevated CaT1 expression by 5-fold at 4 h, with progressively higher expression (up to 14-fold) at the later time points.
Early Temporal Response of CaT1 mRNA to 1,25-dihydroxyvitamin D in Caco-2 Cells. Caco-2 cells were treated with 1,25-dihydroxyvitamin D (10 -7 mol/L) for 0, 2, 4, 8 or 24 h. The bar graph illustrates the mean (SEM) fold-increase in CaT1 mRNA expression at each of the time points from 6 independent time course studies. There was a significant (p < 0.0001, ANOVA) time-dependent induction of CaT1 mRNA expression by 1,25-dihydroxyvitamin D with evidence of increased induction as early as 2 h and clear and progressive up regulation of CaT1 expression between 4 h and 24 h.
The relationship between the dose of 1,25-dihydroxyvitamin D treatment and CaT1 mRNA response was investigated in a third set (n= 3 studies) of experiments at 10-9, 10-8, and 10-7 mol/L 1,25-dihydroxyvitamin D for 24 h. In these studies there was a significant increase in CaT1 mRNA levels at the lowest dose (10-9 mol/L) of 1,25-dihydroxyvitamin D studied (Figure 4), while a slightly higher level of CaT1 mRNA expression was observed at 10 -8 mol/L 1,25-dihydroxyvitamin D, no further effect on CaT1 mRNA was seen with another 10-fold increase in 1,25-dihydroxyvitamin D concentration.
Effect of 1,25-dihydroxyvitamin D Dose on CaT1 mRNA in Caco-2 cells. Cells were treated for 24 h with increasing doses (1–100 nM) of 1,25-dihydroxyvitamin D. The mean effect of 1,25-dihydroxyvitamin D (n= 3 studies) on CaT1 mRNA was evident at the lowest dose (10 -9 mol/L) investigated. In these experiments there was a further increase in expression at 10 -8 mol/L 1,25-dihydroxyvitamin D, but no further increase in expression at a 10-fold higher dose of 1,25-dihydroxyvitamin D.
The rapid time course of induction and sensitivity of CaT1 mRNA response to 1,25-dihydroxyvitamin D treatment supports the idea that CaT1 could play a role as a gatekeeper of intestinal calcium absorption. We have previously reported that it takes from 12–16 h following 1,25-dihydroxyvitamin D treatment of Caco-2 cells to first appreciate an increase in saturable transcellular calcium transport [19]. This temporal response is similar to the 8–16 h observed for the achievement of maximal rates of intestinal calcium absorption in vivo following 1,25-dihydroxyvitamin D treatment of vitamin D-deficient chickens [28]. Future studies of the time course of CaT1 protein expression and localization of this protein to the brush border will be necessary to confirm this point, but must await the development of a suitable CaT1 antibody.
CaT1 induction precedes the increase in calbindin D9k mRNA expression
It has been argued [7] that the cytosolic vitamin D-dependent calcium binding protein calbindin D9K is the likely rate-limiting step in vitamin D-induced intestinal calcium absorption. However, despite a considerable amount of associative evidence [7], there has long been some controversy on this issue [29-31] because of apparent discrepancies in the temporal pattern of calbindin D9k induction compared to calcium absorption following 1,25-dihydroxyvitamin D treatment in experimental animals. Thus, in the current study, to evaluate a possible gatekeeper role for CaT1 in vitamin D-mediated calcium transport we compared the time course of 1,25-dihydroxyvitamin D induction of calbindin D9k mRNA and CaT1 mRNA in Caco-2 cells. In Caco-2 cells, the mRNA for both calbindin D and CaT1 can be detected by PCR in the absence of vitamin D treatment. The concentration of 1,25-dihydroxyvitamin D in the basal cell culture medium is approximately 10 -12 mol/L. As illustrated in Figure 5, the time course of vitamin D-dependent mRNA induction of the two genes clearly differs with CaT1 expression responding much earlier than calbindin D9k. There is no evident induction of calbindin D mRNA by 1,25-dihydroxyvitamin D in Caco-2 cells up to 8 h post treatment. In these experiments calbindin D mRNA was 3-fold greater than control by 24 h. In contrast, CaT1 mRNA was clearly increased (>5-fold) by 4 h post 1,25-dihydroxyvitamin D treatment in the same cells. These observations will need to be expanded in the future to include measurement of the relevant proteins. However, such association studies will not completely settle the question of the relative importance of either of these vitamin D-induced genes in the calcium transport response to 1,25-dihydroxyvitamin D. Those questions will be likely answered more directly in the future by individually altering the expression of these calcium transport genes in transfected cell lines or in transgenic animal models.
Comparison of the Time Course of CaT1 mRNA and Calbindin D 9K mRNA Induction by 1,25-dihydroxyvitamin D in Caco-2 Cells. Caco-2 cells were treated with 1,25-dihydroxyvitamin D (10 -7 mol/L) for 0, 2, 4, 8 or 24 h. The bar graph indicates the mean (SEM) fold-increase in CaT1 mRNA and calbindin D mRNA expression at each of the time points. The early induction of CaT1 mRNA by 4 h and 8 h following 1,25-dihydroxyvitamin D treatment is not evident for calbindin D mRNA, which did not increase above control until the 24 h time point. The data for CaT1 is the same as presented in Figure 3.
Given what is already known about the characteristics of the epithelial calcium channels and calbindin D, it is a plausible hypothesis that CaT1 and calbindin D could work in tandem to maintain high rates of calcium absorption. Studies in human duodenal biopsy specimens have found a significant correlation between CaT1 mRNA and calbindin D mRNA expression [16]. Thus, as a working hypothesis, we propose that during times of dietary calcium stress, or during periods of high calcium demand, CaT1 expression in the enterocyte is increased by a vitamin D-dependent mechanism. This regulatory genomic response of the enterocyte would result in an increase in the brush border membrane content of CaT1 calcium influx channels that could function as a gatekeeper mechanism to augment the rate of calcium entry across the apical membrane [9]. However, given that there is a Ca-dependent negative feedback mechanism to limit calcium entry through the channel by maintaining the channel in the closed position [14,15] there must be a way to remove calcium from the vicinity of the cytosolic face of the channel to optimize and sustain calcium influx rates. The coordinated vitamin D-dependent induction of the cytosolic calcium-binding protein calbindin D could serve this function when the need for transcellular calcium flux is high by acting directly or indirectly as a molecular sponge and selectively removing calcium ions from the CaT1 channel and/or to buffer increases in intracellular calcium, thereby preventing premature inactivation of this ICRAC channel [15]. Calbindin D could also play its traditionally conceived role as a calcium chaperone protein to facilitate the transfer of calcium through the cytosol [32] and perhaps delivery of Ca++ to the basolateral ATP-dependent calcium pump for extrusion out of the enterocyte [33].
Conclusions
These observations are the first to demonstrate regulation of CaT1 expression by 1,25-dihydroxyvitamin D in the human enterocyte and are consistent with a new model of intestinal calcium absorption wherein vitamin D-mediated changes in brush border membrane CaT1 levels could be the primary gatekeeper regulating homeostatic modulation of intestinal calcium absorption efficiency.
Materials and MethodsConditions of cell culture
Caco-2 cells (HTB 37, ATCC, Rockville, MD) were propagated and maintained as described previously [17]. Caco-2 cells were studied between passages 30–40. Cells were seeded into six-well dishes (35-mm diameter; Costar, Cambridge, MA) and grown for 15 days. Cell culture medium, nutrients and antibiotics were purchased from BioWhittaker (Walkersville, MD). FBS was purchased from HyClone Laboratories (Logan, Utah). Chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) and 1,25-dihydroxyvitamin D was purchased from Biomol Research Laboratories (Plymouth Meeting, PA).
Cell treatments
In time-course studies, cells were treated with 100 nanomol/L 1,25-dihydroxyvitamin D diluted in DMEM + 5% FBS for 0, 2 h, 4 h, 8 h and 24 h. All treatments ended at the 24 h time point. Control treatments were treated with ethanol vehicle alone. In dose-response studies, cells were treated for 24 h with 0,1,10, or 100 nanomol/L 1,25-dihydroxyvitamin D diluted in DMEM + 5% FBS. Control cells were treated with ethanol vehicle.
Following experimental treatments, cells were harvested and total RNA was isolated according to manufacturer's instructions (Tri Reagent, Molecular Research Center, Cincinnati, OH). One microgram of total RNA was made into a cDNA library by a reverse transcription process using the following conditions: 0.1 ug of oligothymidine, 100 U of reverse transcriptase, 5 ug bovine serum albumin (BSA), 10 U RNAsin (Promega, Madison WI) in a final volume of 10 ul buffer containing 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, 0.5 mM each of CTP, ATP, GTP, TTP, incubated for 2 h at 37°C, and denatured at 95°C for 10 minutes. The cDNA solution containing 0.1 μg equivalent RNA was analyzed by PCR for GAPDH, human calbindin D9K, and human CaT/ECaC. Primers sets for the epithelial calcium channel were as follows: sense primer 5'TGAACCTGGTGCGCGCACTGC3' (GenBank accession number AJ271207.1 nucleotides 480–500) and antisense primer 5'CCCAGGGAGTCCTGGGCCCGGA3' (nucleotides 657–678). The PCR primers used for GAPDH and calbindin D were as previously described [22]. PCR conditions used were: denature at 94°C for 40 s, annealing at 55°C for 50 s with an extension at 72°C for 50 s. All reactions were run within the linear range of number of cycles to product formed. GAPDH was amplified for 21 cycles, CaT1 for 28 cycles and calbindin D9K for 28 cycles. A blank PCR was run using water for each gene tested. PCR products were electrophoresed on a 2% agarose gel containing ethidium bromide. Relative amounts of amplified PCR product from each experimental condition were visualized under ultraviolet light and digitized with the Gel Doc 2000 gel documentation system (Bio-Rad Laboratories, Hercules, CA). Relative amounts of product were estimated by digital densitometry using Quantity One (version 4) quantitation software (Bio-Rad). CaT1 and calbindin D9K mRNA expression were normalized relative to the expression of GAPDH mRNA, a constitutively expressed gene whose levels were not affected by 1,25-dihydroxyvitamin D treatment.
Sequencing and RFLP analysis to establish the identity of CaT1 in Caco-2 Cells
PCR products from the epithelial calcium channel were generated with the primers described above, subjected to electrophoresis on a 1% low-melting-point agarose gel, and purified for sequence verification. The DNA sequencing facility at Tufts University sequenced the PCR product to identify whether the amplified PCR product from Caco-2 cells was either CaT1 or the closely related ECaC1 sequence. Because of the high nucleotide similarity of CaT1 and ECaC1 (97%), an additional identification step based on restriction fragment length polymorphism (RFLP) analysis was also conducted to confirm the identity of CaT1 in Caco-2 cells. PCR products were generated using the following primer set: sense primer (5'TGACATCTGAGCTCTATGAGGGT3', GenBank AF365927 nucleotides 545–567) and antisense primer 5'CCCAGGGAGTCCTGGGCCCGGA3' (nucleotides 782–803). These primers yield a 259 bp PCR product with a specific restriction site that makes it possible to clearly distinguish between CaT1 and ECaC1 using the restriction enzymes Bgl1 (New England Biolabs). In this PCR amplified product, there is no predicted Bgl1 restriction enzyme site in the corresponding ECaC1 sequence. In contrast, the presence of this Bgl1 cut site in the CaT1 sequence will result in two smaller bands (≅ 180 and ≅ 80 bp) after digestion with the endonuclease. The endonuclease digested and undigested PCR samples were electrophoresed on a 2% agarose gel and the bands were visualized with ethidium bromide staining. The size of the cut bands was estimated with standard 50 pb markers to identify whether the PCR product amplified from Caco-2 cells was being cut into the appropriately sized bands with Bgl1.
Data analysis
Data are reported as mean ± the standard error of the mean (SEM). Experiments were repeated from 3 to 10 times. Treatment effects were evaluated by ANOVA with a post hoc Tukey test for individual treatment comparisons (SYSTAT version 9, SAS Institute, Cary NC) where applicable.
Acknowledgements
This work was supported from funds from the United States Department of Agriculture, Agricultural Research Service cooperative agreement 1950-51520-006-00D. The contents of the publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.
Some of the data presented here were reported in preliminary form at the annual meeting of the federation of American Societies for Experimental Biology (Experimental Biology 2001) in Orlando FL (FASEB J 2001; 15: A59).
In perfused hearts, high calcium-induced inotropy results in less developed pressure relative to myocardial oxygen consumption compared to the β-adrenergic agonist dobutamine. Calcium handling is an important determinant of myocardial oxygen consumption. Therefore, we hypothesized that this phenomenon was due to reduced myofilament responsiveness to calcium, related to protein kinase C activation.
Results
Developed pressure was significantly higher with dobutamine compared to high perfusate calcium of 3.5 mM (73 ± 10 vs 63 ± 10 mmHg, p < 0.05), though peak systolic intracellular calcium was not significantly different, suggesting reduced myofilament responsiveness to intracellular calcium with high perfusate calcium. The ratio of developed pressure to myocardial oxygen consumption, an index of economy of contraction, was significantly increased with dobutamine compared to high perfusate calcium (1.35 ± 0.15 vs 1.15 ± 0.15 mmHg/μmoles/min/g dry wt, p < 0.05), suggesting energetic inefficiency with high perfusate calcium. The specific protein kinase C inhibitor, chelerythrine, significantly attenuated the expected increase in developed pressure when increasing perfusate calcium from 2.5 to 3.5 mM (3.5 mM: 64 ± 8 vs 3.5 mM + chelerythrine: 55 ± 5 mmHg, p < 0.05), though had no effects on dobutamine, or lower levels of perfusate calcium (1.5 to 2.5 mM).
Conclusions
By measuring intracellular calcium, developed pressures and myocardial oxygen consumption in perfused mouse hearts, these results demonstrate that high perfusate calcium positive inotropy compared to dobutamine results in reduced myofilament responsiveness to intracellular calcium, which is associated with energetic inefficiency and evidence of protein kinase C activation.
Background
During positive inotropy there is an increase in ATP utilization and thus myocardial oxygen consumption related to increased cross-bridge cycling and calcium handling [1]. Heat measurements of isolated heart muscle have assigned approximately 50% of total energy consumption to cross-bridge cycling and 20% to calcium cycling [2]. However, recently, studies in both isolated hearts and myocytes [3,4], examining the energetic effects of increasing inotropy with the calcium sensitizing agent EMD 57033 have suggested that the energetic cost associated with increasing inotropy is largely related to calcium cycling. In these studies EMD 57033 resulted in only small increases in myocardial oxygen consumption and this was felt related to the ability of EMD 57033 to increase inotropy without increasing calcium cycling (though intracellular calcium was not directly measured). Consistent with this Brandes et al [5] have studied the contribution of mechanical work and calcium cycling to total mitochondrial ATP hydrolysis as measured by NADH levels using fluorescent spectroscopy of rat cardiac trabeculae. They concluded that there was an equal contribution of mechanical work and calcium cycling to total ATP hydrolysis. In the present study, we hypothesized that energy consumption during altered inotropy is largely to related calcium handling. To study this we directly measured developed pressures, myocardial oxygen consumption and used a newly developed technique to measure intracellular calcium in perfused mouse hearts with the calcium sensitive fluorescent dye rhod-2 [6,7] during calcium-induced inotropy and with dobutamine. In addition we studied the perfused mouse heart, as the ability to manipulate the murine genome will likely provide important information about the molecular determinants of energy consumption in the heart.
The mechanisms controlling calcium-induced inotropy are probably several fold, and include activation of protein kinase C [8], of which several isoforms are calcium dependent. Protein kinase C decreases maximal actomyosin MgATPase activity through phosphorylation of troponin I [9]. This would be predicted to result in a reduction in force (or developed pressure) relative to intracellular calcium resulting in an energetically inefficient state. Thus, we predicted that high perfusate calcium positive inotropy would result in lower developed pressures relative to both intracellular calcium and myocardial oxygen consumption compared to dobutamine, and that high perfusate calcium would be associated with evidence of protein kinase C activation, though not dobutamine.
Results
During increases in perfusate calcium (1.5 – 3.5 mM) there were stepwise increases in developed pressure, and developed pressures with dobutamine were significantly higher than perfusate calcium 3.5 mM (Figure 1A). Peak systolic intracellular calcium increased minimally between 1.5 and 2.5 mM despite significant increases in developed pressures indicating increased myofilament responsiveness to calcium at perfusate calcium 2.5 compared to 1.5 mM (Figure 1B). Also, indicating reduced myofilament responsiveness to perfusate calcium 3.5 mM compared to dobutamine, levels of peak intracellular calcium were similar for these 2 interventions despite higher developed pressure with dobutamine. Increases in myocardial oxygen consumption reflected the changes in peak intracellular calcium, in that there were minimal changes between 1.5 and 2.5 mM perfusate calcium, and myocardial oxygen consumption was not significantly different between perfusate calcium 3.5 mM and dobutamine (Figure 1C). Thus, the ratio of developed pressure to myocardial oxygen consumption, an index of economy of contraction, significantly increased between 1.5 and 2.5 mM perfusate calcium and was significantly higher with dobutamine compared to high perfusate calcium (Figure 1D).
A. Developed pressure at perfusate calcium 1.5 – 3.5 mM and with dobutamine (dob, 0.9 μM). B. Peak systolic intracellular calcium, C. Myocardial oxygen consumption (MVO2), D. Ratio of developed pressure to myocardial oxygen consumption (DP/ MVO2). Symbols for Figure 1A: * p < 0.005 vs perfusate calcium 2.0, 2.5, 3.5 mM and dobutamine; † p < 0.05 vs dobutamine and perfusate calcium 3.5 mM; §p < 0.0001 vs dobutamine; ** p < 0.05 vs dobutamine; Symbols for Figure 1B:* p < 0.005 vs perfusate calcium 3.5 mM and dobutamine; † p < 0.05 vs perfusate calcium 3.5 mM and dobutamine; Symbols for Figure 1C:† p = 0.05 vs dobutamine; § p < 0.05 vs dobutamineSymbols for Figure 1D:* p < 0.0001 vs perfusate calcium 2.0, 2.5, 3.5 mM and dobutamine; † p < 0.0001 vs dobutamine; §p < 0.0005 vs dobutamine; ** p < 0.05 vs dobutamine
To further explore the differences between dobutamine and high calcium positive inotropy, analysis of pressure and calcium transient waveforms and phase-plane plots relating pressure to calcium was performed. Pressure waveforms were significantly abbreviated with dobutamine compared to high perfusate calcium (total waveform duration: 100 ± 15 vs 125 ± 0 ms, p < 0.01). The time from onset of the pressure waveform to peak pressure was not significantly different between dobutamine and high calcium (45 ± 3 vs 52 ± 7 ms, p=NS), though the time from peak pressure to fall to baseline was significantly shorter with dobutamine compared to high calcium (55 ± 12 vs 73 ± 7 ms, p < 0.05). An example of a normalized pressure waveform is shown in Figure 2. No significant differences were seen with calcium transient durations at any level of perfusate calcium or dobutamine (examples Figure 3A,B and 3C). Total transient duration at 3.5 mM perfusate calcium was 106 ± 20 ms and with dobutamine was 93 ± 22 ms (p=NS), and there was also no significant difference in the time from onset of the transient to peak (3.5 mM perfusate calcium 33 ± 8 vs dobutamine 38 ± 14 ms, p=NS), or difference in the duration between the peak of the transient to return to baseline (3.5 mM perfusate calcium 72 ± 25 vs dobutamine 60 ± 25 ms, p=NS). Phase-plane plots illustrate that high perfusate calcium results in a similar increase in intracellular calcium compared to dobutamine though relatively smaller increase in pressure (Figure 3D,E, and 3F). Thus, the ratio of pressure to calcium at the point of peak systolic pressure is significantly higher with dobutamine compared to high perfusate calcium (0.103 ± 0.009 vs 0.081 ± 0.012 mmHg/nM, p < 0.05, Figure 4). The increase in the 'dynamic trajectory' of the phase-plane plot between peak calcium and peak pressure may relate to underlying changes in the steady-state force-calcium relationship as proposed by Hunter [10].
Example of normalized pressure waveforms for dobutamine and 3.5 mM perfusate calcium, illustrating abbreviation of the dobutamine waveform compared to 3.5 mM perfusate calcium.
A: Averaged calcium transients (6–8 cycles) at perfusate calcium 1.5 and 2.5 mM, B: perfusate calcium 2.5 and 3.5. mM, and C: perfusate calcium 2.0 mM and with dobutamine. D. Phase-plane plots of pressure and calcium at perfusate calcium 1.5 and 2.5 mM, E: perfusate calcium 2.5 and 3.5. mM, and F: perfusate calcium 2.0 mM and with dobutamine.
Averaged phase-plane plots for high perfusate calcium and dobutamine. While the increase in intracellular calcium is similar for both, at peak pressure the ratio of pressure to intracellular calcium is significantly higher with dobutamine (* p < 0.05).
The specific protein kinase C inhibitor chelerythrine significantly inhibited high calcium positive inotropy (2.5 to 3.5 mM, p < 0.05), though had no effects on dobutamine, or with lower levels of perfusate calcium (1.5 to 2.5 mM) (Figure 5A). The effects of chelerythrine were mediated through preventing the expected increase in systolic and diastolic intracellular calcium (figure 5B). The difference in myocardial oxygen consumption between 2.5 and 3.5 mM perfusate calcium was significantly reduced with the addition of chelerythrine (Δ 2.5 to 3.5 mM alone: 9.1 ± 2.8 vs Δ 2.5 to 3.5 mM + chelerythrine: 5.4 ± 1.2 μmoles/min/g dry wt., p < 0.05). Also, chelerythrine had no significant effects on developed pressures at a constant perfusate calcium of 2.5 mM (53 ± 9 to 50 ± 10 mmHg, p=NS).
A. Effects of protein kinase C inhibitor (PKCi) chelerythrine on inotropic responses (developed pressure) of perfusate calcium 1.5 to 2.5 mM (left), 2.5 to 3.5 mM (middle), and dobutamine (right). B. Effects of PKCi with perfusate calcium 2.5 to 3.5 mM (left) on peak systolic intracellular calcium, and basal diastolic calcium (right). Baseline controls values (perfusate calcium 1.5, 2.0 or 2.5 mM) for these experiments include those experiments with addition of chelerythrine to the higher perfusate calcium or dobutamine after control measurements, combined also with those experiments without chelerythrine. * p < 0.05, and ** p < 0.005.
Discussion
In the present study we have directly related developed pressures, myocardial oxygen consumption and intracellular calcium in perfused mouse hearts. The results demonstrate that there are significant differences between high perfusate calcium and the predominant β-adrenergic agonist dobutamine in the perfused mouse heart. High perfusate calcium results in less myofilament calcium responsiveness and reduced energetic efficiency compared to dobutamine and also is dependent in part on protein kinase C activation. There are also significant differences in myofilament responses to calcium and energetic efficiency between different levels of perfusate calcium. These data support the hypothesis that intracellular calcium cycling is responsible for a significant proportion of total myocardial oxygen consumption during changes in inotropy [11], and thus alterations in myofilament protein calcium responsiveness may result in a significant effect on the energetic characteristics of a particular inotrope.
Protein kinase C phosphorylation of troponin I (predominantly at serines 43/45) results in a decrease in maximal actin-myosin ATPase activity [9], and also a decrease in the calcium sensitivity of the contractile apparatus through cross-phosphorylation of protein kinase A sites (serines 23/24) on troponin I [9]. Dobutamine, which is predominantly a β-adrenergic agonist, activates protein kinase A, which mediates a reduction in the calcium sensitivity of the contractile apparatus through phosphorylation of troponin I, though it does not effect maximal actin-myosin ATPase activity [12]. The findings suggesting that protein kinase C is activated at higher levels of perfusate calcium (though not with dobutamine) may indicate a role for protein kinase C in the energetic inefficiency associated with high calcium. That force does not increase proportionate to the rise in intracellular calcium during calcium-induced inotropy maybe related to 'down stream' effects of protein kinase C, such as phosphorylation of troponin I reducing maximal actin-myosin ATPase activity [9]. Indeed, recently we have shown that in a transgenic mouse expressing mutant troponin I lacking protein kinase C phosphorylation sites that during high perfusate calcium inotropy that there is a significant elevation in the ratio of developed pressure to myocardial oxygen consumption relative to wild type controls [13]. This data together with the data in the present manuscript provides reasonable evidence that activation of protein kinase C during high calcium-induced inotropy reduces maximal myofilament calcium responsiveness resulting in a decrease in developed pressure relative to peak systolic calcium, and energetic inefficiency. Also, conditional expression of protein kinase C β in mice [14], which is well suited to study the immediate effects of protein kinase C activation, is associated with increased contractility and calcium transients and decreased maximal tension development by myofilaments when expression is turned on. This compares favorably with our data showing that protein kinase C inhibition results in decreased contractility and reduced calcium transients.
The abbreviation of the pressure waveforms with dobutamine compared to high perfusate calcium is another marked difference between these 2 forms of positive inotropy. Differences in phosphorylation of proteins involved in relaxation such as troponin I and phospholamban may account for this [15,16]. We have not found any difference in the calcium transient duration to accompany the pressure waveform findings. This may relate to the relatively limited number of sampling points (N = 8) used in obtaining calcium transients or motion artifact related to the inotropic state. Recent improvements in acquisition software and recording capabilities will likely improve our ability to analyze calcium transients in detail.
Measurement of intracellular calcium and myocardial oxygen consumption in the perfused mouse heart
Isolated perfused mouse heart measurements of intracellular calcium and energetics have been performed using other methods by other investigators [17,18]. The unique aspects of the present study are that we use the calcium sensitive fluorescent dye rhod-2 to measure intracellular calcium [6,7], and that we have related these measurements to both myocardial oxygen consumption and developed pressure. An appreciation of the strengths and limitations of these techniques is required to critically evaluate the results. For instance, measurements of intracellular calcium are taken from the outer layers of the left ventricular free wall, though our measurements of myocardial oxygen consumption and developed pressures are global measurements. Differences between the epicardium and endocardium have been described [19] and our measurements of intracellular calcium do not take this into account, and regional differences in myocardial strains and oxygen consumption are also not accounted for. However, one advantage of using rhod-2 is that its long excitation and emission wavelengths relative to other calcium sensitive dyes results in minimal filtering of excitation and emission light and autofluorescence [7,20]. At these visible range wavelengths light may penetrate between 0.3 and 0.4 mm into the myocardium [21] which may be significantly deeper than shorter wavelength dyes. This is approximately 50% of the total wall thickness in the mouse [22], and thus our measurements are from the outer 50% of the myocardium.
Measurements of intracellular calcium and myocardial oxygen consumption have been performed in different hearts due to the difficulty in obtaining these measurements in a single experiment, and this may introduce variability into the results. We have also performed only 2 interventions per heart. The time taken to obtain stable measurements of effluent oxygen saturations used in the calculation of myocardial oxygen consumption is approximately 30 minutes and the mouse heart preparation is usually not stable beyond 80 minutes, so that only 2 interventions can be obtained. Furthermore, when comparing different inotropes such as high calcium and dobutamine, it is desirable not to use the same heart. As we have postulated that different signaling pathways may be activated by these inotropes, serially administering these to a single heart may result in persistent activation of a protein kinase even after washout of that agent.
Coronary flow does not increase with inotropes in this study, though other studies with perfused mouse hearts have shown an increase [23]. The reason for this is unclear, though oxygen delivery to the heart is not compromised. We have the advantage with the use of the absorbance spectrum between 500 and 600 nm, used in the calculation of intracellular calcium, that this is also sensitive to the presence of deoxy- or oxygenated myoglobin [7]. With our technique the myoglobin spectrum is invariably consistent with full oxygenation and any (rare) hearts that do not show this pattern are discarded. By using physiological heart rates and temperatures we have attempted to reproduce normal murine physiological conditions. As the mouse heart has a negative force-frequency response at physiological heart rates [24], developed pressures in our studies are lower than might be expected with lower heart rates. Nevertheless, the developed pressures in this study are quite consistent with other studies [17,24,25]. That pressures in perfused mouse heart studies are generally low may relate to the resistance of the murine myocardium to extracellular calcium [26]. Diastolic calcium levels are relatively high with this method especially with dobutamine or 3.5 mM perfusate calcium (table 1). This appears at least in part related to the limited sampling of 8 time points per beat in the current paper, and may also relate to motion artifact associated with the positive inotropes. These limitations are especially prominent in the small and rapid beating mouse heart. Recently, improved analogue to digital capabilities and software acquisition has significantly improved our ability to record many more time points during the cardiac cycle, and with this we have noted slightly higher differences between peak and diastolic calcium levels (MacGowan and Koretsky, unpublished data). It appears unlikely that this is related to a calcium overload state as diastolic pressures do not rise with high perfusate calcium or dobutamine, and pressure waveforms abbreviate with dobutamine.
Additional functional data, not included in Figure 1.
Perfusate Ca2+ (mM):
Dias. P
Volume
Flow
Dry heart
Diastolic
(mmHg)
(μL)
(mL/min)
weight (g)
[Ca2+]i (nM)
1.5
3 ± 2
19.9 ± 0.8
2.2 ±0.1
0.029 ± 0.002
353 ± 60*
2.0
3 ± 3
23.3 ± 1.9
2.0 ± 0.3*
0.029 ± 0.002
375 ± 17†
2.5
5 ± 4
20.2 ± 0.6
2.3 ± 0.2†
0.025 ± 0.002
346 ± 93*
3.5
4 ± 4
20.4 ± 0.9
2.1 ± 0.2
0.025 ± 0.001
532 ± 53
3.5+Chelerythrine
2 ± 1
20.7 ± 1.6
2.2 ± 0.4
0.027 ± 0.001
369 ± 46*
Dobutamine
3 ± 3
23.3 ± 1.9
1.8 ± 0.2
0.029 ± 0.002
587 ± 41
Symbols for coronary flow column: * p < 0.05 vs 2.5 mM; † p < 0.05 vs dobutamine Symbols for diastolic calcium column: * p < 0.005 vs perfusate calcium 3.5 mM and dobutamine; † p < 0.05 vs perfusate calcium 3.5 mM and dobutamine.
Conclusions
In the perfused mouse heart, high perfusate calcium positive inotropy compared to dobutamine results in reduced myofilament calcium responsiveness, decreased energetic efficiency and is dependent in part on protein kinase C activation. It must be recognized that agents that increase the amplitude of the calcium transient through activation of protein kinase A, such as phosphodiesterase inhibitors are associated with adverse outcomes in congestive heart failure patients [27], and so our findings with dobutamine are unlikely to translate into clinical benefit. Nevertheless, these results illustrate the importance of calcium handling in the energetic efficiency of an inotrope, and these methods will be useful in studying genetically engineered mouse models with altered myofilament calcium responsiveness.
Materials and methodsThe isolated perfused mouse heart and myocardial oxygen consumption
Experiments were performed in 2 groups. In one group, developed pressure and myocardial oxygen consumption were measured in the perfused mouse heart. In another group of perfused hearts intracellular calcium and developed pressure were measured. Male mice (129/SV ter strain, mean weight 26.9 ± 2.6 g) were used. Anesthesia was induced with 1.5 to 3.0 mg of intraperitoneal pentobarbital sodium, and the animal was anticoagulated with 100 units of heparin. The heart was removed from the chest and the aorta was cannulated with a 23 gauge needle. Retrograde coronary perfusion at a constant perfusion pressure of 55 mmHg with oxygenated modified Krebs solution was started. The Kreb's solution consisted of (mM): NaCl 112, KCl 4.7, MgSO4 1.2, Na-EDTA 0.5, NaHCO3 28.0, Glucose 5.5, Pyruvate 5.0, CaCl2 2.5, 50 μM octanoate, and pH was adjusted to 7.4. For some studies the concentration of calcium was varied to produce an inotropic effect. The concentration of NaCl was adjusted accordingly for these experiments to maintain osmolality. To eliminate the effects of cathecholamine release by pacing esmolol (0.1 μM) (Du Pont, Wilmington, DE) was added to the perfusate. The temperature of the perfusate entering the heart was set at 37°C. Flow was monitored with a glass flow meter. A rubber latex balloon on the end of a plastic cannula, was inserted into the left ventricle through an incision in the left atrium. This was then connected to a Gould pressure recorder (Gould, Cleveland, Ohio) for measurement of left ventricular pressure. The left ventricular diastolic pressure was set at 0 – 5 mm Hg using a micro-syringe. Hearts were paced at 8 Hz which is physiological heart rate for the mouse, using a stimulator with one lead inserted into the right ventricle and the other adjacent to the epicardium. In the myocardial oxygen consumption experiments, the perfused mouse heart was placed in a glass, water jacketed chamber, which was sealed at the top. Myocardial oxygen consumption was determined from influent and effluent oxygen content (measured by blood gas analyzer, ABL-30, Copenhagen, Denmark) and flow rate.
Measurement of Intracellular Calcium
The methods used to measure intracellular calcium with rhod-2 in perfused hearts have been previously extensively described [6,7,20]. Rhod-2 (Molecular Probes, OR, 100 μg) was dissolved with dimethylsulfoxide (DMSO, 4 μL) and dH2O (200 μL), and loaded through the coronary perfusate. After the washout period serial measurements of fluorescence alternating with absorbance were taken. Fluorescence scans were taken at high time resolution to allow quantification of changes in fluorescence during the cardiac cycle. Excitation at 524 nm and emission at 589 nm was used for fluorescence measurements.
Quantification of the relative amount of rhod-2 in the heart using absorbance measurements, was done by taking the ratio of absorbance at 524 nm (rhod-2 sensitive) to 589 nm (rhod-2 insensitive) which eliminated the effect of motion as both wavelengths would be equally affected by motion, though only 524 reflected the concentration of rhod-2 [7]. These wavelengths (524, 589 nm) were chosen as these were isosbestic points not effected by changes in absorbance of myoglobin induced by oxygen desaturation [7]. In solution maximal rhod-2 absorbance is at 554 nm. However, this wavelength is effected by changes in oxygen saturation, the relative absorbance compared to 524 nm is decreased due to inner filter effects and it does not adequately correct for changes in scattering [7]. Dye absorbance (Arhod2) was calculated according to the formula:
Arhod2 = log{(R524/R589)0(R524/R589)rhod2} (1)
where R524 is the reflectance intensity at the rhod-2 sensitive point of 524 nm, and R589 is the rhod-2 insensitive point, before ()0 and after ()rhod2 loading.
At the end of the perfusion protocol, maximal fluorescence, used in the calculation of [Ca2+]i was determined by tetanizing the heart with a bolus of calcium chloride (20 mM) without any energy substrate, and with cyclopiazonic acid (Sigma Chemical Co., 10 μM), which is a potent inhibitor of Ca2+-ATPase and thus blocks calcium uptake by the sarcoplasmic reticulum [28]. Fluorescence and pressure were monitored continuously, and the point of maximal fluorescence was taken as the point were pressure stabilized at a steady state. To account for changes in light scattering properties from the heart during tetanization, the maximal fluorescence was corrected by multiplying by the ratio of R524 pre-tetanization to R524 during tetanization [7].
Intracellular calcium was calculated using the formula:
[Ca2+]i = Kd(Ft - Fo)/(Fmax - Ft) (2)
where Kd is the dissociation constant for rhod-2 and calcium (determined by in-vitro calibration with rhod-2 and myoglobin by del Nido et al, ref [20], and confirmed by in-vivo manganese quenching, ref [29]) and is 710 nM, Ft = fluorescence at time t, Fmax =maximal fluorescence from tetanized heart, and the fluorescence from the heart assuming rhod-2 had no calcium bound is given by F0 = Fb + a(Fmax - Fb), where Fb is the background counts from the heart prior to dye loading, and a = rhod-2 fluorescence in the absence of calcium / rhod-2 fluorescence in the presence of saturating calcium. For rhod-2 the value of a is approximately 0, thus for rhod-2, F0 was assumed to be equal to Fb.
To account for changes in dye concentration, formula 2 needs to be modified to account for changes in absorbance (Arhod-2, formula 1) due to dye leakage.
where At = dye absorbance at time t, Amax is dye absorbance just prior to tetanizing the heart. Amax is not determined when heart has tetanized because of the marked influence of the shape change and desaturation of myoglobin on the reflectance spectrum.
Interventions
For each heart only a baseline and one intervention was used, and N = 4–6 for all groups. The time to achieve a stable state (particularly with respect to effluent PO2 concentrations) or load rhod-2, and then acquire baseline data and perform 1 intervention is 60–90 minutes, beyond which the preparation is not stable and so more than 1 intervention per mouse heart was not undertaken. In the developed pressure/oxygen consumption experiments, the following interventions were used: 1) varying perfusate calcium from baseline 2.5 mM to 3.5; 2) baseline perfusate calcium 1.5 mM to 2.5 mM; 3) baseline perfusate calcium 2.0 mM and then addition of dobutamine 0.9 μM.; 4) to antagonize the effects of protein kinase C, the specific antagonist chelerythrine (ref 30; 10 nM, Calbiochem-Novabiochem Intl, La Jolla, CA) was added to the perfusate when inducing positive inotropy. In all of these the chelerythrine was added to the perfusate containing the higher level of perfusate calcium or dobutamine. This dose was chosen to reduce the expected increase in pressure with perfusate calcium 3.5 mM by 50% after dose ranging studies (1 nM to 1000 nM). In addition in another group of experiments during constant infusion of perfusate calcium 2.5 mM, chelerythrine was added without changing perfusate calcium concentration. In all experiments, the control period was followed by the intervention, and repeated measurements of both developed pressure and myocardial oxygen consumption were taken for a period of approximately 30 minutes. Measurements of intracellular calcium were performed 1) at perfusate calcium concentrations at baseline of 2.5 mM increasing to 3.5 mM, 2) decreasing from 2.5 to 1.5 mM, and 3) 2.0 mM baseline with subsequent addition of dobutamine 0.9 μM. The effects of chelerythrine on intracellular calcium was only determined for those interventions with which chelerythrine had significant effects on myocardial oxygen consumption and developed pressure.
Data and Statistical Analysis
One way analysis of variance with the Scheffe test was used to test significance of differences between groups. Developed pressure data from the myocardial oxygen consumption and intracellular calcium experiments were combined as this was obtained under the same experimental conditions. The ratio of developed pressure to myocardial oxygen consumption was derived as an index of economy of contraction. Data are expressed as mean ± standard deviation.
Acknowledgements
Supported by AHA PA Affiliate Grant-in-Aid Beginning (B98452P) and NIH grant (HL-03826) to G.A.MacG, and National Institutes of Health Grant RR-03631 to the Pittsburgh NMR Center for Biomedical Research, and NIH Grant (HL-40354) to A.P.K., and the NINDS intramural research program.
Venous return from the posterior region of amphibians travels by either two renal portal veins to the kidney or a central abdominal vein that drains into the hepatic portal system. The relative proportions of blood flow in these vessels has never been measured nor has a modification of flow been determined when venous return increases by changes in blood volume during hypervolemia or during increased volume input from the posterior lymph hearts.
Results
Venous return from the posterior region of Bufo marinus was measured under resting conditions and in response to a systemic hypervolemia. Doppler flow probes were positioned on the renal portal and ventral abdominal veins, and flow was recorded as injections of artificial plasma equaling 100% of the animal's plasma volume were administered through the sciatic artery. Resting flow was found to be 5.54 ± 2.03 ml min-1 kg-1 in the paired renal portal veins, and 7.31 ± 0.89 ml min-1 kg-1 in the ventral abdominal vein. While renal portal flow was found to increase by a factor of 2.4 times during the first 10 min of hypervolemia, ventral abdominal flow only increased by a factor of 1.3.
Conclusions
Our results quantify the contribution to circulation from posterior venous return in the toad Bufo marinus. A preferential movement of excess fluid through the renal portal pathway was also demonstrated, supporting the possibility of water elimination via the renal portal circulation, especially during periods of high water influx into the animals.
Background
The distribution of blood flow through the amphibian body is crucial to many vital homeostatic functions, including those involving respiration, nutrition and elimination of unwanted materials. Although blood flow has been quantified for the primary arterial components of the circulatory system, the magnitude and partitioning of venous return from the posterior regions through several major organs is unknown. Blood returning from this area in an anuran amphibian may pass through one of two major pathways; the first carrying blood through the hepatic portal system, and the second traversing the kidneys via the renal portal system, a component of circulation of relatively unknown function [1].
Through direct measurement of arterial flow in Bufo marinus, West [2] determined cardiac output to be 57.2 ml min-1 kg-1. Aortic blood flow represented 44.9% of this value, with the pulmocutaneous artery carrying 48.4% and the carotid artery 8.0% of cardiac output [2]. Of blood pumped to the posterior of the animal through the aorta, a significant amount passes through vessels surrounding the intestine and through the kidneys via the renal arteries. The remaining blood bathes tissues of the abdomen and hind limb region and returns to the heart through the two main pathways, the renal portal and ventral abdominal veins [1].
The common pelvic and femoral veins join to form the renal portal veins, just distal to the drainage site for the posterior lymph hearts into the circulation (Fig. 1) [1]. The renal portal veins run anteriorly along the kidneys, producing networks of vessels that cover the dorsal surface of the organs [3]. As the branched vessels traverse the kidneys, they give rise to peritubular vessels, which surround the kidney nephra. Deeds et al.[4] found that in the perfused bullfrog kidney, the renal portal circulation contributed up to 72% of the perfusate, indicating the apparent importance of this portal circulation to renal blood flow.
General scheme of circulation in Bufo marinus showing major pathways of venous return from the posterior region. Placement of Doppler flow probes is indicated. Figures in brackets represent percentages of flow returning through the specified pathways from the posterior systemic circulation. Abbreviation: vein (V.).
The perfusion of the renal portal vessels through the kidneys is still uncertain. However, it is generally accepted that the efferent arterioles from the glomerular capillaries empty into the peritubular capillaries, serving as a link between arterial and portal circulations [5]. Although it has been suggested that this close association may allow for filtration of renal portal blood, Deeds et al.[4] found that in the perfused bullfrog kidney, only 0.52% of the filtrate was derived from the portal perfusate, likely due to the high resistance bridge between the two circulations. Efferent renal veins eventually collect the bulk of renal blood from the peritubular capillaries, and join to form the posterior vena cava [1].
The ventral abdominal vein represents an alternate pathway for venous return from the posterior region (Fig. 1). This vein runs anteriorly from the junction of the common pelvic veins, and joins the hepatic portal system as it enters the liver [1]. The common pelvic and femoral veins have interconnections, providing a link between the two venous return pathways [1]. Ohtani and Naito [3] have suggested that the connection may serve as a means of delivering blood from the ventral abdominal vein to the kidneys.
The delicate balance between water loss and uptake in amphibians makes an understanding of venous return from this region even more significant. Terrestrial amphibians show a behavior known as the water absorption response, using a portion of their ventral skin called the pelvic patch to take up extraneous water [6]. Although it has long been assumed that this external water moves into the circulatory system, one study shows movement of tritiated water directly into the lymph sacs from the environment [7]. The connection of lymphatic and circulatory systems through the renal portal veins has elicited suggestions of rapid elimination of excess incoming fluid via the renal portal vessels [8].
This study will measure flow through both the renal portal and ventral abdominal veins in the cane toad, Bufo marinus, using Doppler flow probes, to determine relative contributions of the two main pathways to venous return from the posterior regions (Fig. 1). An arterial volume load will also be performed in order to examine possible physiological consequences of volume stress on venous return.
ResultsResting blood flow
Values for resting renal portal and ventral abdominal venous blood flow were determined using 5 min pre-injection levels. Measured flow from a single renal portal vessel was doubled for each individual. Throughout the text, all mention of renal portal blood flow refers to doubled values. Renal portal venous flow was 5.5 ± 2.0 ml min-1 kg-1, while flow in the ventral abdominal vessel was 7.3 ± 0.9 ml min-1 kg-1.
Blood flow during hypervolemia
Following an injected 100% increase in plasma volume there were significant increases in blood flow in ventral abdominal and renal portal veins during the first 10 min period but not in the time following 10 min post infusion (Fig. 2). Flow in the renal portal veins increased by a factor of 2.4 (t14 = 6.6; p < 0.001) and in the ventral abdominal vein by a factor of 1.3 (t14 = 3.36; p = 0.0047). Following the 10 min post infusion period blood flow in all veins decreased rapidly and flows did not differ significantly from the pre infusion level (p > 0.29).
Effect of systemic hypervolemia on mean relative renal portal blood flow and relative ventral abdominal blood flow as compared with control values of Bufo marinus (n = 8). Filled circles denote mean relative renal portal blood flows and open circles show mean relative ventral abdominal flows. Injection was administered at t = 0 and lasted 4.5 to 5.5 min. All values are expressed as means ± S.E. The 5 min pre-injection values were used as control values for both relative blood flow determination and statistical analysis.
Deviation from resting renal portal blood flow pattern
Blood flow in approximately 3/4 of the renal portal venous flow tracings, from 30 to 90 min post injection, showed periods where a deviation from the normal pattern was apparent (Fig. 3). This pattern involved sharp pulsatile increases in blood flow, distinguishing these tracings from the normal, slow wave pulsations seen in the simultaneous ventral abdominal tracings and in the resting renal portal and ventral abdominal venous tracings. The rate of renal portal pulsations was 48.0 ± 6.3 beats min-1. The rate of slow wave pulsations in the ventral abdominal vein during the same time period was 23.3 ± 4.4 beats min-1. The resting rate of slow wave pulsation, identical in both vessels, was 19.5 ± 3.0 beats min-1, calculated using 10 min pre-injection values. The sharp pulsatile increases above baseline renal portal venous levels generated flows of 6.2 ± 2.9 ml kg-1 h-1, and stroke volumes of 0.0018 ± 0.0006 ml kg-1.
Flow tracings from the renal portal vein of Bufo marinus showing sharp extra pulsations at: A) 50 min post injection, flow values shown for renal portal vein; and B) 55 min post injection with recording expanded, flow values shown for pulsations above resting renal portal flow. The injection was performed at t = 0 min. Chart speeds are: A) 1 mm sec-1; and B) 10 min sec-1.
Discussion
The pathway of venous return from the posterior regions of Bufo marinus is via either the renal portal or hepatic portal systems before returning to the heart [1]. Resting blood flow in the two renal portal vessels combined was quite similar to that of the single ventral abdominal vein (which drains into the hepatic portal vein), with a slightly higher value observed for the ventral abdominal vessel. Total flow in the two vessel systems was 12.85 ml min-1 kg-1, which comprises the total blood flow returning from the hind limb region of the toad. Ventral abdominal flow accounted for 57% of this venous return, with renal portal flow contributing the remaining 43% (Fig. 1).
West [2], using similar techniques to those used in this study, found cardiac output in Bufo marinus to be 57.2 ml min-1 kg-1, with aortic blood flow being 26.8 ml min-1 kg-1[2]. Using values obtained in this study, venous return from the hind limbs of the toad would thus represent approximately 1/4 of cardiac output.
When blood flow was examined following an infused doubling in plasma volume, both renal portal and ventral abdominal blood flow increased significantly during the 10 minute post infusion period but flow in the vessels leading to the kidneys was greater. The site of injection, the existence of a vascular connection between the two flow pathways and the large volume of fluid injected would support an equal distribution of excess fluid in both pathways. Our results strongly suggest a preferential movement of fluid through the renal portal system.
A favored movement of fluid through the renal portal veins would support the hypothesis that, in a natural aquatic setting, the rapid elimination of incoming water could be made possible by the renal portal system [8]. Carter [8] postulates that this excess water may move from the peritubular capillaries into the tubular lumen due to osmotic concentration differences, and therefore be excreted as urine. In this case, fluid would move in a preferred direction such that its rapid elimination is maximized. It has also been suggested that the femoral vein may serve as a functional connection between the renal portal and ventral abdominal veins [3]. This would seem to be supported if increased venous flow in the hind limbs is primarily directed through the renal portal veins. This connection could then serve as a means for delivery of increased volumes of blood to the kidneys, allowing for the elimination of excess fluid.
Interestingly, the pattern of venous blood flow observed in both of the vessels throughout the experiments was a regular, slow-wave type pulsation. Previous work has shown heart rate in Bufo marinus to be 39.9 beats min-1 and lung ventilation rate to be 16.3 ventilations min-1 in normoxic conditions [9]. The calculated resting rate of 19.5 ± 3.0 pulsations min-1 in the vessels appears closer to values for lung ventilation, however the rhythmic nature of the pulsation pattern seems consistent with changes due to heart contraction. Although we did not monitor lung ventilation, it is possible that ventilatory pressure changes within the toad result in regular oscillations in venous flow. Pulsations in flow due to negative pressure from the heart, however, seem unlikely, as both of the veins studied branch into portal systems before emptying into the vena cava.
Much of the speculation on pulsation patterns arose directly as a result of the observation of flow "spikes" in the renal portal tracings, which appeared as flow returned to normal after hypervolemia. The sharp extra pulsations observed were less rhythmic than the resting pulsations and contrasted with the flow tracings recorded simultaneously in the ventral abdominal vein. The pattern of sharp, irregular peaks corresponds remarkably with tracings observed when lymph flow is monitored directly from lymph heart efferent vessels using similar techniques [10]. Williams et al.[11] found that 50 min after a doubling in plasma volume, lymph heart rate was approximately 50 beats min-1, consistent with the "spike" rate of 48.0 ± 6.3 pulsations min-1 calculated in this study during a similar time frame. It would appear that these tracings reflect contributions to renal portal blood flow by lymph, which is ejected from the nearby posterior lymph hearts.
Previous measurement of lymph heart function in response to systemic hypervolemia has shown lymph flow from a single posterior heart to be 20.7 ml kg-1 h-1 and stroke volume to be 0.0074 ml kg-1 at the 50 min post-injection mark [11]. Our values, determined through measurement of flow peaks above baseline renal portal flow, are considerably lower, with lymph flow being 6.2 ml kg min-1, and stroke volume 0.0018 ml kg-1. These measurements of lymph flow may vary due to a more "downstream" location and the indirect method of lymph flow measurement used in this study. In addition lymph heart flow "spikes" in the renal portal veins were not visible under normal resting conditions. Williams et al.[11] found that, 50 min following a 100% increase in plasma volume, lymph flow was increased, although this change was not significant. It is possible that these sharp pulsatile increases in flow are peaks of the lymph heart systolic output, which has been increased such that it becomes visible, superimposed upon the larger flow of the renal portal vein.
The contribution of lymph flow to renal portal flow can also be considered with respect to resting state. Jones et al.[10] found lymph flow in Bufo marinus under normal conditions to be 25.9 ml kg-1 h-1 from a single posterior lymph heart. Using renal portal flow values measured in this study, it is estimated that almost one-sixth of the renal portal circulation consists of fluid originating in the posterior lymph hearts.
Conclusions
The shift in distribution of blood flow from the posterior end of the toad during volume stress demonstrates the dynamic nature of the anuran circulatory system. The possibility of the elimination of excess water via the renal portal system illustrates the necessity of a greater understanding of the association between lymphatic and circulatory systems. This association was clearly emphasized by the quantification of lymphatic function within the circulatory system that was demonstrated in this study.
Material and methodsAnimals
Cane toads, (Bufo marinus, L.; 245–326 g hydrated mass), were obtained from commercial suppliers (Boreal Laboratories, St. Catherine's, ON, Can.; Charles D. Sullivan Co. Inc., Nashville, TN, USA). Animals were maintained in fiberglass aquaria (0.9 m × 0.6 m × 0.6 m)filled with sand to a depth of approximately 5 cm, and allowed free access to water. Toads were force fed raw beef liver once a week, the water was changed weekly and the sand was replaced every second week.
Surgical procedure
Toads of either sex were chosen for experimentation and anaesthetized in a solution of 2 g l-1 of aminobenzoic acid ethylester (MS-222, Sigma Chemical Co.) and 2 g l-1 of NaHCO3 in tapwater until the corneal reflex was absent.
At a point on the median line of the ventral surface of the toad, 2 cm anterior to the juncture of the torso with the hind limb, a transverse incision of about 2 cm was made and the skin was separated from the underlying tissue. A longitudinal incision of about 0.5 cm was made through the muscle at a position 0.5 cm lateral to the ventral abdominal vein, and a small length of this vein was separated from the abdominal wall by removal of the connective tissue. The vein was fitted with a Doppler flow probe (silastic cuff, i.d. 1.0 mm; Iowa Doppler Products, Iowa City, LA, USA) and the probe was secured to the musculature between the ventral abdominal wall and the skin using 5–0 gauge silk sutures. The leads of the probe were tied down to the skin in 3 locations, and the initial incision was closed using 3–0 gauge silk sutures.
On the dorsal surface an incision was made just above the junction of the renal portal vein and the kidney, just anterior to a posterior lymph heart. As with the ventral abdominal vein, the renal portal vein was dissected free and fitted with a Doppler flow probe (silastic cuff, i.d. 0.8 mm; Iowa Doppler Products, Iowa City, IA, USA), which was placed as close to the point of entry of the vein into the kidney as possible without restriction of flow. The probe was secured to the surrounding musculature, the leads of the probe were tied down to the skin in 3 locations, and the incision was closed. For arterial infusion purposes, the sciatic artery was then cannulated using previously described methods [12].
Following surgery the animals were placed under running water to recover, and then moved to 6 L covered plastic boxes containing 2–3 cm of dechlorinated tap water and several air ventilation and lead exit holes cut in the top. Animals were allowed to recover for a minimum of 18 h following surgery. During all experiments animals were freely moving and unanaesthetized in the chambers.
Experimental procedure
For measurement of venous flow, the leads of the Doppler flow probes were connected to a pulsed Doppler flowmeter (model 545c-4; Bioengineering, University of Iowa), which was then connected to a two-channel Gould chart recorder.
Resting renal portal and ventral abdominal venous flows were recorded for 15 min, at a chart speed of 1 mm s-1. Following this period, the animals were volume loaded through the sciatic artery cannula with a solution of Bovine Albumin (3.60 g 100 ml-1; [13]) in MacKenzie's saline [14]. Plasma volume was doubled, assuming 7.4 ml of plasma for 100 g of animal [15]. An infusion pump was used to inject fluid over a period of 4.5 to 5.5 min. Measurements were recorded for 90 min from the onset of injection, and flow was determined at time 0, then every min for 10 min, every 5 min until 45 min and at 60, 75 and 90 min. For those animals in which the flow probes remained stable for a 24 h period following the initial trial, a second injection was performed. This injection was identical to the initial one, and measurements were taken at the times outlined for the first injection.
Following experimentation, each Doppler flow probe was calibrated by removal of the vessel from the animal with the probe still in place. The vessel was then cannulated with polyethylene tubing (PE-100 for the ventral abdominal vein; PE-50 for the renal portal vein), and whole blood from an exsanguinated toad was perfused through the vessel using an infusion pump. This allowed determination of absolute flow in ml min-1 for each probe and animal.
Statistical analysis
To correct for large differences among toads in the basal flow rates for both arteries, we standardized all observations by dividing by the flow rate at t = -5 min. This helps to correct for differences resulting from animal size and physical condition at the time of experimentation. We then further simplified the analysis by dividing the set of experiments into three groups – prior to the infusion, 0–10 min post infusion, and >10 min post infusion, and calculated the mean flow rate for each artery, for each toad, over each period.
Data were analyzed using a linear mixed-effects model in Splus. Toads were considered random effects, and the period of time was considered a fixed effect. Differences between the two levels of treatment (immediately post infusion, and > 10 min post infusion) were assessed by fitting "treatment" contrasts (with the pre-treatment mean being considered a control), and assessing the significance of the two parameter estimates. Separate models were fit for the two veins.
Acknowledgments
Financial support for this work was provided by an NSERC operating grant to D.P.T. and an N.S.E.R.C. Undergraduate Student Research Award to E.E.K. Thanks are expressed to Judy Jones for technical assistance and Philip Taylor for statistical assistance.
Little information is available on the circadian sequela of an immune challenge in the brain of aged rats. To assess them, we studied 24-hour rhythms in hypothalamic and striatal norepinephrine (NE) content, hypothalamic and striatal dopamine (DA) turnover and hypophysial NE and DA content, in young (2 months) and aged (18–20 months) rats killed at 6 different time intervals, on day 18th after Freund's adjuvant or adjuvant's vehicle administration.
Results
Aging decreased anterior and medial hypothalamic NE content, medial and posterior hypothalamic DA turnover, and striatal NE concentration and DA turnover. Aging also decreased NE and DA content in pituitary neurointermediate lobe and augmented DA content in the anterior pituitary lobe. Immunization by Freund's adjuvant injection caused: (i) reduction of DA turnover in anterior hypothalamus and corpus striatum; (ii) acrophase delay of medial hypothalamic DA turnover in old rats, and of striatal NE content in young rats; (iii) abolition of 24-h rhythm in NE and DA content of neurointermediate pituitary lobe, and in DA content of anterior lobe, of old rats.
Conclusions
The decline in catecholamine neurotransmission with aging could contribute to the decrease of gonadotropin and increase of prolactin release reported in similar groups of rats. Some circadian responses to immunization, e.g. suppression of 24-h rhythms of neurointermediate lobe NE and DA and of anterior lobe DA were seen only in aged rats.
Background
In 1974, Pittendrigh and Daan first described the changes caused by aging in the circadian timing system of rodents [1]. Since then, a bulk of information has accumulated indicating that reduced amplitude, shorter free-running periods and desynchronization of circadian rhythms are associated with advanced age in rodents as well as in humans (for references see [2]). Both the efficacy of input and output pathways from the central nervous system circadian pacemaker, located in the hypothalamic suprachiasmatic nuclei, and the functioning of the central pacemaker itself, change with advancing age. In addition, some of the decline in overt circadian rhythmicity may be due to deteriorating function of the body's aging effector systems [2].
An aspect of circadian organization in aged subjects less known is the modification in amplitude or phase of circadian rhythms during the different phases of the response to an immune challenge. Since aging is associated with declines in multiple areas of immune function [3], it seems feasible that differences in the circadian response to an immune challenge with age occur.
Adjuvant arthritis is an experimental model for rheumatoid arthritis, induced by the intradermal injection of heat-killed Mycobacterium tuberculosis in incomplete Freund's adjuvant to rats [4]. In the classical adjuvant-induced arthritis model, polyarthritis is accompanied by a widespread systemic disease. In this experimental model we previously reported the effect of aging on circadian organization of plasma prolactin, growth hormone (GH), thyrotropin (TSH), insulin, follicle-stimulating hormone (FSH), luteinizing hormone (LH) and testosterone, studied during the acute phase of inflammatory disease of the joints (18 days after Freund's adjuvant injection) in rats [5]. As a continuation of those studies we now report, in comparable groups of animals, the changes in 24-hour organization of hypothalamic and striatal norepinephrine (NE) content and dopamine (DA) turnover and hypophysial NE and DA content. The results support the view that 24-h rhythms and levels of hypothalamic, striatal and hypophysial catecholamines are age-dependent, as are some of the responses to Freund's adjuvant administration.
Results
Figure 1 shows the 24-h changes in hypothalamic NE content in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier. Rhythm parameters as analyzed by Cosinor are summarized in Table 1. In the three hypothalamic regions examined significant 24-h changes of NE content occurred, as shown by individual oneway analysis of variance (ANOVA) (Fig. 1). In the experimental groups in which data fitted a cosine function, acrophases coincided with the second half of activity span or the first half of rest span (04:06 – 13:08 h. Table 1). A factorial ANOVA taking age as a main factor revealed a significantly lower NE content in anterior and medial hypothalamus of aged rats (F1,126= 8.76, p < 0.004, and F1,122 = 4.83, p < 0.03, respectively). In Cosinor, mesor values of NE content in anterior hypothalamus and amplitude values in posterior hypothalamus of old rats were significantly lower than their respective younger counterparts (Table 1).
Twenty-four h changes of hypothalamic NE content in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier. Groups of 5–7 rats were killed by decapitation at six different time intervals throughout a 24-h cycle. Shown are the means + SEM. The groups in which significant differences among time intervals were detected by a one-way ANOVA are indicated by their F values in the Figure. For further statistical analysis, see text.
Cosinor analysis on 24-h changes in hypothalamic NE content and DA turnover in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier.
Mesor
Amplitude
Acrophase (h, min)
Percent of rhythm
NE CONTENT
Anterior hypothalamus
Young, Freund's adjuvant
1030 ± 99
428 ± 34
12:36 ± 02:01
65 ± 12
Young, adjuvant's vehicle
1073 ± 113
444 ± 56
08:09 ± 02:35
51 ± 6.5
Old, Freund's adjuvant
803 ± 77*
499 ± 66
13:08 ± 02:13
48 ± 6.6
Old, adjuvant's vehicle
731 ± 98*
-
-
-
Medial hypothalamus
Young, Freund's adjuvant
661 ± 76
-
-
-
Young, adjuvant's vehicle
762 ± 55
225 ± 35
08:40 ± 02:08
44 ± 5.9
Old, Freund's adjuvant
618 ± 64
-
-
-
Old, adjuvant's vehicle
557 ± 43
285 ± 43
04:06 ± 02:34
70 ± 6.9
Posterior hypothalamus
Young, Freund's adjuvant
142 ± 32
55 ± 9
06:24 ± 02:44
49 ± 9.7
Young, adjuvant's vehicle
143 ± 13
62 ± 11
08:50 ± 02:03
81 ± 10.1
Old, Freund's adjuvant
143 ± 33
36 ± 4*
10:35 ± 01:33
53 ± 7.8
Old, adjuvant's vehicle
152 ± 27
38 ± 6*
11:26 ± 02:22
70 ± 9.2
DA TURNOVER
Anterior hypothalamus
Young, Freund's adjuvant
0.26 ± 0.01
-
-
-
Young, adjuvant's vehicle
0.27 ± 0.03
0.08 ± 0.01
06:51 ± 01:02
64 ± 3.4
Old, Freund's adjuvant
0.24 ± 0.02
0.05 ± 0.02
13:47 ± 01:47
59 ± 6.4
Old, adjuvant's vehicle
0.30 ± 0.01
0.05 ± 0.01
09:28 ± 02:21
53 ± 7.7
Medial hypothalamus
Young, Freund's adjuvant
1.04 ± 0.01
0.22 ± 0.03
20:17 ± 02:33
62 ± 7.7
Young, adjuvant's vehicle
1.16 ± 0.02
n.s.
n.s.
n.s.
Old, Freund's adjuvant
0.96 ± 0.02
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
0.96 ± 0.03
-
-
-
Posterior hypothalamus
Young, Freund's adjuvant
0.34 ± 0.01
n.s.
n.s.
n.s.
Young, adjuvant's vehicle
0.38 ± 0.02
0.05 ± 0.02
09:20 ± 02:36
29 ± 4.4
Old, Freund's adjuvant
0.32 ± 0.02
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
0.30 ± 0.01
0.07 ± 0.01
04:00 ± 01:35
39 ± 4.9
Shown are the means ± S.E.M. (n = 5–7/group). Mesor and amplitude values are expressed as pg/mg (NE content) or as DOPAC/DA ratios. Percent of rhythm defines the part of variation that could be explained by a cosine function in Cosinor. Asterisks designate significant differences (p < 0.05) as compared to the respective young counterparts in a one-tailed Student's t test. n.s.: not significant daily changes in a oneway ANOVA; (-): not significant changes in Cosinor
Figure 2 depicts the 24-h changes in hypothalamic DA turnover. Individual one-way ANOVA's indicated significant time-of-day changes in the anterior hypothalamus of all experimental groups, in the medial hypothalamus of young rats treated with Freund's adjuvant and of old rats treated with adjuvant's vehicle, and in the posterior hypothalamus of young and old rats treated with adjuvant's vehicle. Acrophases occurred during the second half of activity span or first half of rest span (04:00 – 13:47 h), except for the medial hypothalamus of Freund's adjuvant-treated young rats whose acrophases were at the beginning of the activity span (20:17 h. Table 1). A factorial ANOVA taking age as a main factor indicated that medial and posterior hypothalamic DA turnover was significantly lower in old rats (F1,124 = 5.29, p < 0.01 and F1,121 = 6.41, p < 0.001, respectively). Mean values (pg/mg) of 3,4-dihydroxyphenylacetic acid (DOPAC) and DA in young rats were: 38 ± 2.7 and 147 ± 11 (anterior hypothalamus); 83 ± 5.5 and 72 ± 4 (medial hypothalamus); 10 ± 1 and 27 ± 2.3 (posterior hypothalamus). Mean values (pg/mg) of DOPAC and DA in old rats were: 31 ± 2 and 122 ± 9 (anterior hypothalamus); 65 ± 6.6 and 68 ± 5 (medial hypothalamus); 7 ± 1 and 26 ± 1.7 (posterior hypothalamus). In every hypothalamic region of old rats DOPAC concentration was lower than that of young rats (p < 0.05, Student's t test). When immunization was taken as a main factor in factorial ANOVA, a significant reduction in DA turnover was found in the anterior hypothalamus of Freund's adjuvant-treated rats (F1,123 = 3.99, p < 0.04). Mean values (pg/mg) of DOPAC and DA in the anterior hypothalamus of adjuvant's vehicle- and Freund's adjuvant-treated rats were: 37 ± 2.3 and 125 ± 13, and 33 ± 1.9 and 130 ± 12, respectively.
Twenty-four h changes of hypothalamic DA turnover in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier. Groups of 5–7 rats were killed by decapitation at six different time intervals throughout a 24-h cycle. Shown are the means + SEM. The groups in which significant differences among time intervals were detected by a one-way ANOVA are indicated by their F values in the Figure (n.s.: not significant F). For further statistical analysis, see text.
Figure 3 depicts striatal NE content and DA turnover. Striatal NE content varied significantly on a 24-h basis. Acrophases coincided with the second half of activity span or the first half of rest span (06:29 – 12:16 h), except for young rats treated with Freund's adjuvant in which acrophase occurred late in the rest span (17:15 h) (Table 2). Differences in acrophases between immunized and non-immunized young rats were significant (Table 2). Analyzed as a main factor in a factorial ANOVA, old rats had lower striatal NE concentrations (F1,121 = 8.76, p < 0.003). In every case, old rats showed smaller mesor values of striatal NE content than their younger counterparts (Table 2). Amplitude decreased significantly in old rats injected with Freund's adjuvant (Table 2).
Twenty-four h changes of NE content and DA turnover in corpus striatum of young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier. Groups of 5–7 rats were killed by decapitation at six different time intervals throughout a 24-h cycle. Shown are the means + SEM. The groups in which significant differences among time intervals were detected by a one-way ANOVA are indicated by their F values in the Figure (n.s.: not significant F). For further statistical analysis, see text.
Cosinor analysis on 24-h changes in striatal NE content and DA turnover in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier.
Mesor
Amplitude
Acrophase (h, min)
Percent of rhythm
NE content
Young, Freund's adjuvant
40.1 ± 5.2
21.1 ± 3.6
17:15 ± 01:34
39 ± 4.5
Young, adjuvant's vehicle
49.6 ± 6.6
16.8 ± 2.1
06:29 ± 01:02#
33 ± 5.6
Old, Freund's adjuvant
27.9 ± 4.1*
8.5 ± 2.8**
12:16 ± 02:01
27 ± 5.0
Old, adjuvant's vehicle
30.3 ± 5.4*
12.5 ± 2.2
11:49 ± 02:21
50 ± 7.6
DA turnover
Young, Freund's adjuvant
0.11± 0.03
n.s.
n.s.
n.s.
Young, adjuvant's vehicle
0.19 ± 0.03*
n.s.
n.s.
n.s.
Old, Freund's adjuvant
0.04 ± 0.03
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
0.15 ± 0.05#
n.s.
n.s.
n.s.
Shown are the means ± S.E.M. (n = 5–7/group). Mesor and amplitude values are expressed as pg/mg (NE content) or as DOPAC/DA ratio. Percent of rhythm defines the part of variation that could be explained by a cosine function in Cosinor. Asterisks designate significant differences (*p < 0.05; **p < 0.01) as compared to their respective young counterparts; #p < 0.05 as compared to their respective Freund's adjuvant-treated counterparts (one-tailed Student's t test), n.s.: not significant daily changes in a one-way ANOVA.
Time-of-day changes in striatal DA turnover were not significant as indicated by individual one-way ANOVA's (Fig. 3). In a factorial ANOVA, the effects of both main factors "age" and "immunization" were significant, aging and Freund's adjuvant decreasing striatal DA turnover (F1,122 = 5.97, p < 0.01 and F1,123 = 13.4, p < 0.00001, respectively). Mean values (pg/mg) of striatal DOPAC and DA in young and old rats were: 437 ± 52 and 3123 ± 225, and 189 ± 21 and 2631 ± 202, respectively. Differences in striatal DOPAC concentration between young and old rats were significant (p < 0.01, Student's t test). Mean values (pg/mg) of striatal DOPAC and DA in adjuvant's vehicle-and Freund's adjuvant-treated rats were: 411 ± 54 and 2417 ± 256, and 215 ± 25 and 2152 ± 234, respectively. The decrease in striatal DOPAC concentration after Freund's adjuvant administration was significant (p < 0.01, Student's t test).
Hypophysial NE and DA levels are depicted in Fig. 4. NE levels of the neurointermediate lobe showed significant 24-h variations in all groups except for old rats receiving Freund's adjuvant. Acrophases were at 02:31 – 03:34 h (Table 3). In a factorial ANOVA in which age was analyzed as a main factor, a significant depression of NE content was detected in aged rats (F1,123 = 10.1, p < 0.0001). Mesor values were significantly lower in old rats. Amplitude in old rats receiving adjuvant's vehicle was significantly lower than that of their respective young counterparts (Table 3).
Twenty-four h changes of NE and DA content in hypophysial neurointermediate lobe and of DA content of hypophysial anterior lobe in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier. Groups of 5–7 rats were killed by decapitation at six different time intervals throughout a 24-h cycle. Shown are the means + SEM. The groups in which significant differences among time intervals were detected by a one-way ANOVA are indicated by their F values in the Figure (n.s.: not significant F). For further statistical analysis, see text.
Cosinor analysis on 24-h changes in hypophysial NE and DA content, in young and old rats injected with Freund's adjuvant or its vehicle 18 days earlier.
Mesor
Amplitude
Acrophase (h, min)
Percent of rhythm
NE CONTENT
Neurointermediate lobe
Young, Freund's adjuvant
208 ± 32
117
03:34 ± 01:09
40 ± 7.1
Young, adjuvant's vehicle
216 ± 26
102 ± 14
02:31 ± 02:02
51 ± 6.5
Old, Freund's adjuvant
125 ± 15*
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
149 ± 19*
69 ± 8*
02:34 ± 00:56
37 ± 4.4
DA CONTENT
Neurointermediate lobe
Young, Freund's adjuvant
277 ± 43
125 ± 23
04:23 ± 02:00
30 ± 6.2
Young, adjuvant's vehicle
250 ± 40
171± 48
02:47 ± 01:54
86 ± 12
Old, Freund's adjuvant
174 ± 20*
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
112 ± 42*
49 ± 22*
03:25 ± 01:54
61 ± 11
Anterior hypophysis
Young, Freund's adjuvant
20 ± 3
-
-
-
Young, adjuvant's vehicle
21± 4
11± 3
10:31 ± 01:05
75 ± 11
Old, Freund's adjuvant
30 ± 6*
n.s.
n.s.
n.s.
Old, adjuvant's vehicle
39 ± 7*
-
-
-
Shown are the means ± S.E.M. (n = 5–7/group). Mesor and amplitude values are expressed as pg/mg. Percent of rhythm defines the part of variation that could be explained by a cosine function in Cosinor. Asterisks designate significant differences (p < 0.05) as compared to the respective young counterparts in a one-tailed Student's t test. n.s.: not significant daily changes in a one-way ANOVA; (-): not significant changes in Cosinor
Significant 24-h changes in DA content of the neurointermediate lobe occurred in the 4 experimental groups, except for old rats receiving Freund's adjuvant (Fig. 4); acrophases were at 02:47 – 04:23 h (Table 3). A factorial ANOVA indicated a depressive effect of aging on neurointermediate lobe DA content (F1,126 = 28.4, p < 0.00001). In Cosinor, mesor and amplitude values in old rats receiving adjuvant's vehicle were significantly lower than those of their respective young counterparts (Table 3).
Figure 4, lower panel, depicts DA levels in the anterior hypophysis. Significant 24-h changes were observed in all groups except for old rats administered with Freund's adjuvant. Only changes in young animals injected with adjuvant's vehicle fit a cosine function, with acrophase at 10:31 h. In a factorial ANOVA, aging augmented anterior hypophysial DA levels (F1,122 = 37.8, p < 0.00001).
Discussion
The present study, performed in rats sacrificed at 6 different time intervals during a 24-h cycle, documented the following effects of aging on hypothalamic, striatal and hypophysial catecholamines: (i) lower NE content in anterior and medial hypothalamus; (ii) lower DA turnover in medial and posterior hypothalamus; (iii) lower striatal NE concentration and DA turnover; (iv) lower NE and DA content of the neurointermediate pituitary lobe; (v) augmented anterior hypophysial DA content. In addition, the identified effects of Freund's adjuvant administration were: (i) reduction of DA turnover in the anterior hypothalamus; (ii) acrophase delay of DA turnover in medial hypothalamus of old rats; (iii) acrophase delay of striatal NE content of young rats; (iv) decreased striatal DA turnover; (v) abolished 24-h variations of neurointermediate lobe NE and DA content, and of anterior lobe DA content, in old rats.
Most previous studies on the effect of aging on brain and peripheral catecholamines in rats have been obtained as single time points in a 24-h cycle. Decreases in NE content in limbic areas, spinal cord, medulla oblongata and pons, and less consistently, in the hypothalamus and corpus striatum of old rats were identified [6-12]. In aging male rats, levels of hypothalamic NE either decreased [6,13-16] or remained unchanged [9,17,18].
Our present results indicate that aged rats had lower NE content in the anterior and medial hypothalamus. In addition, hypothalamic NE levels showed significant 24-h changes with acrophases at the second half of activity span or first half of rest span. In the posterior hypothalamus of old rats, amplitude of rhythm was significantly lower than in their respective younger counterparts. Since noradrenergic neurons stimulate LH releasing hormone release [19], the decline in noradrenergic stimulation with aging could contribute to the age-related decrease of FSH and LH described previously in similar groups of rats [5]. It should be noted that a strong positive correlation did exist between the rate constant of NE loss measured in rats and the magnitude of the age-related depletion in NE concentrations within specific brain regions [20].
The present results also support the existence of a lower DA turnover in medial and posterior hypothalamus of aged rats. This was the consequence of a decrease of both DA and DOPAC concentration, being most marked in the case of the latter. DA released by tuberoinfundibular dopaminergic neurons in the hypothalamic periventricular and arcuate nuclei tonically inhibits pituitary prolactin secretion. In turn, prolactin stimulates DA secretion from dopaminergic neurons, forming a feedback loop [21,22]. This autoregulatory feedback control of prolactin is altered in the aged rat, as evidenced by increased circulating concentrations of prolactin and decreased activity of these neurons [22,23]. In addition, tuberoinfundibular DA neurons of old rats failed to respond to exogenous prolactin administration [24].
Prolactin increases occur with aging in male rats of most strains, including Wistar [25] and Long-Evans [26] rats, but not in all, e.g. Sprague-Dawley [27] rats. Most studies have reported a decrease in DA content in the hypothalamus and median eminence of aging male rats, including Wistar [14,15], Long-Evans [28], Sprague-Dawley [9] and F344 [16] rats. In the present study, medial hypothalamic, as well as median eminence DA levels (data not shown), decreased with age. Since the decreases in DOPAC levels exceeded those of DA, DA turnover rate also decreased. Our results agree with measurements of hypothalamic DA turnover in aged Wistar [14] and Long-Evans [28] rats at single time-points in a 24-h cycle.
In vivo, secretion of DA into hypophysial portal blood has been reported to decrease [29] or to increase with aging [30]. Long-term (12 months) testosterone replacement in 24-month-old male Wistar rats increased DA release in the medial preoptic area nearly to levels observed in 3-month-old rats [31]. These findings suggest that DA is involved in the age-related decline in male reproductive function, as indicated by the low FSH, LH and testosterone levels reported in our previous study [5].
In rats, the concentrations of DA in the adenohypophysis increases progressively with age. The increase in the DA content is not a consequence of reduced metabolism of DA, nor is it a static pool because as in young rats, adenohypophysial DA is rapidly decreased by pharmacological treatments that reduce the activity of tuberoinfundibular DA neurons. In aged rats, the amount of DA associated with light particles was 5 times that found in young rats, whereas the amount of DA associated with heavy particles was the same as that in young rats [32]. Therefore, an increased concentration of DA in anterior hypophysis, like that reported in the present study in old rats, does not necessarily result in inhibition of prolactin secretion. It is interesting that, in addition to DA, impairments are observed in the processing (binding, accumulation and intracellular distribution) of hypothalamic hormones in the adenohypophysis of old rats. Taken together, these observations are supportive of the view that the neuroendocrine/endocrine changes appearing with age result from a complex balance of functional alterations occurring at each level, central and peripheral, of the axis.
In the present study, striatal DA and DOPAC levels and turnover rate decreased significantly in aged rats. The results agree with the bulk of information obtained in aged rodents indicating decreased levels and turnover rate of DA [8,9,16,28,33-36]. Generally, the decreased levels of midbrain DA and DOPAC detectable in aged rats have little correlation with age-related changes in the density of dopaminergic receptor binding or the density of DA uptake sites [37].
Significant amounts of NE and DA are present in the neurointermediate lobe of the hypophysis. DA nerve terminals belong to centrally located neurons whereas the origin of NE fibers is in part peripheral, as shown by the 40–50% decrease of posterior pituitary NE content after the bilateral removal of the sympathetic superior cervical ganglia [38,39]. Activation of central noradrenergic input to the magnocellular nuclei augments arginine vasopressin release [40] whereas that of peripheral noradrenergic input decreased arginine vasopressin release [39]. Data presented herein indicate that reduced concentrations of DA occurred in the neurointermediate lobe of the pituitary gland of old rats as compared to those of young rats. Our present results also indicate that NE levels in neurointermediate lobe showed significant 24-h variations in all groups except for old rats receiving Freund's adjuvant. Acrophases were at about the middle of the activity span (at 02:47 – 04:23 h), in concordance with the previously reported maxima in tyrosine hydroxylase activity in the superior cervical, stellate and celiac-superior mesenteric ganglia of young and old rats [41]. Therefore, the central and peripheral sources of noradrenergic innervation of hypophysial neurointermediate lobe seem to have similar 24-h patterns of activation. NE content in neurointermediate lobe and tyrosine hydroxylase in sympathetic ganglia [41] was lowest in old rats.
A number of studies have indicated an age-related reduction in immune function associated with cell mediated immunity in both experimental animals and humans. In advancing age alterations were mainly observed in T cell mediated immunity including decreased proliferative responsiveness of T cells to mitogens, decreased T cell-dependent humoral immune responses, lowered resistance to tumor cell challenge, decreased graft-vs.-host reactivity, delayed skin allograft rejection time, impaired delayed hypersensitivity, reduced cytolytic immune response, altered cytokine production after stimulation, and decreased natural killer cell activity (for references see [42]). Most of the studies were performed at single time points in a 24-h cycle, thus overseeing any effect the circadian system can have on the responses.
In the present study, the injection of Freund's adjuvant to young and old rats was employed as an antigenic challenge. The adjuvant-induced arthritis that follows is considered to be a model for T-lymphocyte-dependent, autoimmune diseases [43] and the central symptoms found partly constitute the "sickness behavior", i.e., the behavioral changes that accompany the immune reaction [44]. One feature of sickness behavior addressed experimentally in the present study was the modification of the 24-hour pattern of hypothalamic, striatal and hypophysial catecholamines occurring 18 days after the injection of complete Freund's adjuvant to young and old rats.
Changes in circadian rhythms are apparent at an early phase of experimental arthritis in rats and persist thereafter [45-49]. In the present study Freund's adjuvant administration brought about, 18 days later, a decrease of DA turnover in the anterior hypothalamus and striatum. It also caused a number of circadian alterations, including acrophase delay of DA turnover in medial hypothalamus of old rats, acrophase delay of striatal NE content of young rats and abolished 24-h variations of neurointermediate lobe NE and DA content, and anterior lobe DA content, in old rats. Experimental evidence suggests that symptomatology after Freund's adjuvant administration are a part of a defense response to antigenic challenge and are mediated by the neural effects of cytokines like interleukin (IL)-l, IL-6, IL-2, granulocyte-macrophage colony-stimulating factor and interferon-α [44,50]. Since we previously reported that immunosuppression restored rhythmicity of several of the neuroendocrine parameters examined in Freund's adjuvant-injected rats [51], the immune-related nature of the studied phenomena seems to be warranted. Suprachiasmatic nuclei themselves may be sensitive to immune-derived signals. Presumably, the chronic stress condition given by mycobacterial adjuvant injection is instrumental in inhibiting a number of circadian rhythms at early and late phases of disease.
Conclusions
Aged rats had lower NE content in the anterior and medial hypothalamus, smaller amplitude of 24-h rhythm in posterior hypothalamic NE content, lower DA turnover in medial and posterior hypothalamus, augmented adenohypophysial DA and decreased NE and DA content in pituitary neurointermediate lobe. These results are probably instrumental for the neuroendocrine/endocrine changes appearing with age. Aged rats had also significantly decreased striatal DA and DOPAC levels and turnover rate when measured at 6 different time points along a 24-h cycle, thus agreeing with the bulk of information obtained at single time points on the occurrence of age-related decreased levels and turnover rate of DA in corpus striatum.
With advancing age alterations were mainly reported on T cell mediated immunity. We hereby reported that during Freund's adjuvant-induced arthritis (a T-lymphocyte-dependent autoimmune disease) a decrease of DA turnover occurred in the anterior hypothalamus and striatum. We also observed that responses to immunization like suppression of 24-h rhythms of neurointermediate lobe NE and DA and of anterior lobe DA were only seen in aged rats. Collectively, our results indicate that 24-h rhythms and levels of hypothalamic, striatal and hypophysial catecholamines are age-dependent, as are some of the responses to Freund's adjuvant administration.
Materials and MethodsAnimals
Experiments were carried out in adult male Wistar rats, kept under light between 0800 and 2000 h daily. Light intensity at the level of the animal cages was about 200 lux. Rats had access to food and water ad libitum. Adequate measures were taken to minimize pain or discomfort, in accordance with the principles and procedures outlined in European Communities Council Directives (86/609/EEC).
Groups of young (2 months) and aged (18–20 months) rats were injected s.c. with Freund's complete adjuvant (0.5 mg heat-killed Mycobacterium butyricum/rat) or its vehicle (0.5 ml paraffin oil containing 15% mannide monooleate) 3 h after light on (HALO) (i.e., at 11:00 h). The effect of varying the time of Freund's adjuvant injection on day-night differences of submaxillary lymph node ornithine decarboxylase activity (an index of lymph node proliferative response) was examined in a previous work [52]. Immunization performed during daylight (5 HALO) or at night (18 HALO) resulted in similar day-night differences in ornithine decarboxylase activity, indicating that changes in lymph node proliferative responses were relatively independent of the biological time of mycobacterial antigen exposure [52]. We maintained a similar injection schedule as in several previous studies conducted on Freund's adjuvant effects on immune and endocrine responses [53,54]; thus, rats included in the present study were injected with Freund's adjuvant at 3 HALO (11:00 h).
Although arthritis is induced most easily in inbred Lewis rats, it is also produced, to a milder extent, in Wistar rats [55-58]. Rats injected with Freund's adjuvant vehicle were included as a control of any inflammatory reaction the adjuvant's oil alone might cause [59-61]. The course of adjuvant-induced arthritis was followed by behavioral observations including those of spontaneous behavior-mobility, exploring, rearing and scratching [4,43]. Eighteen days after Freund's adjuvant injection a lack of mobility and exploring behavior, an increase in scratching behavior and signs of hyperalgesia were clearly established in young and old rats as compared with their respective adjuvant's vehicle-injected groups. As reported previously by using plethysmography [62], old rats exhibited less behavioral signs of inflammation (spontaneous behavior-mobility, exploring, scratching) than young rats.
On day 18 after injection, groups of 5–7 rats were killed by decapitation at 6 different time intervals throughout a 24-hour cycle. The brains were quickly removed and the hypothalamus, corpus striatum and pituitaries were taken out. The hypothalamus was further sectioned in the frontal plane, the anterior and posterior regions comprising one-third of the block each [63].
Catecholamine assays
Tissue was weighed and homogenized in chilled (0–1°C) 2 M acetic acid. After centrifugation (at 15 000 × g for 30 min, at 5°C), the samples were analyzed by high performance liquid chromatography using electrochemical detection (Coulochem, 5100A, ESA; USA). AC-18 reverse phase column eluted with a mobile phase (pH 4, 0.1 M sodium acetate, 0.1 M citric acid, 0.7 mM sodium octylsulphate and 0.57 mM EDTA containing 10% methanol, v/v), was employed. Flow rate was 1 ml/min, at a pressure of 2200 psi. Fixed potentials against H2/H+ reference electrode were: conditioning electrode: -0.4 V; preoxidation electrode: +0.10 V; working electrode: +0.35 V. Catecholamine concentrations were calculated from the chromatographic peak heights by using external standards. The linearity of the detector response for NE, DA and DOPAC (a major catabolite of DA) was tested within the concentration ranges found in hypothalamic supernatants. The turnover of DA was assessed by the DOPAC/DA ratio [64]. Although DOPAC concentration depends on the balance between catabolite synthesis and clearance, its acidic nature slows substantially its clearance from tissue, so that DOPAC concentration reflects an integral of past DA release.
Statistical analysis
Statistical analysis of results was performed by a one-way ANOVA, a two-way factorial ANOVA or one-tailed Student's t test, as stated. Cosinor analysis was used to analyze general rhythmic parameters, i.e., acrophase (the maximum of the cosine function fit to the experimental data), mesor (the statistical estimate of the 24-h time series mean) and amplitude (half the difference between maximal and minimal values of the derived cosine curve). Percent of rhythm defined the part of variation that could be explained by a cosine function. Statistical analysis of Cosinor parameters was carried out by standard procedures [65]. Statistical significance of the derived cosine curves was tested against the null hypothesis (i.e., amplitude = 0) [66]; p values lower than 0.05 were considered evidence for statistical significance.
Acknowledgements
This work was supported by grants from DGES, PB97-0257, Spain, University of Buenos Aires, Beca Ramón Carrillo – Arturo Oñativia, Ministerio de Salud, Argentina and Agencia Nacional de Promoción Científica y Tecnológica, Argentina (PICT 6153).
The L-type Ca2+ channel formed by the dihydropyridine receptor (DHPR) of skeletal muscle senses the membrane voltage and opens the ryanodine receptor (RyR1). This channel-to-channel coupling is essential for Ca2+ signaling but poorly understood. We characterized a single-base frame-shift mutant of α1S, the pore subunit of the DHPR, that has the unusual ability to function voltage sensor for excitation-contraction (EC) coupling by virtue of expressing two complementary hemi-Ca2+ channel fragments.
Results
Functional analysis of cDNA transfected dysgenic myotubes lacking α1S were carried out using voltage-clamp, confocal Ca2+ indicator fluoresence, epitope immunofluorescence and immunoblots of expressed proteins. The frame-shift mutant (fs-α1S) expressed the N-terminal half of α1S (M1 to L670) and the C-terminal half starting at M701 separately. The C-terminal fragment was generated by an unexpected restart of translation of the fs-α1S message at M701 and was eliminated by a M701I mutation. Protein-protein complementation between the two fragments produced recovery of skeletal-type EC coupling but not L-type Ca2+ current.
Discussion
A premature stop codon in the II-III loop may not necessarily cause a loss of DHPR function due to a restart of translation within the II-III loop, presumably by a mechanism involving leaky ribosomal scanning. In these cases, function is recovered by expression of complementary protein fragments from the same cDNA. DHPR-RyR1 interactions can be achieved via protein-protein complementation between hemi-Ca2+ channel proteins, hence an intact II-III loop is not essential for coupling the DHPR voltage sensor to the opening of RyR1 channel.
Background
The dihydropyridine receptor (DHPR) of skeletal muscle consists of α1S, α2, β1a and γ1 subunits [1]. The α1 subunit is a large four-repeat transmembrane protein of ~220 KDa that contains the basic functional elements of the L-type Ca2+ channel, including the Ca2+ selective pore and S4 "voltage-sensing" transmembrane segments in each of the four internal repeats [2]. β subunits are ~65 kDa cytosolic proteins essential for membrane trafficking, modulation of channel kinetics, and for excitation-contraction (EC) coupling [3,4]. The α2 subunit is a highly glycosylated ~175 kDa protein formed by two disulfide-linked peptides [5], whereas the γ1 subunit is a ~32 kDa skeletal muscle-specific protein of four presumptive transmembrane domains with almost unknown function [6,7].
Skeletal muscle cells utilize the voltage sensors formed by the S4 segments to trigger a rapid elevation of cytosolic Ca2+, thus coupling membrane excitation to muscle cell contraction. Subsequent to charge movements in the voltage sensors, a conformational change in the DHPR is transmitted to the ryanodine receptor (RyR1), presumably, via protein-protein interactions [8]. Ultimately, there is a brief opening of the RyR1 channel resulting in the release of Ca2+ from the sarcoplasmic reticulum (SR). Numerous observations have lent support to this view, and especially significant are the functional expression studies in dysgenic myotubes lacking α1S. The dysgenic myotube is devoid of L-type Ca2+ current, charge movements and EC coupling. All three are restored in the dysgenic myotube by expression of α1S[9-11]. These results corroborated the essential role of α1S in the mechanism of EC coupling of skeletal muscle cells.
The mechanism by which the DHPR signals the RyR1 is poorly understood [12,13]. Domains in the cytoplasmic linker between repeats II and III have been clearly implicated [14-19], and some regions such as Thr671-Lue690 were suggested to trigger RyR1 opening by binding to RyR1 [15]. However, extensive deletions within the II-III linker that eliminate the RyR1 binding region, and other suggested signaling regions in the II-III loop [16], do not entirely eliminate EC coupling [20,21]. Hence additional domains of α1S and/or other DHPR subunits appear to be engaged by the voltage sensor and contribute to an EC coupling signal. In this respect, the contribution of the β1a subunit of the DHPR to EC coupling in skeletal muscle cells has been extensively documented [4,22-24].
In the present report, we characterized a frame-shift mutant of α1S that expresses two complementary fragments of α1S. Complementation between the two α1S fragments produced recovery of EC coupling in dysgenic muscle cells lacking α1S. The results suggest the EC coupling voltage sensor of skeletal muscle is modular in function and can be assembled from separate hemi-Ca2+ channel fragments.
Results and DiscussionExpression of a frame-shift mutation of α1S in dysgenic myotubes
Primers for the frame-shift mutant, fs-α1S, were originally designed to delete the 20-mer Thr671-Leu690 in the cytosolic loop between repeats II and III of α1S and to generate a full-length α1S carrying this internal deletion. A proofreading error during a PCR reaction resulted in an amplified DNA with the desired deletion but also containing an additional thymidine following the TTG codon for Leu670 (Fig. 1A). The one-base shift in reading frame introduced a serine at position 671 followed immediately by a stop codon (Fig. 1B). This frame-shift mutation was re-ligated into an otherwise full-length α1S, subcloned into the mammalian expression vector pSG5, and transfected into dysgenic (α1S null) myotubes. Fs-α1S was abundantly expressed in primary dysgenic myotubes in culture (Fig. 1C) and produced the expected truncated α1S protein (Fig 1D). Western blots using N-terminus T7-tagged fs-α1S and T7-tagged full-length α1S showed that the expressed full-length α1S protein migrated with an apparent molecular weight of approximately 185 KDa under reducing conditions. This result is consistent with the mobility of the native purified skeletal muscle α1S subunit [25]. The fs-α1S migrated with a molecular weight of approximately 90 KDa which is entirely consistent with the theoretical molecular weight of the expressed fragment which was 85.6 KDa. Furthermore, 5-fold overloading of the SDS-PAGE gel failed to detect any fragment of a size comparable to full-length α1S (not shown).
Nucleotide sequence and protein expression of fs-α1S. A) Nucleotide sequence of wt-α1S and fs-α1S in the region of the frame-shift. B) The three reading frames of fs-α1S in the region of the frame-shift are shown. Translation of the C-terminal half of α1S is explained by a restart of translation at the indicated ATG codon which is located 25 bases downstream from the termination codon indicated by the asterisk. C) Confocal images (calibatrion bar is 10 microns) show details of the intracellular distribution of the expressed proteins. Cells were transfected with the CD8 cDNA plus wt-α1S or fs-α1S. Cells were incubated with CD8 antibody beads, fixed, and stained with T7 primary/fluorescein-conjugated secondary antibodies. Pixel intensity was converted to a 16-level inverted gray scale with high-intensity pixels in black color. Asterisks show on-focus CD8 antibody beads (diameter 4.5 microns) bound to cells. NT indicates a non-transfected myotube in the same focal plane of the transfected cell. D) Immunoblots using anti T7 antibody of cultures of dysgenic myotubes expressing wt-α1S and fs-α1S. Indicated are 3 of 7 molecular weight markers run in the same gel.
Recovery of EC coupling by the frame-shift a1S cDNA
EC coupling was investigated in voltage-clamped myotubes with simultaneous monitoring of intracellular Ca2+ using confocal fluorescence of fluo-4 [26]. Controls shown in Fig. 2A indicated that the overwhelming majority of non-transfected dysgenic muotubes (13 of 15 cells) did not produce detectable Ca2+ transients (<0.1 ΔF/Fo) or Ca2+ currents (<20 pA/cell) in response to depolarization under voltage-clamp. This is shown in the line-scan images of fluo-4 fluorescence in Fig. 2A and the corresponding traces of ICa2+ during a 50-ms depolarization to +30 mV and +90 mV delivered at the start of the line scan in the same cell. However, in two cells (2 of 15 cells) we observed Idys, the low-density endogenous Ca2+ current previously described in dysgenic myotubes [27,28]. The reason for the low abundance of this current in these cultured myotubes is unknown. Ca2+ currents and stimulated fluorescence for one of the cells expressing Idys is shown in Fig. 2B. We observed a peak ICa2+ density of approximately 0.8 pA/pF and a barely detectable fluorescence signal which in Fig. 2B is indicated by the arrow in the trace of integrated fluorescence at +30 mV. This small fluorescence signal disappeared entirely at +90 mV, suggesting it might be contributed directly by Idys or might be due to SR Ca2+ release induced by Idys. The voltage dependence of the fluorescence signal and ICa2+ are compared in Fig. 2C for the two cells expressing Idys and for the vast majority of cells which altogether did not express intracellular Ca2+ transients or ICa2+. The maximum fluorescence signal contributed by Idys, when Idys was present, was <0.2 ΔF/Fo units. Furthermore, the shape of the fluorescence vs. voltage relationship was bell-shaped and a mirror image of the ICa2+ vs. voltage curve. These controls indicated that non-transfected dysgenic myotubes are low-background cells that do not express voltage-activated Ca2+ signals of major consequence for the present studies.
Absence of EC coupling in non-transfected dysgenic myotubes. The confocal line-scan images in color show fluo-4 fluorescence across myotubes in response to a 50-ms depolarization from a holding potential of -40 mV. Line scan images have a constant temporal dimension of 2.05 s (horizontal) and a variable spatial dimension (vertical) depending on the cell size. Traces immediately above each line scan show the time course of the fluorescence change in resting units (ΔF/Fo). The amplitude and the timing of the depolarization are indicated under each line-scan. Arrow indicates a small Ca2+ transient elicited in 1 of 2 cells found to express Idys. Traces next to lines-cans show ICa2+ during the 50 ms depolarization used to stimulate the Ca2+ transient. Current calibration bars are 10 ms and 1 pA/pF. A) Absence of Ca2+ transients and ICa2+ in a typical dysgenic cell. B) Minor Ca2+ transient and ICa2+ in a cell expressing Idys. Note that fluorescence calibration bar is 0.5 ΔF/Fo. A 16-color calibration bar in ΔF/Fo units is included in Fig. 3 for visual reference. C) Voltage dependence of the mean (± SEM) ICa2+ and mean peak Ca2+ transient (± SEM) in 13 cells not expressing ICa2+ and 2 cells expressing Idys (mean only).
Fig. 3 shows that fs-α1S recovered a significant fraction of the voltage-activated Ca2+ transient compared to that express by full-length wt-α1S. The magnitude of the fluorescence signal expressed by fs-α1S was approximately 5-fold larger than the largest Ca2+ transient detected in non-transfected myotubes expressing Idys, >20-fold larger than the average Ca2+ transient detectable in non-transfected cells, and about 1/3 of the maximum SR Ca2+ release expressed by the control wt-α1S construct. Thus, we are confident that the voltage-evoked Ca2+ transient in cells transfected by fs-α1S cells was a direct consequence of the expressed protein. Also shown in Fig. 3 is ICa2+ activated by the 50-ms depolarization used to activate the Ca2+ transient. fs-α1S did not express L-type Ca2+ current even though it was consistently able to activate the Ca2+ transient in 15 of 15 cells. Absence of ICa2+ was further verified using longer 500-ms depolarizing pulses (not shown). The skeletal nature of the EC coupling expressed by fs-α1S is shown in Fig. 4A. The peak Ca2+ vs. voltage relationship expressed by fs-α1S, like that of wt-α1S, was sigmoidal in shape reaching a maximum at large positive potentials (>50 mV), a range in which ICa2+ is progressively small. The line scans of Fig. 4B further confirmed that a Ca2+ transient of similar shape and magnitude was observed in a fs-α1S transfected myotube in the absence of external Ca2+. Hence the signaling mechanism, like that reported in normal myotubes and dysgenic myotubes expressing wt-α1S, was Ca2+ entry independent [29]. We also expressed fs-α1S in cultured myotubes from two available gene knock-out (KO) mice, lacking the endogenous β1a isoform of the skeletal muscle DHPR [30] and lacking RyR1 [31]. As shown in Fig. 4B, we failed to detect EC coupling in these KO cells transfected with fs-α1S. In summary, the EC coupling expressed by fs-α1S is strictly skeletal-type and requires RyR1 and DHPR β1a.
Ca2+ transients in dysgenic myotubes transfected with fs-α1S. The confocal line-scan images in color show fluo-4 fluorescence across myotubes in response to a 50-ms depolarization from a holding potential of -40 mV to +30 mV (top) and +90 mV (bottom). Line scan images have a constant temporal dimension of 2.05 s (horizontal) and a variable spatial dimension (vertical) depending on the cell size. Traces immediately above each line scan show the time course of the fluorescence change in resting units (ΔF/Fo). Traces under lines cans show ICa2+ during the 50 ms depolarization used to stimulate the Ca2+ transient. The amplitude and the timing of the depolarization are indicated under each line scan. Note that fluorescence calibration bar is 1 ΔF/Fo. A 16-color calibration bar in ΔF/Fo units is included for visual reference.
Skeletal-type EC coupling expressed by fs-α1S. A) Voltage-dependence of peak Ca2+ for 5 control myotubes expressing wt-α1S and 5 myotubes expressing fs-α1S. Ca2+transients for 15 non-transfected myotubes (NT) are included for reference. The sigmoidal lines are a Boltzmann fit with parameters ΔF/Fo max = 2.9, 1.4 ΔF/Fo; V1/2 = 11.7, 20 mV; k = 8.7, 13.2 mV, for wt-α1S and fs-α1S respectively. B) Line-scans (horizontal dimension is 2.05 seconds) and traces of integrated fluorescence in ΔF/Fo units for depolarizations to +30 mV. Top line-scans are for the same fs-α1S transfected myotube in standard external solution (10 mM CaCl2) and the same solution without added CaCl2 (0 Ca2+). Bottom line-scans show fs-α1S transfected KO myotubes lacking DHPR β1a or lacking RyR1 in standard external solution.
fs-α1S expresses two complementary protein fragments
The EC coupling recovered by fs-α1S could be due either to the activity of the N-terminal half of α1S alone or to protein-protein complementation between the N-terminal half and a fragment expressing the C-terminal half of α1S. The C-terminal half of α1S could have been translated by the fs-α1S expression vector if the ATG codon (Met701), which is downstream from the TGA termination codon and is in-frame with the wild-type message (Fig 1B) served as open reading frame for translation of the second half of the wt message. Although this would be unusual, the fact that the codon for Met701 is only 25 bases downstream from the termination codon could have substantially increased the possibility of a re-start of the translation of the second half of the message at Met701. This phenomenon has been described in eukaryotic cells and in viral-infected mammalian cells and is known as translation by leaky ribosomal scanning [32,33]. To test this explanation, the presumptive restart condon, Met701, was mutated to Ile701 in the fs-α1S template. If fs-α1S recovered EC coupling by virtue of expressing a single protein fragment, then fs-α1SM701I should also recover EC coupling since the mutation was introduced downstream from the stop codon. Fig. 5 shows that this was not the case. Fs-α1SM701I did not recover Ca2+ transients in 9 of 9 tested cells, consistent with leaky ribosomal scanning. As a positive control, we coexpressed fs-α1SM701I and the C-terminus half of α1S, namely α1SΔ1–700, cloned into a separate pSG5 vector. The results in Fig. 5 indicated that α1SΔ1–700 alone was inactive. However, when myotubes were cotransfected with fs-α1SM701I and α1SΔ1–700, each in a separate pSG5 vector, there was a robust recovery of Ca2+ transients in 5 of 5 cells. Fig. 6A shows fluorescence vs. voltage relationships for the fs-α1SM701I mutant and for this mutant coexpressed with α1SΔ1–700. The combined expression of the two complementary fragments of α1S resulted in a robust recovery of EC coupling with sigmoidal Ca2+ release vs. voltage characteristics. A summary of the maximum fluorescence during the Ca2+ transient in response to a depolarization to +90 mV is shown in Fig. 6B. The magnitude of the Ca2+ transient expressed by fs-α1SM701I + α1SΔ1–700 was indistinguishable from that of wt-α1S (t-test significance p = 0.671, see figure legend). To confirm expression of the C-terminus half of α1S in cells transfected with fs-α1S, we used the II-III loop polyclonal antibody SKI [34] directed against epitope Ala739-Ile752 which is downstream from Met701. Fig. 6C shows that the II-III loop antibody recognized the C-terminus half when cells were transfected with fs-α1S but not when myotubes were transfected with fs-α1SM701I. The C-terminal protein migrated with a molecular weight of approximately 126 KDa which is consistent with the theoretical molecular weight of 132 KDa. Finally, Fig. 6D shows that fs-α1SM701I was abundantly expressed in myotubes in the absence or presence of the C-terminal fragment. This indicated that the absence of EC coupling observed in myotubes expressing fs-α1SM701I was not due the production of a labile protein. In summary, the recovery of EC coupling by coexpression of two functionally inactive proteins (Fig. 5) taken together with the immunoblots (Fig. 6C) favor the explanation that 1) fs-α1S recovers DHPR function by virtue of expressing two complementary fragments of α1S and 2) the expression of the C-terminal half of α1S by fs-α1S is likely to occur by leaky ribosomal scanning.
EC coupling generated by two complementary fragments of α1S. Line scans (horizontal dimension is 2.05 seconds) of fluo-4 fluorescence show Ca2+ transients in response to the indicated 50-ms depolarization from a holding potential of -40 mV. Trace of integrated fluorescence in ΔF/Fo units is shown for each line scan. Each set of depolarizations is from a separate dysgenic myotube expressing the α1S construct(s) indicated at the top of each column. A 16-color calibration bar in ΔF/Fo units is included in Fig. 3 for visual reference.
Expression of C-terminal fragment of α1S is essential for EC coupling. A) Voltage dependence of peak Ca2+ during the Ca2+ transient for dysgenic myotubes transfected with the indicated constructs. The sigmoidal curve is a Boltzmann fit with parameters ΔF/Fo max = 2.45; V1/2 = 15.4 mV; k = 9.3 mV for 5 cells coexpressing fs-α1SM701I + α1SΔ1–700. Absence of response is shown for 9 cells expressing fs-α1SM701I alone. B) ΔF/Fo max (mean ± SEM) obtained from a depolarization to +90 mV is shown for the indicated number of cells. NT denotes non-transfected dysgenic myotubes. Compared to wt-α1S (control), the statistical significance in unpaired t-Student test was p = 1.6 × 10-6 (non-transfected, NT); 0.014 (fs-α1S); 0.0002 (fs-α1SM701I); 0.0008 (α1SΔ1–700); 0.671 (fs-α1SM701I + α1SΔ1–700). Compared to fs-α1S, the statistical significance was p = 1 × 10-6 (non-transfected, NT); 0.014 (wt-α1S); 0.00019 (fs-α1SM701I); 0.0008 (α1SΔ1–700); 0.013 (fs-α1SM701I + α1SΔ1–700). C) Immunoblots using a polyclonal antibody directed to the II-III loop epitope Ala739-Ile752 [34] in cultures of dysgenic myotubes expressing fs-α1S and fs-α1SM701I. Indicated are 3 of 7 molecular weight markers run in the same gel. D) Confocal images of cells transfected with the CD8 cDNA plus T7-tagged fs-α1SM701I or T7-tagged fs-α1SM701I + untagged α1SΔ1–700. Cells were incubated with CD8 antibody beads, fixed, and stained with T7 primary/ fluorescein-conjugated secondary antibodies Pixel intensity was converted to a 16-level inverted gray scale with high-intensity pixels in black color. NT indicates a non-transfected myotube. CD8 antibody beads have a diameter of 4.5 microns. Calibration bar is 10 microns.
Implications for EC coupling in skeletal myotubes
Except for the magnitude, the SR Ca2+ release signal expressed by fs-α1S was entirely typical of skeletal myotubes with sigmoidal voltage-dependence, proceeding in the absence of external Ca2+ and requiring RyR1. A comparison of the maximum fluorescence (ΔF/Fo max) at +90 mV (Fig. 6B) shows that the signal generated by fs-α1S was significantly smaller than that generated by the control construct (wt-α1S vs. fs-α1S t-test significance p = 0.014) and smaller than that generated by the two coexpressed fragments (fs-α1S vs. fs-α1SM701I + α1SΔ1–700 t-test significance p = 0.013). These observation suggests that the magnitude of the Ca2+ release appears to be limited by the low yield of expression of the C-terminal fragment achieved by leaky ribosomal scanning of the second half of the fs-α1S message. To test this explanation further, we coexpressed fs-α1S and the C-terminal half of α1S each in a separate pSG5 vector. We found that fs-α1S and α1SΔ1–700 together expressed Ca2+ transients with a ΔF/Fo max similar to wt-α1S control (not shown). Thus we are certain that the EC coupling expressed by fs-α1S is mechanistically similar to control skeletal-type EC coupling but limited in magnitude by a comparatively lower density of functional DHPRs that are assembled in cells expressing fs-α1S. It is conceivable that the functional integrity of the fragmented α1S protein is maintained in part by the β subunit of the DHPR which spans both halves of the α1 pore subunit by binding to the I-II loop and the C-terminus [35,36]. Consistent with this explanation, we failed to detect EC coupling recovery when fs-α1S was expressed in β1-null myotubes.
Earlier studies in the voltage-gated Na+ channel had shown that pore function was not compromised when the II-III linker or III-IV linker was cut and the two fragments were coexpressed each in a separate vector [37]. We would thus conclude that in the case of the Ca2+ channel, an intact II-III loop is essential for this function since neither fs-α1S nor the combined expression of the two truncated fragments (not shown) was able to rescue Ca2+ current. This result is entirely consistent with the identification of the II-III loop as critical for enhancement of L-type Ca2+ current expression by the RyR1 [38]. However EC coupling, per se, can clearly proceed with a cut in the II-III loop. This was shown here by the behavior of the fs-α1S construct and elsewhere by expressing the N-terminal half (α1S1–670) and the C-terminal half (α1S701–1873), each with entirely wt sequence and each in a separate expression vector [20]. The fact that Thr671-Leu690 region known as Peptide A [15] is missing in fs-α1S suggests this 20-mer domain is not critical for the conformation change transmitted from the DHPR to the RyR1. To test this further, we generated an in-frame deletion of this region, as originally intended, that showed normal function [20]. Another domain critical for EC coupling is Csk35 downstream from Peptide A (Leu720-Gln765). This region was identified using chimeras of α1S and α1C[16]. Since Csk53 is present in α1SΔ1–700 and was detected by the II-III loop antibody which is directed against the center portion of Csk53 [34], the participation of this domain in the EC cannot be ruled-out.
In eukaryotic cells, translation starts at the AUG codon nearest to the 5' end of the mRNA, and this initiator site is found by sequential ribosomal scanning in the 5' to 3' direction [32]. However, translation initiation at internal AUGs due to leaky ribosomal scanning has been documented, especially for mRNAs consisting of a short leader ORF upstream from the main ORF. In some cases, reassembly of a new ribosomal initiation complex after the terminator codon of the leader ORF serves to re-initiate translation at the initiator AUG codon for the main ORF, and thus two proteins are generated [39]. In other cases, leaky scanning by the ribosomal initiation complex bypasses the 5'-located leader ORF entirely, and only the "internal" ORF is translated [32]. In the case of fs-α1S, we are not entirely certain which mechanism best applies. A bypass of the ORF at the 5' end of the α1S mRNA in favor of a presumed ORF at Met701 is unlikely because a C-terminal fragment of the size expressed by α1SΔ1–700 has not been detected in skeletal muscle with the SKI antibody [34] or other antibodies [25]. Thus Met701 is not an internal initiation site under normal circumstances. Entry of a new ribosomal complex at Met701 seems a more likely explanation. However, re-initiation after a stop has only been described for cases in which the leader ORF is no longer than 30 codons because initiation factors fall off shortly after recognition of the initiator AUG [32]. In the case of fs-α1S, what can be considered the "leader" ORF encodes for a protein of 671 residues. Hence we are not certain if re-initiation of translation after a stop signal, as currently described in the literature, would apply here. At the same time, it is important to point out that in the present studies, fs-α1S expression is under the control of a viral promoter and that in this hybrid viral-mammalian expression system, the rules pertaining to leaky ribosomal scanning may be different. The mechanism of translation of the fs-α1S clearly deserves closer scrutiny in the future.
Conclusions
The present studies show EC coupling recovery by a frame-shift mutant of α1S due to protein-protein complementation of the N-terminal and C-terminal halves of α1S. The N-terminal half houses repeats I and II with the adjoining cytosolic loop and the C-terminal half houses most of the II-III loop, along with repeats III and IV with the adjoining loop. Protein-protein complementation between the N-terminal and C-terminal fragments produced a DHPR capable of functioning as EC coupling voltage sensor, thus suggesting the presence of at least two functional modules within α1S. Recent evidence suggests that the four internal repeats of the voltage-gated Na+ channel, which is closely related to the L-type Ca2+ channel encoded by the DHPR, have non-equivalent functional roles because the S4 segments of repeats I and II move much faster than those of repeats III and IV [40]. By analogy, the "fast-moving" module of the DHPR would be represented by the N-terminal fragment and the "slower-moving" module by the C-terminal fragment. Interactions between these two modules are likely to be critical for intramembrane charge movements in the assembled four-repeat channel and for coupling the movement of the S4 gating charges to the opening of the RyR1 channel. Future studies of gating currents in each hemi-Ca2+ channel fragment should provide valuable information on how the "fast" and "slow" gating modules interact during EC coupling in skeletal muscle.
The C-terminal fragment was generated by an unusual restart of translation of the fs-α1S message at M701, presumably by leaky ribosomal scanning, and was eliminated by a M701I mutation. Hence, a premature stop codon in the II-III loop upstream of M701 may not necessarily cause a loss of DHPR function because in these cases, function would be recovered by complementation between protein fragments expressed by the same cDNA. From a methodological perspective, leaky scanning could be further used as a means to control protein expression to desired levels, since restart of translation after a premature stop codon is sensitive to the number of nucleotides separating the stop and restart codons [39]. By changing the position of the restart methionine relative to the premature stop codon, it might be possible to significantly change the level of expression of the distal protein fragment and hence functional protein as a whole. Thus, leaky scanning remains as an attractive possibility for boosting or depressing protein levels in a transfected cell.
Materials and MethodsPrimary cultures of mouse myotubes
Primary cultures were prepared from hind limbs of day 18 embryos (E18) as described previously [23]. cDNAs of interest and a separate expression vector encoding the T-cell membrane antigen CD8 were subcloned into the mammalian expression vector pSG5 (Stratagene, CA) and were mixed and cotransfected with the polyamine LT-1 (Panvera, WI). Whole-cell recordings and immunostaining were done 3–5 days after transfection. Cotransfected cells were recognized by incubation with CD8 antibody beads (Dynal, Norway). The coincidence of expression of CD8 and a cDNA of interest was >85%.
α1S cDNA constructs
All cDNA constructs were sequenced twice or more using BigDye technology (Perkin Elmer, CA) at a campus facility. For epitope tagging and expression in mammalian cells, the unmodified full-length rabbit α1S cDNA encoding residues 1–1873 (Genebank #M23919 nucleotide coordinates nt 226 to nt 5847) was fused in frame to the first 11 amino acids of the phage T7 gene 10 protein in pSG5 using AgeI and NotI cloning sites. All constructs were made using the T7 tagged α1S as template in PCR-based strategies, some previously described [20,21]. All primers were HPLC-purified (Operon, CA) and a phosphate was tagged to the 5'-end of the sense primer. Genebank #M23919 nucleotide coordinates are used below to describe primers.
pSG5 wt-α1S
A unique silent HindIII site was introduced by PCR at nt 2228 in the full-length α1S template and cloned into the T7-α1S pSG5 vector using AgeI and XhoI sites. The HindIII-XhoI fragment (nt 2228 to nt 2878) encompasing the II-III loop was subcloned into pCR 2.1 TOPO TA (Invitrogen, CA) and this plasmid was further used for PCR reactions.
pSG5 fs-α1S
PCR reactions for deletion of residues 671–690, consisted of 10 nanograms pCR 2.1 TOPO/HindIII-XhoI insert, 15 pmoles of each primer, 0.5 mM dNTPs, 1X cloned Pfu buffer (Stratagene) and 2.5 U cloned Pfu DNA polymerase (Stratagene). The antisense primer was complementary to nt 2202 to nt 2235 and the sense primer was nt 2296 to nt 2326. Amplification was carried out for 30 cycles at 95°C for 45 seconds, 60°C for 2 minutes and 72°C for 2 minutes/kb of plasmid. The PCR reaction was treated with 10 U of DpnI (Stratagene) and recircularized with T4 DNA ligase (Stratagene). Once amplified by PCR, the HindIII-XhoI digest was ligated into the T7-α1S pSG5 vector using the same restriction sites.
pSG5 fs-α1SM701I
The construct was produced by a two-step PCR reaction using fs α1S as template. Using conditions as above, the sense primer (nt 1932 to nt 1959) was paired with antisense primer 5'TCCAGCTTCTTGGCGATCACAGACTTCTCC3' carrying the point mutation. In a seperate reaction, sense primer 5'GGAGAAGTCTGTGATCGCCAAGAAGCTGGA3' was paired with antisense primer (nt 3100 to nt 3081). The two PCR products were diluted 1:500 in ddH2O and then hybridized to each other for 2 minutes at 95°C, 1 minute at 43°C and 1 minute at 72°C for 4 cycles. 15 pmol of nt 1932 to nt 1959 primer and 15 pmol of nt 3100 to 3081 primer were added and further cycled for 4 minutes at 95°C, then 30X of 2 minutes at 95°C, 1 minute at 60°C, 1 minute at 72°C and finally 10 minutes at 72°C. The PCR product was then cloned into the fs α1S construct using HindIII and XhoI sites.
pSG5 a1SΔ1–700
The construct was produced by cutting pSG5 fs α1S with AgeI and HindIII enzymes and filling-in the overhangs with klenow fragments. The plasmid was religated using DNA T4 ligase.
Whole-cell voltage-clamp
Whole-cell recordings were performed as described previously [22] using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). All experiments were performed at room temperature. Patch pipettes had a resistance of 1–2 MΩ . The external solution was (in mM) 130 TEA-Methanesulfonate, 10 CaCl2, 1 MgCl2, 10 HEPES-TEA(OH), pH 7.4. The pipette solution was (in mM) 140 Cs-aspartate, 5 MgCl2, 0.1 EGTA (for Ca2+ transients) or 5 EGTA (for Ca2+ current), 10 MOPS-CsOH, pH 7.2. The voltage dependence of peak intracellular Ca2+ (ΔF/Fo) was fitted according to a Boltzmann distribution (Eqn. 1) A = Amax/(1+exp(-(V-V1/2)/k)). Amax is ΔF/Fomax; V1/2 is the potential at which A = Amax/2; and k is the slope factor. ΔF/Fo=(F-Fo)/Fo where F is the fluorescence during a Ca2+ transient and Fo is the resting fluorescence of the cell immediately before the stimulation.
Confocal fluorescence microscopy
Line-scans were performed as described [26] in cells loaded with 4 mM fluo-3 AM (fluo-3 acetoxymethyl ester, Molecular Probes, OR) for ~30 minutes at room temperature. Cells were viewed with an inverted Olympus microscope with a 20X objective (N.A. 0.4) and a Fluoview confocal attachment (Olympus, NY). Excitation light was provided by a 5 mW Argon laser attenuated to 20% with neutral density filters. For immunofluorescence, confocal images had a dimension of 1024 by 1024 pixels (0.35 microns/pixel) and were obtained with a 40X oil-immersion objective (N.A. 1.3).
Immunostaining
Cells were fixed and processed for immunofluorescence as described [4,20]. The N-terminal fragment expressed by fs-α1S or wt-α1S was identified with a mouse monoclonal antibody against a T7 epitope fused to the N-terminus of α1S. The anti-T7 antibody (Novagen, WI) was used at a dilution of 1:1000. Secondary antibodies were a fluorescein-conjugated goat anti mouse IgG (Boehringer Mannheim, IN) used at a dilution of 1:1000 and a fluorescein-conjugated donkey anti-rabbit IgG (Chemicon, CA) used at a dilution of 1:1000.
Western blots
The C-terminal fragment was identified with SKI, a rabbit polyclonal antibody against the II-III loop of α1S (Ala739-Ile752) previously characterized [34]. Cells were scrapped from tissue cultures dishes with cold PBS plus protease inhibitors and spun in a cold table-top centrifuge. Cells were homogenized in a glass-teflon homogenizer in a minimal volume of PBS and diluted 1:1 (vol:vol) with SDS-gel loading buffer composed 100 mM Tris-Cl (pH 6.8), 200 mM dithiothreitol, 4% SDS, 0.2% bromophenol blue and 20% glycerol. Samples were incubated at 100°C for 20 minutes. Approximately 10 mg of total protein was applied to a 5–15% SDS polyacrylamide gel and electrophoresed for 2 hours at 40 mA. Proteins were transferred to PVDF membranes and analyzed with either anti-T7 or SKI antibodies and the appropriate secondary antibodies. The subunits were visualized using SuperSignal ECL reagent (Pierce, Rockford, IL). The images were captured on a Chemi-Imager (Alpha Innotech, San Leandro, CA) set to a level just below saturation.
Supported by National Institutes of Health Grants HL-47053, AR46448 and by a predoctoral fellowship from Wisconsin Heart Association to C.A.A. We are grateful to Dr. Patricia Powers (University of Wisconsin Biotechnology Center) for suggesting leaky ribosomal scanning as a mechanism consistent with the data.
Iron deficiency (ID) results in ventricular hypertrophy, believed to involve sympathetic stimulation. We hypothesized that with ID 1) intravenous norepinephrine would alter heart rate (HR) and contractility, 2) abdominal aorta would be larger and more distensible, and 3) the beta-blocker propanolol would reduce hypertrophy.
Methods
1) 30 CD rats were fed an ID or replete diet for 1 week or 1 month. Norepinephrine was infused via jugular vein; pressure was monitored at carotid artery. Saline infusions were used as a control. The pressure trace was analyzed for HR, contractility, systolic and diastolic pressures. 2) Abdominal aorta catheters inflated the aorta, while digital microscopic images were recorded at stepwise pressures to measure arterial diameter and distensibility. 3) An additional 10 rats (5 ID, 5 control) were given a daily injection of propanolol or saline. After 1 month, the hearts were excised and weighed.
Results
Enhanced contractility, but not HR, was associated with ID hypertrophic hearts. Systolic and diastolic blood pressures were consistent with an increase in arterial diameter associated with ID. Aortic diameter at 100 mmHg and distensibility were increased with ID. Propanolol was associated with an increase in heart to body mass ratio.
Conclusions
ID cardiac hypertrophy results in an increased inotropic, but not chronotropic response to the sympathetic neurotransmitter, norepinephrine. Increased aortic diameter is consistent with a flow-dependent vascular remodeling; increased distensibility may reflect decreased vascular collagen content. The failure of propanolol to prevent hypertrophy suggests that ID hypertrophy is not mediated via beta-adrenergic neurotransmission.
Background
The iron status of an individual may play an important role in cardiovascular health, with either an excess of iron (mainly the storage form ferritin) or iron deficiency leading to significant problems. Sullivan [1] has proposed that excess iron (i.e., any iron in the ferritin form) leads to formation of free radicals that can worsen ischemic myocardial injury and contribute to atherogenesis. This hypothesis has considerable experimental [2,3] and epidemiological [4,5] support. However, the "iron hypothesis" has not been universally accepted [6], and considerable debate on the role of iron overload on cardiovascular disease continues [7-10].
Iron deficiency has also been shown to lead to ventricular hypertrophy in developing rats [11-17]. While the mechanisms responsible for this hypertrophy have received little research attention, studies have documented an eccentric hypertrophic pattern [15-17] which has led investigators to suspect a ventricular volume overload at end diastole as a primary hypertrophic stimulus. The hypothesized volume load, if it occurs, does not result from a change in blood volume [18]. It has also been hypothesized that a chronic elevation in sympathetic nervous system activity is involved in iron deficiency-induced hypertrophy. Evidence to support this view includes the observations of increased levels of norepinephrine in plasma and urine, and decreased norepinephrine content in the iron deficient heart [19-22]. Rossi [23], after finding that reserpine administration prevented the development of this hypertrophy, even suggested that norepinephrine is the causal agent in the pathology. While little published research has subsequently focused on iron deficiency hypertrophy, there have been numerous studies investigating the role of sympathetic neurotransmission with cardiac hypertrophy induced via various methods (coronary artery ligation, chronic pressure overload, transgenic models, etc.). Most have demonstrated alterations in either plasma or heart norepinephrine content, and a de-sensitization of the beta-adrenergic receptors of the heart [24-33]. Recently, the alpha-adrenergic receptor has been shown in vitro to be an important modulator of ventricular hypertrophy. However, even in transgenic animal models overexpressing alpha receptors, the beta-adrenergic receptor appears to play an important role in hypertrophy development and the transition to heart failure [34,35]. Regardless of which receptors are involved in the various forms of hypertrophy, the sympathetic nervous system does appear to play a role in most, if not all, forms of cardiac hypertrophy, and much remains to be done in this area [36].
Almost no attention has been paid to the peripheral vascular consequences of iron deficiency [37]. With iron deficiency, the poor oxygen-carrying capacity of the blood must be offset in order for the animal to develop to full maturity. Therefore, the chronic sympathetic activation associated with iron deficiency should result in increased cardiac output. In turn, flow through the major arterial network would be enhanced in an attempt to maintain near normal oxygen delivery. Kamiya and Togawa [38] have found that increased flow stimulates remodeling of an artery to a larger diameter through an endothelial dependent mechanism. It has yet to be demonstrated whether iron deficiency will stimulate a similar arterial remodeling.
Collagen is an integral component of the arterial wall, providing rigidity and support. Iron is necessary in the synthesis of collagen [39], and collagen content has been shown to be reduced in iron deficient hearts [40]. If this is similarly true for arteries in iron deficient animals, a significant increase in distensibility would result, altering the normal pressure-volume relationship.
We investigated three hypotheses using anesthetized rats. First, we hypothesized that in iron deficient rats, intravenous norepinephrine would cause an altered cardiovascular response, either in heart rate or rate of arterial pressure generation (contractility, dp•dt-1). Second, we hypothesized that iron deficient arteries would be larger and more distensible than in iron replete controls. Finally, we hypothesized that a daily injection of the beta-adrenergic antagonist propanolol would significantly reduce the development of ventricular hypertrophy.
Our results suggest that iron deficient, hypertrophic hearts display a hyper-sensitive inotropic, but not chronotropic, response to norepinephrine. Second, abdominal aortas from iron deficient rats display not only increased diameters, but also significant increases in distensibility. Finally, propanolol injections do not prevent the development of iron deficient cardiac hypertrophy, drawing into question the potential causal relationship between beta-adrenergic neurotransmission and this form of hypertrophy.
Results and DiscussionBody weight, heart weight, and hematocrit (Experiments 1 and 2)
Rat growth rate is shown in figure 1. There were no differences between any of the experimental groups (see Materials and Methods) until the animals reached 45 days of age, after which, iron deficient rats are of less mass than controls (*, p < 0.05). Figure 1 also shows the final mean body mass of the four experimental groups. Iron deficient and control groups more than doubled their body mass from 1 week to 1 month on the respective diets (†, p < 0.0001). The control rats also had a significantly greater mass than the iron deficient rats (p = 0.0004) in the one month group (*, p = 0.0006), but not at one week (p = 0.1530). The control rats in this study increased mass at a rate similar to that published for the ad libitum fed CD rat [41,42], suggesting growth was inhibited by iron deficiency. It is likely that growth inhibition is attributable to metabolic causes arising from both oxygen delivery and mitochondrial insufficiencies associated with iron deficiency [14]. The growth data for both the control and iron deficient groups are also similar to that reported for male Wistar rats fed an iron and copper deficient diet [15], but the iron deficient rats at 52 days of age were larger than reports of 10 week old Harlan Sprague-Dawley rats fed an AIN-76 iron deficient diet [17].
Rat Growth Rate and Final Body Mass
Heart mass increased significantly from 1 week to 1 month of the diets (figure 2, †, p < 0.0001). Diet group differences were not apparent at 1 week (p = 0.8674), but the hearts of iron deficient rats were of greater mass at 1 month (*, p = 0.0357), indicating that the heart has undergone hypertrophy by this time. These differences remained significant after normalization of heart weight to body weight (data not shown). This time frame for development of hypertrophy is similar to previous findings with iron or copper deficiency [14-17,40,43].
Heart mass and Hematocrit
Mean hematocrit levels (figure 2 bottom) of iron deficient animals were significantly less than controls after 1 month (*, p < 0.0001), but not 1 week (p = 0.3102), of the respective diets. Many studies have documented significant hematocrit decreases after one month of an iron deficient diet [14-17,40,43]. However, few of these report data for 1 week, though one published study indicated a small but significant difference [15], and we have observed a group difference after 1 week of iron deficiency in another study (Mullendore, et. al., manuscript in progress). We speculate that the ability to create significant iron deficiency in this short time is dependent upon the purity of the iron deficient diet as well as individual animal differences. Nonetheless, examination of several time points during the progression of iron deficiency can potentially provide important insights into the development of resulting physiological alterations.
Experiment 1: Cardiodynamics
In these experiments, an infusion of intravenous saline was used to ensure that the delivery method employed did not evoke a cardiovascular response. In all variables measured, which include heart rate and ventricular contractility (figure 3), systolic and diastolic pressures (figure 4), there were no differences between pre-infusion and peak responses to saline. Therefore, any differences seen with intravenous norepinephrine were attributable to the catecholamine and not the infusion per sé. Prior to the infusion of saline, group differences were already apparent between iron deficient and control rats at 1 month in heart rate (*, p = 0.0025), systolic (*, p = 0.0041) and diastolic pressure (*, p = 0.0020), but not contractility (p = 0.6373). We believe these pressure changes are associated with vascular remodeling, as discussed below.
Heart Rate and Contractility and Response to Norepinephrine
Systolic and Diastolic Pressures and Response to Norepinephrine
Figure 3 also shows heart rate and contractility immediately prior to, and at peak response to, intravenous norepinephrine infusions. Prior to norepinephrine infusion, there were no differences between iron deficient and control groups within each duration in either heart rate or contractility (p > 0.05). Heart rate was unchanged from pre-norepinephrine infusion to peak response (p = 0.3905). The only difference in peak heart rate response to norepinephrine was the iron deficient group at 1 month (#, p = 0.0272), which showed a significant decrease. Contractility was significantly increased for both control (#, p = 0.0259) and iron deficient (#, p = 0.0008) groups from pre-infusion to peak response. At one month, contractility for iron deficient rats increased 43.71% upon norepinephrine infusion, while the control group at 1 month increased only 10.0%.
Figure 3 demonstrates a unique response pattern to the sympathetic nervous system neurotransmitter, norepinephrine. For both control and iron deficient groups, the response to norepinephrine is to increase contractility in the absence of chronotropic increases. Cardiac output, therefore, would be expected to increase due to enhanced stroke volume, with an uncompromised filling time and a more powerful ventricular contraction. In fact, this pattern is even more exaggerated in the iron deficient animals at 1 month. Heart rate in this group was decreased, which allows the animal two potential advantages. First, ventricular filling time would be increased, potentially allowing for greater end diastolic and stroke volumes. This would be expected chronically to result in a eccentric hypertrophic pattern, a condition that we and others [15,17] have found in iron deficient rat hearts. Second, the decrease in heart rate in iron deficient hearts when norepinephrine was infused would allow for longer myocardial perfusion time, which would compensate for the decreased oxygen carrying capacity of the blood and impaired mitochondrial function [14]. The iron deficient norepinephrine response compliments the finding of an increase in capillary lumenal volume in iron deficient hearts [15], which should also allow for better myocardial perfusion during diastole. Norepinephrine infusion resulted in a dramatic enhancement of contractility (43.71%) in 1 month iron deficient rats when compared to 1 month control rats (10.0%). By utilizing a hyper-sensitive response to sympathetic stimulation, coupled with an unchanged or lower heart rate, the iron deficient heart can potentially offset some of the oxygen delivery deficiencies of anemic blood without risking myocardial ischemia. These appear to be positive adaptations, at least short term, in response to a serious nutritional challenge. Iron deficient cardiac hypertrophy is not unique in this regard, as Lin has recently documented enhanced contractility without increased heart rate in hyper-sympathetic transgenic mice [34]. In contrast, there is a large body of evidence which suggests that cardiac hypertrophy which has progressed to the point of pathology, ultimately results in a decrease in contractility [25-27,29,30,32,33,44,45]. We have not yet investigated whether the iron deficient heart will similarly lose the augmented contractility response in time.
It has been suggested in several papers that changes in sympathetic nervous system responsiveness may begin as a positive adaptation, but ultimately results in pathologic failure of the myocardium. For example, Engelhardt [46], in a study of mice genetically bred with an increase in beta-adrenergic receptor density, showed that increased contractility was an initial response, but by 35 weeks contractility was reduced by 50% and ejection fraction by 20%. The genetically altered, hyper-adrenergic heart ultimately failed after initial enhancement of cardiac function.
Mukherjee and Spinale [47], in a recent review, concluded that with hypertrophy comes a decrease in cardiac contractility as a result of changes to the L-type Ca2+ channel. However, the authors noted that this adaptation, while generally present in a failing hypertrophied myocardium, is not necessarily present in less serious cardiac hypertrophies. Similarly, Tse [48] found that in compensated hypertrophy the population of beta-adrenergic receptors is relatively normal, but in heart failure, the beta receptors have both down-regulated and de-sensitized. Sheridan [49] and Iaccarino [35] came to a similar conclusion regarding heart failure and adrenergic down-regulation.
Figures 4 shows systolic and diastolic pressures before and after the norepinephrine infusion. There was a significant increase in systolic pressure for both the control (*, p = 0.0020) and the iron deficient (*, p = 0.0001) groups from pre-infusion to peak response. Diastolic pressure similarly increased from pre-infusion to peak response for both control (#, p = 0.0040) and iron deficient (#, p < 0.0001) groups when norepinephrine was infused. These observations are consistent with the vasoconstrictive effect of norepinephrine. Also, increased heart contractility is expected to increase peak systolic pressure. The diastolic pressure of iron deficient rats was significantly lower than controls both before (*, p = 0.0131) and after (*, p = 0.0004) norepinephrine infusion. This is consistent with the idea of vascular remodeling to a larger arterial diameter with iron deficiency (discussed below), which could reduce the afterload against which the heart must work to eject blood.
Experiment 2
Representative aortas from control and iron deficient rats after one month on the respective diets are shown in figure 5, and mean aorta external diameter at 100 mmHg pressure is illustrated in figure 6. There was a significant duration effect (†, p < 0.0001), likely due to body growth. Overall, the diameter of aortas from iron deficient rats was greater than from control rats (*, p = 0.0280). The regulation of lumenal diameter in large arteries is accomplished through the maintenance of shear stress along the arterial wall [38,50-54]. As blood flow through an artery increases, shear is also increased, and the endothelium releases nitric oxide for vasodilation [55]. Chronically, the arterial wall will remodel to a larger diameter to accommodate the increased flow [38]. While we did not measure arterial flow through these iron deficient arteries, the enhanced inotropic compensation to iron deficiency, combined with the enlarged aortic diameter, leads us to suspect a flow dependent mechanism for remodeling. Future investigations will be necessary to confirm this assertion.
Control and Iron Deficient Aorta Images
Aorta External Diameter and Distensibility
Aortic distensibility, the increase in external diameter during stepwise internal pressure increases, is also shown in figure 6. There is a significant group difference (*, p = 0.0101) with the iron deficient aortas more compliant than the controls overall. A decrease in collagen production with iron deficiency could account for this enhanced distensibility. It has been reported that collagen production is reduced in iron deficient hearts [40], but we are aware of no studies that have specifically examined alterations in collagen production in iron deficient arteries. Further research regarding this issue is warranted.
Experiment 3
Figure 7 shows the heart weight to body weight ratio for rats that received a daily intra-peritoneal injection of either the beta-blocker propanolol or saline for 1 month. This ratio was significantly greater in rats fed the iron deficient diet than controls (*, p = 0.0002). Clearly, iron deficient cardiac hypertrophy is not caused by chronic stimulation of beta-adrenergic receptors in the heart.
Heart Mass/Body Mass After Beta-blocker Treatment
Many studies have shown that chronic intravenous infusion of norepinephrine is sufficient to cause hypertrophy, and that various forms of hypertrophy are linked to a decrease in either beta-adrenergic receptor density or a decreased responsiveness to beta-adrenergic stimulation [46,48,49,56-63]. Barth, for example, showed that norepinephrine infusion induced left ventricular hypertrophy that could be prevented by an adrenergic-receptor blocker [58]. It has been suggested that ornithine decarboxylase is a link between beta-adrenoreceptors and stimulation of tissue growth factor, which results in hypertrophy [64]. However, recent studies have focused on alpha-adrenergic reception as a mediator of cardiac hypertrophy [35], although some question the role of the alpha-adrenergic receptor as a hypertrophic mediator in vivo [34]. To date, no published studies of which we are aware have examined alpha-adrenergic reception and the development of cardiac hypertrophy with iron deficiency.
The literature surrounding norepinephrine and its role in the development of iron deficient hypertrophy documents the following facts: 1) prolonged iron deficiency causes cardiac hypertrophy [11-17,23], 2) the pool of stored norepinephrine in the heart is decreased with iron deficiency [17,21], 3) plasma and urine concentrations of norepinephrine are increased with iron deficiency [19-22,68], and 4) chronic administration of reserpine (which depletes norepinephrine) prevents the development of iron deficient hypertrophy [12], but not hypertrophy that results from aortic banding [23]. This, combined with other studies that relate chronic adrenergic stimulation to hypertrophy, led Rossi to the conclusion that iron deficiency hypertrophy was caused by chronic sympathetic stimulation. However, our finding that a daily injection of the beta-blocker propanolol does not prevent the development of iron deficient cardiac hypertrophy suggests that beta-adrenergic sympathetic stimulation may not be a causal agent in the development of hypertrophy. Future investigation is warranted into the potential alpha-adrenergic role in this hypertrophy.
Conclusions
The purpose of this study was to investigate cardiac and vascular responses associated with the development of iron deficiency. In our first experiment, we tested the hypothesis that the iron deficient heart would display an altered response to norepinephrine, the sympathetic nervous system neurotransmitter. In contrast to what has been seen with most cardiac disease states, we found that the iron deficient heart is hyper-sensitive to norepinephrine. This suggests an adaptive compensation to the depletion of the neurotransmitter stores in the heart sympathetic nerve terminals. Further, we found that contractility was enhanced, but heart rate was not. This would allow for the iron deficient animal to increase cardiac output through enhanced stroke volume, while maintaining the amount of time available for cardiomyocyte perfusion.
Our second experiment was designed to investigate whether vascular morphology would be altered with iron deficiency. Our results show an increase in abdominal aorta diameter, suggesting that a flow-dependent remodeling of the arterial wall occurred. Also, distensibility was significantly increased, which we suggest may be due to a reduction in arterial collagen. These morphological adjustments can be seen as a positive adaptation to simultaneously reduce afterload on the heart and improve blood flow to peripheral tissues.
Our final experiment was a simple test to determine if the beta-adrenergic antagonist propanolol would prevent hypertrophy. The failure of the beta-blocker to prevent hypertrophy suggests that, if the sympathetic nervous system is a causal agent in the development of this form of hypertrophy, the signaling mechanism is not mediated by the beta-adrenergic receptor.
In summary, much remains to be learned about this form of hypertrophy. Further experimentation may serve to elucidate not only the pathology of iron deficient cardiac hypertrophy, but other forms of this pathology as well.
Materials and methods
All procedures conducted in this experiment were consistent with "The American Physiological Society Guiding Principles for the Care and Use of Animals" . 40 male CD (Sprague Dawley) rats (Charles River Laboratories, Raleigh, NC, were randomly divided into two groups and placed on either a control (iron-replete) or iron-deficient diet (Dyets Inc., Bethlehem, PA, ) at 24 days of age. The control diet was a pelleted AIN-93G purified rodent diet [69]. The iron deficient diet was based on the AIN-93G standard, but with microcrystalline cellulose and a reagent grade mineral mixture free of ferric citrate. This diet provides < 5 ppm iron. All rats were given constant access to the respective diets and distilled water, kept on a 12 h light cycle, and weighed weekly and just before each was sacrificed.
A total of 40 animals (20 iron deficient and 20 control) were utilized for three experiments. Experiments #1 and 2 used 30 rats after either 1 week or 1 month of the dietary manipulations, while experiment #3 used the remaining 10 rats after 1 month on the diets.
Experiment 1: cardiovascular response to norepinephrine
The purpose of this experiment was to determine whether iron deficient rats exhibit altered cardiac and vascular pressure responses to intravenous infusions of the sympathetic nervous system neurotransmitter norepinephrine. Intravenous infusions of physiological saline solution were used as a control. This experiment was a 2 × 2 factorial design, with diet (iron deficient vs. control) and duration (1 week vs. 1 month on the diets) as variables.
Experimental procedure
Rats were sedated by intra-muscular injection (right hamstring) of 65 mg· kg-1 body mass ketamine hydrochloride. Three minutes later, they were anesthetized with 6.5 mg· kg-1 body mass xylazine. Adequate anesthesia was demonstrated by little or no withdrawal reflex to a toe pinch. Once anesthetized, one PE-50 (Becton-Dickinson) catheter filled with heparinized saline solution (131.9 mM NaCl, 4.7 mM KCl, 2.0 mM CaCl2, 1.17 mM MgSO4, 17.4 mM NaHCO3, 50 IU heparin sodium) was surgically implanted into the right carotid artery, and another into the left jugular vein.
Following catheterization and closure of the incision, the carotid catheter was connected to a pressure transducer (Cobe Model CDX III) and analog to digital conversion system (ETH-250 bridge/bio-amplifier used in bridge mode with an input gain of 100× and a 50 Hz filter, C.B. Sciences, Inc. and MacLab/4E analog to digital converter, A.D. Instruments, Inc.). The converter was connected via SCSI interface to an accelerated Power Mac 6100/66 computer (Apple Computer, Inc., Cupertino, CA); Chart software (Version 3.4.2, sampling every 0.01 s and set to accept a maximum input of 1 V, A.D. Instruments, Inc.) was used to visualize the output. The pressure transducer system was calibrated by mercury manometer at the beginning of each experiment, and was found to be linear and consistent throughout the range of pressures measured.
One hundred microliters of saline solution was infused into the jugular vein catheter, and the cardiovascular response was recorded via the carotid pressure transducer until pressure stabilized near its baseline value. This procedure was repeated using 100 μL of 1.52 mM norepinephrine (in saline solution), and the arterial trace recorded.
Data Treatment and Statistical Analysis
Cardiovascular variables were analyzed immediately prior to infusion (pre-infusion) and at the maximal pressor response (peak response), 20 s after jugular infusion. Chart software was used to determine minimal (diastolic) and maximal (systolic) pressures and heart rate. Contractility (dP•dt-1) was measured as the average slope between each pair of data points within the systolic pulse. Pulse pressure was calculated as the difference between systolic and diastolic pressures.
Analysis of Variance was performed using StatView software (Version 5.0.1, SAS Institute, Cary, NC), to determine diet (iron deficient vs. control) and duration (1 week vs. 1 month) effects. Paired t-tests were employed to determine significant changes from pre-infusion to peak response. Dependent variables were heart rate, contractility, systolic pressure, and diastolic pressure. Probabilities below 0.05 were accepted as significant for all statistical procedures throughout all experiments in this study.
Experiment 2: remodeling of abdominal aorta
The purpose of this experiment was to explore alterations in vascular morphology of iron deficient rats, using the abdominal aorta as a representative artery. This experiment used the same rats as in experiment #1, and was also 2 × 2 design with diet and duration factors.
Experimental Procedures
After experiment 1, the rats were decapitated. The abdominal aorta was carefully exposed, and two catheters were surgically implanted. The first was inserted just caudal to the superior mesenteric artery, and the other just cranial to the inferior mesenteric artery. The superior catheter was used for stepwise saline infusions at a range of pressures from 0 – 120 mmHg. Exact pressures were measured via a pressure transducer attached to the inferior catheter. Digital images of the abdominal aorta were taken at each pressure with a Kodak DC120 digital camera coupled to a dissecting microscope (Leica MZ6). The external diameter of the abdominal aorta was measured using NIH image software (Version 1.62, National Institutes of Health, Bethesda, MD) on an accelerated Power Mac 7500/100 (Apple Computer, Cupertino, Ca). The system was calibrated by using the digital image of a ruler at the same magnification.
Data Treatment and Statistical Analysis:
Regression analysis was performed using aorta external diameter and inflation pressure as dependent and independent variables, respectively. Distensibility was estimated as the slope of the regression line for each individual animal. The external diameter of the abdominal aorta at 100 mmHg pressure was also estimated from the individual regression equations. Analysis of Variance was then performed on distensibility and 100 mmHg external diameter.
Experiment 3: propanolol effects on heart morphology
The purpose of this experiment was to see if a daily injection of the beta-blocker propanolol inhibits the cardiac hypertrophy that is associated with iron deficiency in weanling rats. This experiment was a 2 × 2 design, using diet and drug (saline vs. propanolol) as factors.
Experimental procedures
Ten rats (5 iron deficient and 5 control) were given a daily intra-peritoneal injection of either propanolol (50 mg•kg-1 body mass per day) or saline solution. After 1 month, the rats were sacrificed by decapitation and the hearts were carefully extracted and weighed.
Data treatment and statistical analysis
The heart mass to body mass ratio was calculated, and Analysis of Variance was performed as above.
Acknowledgments
We gratefully acknowledge the financial support of the Henson School of Science and Technology at Salisbury State University for these experiments.
Maitotoxin (MTX) initiates cell death by sequentially activating 1) Ca2+ influx via non-selective cation channels, 2) uptake of vital dyes via formation of large pores, and 3) release of lactate dehydrogenase, an indication of cell lysis. MTX also causes formation of membrane blebs, which dramatically dilate during the cytolysis phase. To determine the role of phospholipase C (PLC) in the cell death cascade, U73122, a specific inhibitor of PLC, and U73343, an inactive analog, were examined on MTX-induced responses in bovine aortic endothelial cells.
Results
Addition of either U73122 or U73343, prior to MTX, produced a concentration-dependent inhibition of the cell death cascade (IC50 ≈ 1.9 and 0.66 μM, respectively) suggesting that the effect of these agents was independent of PLC. Addition of U73343 shortly after MTX, prevented or attenuated the effects of the toxin, but addition at later times had little or no effect. Time-lapse videomicroscopy showed that U73343 dramatically altered the blebbing profile of MTX-treated cells. Specifically, U73343 blocked bleb dilation and converted the initial blebbing event into "zeiosis", a type of membrane blebbing commonly associated with apoptosis. Cells challenged with MTX and rescued by subsequent addition of U73343, showed enhanced caspase-3 activity 48 hr after the initial insult, consistent with activation of the apoptotic program.
Conclusions
Within minutes of MTX addition, endothelial cells die by oncosis. Rescue by addition of U73343 shortly after MTX showed that a small percentage of cells are destined to die by oncosis, but that a larger percentage survive; cells that survive the initial insult exhibit zeiosis and may ultimately die by apoptotic mechanisms.
Background
Recent studies have shown that maitotoxin (MTX), a potent cytolytic agent isolated from the dinoflagellate Gambierdiscus toxicus, is an important new molecular tool for the study of oncotic (necrotic) cell death [1,2]. In a variety of cell types, MTX initiates a cell death cascade that involves a sequence of cellular events essentially identical to those activated by stimulation of purinergic receptors of the P2Z/P2X7 type. Initially, MTX causes a graded increase in cytosolic free Ca2+ concentration ([Ca2+]i). This is followed closely in time by the opening of cytolytic/oncotic pores (COP) that allow the exchange of large organic molecules of molecular weight less than ~800 Daltons across the plasma membrane. COP activation can be monitored by the cellular accumulation of ethidium or propidium-based vital dyes, which are normally excluded from the cytoplasm, but gain access to cellular nucleotides via COP and exhibit an increase in fluorescence. In isolated bovine aortic endothelial cells (BAECs), the opening or activation of COP is associated with formation of spherical membrane blebs with a diameter of 3–5 microns [3]. The final stage of MTX-induced cell death is cell lysis as indicated by the release of large cytoplasmic enzymes, such as lactate dehydrogenase (LDH). Using time-lapse videomicroscopy, we have shown that MTX-induced release of LDH from vascular endothelial cells is associated with massive bleb dilation and rapid staining of the nucleus with vital dyes [3].
The initial MTX-induced increase in [Ca2+]i reflects the activation of a Ca2+-permeable non-selective cation channel (CaNSC) [1,4-8]. This channel, which has a reported conductance in the range of 12–40 pS depending on ionic conditions [5,9-11], causes rapid membrane depolarization, which in excitable cells, leads to activation of voltage-sensitive channels. Although it appears that a rise in [Ca2+]i is necessary, but not sufficient for activation of COP [1], the molecular mechanisms by which this occurs remains unknown. Likewise, the subsequent steps leading to membrane blebbing and cytolysis are poorly understood. It is however, well established that MTX causes the hydrolysis of phosphoinositides in some cell types, presumably via activation of phospholipase C (PLC) [12,13]. Activation of PLC by MTX appears to be indirect resulting as a consequence of increased [Ca2+]i. These results suggest that PLC may be involved in activation of COP and/or in the cytolysis phase of MTX action. Thus, the initial purpose of the present study was to determine the role of PLC in MTX-induced cell death. To accomplish this goal, the effect of U73122, a specific inhibitor of mammalian PLC was examined. This compound selectively inhibits mammalian PLC, but has no direct effect on bacterial PLC, bacterial or mammalian phospholipase A2 or adenylyl cyclase [14]. U73343, a structural analogue of U73122 that differs by only one double bond, has no direct effect on PLC and is commonly used as a negative control. However, both compounds have been shown to produce non-specific effects presumably unrelated to inhibition of PLC [15-20]. The results of the present study show, that both U73122 and U73343 inhibit MTX-induced change in [Ca2+]i, ethidium uptake, and LDH release in BAECs. Although these results suggest that blockade of MTX-induced responses by the U-compounds is independent of PLC, they identify these compounds as novel, potent, and rapid blockers of MTX action. Interestingly, in experiments designed to examine MTX reversibility, we discovered a rather stunning change in the pattern of membrane blebbing. Specifically, cells rescued from MTX by subsequent application of U73343 exhibit a blebbing pattern known as "zeiosis". Zeiosis, which comes from the Greek word Zειω meaning "to boil over" [21], is characterized by violent cytokinesis with continuous bleb extension and retraction. Zeiosis has been associated in many cell types with apoptosis [22-24]. The results of the present experiments suggest that following a brief exposure to MTX, U73343 rescues cells from oncotic cell death. Cells that survive the initial insult may ultimately die by apoptosis.
ResultsU73122 and U73343 inhibit MTX-induced change in [Ca2+]i and ethidium uptake in BAECs
To test the hypothesis that MTX-induced cell death requires PLC, the effect of U73122, a specific inhibitor of PLC was examined. Addition of MTX to fura-2 -loaded BAECs, suspended in a cuvette in normal Ca2+-containing buffer at 37°C, produced a time-dependent increase in [Ca2+]i (Fig 1, upper panel). Addition of either U73122 or U73343, the inactive analog, ~3 min before MTX produced a concentration-dependent inhibition of the change in [Ca2+]i. U73343 exhibited a 3-fold greater potency compared to U73122 (Fig 1, insets). Since U73343 has no significant effect on PLC over the concentration range employed [14,25-27], these results suggest that inhibition of the MTX-induced response is independent of PLC.
U73343 and U73122 block MTX-induced Ca2+ influx and ethidium uptake. Upper panel. Fura-2 loaded BAECs were suspended in HBS and the fluorescence ratio was recorded as a function of time. Five traces are shown superimposed. At time 100 sec, U73343 was added at the final concentration indicated to the right of each trace. At time 300 sec 0.3 nM MTX was added to stimulate Ca2+ influx. Concentration-response curves were determined for both U73122 (n = 3) and U73343 (n = 5). Inset shows the % inhibition relative to DMSO control. Lower panel. Ethidium bromide (EB) uptake in BAECs was determined from the increase in fluorescence as a function of time as described in Materials and Methods. U73122 (red trace), U73343 (blue trace), or DMSO (black trace) were added at time 0, EB was added at 50 sec, and MTX was added at 200 sec. Values shown are normalized to the maximum fluorescence obtained by addition of digitonin at the end of each trace. Results shown are representative of 3 independent experiments.
Studies in BAECs have shown that following elevation of [Ca2+]i, MTX causes the activation of large pores (i.e., COP) that allow the flux of ethidium and propidium-based vital dyes into the cell. As previously reported [3], MTX-induced uptake of ethidium in BAECs was biphasic in the absence of the U-compounds (Fig 1, lower panel). The first phase, which extends for ~5 min after addition of MTX, reflects the activation of COP, whereas the second phase is temporally associated with LDH release and thus reflects cell lysis [3]. Addition of either U73122 or U73343, ~3 min before MTX, produced an inhibition of ethidium uptake (Fig 1, lower panel). Both phases of ethidium uptake were attenuated by the U-compounds and U73343 again appeared to have a greater potency compared to U73122. These results suggest that inhibition of MTX-induced change in [Ca2+]i prevents or attenuates both the activation of COP and cytolysis.
To determine if the effects of the U-compounds are reversible, BAECs were pre-treated with U73343 (5 μM) for 5 min, washed and resuspended in normal extracellular buffer in the absence of U73343. As seen in Fig 2, BAECs pre-treated with U73343 partially recover responsiveness to MTX following washout of U73343, but the response failed to recover further over the subsequent 10 min. Thus, the ability of U73343 to block the effect of MTX appears to reflect both a rapidly reversible and slowly reversible component.
Blockade of MTX-induced Ca2+ influx by U73343 is partially reversible. Prior to fluorescence recording, fura-2 loaded BAECs were equilibrated with HBS containing either 5 μM U73343 or 0.1% DMSO for 5 min. The cells were washed to remove the drug which required ~5 min. The cells were resuspended in the absence of U73343, immediately placed in a cuvette, and the fluorescence recorded. Addition of 0.3 nM MTX at 100 sec to the DMSO-treated cells (black trace) resulted in a typical MTX response. Addition of MTX to the U73343 pretreated cells at 100 (red), 200 (green), 300 (yellow), 400 (blue) or 500 sec (purple) are shown superimposed. Results shown are representative of 3 independent experiments.
U73343 partially reverses the effect of MTX
To determine if U73343 could reverse the action to MTX, the compound was added to the cuvette after MTX. As seen in Fig 3 (upper panel), addition of U73343 at various times after MTX, immediately stopped further increases in [Ca2+]i and resulted in an immediate recovery of [Ca2+]i back towards basal resting levels. Over the time-course examined however, [Ca2+]i never fully returned to the level observed before MTX addition. The same experimental protocol was employed for evaluation of ethidium uptake (Fig 3, lower panel). Again, MTX-induced uptake of ethidium was biphasic in the absence of U73343 (trace a). Addition of the U-compound shortly after MTX (i.e., at 200 sec; trace e), blocked both phases of the response. Addition of U73343 towards the end of the first phase (i.e., at either 400 or 500 sec; traces b and c) had only a small effect on the magnitude and time course of subsequent ethidium uptake. However, when U73343 was added during the first phase of ethidium uptake (i.e., 300 sec; trace d), dye accumulation in the cells was immediately blocked. After a short delay, dye uptake again increased with time to a value that was substantially less than maximum. These results suggest that U73343 is able to rapidly block both the MTX-induced change in [Ca2+]i and COP formation when added shortly after MTX. When U73343 is added at longer times, a large proportion of the cells challenged by MTX have apparently passed a point-of-no-return and are destined to die by oncosis. However, the percentage of cells that either die or survive depends on the time of U73343 addition after MTX.
U73343 rapidly blocks and partially reverse the effects of MTX. Five traces are shown superimposed in each panel. Upper panel. Fura-2 loaded BAECs were stimulated with MTX at 100 sec. U73343 (5 μM) was subsequently added at either 200 (red), 300 (green), 400 (yellow), or 500 sec (blue). Lower panel. EB uptake in BAECs with the same protocol and trace colors as the upper panel are shown. Results shown are representative of 3 independent experiments.
U73343 rescues cells from MTX-induced oncotic cell death
To directly determine the effect of U73343 on MTX-induced cell lysis, LDH release was correlated with the change in [Ca2+]i and ethidium uptake in paired experiments. As previously reported [3] and as seen in Fig. 4 (bottom panel, black circles), MTX-induced LDH release in the absence of U73343, correlates in time with the second phase of ethidium uptake, demonstrating that the second phase reflects cell lysis. Addition of U73343 during the first phase of the MTX-induced response, immediately reversed the rise in [Ca2+]i and blocked dye uptake. However, at ~400 sec, ethidium uptake resumed and increased to an intermediate value as was seen in Fig 3. This secondary uptake of ethidium correlated in time with a partial LDH release (Fig 4, bottom panel, red triangles). U73343 alone had no effect on LDH release (Fig 4, bottom panel, green squares). These results directly demonstrate that a percentage of cells within the population died by oncosis, but perhaps more importantly, that a significant number of the cells are rescued from MTX-induced oncotic cell death when U73343 is added within a short time-window after MTX.
U73343 rescues the cells from MTX-induce lytic cell death. MTX-induced change in [Ca2+]i (upper panel) or EB uptake (lower panel) were assayed in BAECs. MTX and U73343 were added at the concentrations and time indicated. LDH release was measured as described in Materials and Methods and plotted in the lower panel using color matched symbols (see inset). All experiments were performed in a paired fashion on the same batch of BAECs. Results shown are representative of 4 independent experiments.
U73343 alters MTX-induced bleb formation
Our previous studies showed that MTX causes a specific change in cell morphology. In particular, MTX causes dramatic time-dependent membrane blebbing in BAECs [3]. An example is shown in Fig. 5. BAECs, attached to glass coverslips and superfused with a solution containing ethidium bromide, were simultaneously monitored by phase and fluorescence microscopy. Selected images pairs (merged phase/fluorescence images) are shown in Fig 5. At time 5 min, the cells were superfused with a solution containing MTX. No change in cell morphology was noted at 10 min, but by 25 min (i.e., 20 min after addition of MTX) large membrane blebs were observed on essentially all the cells within the field of view (Fig 5A, see examples indicated by blue arrows). Additionally, clear staining of the nucleus and cytoplasm by ethidium was observed at this time. The graph shown in Fig 5A quantifies the fluorescence change in individual cells as a function of time. Note that uptake of ethidium in response to MTX is biphasic at the single cell level and that heterogeneity between cells primarily reflects the time of entry into the second phase of dye uptake, i.e., the cell lysis phase. The dynamic nature of the MTX-induced bleb formation, bleb dilation, and ethidium uptake can be observed by viewing the time-lapse movie associated with this experiment (see Additional File 1). As previously reported, blebs form during the first phase of dye uptake, whereas rather dramatic bleb dilation is seen during the cytolysis phase of rapid dye uptake. In the continuous presence of MTX, blebs continue to swell and in many cases seem to disappear as they dilate and extend out of the focal plane.
U73343 alters MTX-induced membrane blebbing. BAECs cultured on circular glass coverslips were visualized using an inverted microscope equipped for epifluorescence as described in Materials and Methods. The cells were maintained at 35–37°C and both phase and fluorescence images were acquired at 30-second intervals. The graph in each panel shows normalized single cell fluorescence intensity as a function of time for cells pretreated with either 0.2% DMSO (Panel A; n = 38) or U73343 (Panel B; n = 52) and subsequently challenged with MTX at the times indicated. Merged phase and fluorescence images (red pseudocolor) at time 10, 25, and 40 min are shown above each graph. Examples of large membrane blebs are highlighted by the blue arrows in Panel A at time 25 min. Examples of cells with rounded morphology are indicated by the blue arrows in Panel B at time 40 min. Note the absence of large dilated blebs and lack of fluorescence in the U73343-treated cells (Panel B). Results shown are representative of 9 and 6 independent experiments for Panels A and B, respectively.
Addition of U73343 (10 μM) 2–3 min before MTX (0.3 nM) completely blocks ethidium uptake and the associated change in cell morphology. As seen in Fig 5B, essentially no blebs are observed at 25 min (i.e., 20 min after MTX) and little or no dye uptake is seen out to 40 min. Thus, the major morphological changes associated with MTX challenge are blocked by U73343. However, by time 40 min, some of the cells (11 out of 53) detach from the glass coverslip and round-up (Fig 5B, see examples indicated by blue arrows). The time-lapse movie of this experiment revealed that these detached cells were actually undergoing a novel type of membrane blebbing called zeiosis (see Additional File 2). As can be seen from the movie (Additional File 2), zeiosis is characterized by cytokinesis with time-dependent bleb extension and retraction. Note that during zeiosis, no ethidium uptake is observed, i.e., there is little or no change in cell fluorescence. This result suggests that COP is not activated and the cells do not undergo a gross change in plasmalemmal permeability during zeiosis. At 1 nM MTX, the number of BAECs that undergo zeiosis in the presence of U73343 is increased (see Additional File 3). As seen in this movie (Additional File 3), the majority of the cells in the field of view (16 out of 24) undergo zeiosis, yet there is no ethidium uptake over the time frame examined, despite the higher concentration of MTX employed. Little or no zeiosis was observed in the absence of MTX, i.e., in the presence of either U73343 alone (4 zeotic cells out of 113) or vehicle alone (DMSO; 5 zeotic cells out of 146) examined for 40 min. Thus, the ability of MTX to cause zeiosis is specific, but requires the presence of U73343, presumably to block or prevent rapid oncotic cell death.
The cuvette experiments suggest that BAECs challenged with MTX can be rescued from oncotic cell death if U73343 is added to the bath solution within a narrow time frame after the addition of MTX. To further test this hypothesis, paired experiments were performed at the single cell level. BAECs, attached to coverslips were first challenged with MTX and subsequently rescued by addition of U73343. The percentage of a) dead cells (indicated by nuclear staining with ethidium), b) cells undergoing zeiosis, and c) surviving cells (indicated by the lack of membrane blebbing and lack of nuclear staining with ethidium) was quantified at time 40 min (Fig 6). All of the cells treated with vehicle alone (DMSO) before MTX exhibited the typical blebbing profile and nuclear staining indicative of rapid oncotic cell death. BAECs treated with U73343 before MTX were essentially protected from cell death, but ~35% of the total population exhibited zeiosis. The same pattern was observed when U73343 was added 4 min after MTX, i.e., full protection from oncosis with ~40% zeiosis. However, ~15% of the cells died by oncosis when U73343 was added at 6 min after MTX (see Additional File 4) and this percentage increased to >90% when U73343 was added at 7 min after MTX (see Additional File 5). Addition of U73343 10 min after MTX was the same as vehicle control, i.e., all the cells lyse by time 40 min. These results are consistent with the cuvette studies and directly demonstrate that within a narrow time window after MTX challenge, U73343 can rescue BAECs from oncotic cell death.
U73343 rescues BAECs from MTX-induce lytic cell death: Evaluation at the single cell level. BAECs, cultured on circular glass coverslips, were visualized using an inverted microscope equipped for epifluorescence as described in Materials and Methods. The cells were maintained at 35–37°C and both phase and fluorescence images were acquired at 30-sec intervals. Each bar of the histogram shows 1) the percentage of dead cells (black; indicated by uptake of ethidium and presence of large membrane blebs), 2) the percentage of cells undergoing zeiosis (red), and 3) the percentage of cells protected (green; indicated by the lack of nuclear staining and membrane blebbing) for the experimental condition given below each bar. The total number of cells examined for each condition is given in parenthesis above each bar. In each experiment 0.3 nM MTX was added at time 5 min; Bar 1, 0.2% DMSO added before MTX; bar 2, 10 μM U73343 added before MTX; bars 3, 4, 5, and 6, U73343 added 4, 6, 7, and 10 min after MTX, respectively. Percentages were determined at time 40 min using time-lapse videos to determine the number of cells undergoing zeiosis.
BAECs rescued from oncosis by U73343 die by apoptosis
Zeiosis is a type of membrane blebbing associated with apoptosis [22-24]. Since U73343 protects cells from MTX-induced oncosis and causes zeiosis, we considered the possibility that the cells surviving the initial insult may ultimately die by apoptosis. To test this hypothesis, BAECs were treated with vehicle alone (DMSO), U73343 alone, or MTX followed by U73343, and specific caspase-3 activity was determined after 48 hrs by the fluorescence substrate assay (Fig 7). A significant increase in caspase-3 activity (p < 0.035; n = 3) was observed in cells challenged with MTX and subsequently rescued by U73343, consistent with increased apoptotic cell death.
BAECs rescued from MTX-induced lytic cell death have elevated caspase-3 activity. BAECs were treated with 0.3 nM MTX for 2.5 min at which time the action of MTX was stopped by addition of 5 μM U73343. Cells were harvested after 48 hr and specific caspase-3 activity was determined using the fluorescence substrate assay (M/U). For control, caspase-3 activity was determined in cells treated with U73343 alone (IU) or vehicle (DMSO). Values represent the mean ± se (n = 3) relative to no-treatment controls.
Discussion
The results of the present study support two major conclusions. First, U73122 and U73343, compounds that are commonly used to examine PLC-dependent mechanisms, potently inhibit MTX-induced change in [Ca2+]i, vital dye uptake, and oncotic cell death in BAECs. The fact that U73343, which has essentially no effect on PLC at the concentrations employed [14,25-27], is more potent than U73122 on MTX-induced responses, clearly shows that the effect of these agents is unrelated to inhibition of PLC. Since U73343 can rapidly block and reverse the increase in [Ca2+]i induced by MTX, it seems likely that these compounds directly block the MTX-activated channels in BAECs, perhaps acting as pore blockers. However, since the effect of U73343 appears to be slowly reversible, there may be additional sites/mechanisms of blocking action for these agents. Another class of compounds known as imidazoles, of which SKF 96365 is the best studied, have also been shown to block MTX-induced responses in many cell types [5,7,12,28]. The imidazoles, which are antifungal agents that inhibit cytochrome P450, are relatively non-selective and are known to block voltage-gated Ca2+ channels and receptor-activated Ca2+ influx through so-called store-operated channels (SOCs) [29], but their mechanism of action in this regard remains unknown.
U73343 appears to be relatively more selective than SKF 96365 with respect to Ca2+ influx pathways. Most studies find that U73343 has no effect on receptor-mediated increase in [Ca2+]i. However, Berven and Barritt [15] report that U73343 partially blocked vasopressin-induced Ca2+ influx, but did not inhibit release of Ca2+ from intracellular stores in hepatocytes. And Wang [16] reported that U73343 suppressed elevation of [Ca2+]i in neutrophils challenged with FMLP in the presence, but not the absence, of extracellular Ca2+ Both reports are consistent with a lack of effect of U73343 on receptor-mediated activation of PLC and a direct inhibitory effect on Ca2+ influx pathways in these cells. The results of the present study therefore, suggest that the MTX-sensitive channel may be responsible for the rise in [Ca2+]i produced by receptor stimulation in some cell types, but not in others. In this regard, Worley et al. [8] showed that SOCs in pancreatic β-cells have characteristics essentially identical to channels activated by MTX in the same cell type. Thus, U73343 may be useful for the ultimate identification and characterization of the physiological role of MTX-sensitive channels. Lastly, because MTX is one of the toxins associated with ciguatera seafood poisoning, the identification of U73343 as a potent and relatively specific blocker of MTX-induced cell death cascade may ultimately lead to improved therapeutic interventions.
The second major conclusion supported by the results of the present study is that MTX apparently has an effect on BAECs that is independent of the rise in [Ca2+]i. We previously showed that MTX causes the appearance of membrane blebs on the surface of the endothelial cells [3]. Initially these blebs are of small diameter (3–5 microns) and bleb formation correlates with the first phase of vital dye uptake. However, during MTX-induced cytolysis, as indicated by the release of LDH, the membrane blebs actually exhibit a dramatic increase in diameter. This bleb dilation phase is clearly associated with rapid, intense staining of the nucleus with vital dyes. It is well established that the acute effects of MTX on [Ca2+]i, uptake of vital dyes, and membrane blebbing require extracellular Ca2+[1,3], but the role of cytosolic Ca2+ remains unknown. Preliminary studies in fibroblasts suggest that vital dye uptake via COP is attenuated in cells loaded with the Ca2+ chelator BAPTA [1], but the effect of BAPTA loading on MTX-induced bleb formation has not been examined. The results of the present study show that despite a blockade of MTX-induced change in [Ca2+]i, addition of U73343 either before, or within a short time after MTX, produced a blebbing profile known as zeiosis. The term zeiosis was first used by Costero and Pomerat [21] 50 years ago to describe the appearance of membrane vesicles on the surface of nerve cells in culture, as visualized using time-lapse cinematography. Although the cellular mechanisms remain poorly defined, it is now clear that zeiosis is associated with apoptotic cell death (for review [24,30]). In static images, the morphological changes associated with apoptosis appear as membrane blebs covering the surface of the cells [31-33]. However, time-lapse images reveal that blebs protrude and retract in a dynamic fashion, a hallmark of zeiosis [22,23]. Dynamic membrane blebbing appears to involve a dramatic change in the cytoskeleton and may be induced by a caspase-3-mediated cleavage of ROCK-I, a Rho-activated serine/threonine kinase known to stimulate actinomycin-based contractions [31-33]. A similar profile can be seen in the videos of MTX-induced cell death in the presence of U73343 (Files 2 and 3). The cells first lose adhesion to the coverslip, retract and roundup. This is followed by zeiotic membrane blebbing. Vital dyes are excluded during this time suggesting that COP is not activated and that there is no gross loss of membrane integrity during blebbing. It is important to note that these MTX-induced events are only observed in the presence of U73343 and occur in the absence of a measurable rise in [Ca2+]i. To our knowledge, this is the first evidence that MTX has effects independent of a rise in [Ca2+]i. Perhaps more importantly, the results of the present study suggest that U73343 converts MTX-induced oncosis into apoptosis and that a rise in [Ca2+]i is necessary for oncotic cell death.
Conclusions
In conclusion, U73343 blocks the MTX-induced cell death cascade in BAECs and converts the normal blebbing profile seen with this toxin into zeiosis, a form of dynamic membrane blebbing commonly associated with apoptosis and characterized by bleb extension and retraction. U73343 may prove useful for identification and characterization of MTX-activated cation channels and for understanding their physiological role in cell signaling and cell death.
Materials and MethodsSolutions and reagents
Unless otherwise indicated, HEPES-buffered saline (HBS) contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 10 mM D-glucose, 1.8 mM CaCl2, 15 mM HEPES, 0.1% bovine serum albumin, pH adjusted to 7.40 at 37°C with NaOH. Fura-2 acetoxymethyl ester (fura-2/AM) and ethidium bromide (EB) were obtained from Molecular Probes (Eugene, OR, USA). Maitotoxin (MTX), obtained from LC Laboratories (Woburn, MA) or Wako Bioproducts (Richmond, VA), was stored as a stock solution in ethanol at -20°C. All other salts and chemicals were of reagent grade.
Cell Culture
Bovine aortic endothelial cells were cultured as previously described [34] using Dulbecco's modified Eagles medium (GIBCO) supplemented with 10% fetal bovine serum (Hyclone, Logan UT), 100 μg/ml streptomycin and 100 μg/ml penicillin (complete-DMEM). All cultures demonstrated contact-inhibited cobblestone appearance typical of endothelial cells grown to confluence.
Measurement of the apparent cytosolic free Ca2+ concentration
[Ca2+]i was measured using the fluorescent indicator, fura-2, as previously described [35]. Experiments were performed with cells in the twelfth to twentieth passage and 2–3 days post-confluency. Briefly, cells were harvested and re-suspended in HBS containing 20 μM fura-2/AM. Following 30 min incubation at 37°C, the cell suspension was diluted ~10-fold with HBS, incubated for an additional 30 min, washed and resuspended in fresh HBS. Aliquots from this final suspension were subjected to centrifugation and washed twice immediately prior to fluorescence measurement. Fluorescence was recorded using an SLM 8100 spectrophotofluorometer; excitation wavelength alternated between 340 and 380 nm and fluorescence intensity was monitored at an emission wavelength of 510 nm. All measurements were performed at 37°C.
Measurement of vital dye uptake
Vital dye uptake was determined as previously described [1-3]. Briefly, an aliquot (2 ml) of dispersed cells suspended in HBS at 37°C was placed in a cuvette. Following addition of ethidium bromide (final concentrations of 5 μM), fluorescence was recorded as a function of time with excitation and emission wavelengths of 302/560 nm, respectively. All ethidium bromide fluorescence values were corrected for background (extracellular) dye fluorescence and expressed as a percentage relative to the value obtained following complete permeabilization of the cells with 50 μM digitonin.
For single cell measurement of vital dye uptake, BAECs in complete-DMEM were sparsely seeded on circular glass coverslips and used within 2–3 days of seeding. The coverslips were mounted in temperature-controlled perfusion chambers and placed on the stage of Nikon Diaphot inverted microscope. The cells were illuminated with light from a 75 watt xenon lamp using a 0–5722 filter cube obtained from Molecular Probes. Epifluorescence was recorded using a SPOT™ camera (Diagnostic Instruments, Sterling Heights, MI) and images were acquired and analyzed using SimplePCI imaging software (Compix Inc., Cranberry Township, PA). During each experiment, phase and fluorescence image pairs were collected at 30 second intervals with shutter controllers switching between light and fluorescent illumination. The fluorescence images were used to quantify dye uptake. A region over the nucleus of individual cells was defined and the average fluorescence intensity of the region was quantified as a function of time. Phase images were contrast enhanced, digitally merged with the corresponding fluorescent images, and time-lapse videos were created using the SimplePCI software.
Measurement of lactate dehydrogenase (LDH) release
Aliquots of dispersed cells (2 ml) were incubated at 37°C for various lengths of time in the presence and absence of MTX. The cells were pelleted by centrifugation for 15 sec at 12,000 rpm in an Eppendorf centrifuge (model 5415 C). The supernatants were removed, and placed on ice. Enzyme activity in aliquots (50 μl) of the supernatants was determined using the LD-L kit from Sigma. All values are expressed as percent LDH released relative to the value obtained following permeabilization of the cells with 50 μM digitonin.
Measurement of specific caspase-3 activity
Extracts were obtained from confluent BAECs following the indicated treatment using a freeze-thaw lysis protocol. BAECs were mechanically harvested, washed, and resuspended in 500 μl lysis buffer (10 mM Tris pH 7.5, 100 mM NaCl, 1 mM EDTA, 0.01% Triton X-100). The cell suspension was frozen using liquid nitrogen and rapidly thawed in a 37°C water bath. Following 5 freeze-thaw cycles, the cell suspensions were subjected to centrifugation for 5 min at 5000 rpm at 4°C in an Eppendorf 5417R centrifuge. Supernatants were collected and stored at -20°C. Protein concentration was determined by the method of Lowry using bovine serum albumin as standard. Cytosolic extracts (300 μg protein) were assayed for caspase-3 activity according to the protocol described in the EnzChek Caspase-3 Assay Kit (Molecular Probes). Specific caspase-3 activity, defined as the activity inhibitable by Ac-DEVD-CHO, was normalized to that of untreated BAECs.
Supplementary Material
Additional file 1
MTX-induced EB uptake and membrane blebbing in single BAECs. The time-lapse video of the experiment shown in Fig 5A was created from the captured images as described in Material and Methods, with a time compression of 3.5 min (i.e., 7 images) per second. The phase and EB fluorescence images, taken every 30 sec for 40 min, were merged into a single video with EB fluorescence shown as red pseudocolor.
Click here for file
Additional file 2
Effect of U73343 pretreatment on MTX-induced EB uptake and membrane blebbing in single BAECs. The time-lapse video of the experiment shown in Fig 5B was created from the captured images as described in Material and Methods, with a time compression of 3.5 min (i.e., 7 images) per second. The phase and EB fluorescence images, taken every 30 sec for 40 min, were merged into a single video with EB fluorescence shown as red pseudocolor.
Click here for file
Additional file 3
Zeiosis induced by a high concentration of MTX. The effect of 1 nM MTX on ethidium uptake and membrane blebbing in the presence of U73343 (10 μM) was performed as described in the legend for Fig. 5B. The time-lapse video was created from the captured images as described in Material and Methods, with a time compression of 3.5 min (i.e., 7 images) per second. The phase and EB fluorescence images, taken every 30 sec for 60 min, were merged into a single video with EB fluorescence shown as red pseudocolor.
Click here for file
Additional file 4
Rescue of MTX-treated cells by U73343. Time-lapse video for the experiment described in Fig 6 with U73343 added at 6 min after MTX, was created from the captured images as described in Material and Methods, with a time compression of 3.5 min (i.e., 7 images) per second. The phase and EB fluorescence images, taken every 30 sec for 40 min, were merged into a single video with EB fluorescence shown as red pseudocolor.
Click here for file
Additional file 5
Rescue of MTX-treated cells by U73343. Time-lapse video for the experiment described in Fig 6 with U73343 added at 7 min after MTX, was created from the captured images as described in Material and Methods, with a time compression of 3.5 min (i.e., 7 images) per second. The phase and EB fluorescence images, taken every 30 sec for 40 min, were merged into a single video with EB fluorescence shown as red pseudocolor.
Click here for file
Acknowledgements
We thank Zack Novince and Justin Weinberg for excellent technical assistance. This work was supported in part by NIH grant GM52019 and grant 9950014N from the National American Heart Association.
Blood leukocytes constitute two interchangeable sub-populations, the marginated and circulating pools. These two sub-compartments are found in normal conditions and are potentially affected by non-normal situations, either pathological or physiological. The dynamics between the compartments is governed by rate constants of margination (M) and return to circulation (R). Therefore, estimates of M and R may prove of great importance to a deeper understanding of many conditions. However, there has been a lack of formalism in order to approach such estimates. The few attempts to furnish an estimation of M and R neither rely on clearly stated models that precisely say which rate constant is under estimation nor recognize which factors may influence the estimation.
Results
The returning of the blood pools to a steady-state value after a perturbation (e.g., epinephrine injection) was modeled by a second-order differential equation. This equation has two eigenvalues, related to a fast- and to a slow-component of the dynamics. The model makes it possible to identify that these components are partitioned into three constants: R, M and SB; where SB is a time-invariant exit to tissues rate constant. Three examples of the computations are worked and a tentative estimation of R for mouse monocytes is presented.
Conclusions
This study establishes a firm theoretical basis for the estimation of the rate constants of the dynamics between the blood sub-compartments of white cells. It shows, for the first time, that the estimation must also take into account the exit to tissues rate constant, SB.
Background
Blood leukocytes are found in two sub-populations constituting the circulating and the marginating pools. The elements of these two sub-populations are interchangeable, i.e., marginated leukocytes return to bloodstream and vice-versa [1-3]. Therefore, a dynamical equilibrium situation occurs, and blood cell counts should reflect the rate constants of margination and return to circulation of those cells.
Several studies addressed the relationships between the sub-populations of white cells within the blood pool. Among other features, the ratio between the sub-populations under normal conditions [4], under altered states [4,5] and the stability of the equilibrium situation (i.e., to resume a previous value after perturbations, [e.g., [6-8]) were approached. It is presently accepted that the blood sub-pools play an important role as a white cell reservoir when increased demands supervene. Increased demands arise in both pathological and non-pathological situations, as during exercise [9,10], burns [11,12], infectious diseases [13], inflammatory processes [5,14], etc.. It is also currently recognized that some hormones acutely alter the dynamics of the blood sub-pools of white cells [6-8]. In this sense, epinephrine is known to possess a demarginating effect that lasts for less than 1 hour, and such an effect is thought to be the result of changes in the rate constants of margination and return [2,6]. Therefore, the interplay between marginating and circulating leukocytes is a relevant issue that should be taken into account in the interpretation of many results.
On the one hand, it is tacitly assumed that the rate constants of margination and return have higher values than other rate constants related to the dynamics of white blood cells (see below). On the other hand, there is no study addressing such an issue in a formal way in order to provide good estimation of these values. The aim of the present study is to provide the theoretical background to perform estimations of these rate constants, which may prove relevant for empirical studies on healthy normal situations as well as under altered states of the organisms.
Constructing the extensive model
In this sub-section, we present a model based mainly on data from monocytes and neutrophils. The model describes the dynamics of three compartments of these cells and also contains the dynamics of specific growth factors. In the next sub-section we show a reduced model from this one, which will then be employed to obtain an improved estimation for the rate constants values. The extensive model is presented to assure consistence of the analysis.
Bone marrow total rate production apparently has three components: a fixed production rate (P1), a production rate dependent on self-regulatory factors (P2) and a production rate dependent on inflammatory/infectious factors (P(I)) [3]. Therefore, the production rate is not a time independent function, and we denote it by the sum of the three terms above. Once the newly produced cell gains the bloodstream, it may marginate (rate constant M). A marginated cell may then return to circulation (rate constant R). Cells in the blood may leave to surrounding solid tissues, and they do that according to a time-independent rate constant SB[1]. Exit to tissues is also a function of rate constants dependent on inflammatory/infectious factors [3]. Thus, let us denote such a total exit rate constant from the circulation by SB + S(t,I). Tissue leukocytes may proliferate locally (rate constant D), produce self-regulatory factors and, eventually, die (rate constant Z) [15]. This set of empirical data suggests the scheme presented in Figure 1 and the following equation system for describing the dynamics of these cells:
Schematic representation of the dynamics of white cells in the organism. The self-reproduction in tissues is represented by "D", without arrows. Crosses denote death or clearance. See text for details.
Where α are self-regulatory factors (e.g., CSF-1) and φi refers to cells that take part in the circulating (c), marginal (m) and solid-tissues (T) pools. G is a constant related to the production of self-regulatory factors, and K is the rate constant of their clearance. Notice that the exit to tissues comes both from the circulating pool, φc, and from the marginated pool, φm. Such an exit has the same rate constant, SB, independently of the sub-pool. This partition of the total exit constant rate to tissues in two components is the basic reasoning that leads to the inference about the existence of two distinct sub-pools within the blood white cells [1]. Therefore, a cell that touches, attaches and passes through the vessel wall cannot return to the circulating pool, even though during a certain time interval this cell was marginated (semantically but not functionally). In this sense, these cells exit to the surrounding tissues coming directly from the circulating pool, and their dynamics is contemplated by the product φcSB of the differential equation. Those marginated cells that can potentially return to circulation have their dynamics accounted for by the rate constant R. On the other hand, part of these marginated cells migrate to surrounding tissues and the product φmSB contemplates this rate. This dynamics is also consistent to the empirical evidence that margination and diapedesis are distinct features arising from different signals and receptors [16].
Reducing to the compact model
The injection of some substances, epinephrine in particular, changes the rate constants of margination and return to circulation (M and R) for very short periods of time (e.g., see [2,6-8]). Under such a condition, the variable values change as the parameter values change. However, since the substance is rapidly metabolized, the parameters return to their previous value and then the variables would also return to some value maintained previously to the perturbation (in the parameters). Thereby, these substances act on the system described by equations 1a-d analogously to an impulsive function [17]. The returning phase is the relaxing period of the system, and it behaves as the perturbation had occurred in the variables instead of in the parameters. During the short time interval of the relaxing period, the production rate can be considered constant, P. Therefore, equations 1a and 1b maintain no loop-connection to the other two equations (1c and 1d), and a compact system can be written as:
ResultsAnalysis of the steady-state condition
The first step in the analysis was to verify whether the extensive model allows the existence of a stable equilibrium point. An equilibrium point means a set of values of the variables (in this case, φc, φm, φT and α) that does not vary with time (if the system is left externally undisturbed). Stability is related to the behavior of the system in face of perturbations in the variables that displace them to some vicinity of an equilibrium point. The equilibrium point is said asymptotically stable if the system returns to the equilibrium point attained previously. Otherwise, the equilibrium point may be neutrally stable, if the system does not return to the equilibrium point but remains somewhere around it, or unstable, if the system leaves that vicinity away (e.g., [18]).
We verified whether the extensive model posses one asymptotically stable equilibrium point as a first step because the biological system modeled seems to behave like this. Once disturbed, it returns to a previous value. The analysis revealed the existence of one equilibrium point, and such a point is asymptotically stable (see Methods). Let * denote the value of the variable in the equilibrium point. Without loss of generality by considering the inflammatory/infectious factors absent and that φT and α are fixed (e.g., for a short time interval or changes in R or M, see Methods), the equilibrium values of the variables that we are interested in the present study are:
As stated before, blood cell counts should reflect the rate constants of margination and return. Let f* denote the ratio φm*/φc* and thus:
Notice that one is not able to know, directly from the ratio φm* /φc*, the values of R and M themselves.
The demarginated state and its return to the equilibrium condition
In a very short time interval succeeding a bolus injection of some substances (e.g., epinephrine) the parameters R and M change and return to their previous values. Therefore, the circulating and marginal pools attain different values from φc* and φm*, respectively. Given the stability of the equilibrium point (see above) these variables return to φc* and φm*. This returning is governed by a second-order differential equation (see Methods). Let be the value attained above (or below) the equilibrium value φc* by the circulating cells after the perturbation (see Methods). In short, (see equation 9). Therefore, taking the equilibrium value of the circulating pool as a reference value, is the difference between this value and the value found at a time t after the beginning of the returning to the steady-state equilibrium condition. This leads to the following equation governing the fast-phase of resuming the steady-state condition:
where φc0 is the peak (or nadir) value of the circulating pool attained after the perturbation, i.e., the initial value of this variable, and λ is the fast component of the process (λ2, see Methods equation 11b). Equation 5 can be linearized to simplify the estimation procedure. Finally, the λ of equation 5 is (see equation 11b):
λ = -(R + SB)(f* + 1)
This value of λ allows a good estimate of R during the fast phase of the returning to the equilibrium point (steady-state of the sub-pools). In the next section, we will work some examples of the application of the present model and its potential relevance.
Worked examples
The present study offers the means to compute the individual values of R and M. As stated before, the ratio f* is obtained by measuring the marginated and the circulating pools in steady-state conditions. This ratio is equal to M/(R+SB) (see equation 4). Therefore, its is important to note that the individual computation of R (and M) assumes that a series of other independent measurements were done: (a) φc* (the steady-state value of the circulating pool); (b) φm* (the steady-state value of the marginated pool); (c) SB (the exit rate constant from the total blood pool to tissues). The knowledge of a general production rate value may also improve the picture, even though its is not a primary need.
The first example intends to compute the R value for mice monocytes under normal conditions. However, the paucity of adequate data prevents a true computation. Therefore, we will compute an approximate value. The other two examples are completely imaginary. They show the usefulness of the model to address changes in the rate constants that would otherwise pass unnoticed.
I. The normal condition
The computations are based on data from van Furth and Sluiter [1]. The report give φc* = 350 monocytes/mm3, φm* = 500 monocytes/mm3 and SB = 9.6 × 10-4 min-1. The f* value is 500/350 = 1.43. An intravenous bolus injection of epinephrine caused a two-fold increase in φc after 10 minutes of the injection. Therefore, (or, φc(0) = 700). After 60 minutes from the injection time (50 minutes after the monocytes peak), the variables resume their normal values. Notice the extremely limited temporal data. Let us suppose that we had a few more points, as illustrated in Figure 2A. From equation 5, the data could be linearized as:
Simulated perturbation in the blood sub-pools of a white cell type (e.g., monocytes) under three different conditions: normal (blue), tumor carrying animal (red), long-term aerobic exercising animal (green); see text for details. (A) "Raw experimental data", the "discrete" time course of φc after a bolus injection of epinephrine under the normal condition. (B) Log-transformed (linearized) data, under the three different conditions. Notice the linear relationship obtained. (C) "Raw "continuous" data" of the φc time course. Data generated in MatLab 5.3 (The MathWorks, Natick, MA) by built-in numerical integration routines of the set of differential equations of the model.
Where T is the discrete sampling time. Figure 2B shows the result of the linearization procedure. A linear regression would then result in λ = -0.113 min-1. Finally, from equation 6, we know that this value must be decomposed as
-0.113 = -(R + SB)(1 + f*)
Thus, R = 0.0456 min-1, a value 47-fold greater than SB. In the data simulation, R = 0.045 min-1, therefore the procedure seems quite adequate. This example highlights the use of the present model as well as the difference between the eigenvalue (λ, obtained directly from the data) and the constant R. Also, as far as we know, this is the first attempt to estimate the value of R (and M) under a complete and formal model.
II. Tumoral effects
Consider, as another example, that mice carrying a certain tumor type have the exit constant rate equal to the normal value (SB = 9.6 × 10-4 min-1). The f* value is 1.3 and these mice have φc* = 490 monocytes/mm3, φm* = 637 monocytes/mm3. Obviously, production rate have increased. Notice the potential confusing situation arising from these data. Does the tumor growth alter the dynamics between the blood pools? The epinephrine experiment (or some analogue) was conducted, and (or, φc(0) = 600). Figure 2C shows both the "normal" and the "tumoral" data. Linear regression of the log-transformed data computes λ = -0.065 min-1 (see Figure 2B) and thus R = 0.0273 min-1 since f* = 1.3. In data simulation, R = 0.027 min-1, again the results are in close agreement with the real parameter value. Therefore, in this constructed situation, one would be able to conclude that the rate constants of the transit between the blood pool were altered by the pathological condition. This could be an important result in tumor immunology, for example, allowing a better understanding of the pathology or devising new treatment strategies.
III. Long Distance Running Effects
Our final imaginary example is related to the effects of long sustained aerobic exercise. Consider that mice trained in this type of exercise are known to have increased production rate of monocytes. During the steady-state condition of the exercise, φc* = 350 monocytes/mm3, φm* = 700 monocytes/mm3, thus f* = 2. SB increased to 8.5 × 10-3 min-1. An epinephrine experiment is performed during the exercise section and samples are taken during the decaying phase after the injection (the exercise section proceeds during the sampling period). The "continuos" count profile is illustrated in Figure 2C. Linear regression of the log-transformed data (see Figure 2B) computes λ = -0.041 min-1 and thus R = 0.0052 min-1 (in data simulation, R = 0.005 min-1). This is a 9-fold decrease in the R value in relationship to the normal condition. On the other hand, the rate constant M decreased less than 2.5 times in relationship to its value in normal conditions (see above). This result might be important in the understanding of many immune suppression/enhancement phenomena related to certain exercise protocols. Notice that the f* value by itself do not tell anything about the change in the individual value of each rate constant.
Discussion
The compartmentalization of blood leukocytes in two sub-pools is an important feature of these cells. For example, the marginal pool may acutely function as an extra source of cells at increased demand conditions. In this sense, knowledge about the rate constants governing the transit between the sub-pools may prove relevant in studies that approach both physiological and pathological situations. In the present manuscript we provide the theoretical background to support empirical studies related to the calculation of these rate constants.
From the general perspective of an extensive model, we first showed that this model corresponds to what is experimentally found in terms of existence and stability of an equilibrium point in the variables (the sub-pools). After that, we reduced the model to a compact one, concerning only the two blood sub-pools of interest here. Within the context of the compact model, it was shown how an impulsive function (e.g., an intravenous epinephrine bolus injection) perturbing the parameters (the rate constants of margination and return to circulation) can be translated to a perturbation in the variables. Then, the returning of the variables to their equilibrium values allows the estimation of the rate constants. This was done by the use of a second-order differential equation (see Methods). From that equation, the fast-decay component was identified and the corresponding eigenvalue contains the rate constants to be calculated.
The present study demonstrates, for the first time, how to adequately and completely estimate the rate constants of margination and return to circulation of white blood cells. Notice that first-order differential equations (as equation 5) were employed before in gross estimates of these rate constants (e.g., [19]). However, the constant then computed comprised the whole eigenvalue, without discriminating its parts. In other words, its was not recognized that the computed value should be partitioned into three components: R, M and SB.
The model also allows to establish the relationship between R and M. This relationship includes the SB constant and the ratio between the marginal and the circulating sub-pools, f* (see equation 4). The need to take into account the f* value in the estimation of the constants is another important aspect for the first time presented. Therefore, the model leads to a more precise evaluation on how imposed disturbances interfere with the dynamics of the blood sub-compartments and thus increase our understanding of the physiology/pathophysiology of many conditions. Two imaginary situations of this kind were constructed and analyzed as examples of the use of the present model (see "Worked Examples" II and III).
Conclusions
This study provides a complete model to approach the estimation of rate constants of margination and return to circulation of with blood cells. It shows, for the first time, how the value empirically found should be partitioned in order to adequately obtain the desired rate constants. In this sense, we were able to recognize that both an exit rate constant to tissues and the ratio between marginated and circulating cells should be taken into account in the computation procedure.
Methods1. Equilibrium point and its stability
The equilibrium point of the system described by equations 1a-d is found by setting all the derivatives to zero and computing the corresponding values of the variables that lead to such a condition. Without loss of generality, the inflammatory/infectious factors can be taken as zero. Therefore, the equilibrium point (denoted by *) corresponds to the following values:
Conditions of existence: (1) Z > D; and (2) K(Z-D) > P2G. The first condition reflects that tissue leukocytes must die at a rate higher than their own local replication otherwise their population would increase forever. The second condition is similar to the first in the sense that it reflects that the self-stimulatory loop must be lower than the loss loop of the system.
The stability of the equilibrium point is verified by the construction of the determinant of the matrix of the coefficients of the variables of the equations of the system [17,18,20]. In order to be asymptotically stable, all the eigenvalues (roots) of the characteristic polynomial of the matrix must have real parts lower than zero. Considering that the system has four equations, the characteristic polynomial is of the 4th-order with the following general formula:
λ4 + a1λ3 + a2λ2 + a3λ + a4 = 0
where λ is an eigenvalue of the equation (note that the equation has 4 roots). The coefficients (ai) come from the parameters of the system. To have all the eigenvalues with real part lower than zero, the Routh-Hurwitz conditions must verify. In short:
1A. All ai must be greater than zero. This condition is true for the system.
1B. a1a2 > a3. This condition is true for the system.
1C. a3(a1a2 - a3) > a4a12. This condition is not easily verified analytically. Therefore, we performed a numerical analysis. Random values were assigned to the parameters (conditions of existence of the equilibrium point verified, see above) and the condition 1C checked (pseudo-random number generator normally distributed built-in function of MatLab 5.3, The MathWorks, Natick MA). This procedure was taken 150,000 times by a routine specifically written for it and the condition always verified true. Therefore, the equilibrium point seems asymptotically stable.
The main point in this sub-section is to realize that both φT* and α* are not affected by R and/or M. This is what assures the results concerning the compact model below.
2. The compact model and its second-order differential equation form
Considering that production rate and inflammatory/infectious factors would not change in a short time interval, the extensive model allows the blood pool to be treated as having no loop connection to the rest of the system (see equations 2a and 2b). The determinant of the Jacobian matrix of the compact model (i.e., the sub-system represented by equations 2a and 2b) is to set to zero:
Consider now a "push" in φc* (and φm*) in such a way that:
is, therefore, the time course of the perturbation around the equilibrium point of φc. Equation 8 is homogeneous for :
Taking into account that M = (R + SB)f* (see equation 4), the eigenvalues of equation 10 are:
Notice that: (a) both eigenvalues are pure real (this means that the equilibrium point is attained without oscillations in the variables); and (b) both eigenvalues are negative (the equilibrium point is asymptotically stable). Given that λ2 < λ1, λ2 is the fast component of the process of resuming the equilibrium value φc* (see sub-section "the demarginated state and its return to the equilibrium condition"). The value of λ2 is -(R + SB)(f* + 1). This value is presented in equation 6.
List of abbreviations
φc: the number of cells (or concentration) in the circulating pool;
φm: the number of cells (or concentration) in the marginal pool;
φT: the number of cells (or concentration) in solid tissues;
α: quantity (or concentration) of self-regulatory factors;
R: rate constant of return to circulation;
M: rate constant of margination;
SB: rate constant of exit to tissues from the blood pool;
P: bone marrow production rate of the cells;
f*: the ratio between the marginal and the circulating pools at steady-state condition.
Acknowledgments
This study was supported by a FAPESP (State of São Paulo Science Foundation) undergraduate fellowship to K.I. (process number 00/01112-7) and by a FAPESP research grant to J.G.C.B. (process number 00/02287-5).
This manuscript was significantly improved by the comments of two anonymous referees.
Aquaporin-1 (AQP1) channels are constitutively active water channels that allow rapid transmembrane osmotic water flux, and also serve as cyclic-GMP-gated ion channels. Tetraethylammonium chloride (TEA; 0.05 to 10 mM) was shown previously to inhibit the osmotic water permeability of human AQP1 channels expressed in Xenopus oocytes. The purpose of the present study was to determine if TEA blocks osmotic water flux of native AQP1 channels in kidney, and recombinant AQP1 channels expressed in a kidney derived MDCK cell line. We also demonstrate that TEA does not inhibit the cGMP-dependent ionic conductance of AQP1 expressed in oocytes, supporting the idea that water and ion fluxes involve pharmacologically distinct pathways in the AQP1 tetrameric complex.
Results
TEA blocked water permeability of AQP1 channels in kidney and kidney-derived cells, demonstrating this effect is not limited to the oocyte expression system. Equivalent inhibition is seen in MDCK cells with viral-mediated AQP1 expression, and in rat renal descending thin limbs of Henle's loops which abundantly express native AQP1, but not in ascending thin limbs which do not express AQP1. External TEA (10 mM) does not block the cGMP-dependent AQP1 ionic conductance, measured by two-electrode voltage clamp after pre-incubation of oocytes in 8Br-cGMP (10–50 mM) or during application of the nitric oxide donor, sodium nitroprusside (2–4 mM).
Conclusions
TEA selectively inhibits osmotic water permeability through native and heterologously expressed AQP1 channels. The pathways for water and ions in AQP1 differ in pharmacological sensitivity to TEA, and are consistent with the idea of independent solute pathways within the channel structure. The results confirm the usefulness of TEA as a pharmacological tool for the analysis of AQP1 function.
Background
Tetraethylammonium is known as a pore-occluding blocker of voltage-gated potassium channels [1], but it also blocks other cationic channels such as calcium-dependent K+ channels [2,3] and the nicotinic acetylcholine receptor [4]. TEA at 0.1 to 10 mM also inhibits osmotic water flux through human AQP1 channels expressed in Xenopus oocytes, decreasing the net swelling rate in hypotonic saline by 30–40% as compared to AQP1-expressing oocytes not treated with TEA [5]. This blocking effect on osmotic water flux was demonstrated to involve AQP1 channels specifically by using site-directed mutagenesis (tyrosine 186 to phenylalanine) to generate a Y186F AQP1 channel that is insensitive to block by TEA, but retains sensitivity to block by mercury. The blocking effect of mercury in AQP1 channels is dependent on a neighboring residue, cysteine 189 [6]. TEA offers attractive advantages over mercury as a reversible and less toxic blocker for AQP1 channels in experimental analyses of water channel function; however, the relevance of TEA as a blocker for AQP1 channels outside the Xenopus expression system has not been examined previously.
Differences in properties (including pharmacological sensitivities) have been observed for ion channels expressed in oocytes as compared to those in native tissues. For example, the Torpedo acetylcholine receptor and the Shaker K+ channel protein differ in the fractions of protein glycosylated, the composition of the oligosaccharide chains, and the degree of protein maturation when expressed in Xenopus oocytes as compared with channels in native tissue or other expression systems [7,8]. Differences in glycosylation patterns can influence the binding of external blocking agents. The presence of tissue-specific targeting signals not recognized in Xenopus oocytes may lead to protein degradation [9]. These discrepancies raised the question of whether AQP1 channels in native tissues are sensitive to TEA, as they are in Xenopus oocytes. Data presented here show that TEA is effective as a blocker of AQP1 channels expressed in a mammalian renal cell line and in native renal epithelial membrane.
AQP1 channels are complex solute conductors. They are constitutively permeable to water, and also function as regulated non-selective cation channels [10] when gated by intracellular cyclic GMP [11,12]. Even though only a small proportion of the AQP1 channels that are present in the membrane appear to be available for cGMP-activation of the ionic conductance (1/56,000), model-based calculations nonetheless suggest that this contribution could be meaningful in a physiological setting [13]. Unlike transporters which move substantial amounts of substrate by design, ion channels function to change membrane potential by the net movement of relatively tiny amounts of charge (~10 picomoles per cm2 membrane to generate a change of 100 mV) [14], hence the low proportion of active AQP1 ion channels is consistent with a functional goal of gated ionic signaling that is distinct from that of massive constitutive osmotic water permeability.
The role of the AQP1 channel in mediating both water and ion fluxes raises interesting questions about whether the pathways are shared or structurally distinct. Expression of AQP1 channels in bilayers has shown that the water permeability but not the ionic conductance is blocked by p-chloromethylbenzenesulfonate, indicating that the two fluxes may involve distinct pathways [12]. Work presented here provides an additional line of evidence that the pathways for water and ions in AQP1 can be distinguished pharmacologically, using the blocker TEA as a probe. In combination with data from three-dimensional imaging analyses [15], these results support the suggestion that the AQP1 channel may contain separate parallel pathways for ion and water fluxes. These findings have potential significance in understanding the complex role for AQP1 in processes of salt and water movement across cell membranes.
Results and DiscussionTEA inhibition of water flux in AQP1-expressing MDCK cells
TEA inhibited significantly the osmotically-driven net movement of water across monolayers of Madin-Darby Canine Kidney (MDCK) cells expressing human AQP1, as compared to AQP1-infected MDCK cells without TEA treatment (Figure 1). MDCK cells were induced to express AQP1 by infection with replication-deficient adenovirus (see Materials and Methods). These cells showed a high rate of water flux, ranging from 4 to 10 μl/cm2/hour, and a significant block by 1 mM TEA, ranging from 32 to 45% in 6 sets of replicate experiments. The TEA block of AQP1-mediated water flux for all data combined was 34.2%. In control MDCK cells infected with empty virus, the water flux was significantly lower (approximately 10-fold), ranging from 0.2 to 1.4 μl/cm2/hour. In these control cells, the net flux rates with and without treatment were not significantly different. Therefore, the enhanced water permeability and concomitant sensitivity to block by TEA were dependent on the expression of AQP1 channels. MDCK cells transfected with empty virus showed no augmentation of transepithelial water movement, confirming previous work that showed water does not flow across tight junctions in MDCK confluent layers [16]. Cells transfected with empty virus also showed no effect of TEA treatment. The magnitude of block of AQP1-mediated water permeability by TEA is consistent with that reported previously for AQP1 channels expressed in Xenopus oocytes [5].
Block by TEA of AQP1-mediated water permeability in monolayers of adenovirus-infected MDCK cells. (A) To account for differences in levels of AQP1 expression between different experimental preparations, data for "Percent water flux" were standardized to the mean water flux measured for AQP1-expressing monolayers without TEA on the same experimental day. Cells infected with empty adenovirus were untreated (Control) or treated with TEA (Con TEA). Cells infected with AQP1 adenovirus and expressing AQP1 were untreated (AQP) or treated with TEA (AQP TEA). Box plots show the distribution of data values; the gray box encloses 50% of the data points, the error bars show the full range, and the horizontal bar shows the modal value. Data for mean, standard error and n value are shown below each column. Significant differences were analyzed by T-test and are reported as p < 0.001 (*), and not significant (N.S.; p < 0.50) (B). Whole cells lysates were prepared from MDCK cells on filters and 10 μg of protein was loaded onto each lane. Shown is the expression of AQP1 protein in MDCK cells infected with empty (CON) or AQP1 adenovirus. No AQP1 protein was seen in CON lanes even after extended exposures to film (i.e., 30 minutes).
TEA inhibition of water flux in renal tubule
Preparations of pure descending thin limbs and pure ascending thin limbs from the inner medulla were segregated by morphological features. Descending thin limbs with their cells with nuclei protruding into the lumen differ markedly in appearance from ascending thin limbs with their large, round, flat nuclei when viewed with DIC optics [17,18]. The differential expression of AQP1 in the two different types of segments (abundant in the descending thin limb and absent in the ascending thin limb), visualized by immunocytochemistry, confirmed the accuracy of the morphological identification (Figure 2). AQP1 provides the predominant pathway for water transport in the descending thin limb [19,20]. In preparations of isolated perfused descending thin limbs, net osmotic water absorption rates (Jv) and water permeability (Pf) were significantly lower in the presence of 10 mM TEA than in the absence of TEA (Figure 3). This sensitivity to TEA correlates with the expression of AQP1 channels in this region. The overall effect of TEA in descending thin limb was to cause a 49% block of JV and a 50% block of Pf. The pairwise comparison of individual tubule preparations, before and after TEA, showed a significant percent change of (-) 52 ± 6 % (mean ± SEM; n = 10) in descending thin limb. In contrast, preparations of ascending thin limbs that lack AQP1 channels showed essentially no net water flux and no effect of TEA under osmotic gradient conditions. For data from ascending thin limbs, the pairwise comparison of individual tubule preparations before and after TEA showed a non-significant percent change of (+) 17 ± 35 % (n = 9). These data show that the blocking effect of TEA on water permeability originally observed in AQP1-expressing oocytes can be reproduced in a physiologically relevant condition. TEA blocks water permeability in a tissue that has abundant native expression of AQP1 channels, whereas it has no effect in tissue lacking AQP1 expression.
Immunocytochemical analysis confirms the expression of AQP1 channels in pure descending but not ascending thin limb preparations. Examples are shown for isolated preparations of rat inner medullary descending thin limb(upper panel) and inner medullary ascending thin limb (lower panel) labeled with a rabbit antibody to the AQP1 carboxy tail domain, and visualized with Cy5-labeled antirabbit antibody immunoglobulin-G. The scale bar is 100 μm.
Analysis of water flux in isolated perfused renal ascending and descending thin limbs of Henle's loop under osmotic gradient conditions.A. Compilation of Jv values obtained for descending thin limbs that were untreated (Desc cont) or exposed to 10 mM TEA (Desc TEA), and for ascending thin limbs that were untreated (Asc cont) or exposed to TEA (Asc TEA). B. Compilation of calculated osmotic water permeabilities (Pf) for the same tubules as analyzed in (A). Box plots show the distribution of data values; the gray box encloses 50% of the data points, the error bars show the full range, and the horizontal bar shows the modal value. Data for group means, standard errors and n values are shown below each column. However, since each tubule served as its own control, differences were analyzed by paired T-test and are reported as p < 0.02 (**), p < 0.05 (*), and not significant (N.S.).
Lack of TEA block of ion channel conductance
Heterologous expression of human AQP1 in Xenopus oocytes, a system that has been well characterized for electrophysiological assays, was used for two-electrode voltage clamp studies to evaluate the effects of TEA on ionic currents. External TEA(10 mM) failed to block the cGMP-dependent ionic current in AQP1-expressing oocytes (Figure 4). Current responses were assessed with voltage steps from +40 to -70 mV, from a holding potential of 0 mV (Fig. 4A). The initial response shows the current-voltage relationship recorded by two-electrode voltage clamp in an AQP1-expressing oocyte in isotonic 100 mM Na+ saline (see Methods). The activated response was measured in the same oocyte at 3–5 minutes after addition of 4 mM sodium nitroprusside (SNP), which elevates endogenous cGMP levels. The effect of 10 mM TEA (with 90 mM NaCl) was assessed by perfusing the recording chamber with TEA saline containing SNP. No appreciable change was observed. For the representative example shown in Figure 4, the conductance (measured as the linear slope of the current-voltage relationship between -50 and +30 mV; Fig. 4B) for the initial condition was 3.0 nA/mV. After channel activation the conductance was 58 nA/mV in Na+ saline and 57 nA/mV in 10 mM TEA saline. The reversal potential of the current (Fig. 4B) shifted from approximately -24 mV in the initial state to about -4 mV in the activated state, consistent with the activation of a non-selective cationic conductance permeable to both Na+ and K+. Similar results were observed for other SNP-activated oocytes (n = 5), as well as for AQP1-expressing oocytes activated by prior incubation in 10 mM 8Br-cGMP (as described in [11]). In AQP1-expressing oocytes, the conductance induced by 8Br-cGMP in control Na+ saline did not change significantly [(+) 11.6 ± 6.2 % (mean ± SD; n = 7)] after perfusion with 10 mM TEA saline, thus confirming the absence of a blocking effect of TEA on the ion conductance by AQP1 channels. The activated AQP1 ion channel is less permeant to TEA than to Na+ or K+[10]; however, because of the high extracellular Na+ concentration (90 mM) in the TEA bath saline in the present study, the measured reversal potential of the activated current was not appreciably different with and without 10 mM TEA.
Tetraethylammonium does not block the ionic conductance in cGMP-activated AQP1 channels expressed in Xenopus oocytes.A. Recordings of current responses to voltage steps analyzed by two-electrode voltage clamp in an AQP1-expressing oocyte before stimulation (initial), after cGMP-dependent activation in 100 mM Na+ saline (activated) and after perfusion with saline containing 10 mM TEA and 90 mM Na+ (+TEA). No block of the ionic current was observed. B. Plot of the current-voltage relationship for data illustrated in (A). Conductance values calculated from linear fits of the data from -50 to +30 mV were 3.0 μS initial, 57.8 μS after activation in Na+ saline, and 57.3 μS with 10 mM TEA.
Results presented here show that TEA serves as an effective, though partial, blocker of the osmotic water permeability mediated by human AQP1 channels expressed in a heterologous mammalian cell line and by native AQP1 channels in isolated rat renal descending thin limbs of Henle's loops. Before the characterization of TEA, mercurial compounds were among the only known blockers of AQP1, affecting both the constitutive water flux and the regulated ion conductance [6,10]. For AQP1 channels expressed in Xenopus oocytes, the block of osmotic water flux is 5% in 0.01 mM TEA, 21% in 1 mM TEA, and 36% in 10 mM TEA, as referenced to the osmotic water flux in control AQP1-expressing oocytes [5]. Data presented here show that the level of block by TEA is comparable or greater for AQP1 channels expressed in mammalian cells. 1 mM TEA produces a 34% block of osmotic water flux in MDCK cells, and 10 mM TEA produces a 50% block in renal descending thin limbs. The effect of TEA on water permeability in cells and tissues that express high levels of AQP1 channels is likely to be significant, given the substantial amount of water that can be moved across AQP1-expressing epithelia (for example, see data on descending thin limbs in Fig. 3). The idea that TEA might be a candidate for a lead compound for the development of drugs with possible clinical applications [5] is supported by the present data that show the effectiveness of TEA in two different mammalian models of AQP1-mediated osmotic water permeability. While the usefulness of TEA in vivo as a research tool may be limited by toxicity and effects on other channels at higher concentrations, it offers a degree of reversibility and selectivity not seen with mercury, an agent which acts by covalent modification of all exposed cysteine residues [21,22].
The regulated ion channel function of AQP1 is not blocked by external TEA. An intriguing question remains regarding the location of the permeation pathway for cations in AQP1 channels. Analysis of the crystal structure of GlpF suggests that potential ion binding sites may be located in the central pore, at the four-fold axis of symmetry in the tetrameric association of subunits that form the channel [15]. Interestingly, this arrangement of a tetramer of subunits to form a central ion pore is a hallmark of the family of voltage-gated ion channels [23], which rely on a tetrameric organization to create a central pore for ions that is lined by sequences contributed from all four subunits. In contrast, the four individual pores for water or glycerol, located within each subunit of AQP1 or GlpF, appear to be lined with hydrophobic residues expected to preclude ion permeability [15,24,25]. Importantly, TEA distinguishes between the two pathways in AQP1, supporting the idea that the ion and water pores are physically distinct although they co-exist in the same channel complex.
The AQP1-mediated ion conductance may contribute measurably to the function of tissues in which AQP1 is abundantly expressed, such as in the mammalian renal proximal tubule [13]. Further work is needed to assess the additional levels of regulatory control that may constrain the availability of the AQP1 proteins to serve as ion channels gated by cGMP, within a background of constitutive water channel activity. The maintenance of an ionic pore, as well as a conserved cGMP-binding domain in the AQP1 carboxy tail for regulating activation [11], suggests that the ion channel function is likely to be essential for physiological regulatory processes that have yet to be fully appreciated.
Materials and MethodsMDCK cellsCell culture
Madin-Darby Canine Kidney (MDCK) cells were obtained as a gift from Dr. R. Lynch, University of Arizona, and used between passages 13–14. Cells were grown and maintained in Dulbecco's Modified Eagle Medium (Life Technologies) that was supplemented with 10% fetal bovine serum (Hyclone), 170 mM glutamine and penicillin/streptomycin (100 μg/ml, Life Technologies) in humidified air containing 5% carbon dioxide at 37°C.
Adenovirus vector
The adenovirus (AV) backbone for the aquaporin-1 (AQP1) sense and antisense AV constructs was a replication-deficient "first-generation" AV with deletions of the E1 and E3 genes [26]. This "empty" AV contains the cytomegalovirus (CMV) promoter and bovine growth hormone polyadenylation (bHG) site separated by a polylinker that was used to clone AQP1 DNA as described previously [27]. A recombinant AQP1 virus was constructed using a plasmid containing the coding sequence for AQP1, pCHIPev [28]. pCHIPev was digested and the AQP1 insert was subcloned into the shuttle vector pSKAC, creating pSKAC/AQP1. pSKAC contains map units 0.0–1.3 of the AV which includes the left terminal repeat of AV, a CMV promoter, an AMV translation enhancer and a polylinker region. DNA fragments containing AQP1 DNA were liberated from pSKAC after restriction and ligated into adenovirus as described previously [27]. Human embryonic kidney cells (293 cells) were transfected with ligation mixture and individual viruses were isolated from cell lysates by two consecutive rounds of plaque purification using an agar overlay as described previously [27]. Individual viruses were amplified in 293 cells and purified over cesium step gradient. Individual AV DNA titers were determined by three different methods: 1) plaque titration on 293 cells, 2) immunofluorescence microscopy of AV protein expression (anti-penton group antigen, clone 143, Biodesign, Kennebunk, ME) in 293 cells infected with serial dilutions of AV and 3) absorbance at 260 nm (pfu/ml = A260 × dilution × 1010).
Osmotic water flux assay
Water permeability was measured as the net fluid movement driven by an osmotic gradient across intact monolayers of adenovirus-infected MDCK cells, by methods similar to those reported previously [27,29]. Cells were seeded onto Transwell filters (Costar, 1 cm2, 0.4 μm pore size) at a density of 1.5 × 105 cells per well. Cells were maintained in humidified air containing 5% CO2 for two weeks to allow for cell-cell junctions to mature. Cells were infected at the apical surface with adenovirus containing AQP1 cDNA or with empty adenovirus at a multiplicity of infection (MOI) of 10. Five days post-infection, all medium was removed carefully and completely from both upper (apical) and lower (basolateral) chambers. Exactly one milliliter of fresh prewarmed isotonic medium (~300 mOsM) was added to the lower chamber and exactly 175 μl of hypertonic medium (~450 mOsM) was added to the upper chamber at time zero. In experimental groups, TEA chloride was present at a 1 mM concentration in both the upper and lower chambers. After four hours of incubation at 37°C, all of the medium from each of the upper chambers was removed carefully and volume was measured using an analytical balance.
Immunoblot analysis
Sodium dodecyl sulfate (SDS) solubilized whole cell lysates containing 5% β-mercaptoethanol were electrophoresed into 12% polyacrylamide gels containing 0.1 % SDS. Fractionated proteins were blotted onto nitrocellulose using the Transblot system as per manufacturer instructions (Biorad, Hercules, CA). The blots were preincubated for 30 minutes at 22°C in Tris-buffered saline containing 5% nonfat powdered milk and 0.2% Tween-20 (TBS-T), and were then probed with affinity purified anti-AQP1 IgG (1:5000) for 2 hr at 22°C. The blots were washed (3 × 15 min) in TBS-T and were incubated for 2 hours with horseradish peroxidase-conjugated secondary antibodies (goat anti-rabbit, 1:5000, Pierce). The blots were washed (3 × 15 min) in TBS-T and specific labeling was visualized following enhanced chemiluminescence (Pierce, Rockford, IL) and exposure (10 seconds) to ECL-Hyperfilm (Amersham, Arlington Heights, IL). Immunoblots were digitized using the UVP gel documentation system and densitometry was performed using LabWorks software (Upland, CA).
Isolated perfused rat inner medullary thin limbsAnimals and dissection of tubules for immunocytochemistry and for in vitro fluid movement measurements
Young male Munich-Wistar rats (Harlan Sprague-Dawley, Indianapolis, In), 90 g average body weight, were maintained on Teklad Mouse/Rat Diet No. 7001 with free access to water. Dissection of inner medullary thin limbs of Henle's loop from fresh rat renal tissue for immunocytochemistry and for in vitro microperfusion was performed as described in detail previously [17]. We used descending thin limb (DTL) and ascending thin limb (ATL) segments from the top 70% of the inner medulla above the pre-bend level. Descending and ascending thin limbs were identified by cell type using an inverted microscope with Nomarski differential interference contrast (DIC) optics at 400× magnification [17,18].
Immunocytochemisty
Rabbit antibody recognizing the carboxy tail domain of AQP1 (from W.D. Stamer and J.W. Regan, University of Arizona) was used for immunocytochemical confirmation of protein expression in isolated thin limbs of Henle as described previously [30]. Tubules were positioned onto a glass microscope slide covered with a layer of Cell-Tak® adhesive (Becton Dickinson, Bedford, MA). Tubules were then fixed in 4% paraformaldehyde for 10 min and permeabilized with 100% methanol for 2 min at -20°C. Primary antibody or non-immune rabbit serum (diluted 1:500) was applied overnight at 4°C., followed by incubation with biotinylated goat anti-rabbit antibody (diluted 1:100, 60 min), and incubation with streptavidin conjugated to Cy5 (60 min). Digital fluorescent images were obtained with a Leica-TCS confocal microscope.
Perfusion and bathing solution for in vitro microperfusions
The Ringer solution used for perfusing and bathing the tubules was that initially used by Chou and Knepper [18] in perfusion of thin limbs from chinchilla kidneys and modified by us. It contained (in mM): 130 NaCl, 20 HEPES, 5 NaHCO3, 2.5 K2HPO4, 2 CaCl2, 1.2 MgSO4, 5.5 glucose, and 5 urea. The osmolality was about 290 mOsm/kg H2O, the pH was adjusted to 7.4 while the solution was gassed with 95 O2/5% CO2.
Perfusion of tubules in vitro
We perfused the dissected tubules in vitro by a technique essentially the same as that originally described by Burg et al. [31] and modified for use in our laboratory [17,32,33]. The perfusion rate was about 15–20 nl/min. The temperature of the bathing chamber was maintained at 37°C during perfusion and the bath was covered with a layer of light paraffin oil. To check for leaks and to measure net fluid movement across the epithelium with an imposed osmotic gradient, [14C]dextran (MW ~70,000) was added to the perfusate.
Determination of net water movement and water permeability of thin limbs perfused in vitro
There is no net transepithelial water movement with solutions of identical osmolalities in the bath and lumen. For this study, however, we determined the occurrence of net water absorption and water permeability when we imposed an osmotic gradient from bath to lumen (see below). Net water absorption, Jv (nl min-1 mm tubule length-1), was determined in each collection period with [14C]dextran in the perfusate using the following relationship: Jv= (Vi - Vo)/L. In this equation, Vi(initial perfusion rate, nl/min) is calculated by dividing the cpm of 14C in the collected fluid by the cpm/nl of 14C in the initial perfusate and by the time of the collection period; Vo (collection rate, nl/min) is measured directly from the collection; and L (mm) is the length of the tubule perfused, measured with an ocular micrometer. For purposes of determining the water permeability of DTL and ATL, we imposed a 100 mOsmol/kg H2O osmotic gradient from bath to lumen with sucrose and determined Jv as just described. Water permeability (Pf, μm s-1) was then determined from the equation Pf = Jv/(As Vw δCosmol) where As = luminal surface area (π DL), Vw = partial molar volume of water (18 ml/M), δCosmol = transepithelial osmolality gradient [18]. When the effect of TEA on water movement was examined, TEA bromide (10 mM) was added to the bathing medium and the NaCl concentration was reduced by 10 mM to maintain the osmolality constant. Collection periods were 5 min in length and each tubule served as its own control for measuring net water flux and water permeability in the presence and absence of TEA. [14C]dextran (sp act 1.08 mCi/g) was obtained from American Radiolabeled Chemicals, St. Louis, MO., USA. Results are summarized as means ± SE. The n value is the number of tubules. Each tubule came from a different animal.
Xenopus oocytesOocyte preparation
Adult female Xenopus laevis were anesthetized with tricaine methane sulfonate (MS-222, Sigma Chemical Co). Stage V-VI oocytes were removed by partial laparotomy and defolliculated by collagenase treatment [11]. Cloned human Aquaporin1 DNA was provided by Dr. P. Agre [28], linearized with XbaI and used as a template for in vitro synthesis of cRNA with T3 RNA polymerase. Prepared oocytes were injected with 50 nl of sterile water (control oocytes) or 50 nl of sterile water containing AQP1 cRNA (~0.2 to 0.5 ng/nl) and were incubated for 2–5 days at 18°C in culture medium (ND96: 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, 5 mM Hepes, 2.5 mM pyruvic acid, 100 U/ml penicillin, and 100 μg/ml streptomycin, pH 7.6) to allow protein expression.
Electrophysiology
Two-electrode voltage clamp recordings were performed at room temperature with electrodes (0.5–3 MΩ) filled with 3 M KCl. Data were recorded with a GeneClamp 500 (Axon Instruments, Foster City, CA), filtered at 2 kHz, digitized at 50 to 2000 μs and analyzed with pClamp software (Axon Instruments). Control recording saline for two-electrode voltage clamp contained (in mM): 100 NaCl, 5 MgCl2, and 5 HEPES, pH 7.3. TEA saline was made with (in mM) 10 TEA.Cl, 90 mM NaCl, 5 MgCl2, and 5 HEPES, pH 7.3. Activation of the ionic current was induced by preincubation of AQP1-expressing oocytes for 10–30 minutes in 10 mM 8Br-cGMP (Sigma) in high K+ saline, containing (in mM): 30 KCl, 80 K gluconate, 5 MgCl2, 5 HEPES, pH 7.3, or by addition of 2–4 mM sodium nitroprusside (SNP), a stimulator of guanylate cyclase and cGMP production [34,35]. Control oocytes treated in the same conditions showed no induction of an ionic current. Higher concentrations of SNP (>10 mM) affected ionic permeability in some batches of control oocytes and were not used. Current voltage relationships were assessed in Na+ control saline and in TEA saline to assess potential blocking effects of TEA.
Chemicals
All chemicals were purchased from standard sources except as specified above and were of the highest purity available.
List of abbreviations
AQP1 Aquaporin-1
Jv Net osmotic water absorption rate
MDCK Madin-Darby Canine Kidney cell
Pf Osmotic water permeability
SNP Sodium nitroprusside (nitric oxide donor)
TEA Tetraethylammonium (chloride salt)
Acknowledgements
We thank Dr. John Regan for antibodies to AQP1, Amy Marble and Dr. Kathryn Bolles for technical assistance, Dr. Heddwen Brooks for helpful discussions, Neil Atodaria and Eileen Ryan for technical assistance with permeability assays using MDCK cells, and Dr. Ron Lynch for the gift of MDCK cells (University of Arizona). Funding was provided by NIH R01 GM59986 (AJY), American Health Assistance Foundation #G2001-026 (WDS), and NIH R01 DK16294 (WHD) and Research to Prevent Blindness Foundation.
Proteorhodopsin (pR) is a light-activated proton pump homologous to bacteriorhodopsin and recently discovered in oceanic γ-proteobacteria. One perplexing difference between these two proteins is the absence in pR of homologues of bR residues Glu-194 and Glu-204. These two residues, along with Arg-82, have been implicated in light-activated fast H+ release to the extracellular medium in bR. It is therefore uncertain that pR carries out its physiological activity using a mechanism that is completely homologous to that of bR.
Results
A pR purification procedure is described that utilizes Phenylsepharose™ and hydroxylapatite columns and yields 85% (w/w) purity. Through SDS-PAGE of the pure protein, the molecular weight of E.-coli-produced pR was determined to be 36,000, approximately 9,000 more than the 27,000 predicted by the DNA sequence. Post-translational modification of one or more of the cysteine residues accounts for 5 kDa of the weight difference as measured on a cys-less pR mutant. At pH 9.5 and in the presence of octylglucoside and diheptanoylphosphotidylcholine, flash photolysis results in fast H+ release and a 400-nm absorbing (M-like) photoproduct. Both of these occur with a similar rise time (4–10 μs) as reported for monomeric bR in detergent.
Conclusions
The presence of fast H+ release in pR indicates that either different groups are responsible for fast H+ release in pR and bR (i.e. that the H+ release group is not highly conserved); or, that the H+ release group is conserved and is therefore likely Arg-94 itself in pR (and Arg-82 in bR, correspondingly).
Background
Proteorhodopsin is a 249-amino acid membrane protein native to several uncultured species of γ-proteobacteria, which are a component of marine plankton [1]. Addition of retinal to E. coli expressing pR was shown to cause a reddish coloration of the bacteria with an absorption maximum near 520 nm. The pR contained in the bacterial membranes was shown to act as a light-activated proton pump, but only when retinal is present. Time-resolved UV/vis studies at pH 8 also revealed that the protein undergoes a photocycle, similar to that of wild type bacteriorhodopsin, but with a predominance of the O intermediate instead of M.
The bR photocycle has been characterized by spectroscopic methods as having six principal photointermediates: bR, K, L, M, N and O. Each intermediate has a distinct absorbance maximum; the most studied are bR (570 nm), M (412 nm), and O (640 nm) since these are the ones that can be produced in the highest concentration at physiological pH values. Monitoring of the absorbance at individual wavelengths after photoexcitation is used to determine the relative concentrations and decay times of each of these photointermediates. The L → M transition in bR is characterized by the deprotonation of the Schiff base to Asp-85, producing the distinctive 412 nm absorbance maximum of M, and by so-called fast proton release, the ejection of a proton from a different (unknown) residue into the external medium on the ~10–100 μs time scale, depending on pH. Reprotonation of the Schiff base from Asp-96 occurs during the M → N transition with an absorbance maximum of 560 nm [2]. The N → O transition involves the reprotonation of the Asp-96 from the cytoplasmic space.
Like bR, pR consists of seven transmembrane α-helices that include in the membrane interior all of the residues conserved among archaeal rhodopsin proton pumps. In particular, analogues of Asp-85, Asp96, Arg-82, and Lys-216 of bR are present in pR. Conspicuously absent are analogues for Glu-194 and Glu-204 of bR. The latter, as well as Arg-82, have been implicated in fast proton release. In particular, mutagenesis of Glu-194 or Glu-204 in bR results in loss of fast proton release [3,4]. The absence of homologs for these residues in pR leaves open the question of whether it carries out fast H+ release.
Experiments described here demonstrate that pR does indeed undergo fast H+ release, at least under elevated pH conditions that resemble somewhat those of the γ-proteobacteria's native open ocean environment. We also demonstrate that there is a post-translational modification of at least one of the three native cysteines when pR is expressed in E. coli. Both of these discoveries were made possible through purification methods for pR described herein.
ResultsPurification
PR was obtained in 85% purity, assuming that values of ε280 and ε546 for pR are the same as for bR solubilized in DMPC/cholate/SDS mixtures at pH 8 (ε280= 7.85 × 104 cm-1 M-1 and ε551 = 4.8 × 104 cm-1 M-1) [5]. This assumption is actually expected to underestimate the purity of pR produced, by up to ~20%, since the pR we expressed has 10 tryptophan and 14 tyrosine residues, as compared to 8 tryptophans and 11 tyrosines in bR from H. salinarum. The absorbance of contaminant proteins was assumed to be 1.1 for a 1 mg/mL solution. By using these assumptions, the relative concentrations of pR and other proteins can be determined from the absorbance spectra of the various fractions (fig. 1). The resulting purity values correlate well with those Coomasie-stained SDS-PAGE gels (see below). The OG extract of cholate-washed membrane pellets starts out at a pR content of 7% total protein (w/w). The Phenylsepharose column increases the purity level to 24%, with approximately 5% loss. The final purification step by hydroxylapatite column chromatography produces pR with ~85% purity and a further loss of ~60%, i.e. the overall yield of the two column procedure was ~30%.
UV/visible absorption spectra of pR in octylglucoside solution (1–3%) at three stages of purification. All three spectra were measured in the presence of octylglucoside at pH 8, and are normalized to the 280-nm protein peak. Spectrum A, the OG extract of cholate-washed E. coli membranes; spectrum B, pooled 546-nm absorbing fractions from Phenylsepharose column; spectrum C, same material after hydroxylapatite column.
Polyacrylamide gel electrophoresis
Relative to protein standards, the apparent molecular weight of bR is 25,000 while the apparent molecular weights of pR-wt and pR-TCM are 36,000 and 31,000, respectively (fig. 2, lanes E and C, respectively). SDS-PAGE (fig. 2) also confirms the estimates of purity level based on the assumed ε280/ε546 ratio identical with that of detergent solubilized bR. Interestingly, the pR appears to be a doublet band whose relative concentrations remain almost unchanged during purification. This doublet is also present in the less-purified sample of pR-TCM, with both bands shifted down by approximately the same amount (fig. 2, lane C).
SDS-PAGE of pR (wild type and pR-triple cysteine mutant). Lane A contains bacteriorhodopsin (bR). Lanes B and F contain BioRad protein molecular weight markers including labeled bands at 21.5 (trypsin inhibitor), 31 (bovine carbonic anhydrase), and 45 (ovalbumin) kDa. Lane C is of the pR triple cysteine mutant (TCM). Lane D contains the Phenylsepharose™-purified pR wild type protein, corresponding to spectrum B of fig. 1. Lane E contains the hydroxylapatite-purified pR wild type protein, corresponding to spectrum C in fig. 1.
Subsequent SDS-PAGE analysis of pR samples that had been stored for periods of time up to several months indicate that after sitting for several weeks in octylglucoside solution at 4°C, the largest post-translational modification on wild-type pR is eliminated – presumably hydrolysed off of the cysteine(s) – leaving only a 31,000-MW band indistinguishable from that seen for pR-TCM (data not shown). Furthermore, after boiling for several min in gel loading solution, this cleaved wild-type protein, as well as the TCM, both give an extra artifactual band near 36,000 dalton. The latter band, a singlet, is coincidentally at almost the same apparent MW as the doublet from the uncleaved post-translationally-modified wild type pR (fig. 2, lane E). These potential artifacts should be taken into consideration in any attempt to reproduce the results in Fig. 2.
Photocycle kinetics and flash-induced proton concentration changes
Photocycle kinetics were measured at 400, 500. and 580 nm in the presence of the short-chain lipid DHPC. This lipid does not support the formation of closed bilayer vesicles, but rather forms micelles like a detergent. The time-resolved measurements showed no positive 400-nm absorbance signals at pH 8.0 or lower (Fig. 3). This is somewhat in disagreement with Béjà et al [1], who detected small 400-nm transient absorbance increases upon photolysis at pH 8.0. However, we observed a transient 400-nm absorbance increase at an elevated pH of 9.5 (fig. 3).
Dependence on pH of the M-like intermediate of pR. Time courses of flash-induced absorbance changes measured at 400 nm and 22°C for pR in 1% DHPC/100 mM NaCl solution at pH 6.5, 8.0 and 9.5. A positive differential absorbance at 400 nm is indicative of the presence of the M intermediate. The logarithmic time scale ranges from 100–107 μs after photolysis by a 10-ns laser pulse at 500 nm, with an energy of 3–6 mJ.
At pH 9.5 in the presence of DHPC, and observing transient changes at 500 nm (fig. 4), pR undergoes a 2-phase decay after the initial unresolved absorbance decrease. Multiexponential fits show that the first decay phase has a time constant of 4 μs, in good agreement with the 4-μs rise time of the 400 nm signal (Fig. 4). The amplitude of this decay represents about 80% of the initial absorbance depletion. The second phase of the 500 nm absorbance decay occurs with a substantially slower time constant of 0.5 s, returning the remaining 20% of initial absorbance change. The slowest decay components of the positive 400-nm signal and the negative 500-nm signal follow similar kinetics, although the amplitudes of these components differ by a factor of 3. At pH 9.5, the 580 nm trace has no significant positive values indicative of an O-like intermediate, although, in agreement with earlier measurements [1], at lower pH values a red-shifted transient is the predominant positive absorbance signal (data not shown).
Photocycle kinetics of pR at selected wavelengths at pH 9.5. Time traces were measured at 400, 500, and 580 nm. The 400-nm trace shows the kinetics of the M intermediate, i.e. the deprotonated Schiff base, as in Fig. 3. The 500-nm trace shows the depletion signal of pR at the earliest times, and then the time course of the N intermediate as well as return of the pR resting state. The 580-nm trace is indicative of an O-like intermediate. The conditions are 1% DHPC, 100 mM NaCl, pH9.5 at 22°C. The laser excitation is as in fig. 3.
Figure 5 shows a different type of time-resolved measurement, probing not the pR chromophore, but rather pH changes in the protein environment. Proton concentration changes in the aqueous bulk phase were measured with the pH sensitive dye cresol red, which has a pKa of 8.2–8.5. The bottom trace in Figure 5 shows the absorbance change of the indicator during the pR photocycle. The negative signal is indicative of a pH decrease, corresponding to transient H+ release from the protein into the solution. The best-fit time constant for the release phase is 6 μs. The positive 400 nm trace in fig. 5 (reproduced from fig. 3) shows that the proton release and uptake follow kinetics very similar to the apparent formation and decay of M, as is typically seen in bR near neutral pH [6,7,21]. However, no proton release signal could be observed for pR at pH 6 or 8 (data not shown).
Comparison of the kinetics of M formation and decay with kinetics of ET release and uptake. The time trace of the M-like intermediate was measured at 400 nm (upper panel). Time-resolved H+ concentration changes (lower panel) were measured with the pH indicator dye Cresol Red. A negative Cresol Red absorbance change at 580 nm is indicative of a transient decrease in the pH of the solution, i.e. of H+ release by pR. Solid lines represent multiexponential fits, with the main rise and decay times indicated for the M intermediate. The H+ release and uptake time constants obtained from the fit are marked with arrows pointing down for release and pointing up for uptake. Sample and excitation conditions are as in fig. 4.
DiscussionPurification of pR
Meaningful comparison of intrinsic physiological properties of pR and bR depends on the purification of pR. The E. coli expression system can easily be used to prepare pR at 85% purity with similar or less effort and time than required for bR (purple membrane) production from S9 H. salinarum.
The initial purity level of the OG-solubilized cholate-extracted membrane is about 7% pR by weight. Phenylsepharose™ column chromatography separates proteins on the basis of hydrophobicity, and has been used previously in hR purification [8-10]. As an initial purification step, the Phenylsepharose™ column achieved a substantial increase in protein purity to 25%, along with removal of most of the lipid. The final purification step, utilizing a hydroxylapatite column, has been previously used with rhodopsin [11]. This column proved to be more efficient as a final step in the purification than as a preliminary one, because large amounts of contaminant protein tended to slow the flow rate drastically. This step of the purification yielded an increase in protein purity to >85%. The overall yield of pR from membrane through the column purifications is ~30%. Most of this loss, ~65%, occurs during the hydroxylapatite column.
Molecular weight differences
Béjà et al. reported a molecular weight for wild type pR of 27 kDa based on the predicted amino acid sequence of the protein [1]. Its 249 amino acids barely exceed the 248 of mature bR, which has a molecular weight of 26,000. However, we observed a significantly higher apparent molecular weight (~36,000) for wild type pR on SDS-PAGE gels (Fig. 2, lanes D and E). Post-translational modification of pR must almost certainly account for some of the observed molecular weight difference between pR and bR. Lipids or sugars covalently bound to the protein surface would not be removed during the purification procedure and could cause a higher apparent molecular weight. Cysteines are frequently a site of lipid association with membrane proteins, (e.g. mammalian rhodopsin, which has two palmitoyl molecules attached to cysteine residues). Therefore we compared the SDS-PAGE gel mobility of wild-type pR to that of a cys-less mutant (pR-TCM). The elimination of the three possible sulfhydryl attachment sites lowers the apparent molecular weight of the pR-TCM by 5,000 (Fig. 2, lane C). From this, we conclude that at least one of the three cysteines in pR is probably modified post-translationally.
However, this by itself does not account fully for the anomalous mobility of pR on SDS-PAGE gels, because pR-TCM is still approximately 5,000 higher in molecular weight than bR according to SDS-PAGE (fig. 2, lanes A & C). Only ~2,600 of this can be accounted for by the V5 eiptope and poly-histidine tail that are appended to the C-terminus of pR by the pBAD-TOPO expression vector that we used [1]. There is undoubtedly a further post-translational modification of unknown nature.
Spectral comparisons
Béjà et al. reported an absorbance maximum for pR of 520 nm in E. coli membranes, using a difference bleaching technique to remove interfering absorbance bands from other membrane components in the impure pR sample [1]. We confirmed this result using crude E. coli membranes without detergent present (data not shown). However, we observed an absorbance maximum of 546 ± 5 nm for pR in OG at pH 7 at all stages of purification (Fig. 1). Small blue shifts were observed for pR samples in OG when measured at pH 8 and 9.5 (8 and 16 nm, respectively; data not shown).
For the pR samples reconstituted in DHPC, which were used for flash photolysis experiments, the absorbance maxima were similar to those measured in OG (spectra not shown). A chromophore absorption maximum near 540 nm was also obtained by using difference spectroscopy of pR in crude E. coli membranes solubilized in OG (spectrum not shown). However, for pR measured directly in crude E. coli membranes, i.e. not solubilized in OG, we obtained the same value (520 nm) as reported previously [1].
Solubilization in detergent presumably leads to structural distortions of the native protein conformation, and therefore a change in the absorbance properties of the chromophore. However, the direction that λmax for pR would have to shift upon solubilization in OG is inconsistent with the pattern for bR, whose λmax decreases when it is solubilized in OG [12]. Furthermore, pR in OG showed resonance Raman spectra (D. Dunmire, R. A. Krebs and M. S. Braiman, unpublished data) indicative of a chromophore structure very close to native light-adapted (i.e. all-trans) bR in purple membrane. However, there is one major difference: pR in OG exhibits an upshifted, doublet C=N Schiff base band consisting of two components of nearly-equal intensity. These appear to correspond to the presence of two distinct subpopulations of pR, at least when expressed in E. coli and solubilized in OG micelles. The different values of λmax for pR in membrane state [1] and OG solution might be related to the presence of these multiple subpopulations, but this connection remains unclear.
Principal photointermediates of pR
Of the six principal photointermediates present in bR, four can be discerned from the time-dependent visible absorbance traces from pR in fig. 4, along with previously published time traces at 600 nm [1]: the resting state (pR), M, N and O. The resting state, with an absorbance maximum of 546 nm (see above), provides the baseline spectrum for the difference time courses reported. The M intermediate of bR has an unprotonated Schiff base group, giving rise to a blue-shifted λmax (400 nm). Likewise in pR, an increase in the 400 nm absorbance should indicate formation of a deprotonated Schiff base, and therefore the presence of an M-like intermediate.
Interpretation of the 500 nm time course trace is more complicated. As in bR, it likely involves decay of M to N, as well as from N back to bR. The absorbance maximum of the N intermediate in bR is 560 nm, corresponding to a protonated Schiff base. This is not very different from λmax for the resting state of pR (546 nm). The likely spectral overlap between pR and its own postulated N photoproduct complicates the determination of amounts of each that are present. The slow (~500 ms) decay observed in both the 400 nm and 500 nm time courses indicates an equilibrium between the M and N intermediates that remains until pR returns to its initial resting state.
Béjà et al. reported a strong positive 580-nm transient absorbance increase in suspensions of membranes prepared from E. coli expressing pR [1]. This positive absorbance difference is indicative of O intermediate formation. We looked for its presence in partially-purified samples reconstituted in DHPC in the pH range 6–11. (Data are shown only at pH 9.5; see Fig. 4). Our observations at pH 9.5 do not show any evidence of O intermediate formation (Fig. 4). Only at lower pH values was a clear positive 580-nm absorption change observed (data not shown). This is in agreement with results on bR, for which O formation is also enhanced at lower pH values, and becomes small or nearly unobservable in the alkaline range.
Dependence of the M intermediate on pH
Deprotonation of the Schiff base linkage of the retinal and Lys-216 is dependent on its pKa which changes between photointermediate states. The Schiff base readily undergoes deprotonation in the M intermediate. However, no M intermediate formation occurs below a pH of ~9 (figure 3); instead the O intermediate predominates [1]. Predominance of O at lower pH values is also observed in bR. However, in bR the M intermediate is detectable at low and high pHs, but has a longer lifetime at higher pH due to a long-lived equilibrium between M and another intermediate, N [13]. It seems likely that an N intermediate of pR is similarly in equilibrium with its M intermediate, based on the fact that the transient positive (400 nm) absorbance increase is smaller than the negative 500 nm bleach, and the time course at 500 nm shows a partial return to baseline on a timescale of ~50 μs.
The difference in pH dependence between the pR and bR photocycles can likely be attributed to differences in the microenvironment of the Schiff base, and is perhaps related to absence of Glu194/204 in pR. These adaptive differences presumably optimize the proteins to operate at maximum efficiency in the niches that their respective organisms occupy. In the case of the λ-proteobacteria, which in the open ocean (pH 7.8–8.0) occupy a signficantly more alkaline environment than halobacteria, perhaps the proton-release group is simply not under any evolutionary pressure to be capable of deprotonating in the M state at neutral pH. In this view, the principal role of E194 and E204 in bR may to modulate the pKa of the H+-release group in the M state to a value lower than the pH of the organism's external environment.
Fast proton release in pR
Under the same conditions where M is observed (pH 9.5 and in 1% DHPC), pR undergoes fast proton release during its photocycle (Fig. 5). The pH indicator dye Cresol Red was used to detect pH changes in the bulk aqueous phase. These turn out to be similar to those observed for bR in the pH range 5.5–10. After photoexcitation, pR (like bR, presumably) ejects a H+ from a residue near its extracellular surface decreasing the pH of the solution. When the N → O transition takes place in bR, H+ is taken up from the medium, raising the pH once again. The H+ signals from pR measured with Cresol Red occur on a time scale similar to that assignable to M and N decay, returning to baseline about 1 s after photolysis.
There is a clear kinetic correlation between M (and/or N) intermediate formation and fast H+ release in pR. The linkage between these two phenomena is further supported by the observation that neither a transient 400-nm absorbance increase, nor fast H+ release, is shown to occur at pH 8.0 and below. Nor is either observed in the absence of a reconstituting lipid (DHPC in these experiments).
In bR, the ejected proton is thought to originate from a triad of amino acids, R82-E194-E204. However, in pR a homolog of only one of these three residues (the arginine) is present. This raises doubts about previous conclusions regarding the specific roles of these 3 residues in fast H+ proton release, in both pR and bR. In particular, the apparently obligatory roles of E204 and E194 in fast H+ release in bR are not matched in pR. Therefore, even in bR it is less likely that these groups themselves change protonation state between bR and M to provide the H+ released to the bulk medium. Instead, it now seems more likely that E204 and E194 merely help to lower the pKa of the H+ release group from above 8, the apparent value for pR, into the vicinity of 6 for bR. It also seems very unlikely that the specific structural configuration of 2 carboxylic acid groups and arginine in bR could be conserved in pR, even if, as suggested previously [1], other surface carboxylic acids in pR could substitute in some ways for the roles of E194 and E204 in bR.
The only way that the H+ release mechanism can be strongly conserved between bR and pR is if arginine itself serves as the principal donor group for fast H+ release in both, with nearby residues (such as E194 and E204) merely modulating the pKa of the arginine in the M intermediate. However, it remains unclear how the pKa of Arg-82 in bR could be made sufficiently low in its M intermediate to serve as the H+ release group at pH values down to 6.0.
Alternatively, it is possible that the fast H+ release we observe from pR at pH 9.5 may differ from that in bR. One possibility is that in pR, the released proton could come directly from the chromophore counterion, Asp-97. This would be consistent with a proposed mechanism for fast H+ release that has been observed above pH 10 in the bR mutant E194Q [14]. In this mutant, Asp-85 was detected by low-temperature infrared difference spectroscopy to be deprotonated only in the N intermediate, and not in M [15]. It is not clear yet whether Asp-85 deprotonation in an N-like state could account for the proton-release kinetics of pR (Fig. 5), or whether the M intermediate itself might have a partially-deprotonated Asp-85.
The reason for the requirement of DHPC in M intermediate formation and fast proton release is unclear. Delipidated bR in octylglucoside is fully capable of M formation and presumably proton release, although with altered kinetics [12,16]. The requirement for pR to be in lipid to show fast H+ release and M formation stems either from a protein/lipid interaction needed to establish a stable, active tertiary structure, or from the need for the phosphate group in DHPC to act as a proton release group. The latter seems unlikely due to the DHPC molecule being zwitterionic at pH 9.5, with no proton on the trimethyl-modified nitrogen of the choline. Hence, the DHPC most likely interacts with the protein to effect minor structural changes needed to place the active site residues in their functional configuration.
Conclusions
A comparison of the primary sequences of pR and bR at first glance seems as to preclude fast H+ release as part of the proton-pumping mechanism of pR due to the absence of residues analogous to Glu194 and Glu 204 of bR. However, fast H+ release is indeed observed in pR under conditions where an M intermediate is formed. Glu194 and Glu 204 in bR play a role in fast H+ release that is apparently not required for the mechanism of the bR family of proton transporters. It is therefore necessary to conclude that either the H+-release groups in pR and bR are non-homologous surface carboxylic acid residues (as suggested previously [1]), or else that a conserved non-carboxylic acid residue, i.e. Arg82/94 or Asp85/97, is the H+-release group in pR. The higher pH requirement for the M intermediate of pR presumably corresponds with adaptation to the more alkaline oceanic environment in which the γ-proteobacteria are found.
The necessity of reconstituting pR with some lipid before it is capable of photocycling shows that the presence of lipids facilitates pR in assuming its fully active structure. E.-coli-expressed pR has post-translational modifications, including ~4000 daltons of substituents at one or more of its three cysteines. Such post-translational modifications might also play a role in explaining the different physiological properties of pR and bR.
Materials & methodsProtein expression and detergent extraction
Proteorhodopsin was expressed by E. coli strain UT5600 containing an additional plasmid encoding for the first-reported pR gene (accession #AF279106, obtained from the uncultured proteobacterium EBAC31A08 clone BAC [1]) with an Ara promoter and ampicillin resistance (kindly provided by O. Béjà). Single colonies were selected and grown overnight in LB/amp media (200 ml, 37°C, 300 rpm). This culture was then diluted 10× into several 500-ml cultures. After a further 2 h incubation in the shaker bath, a stock solution of 20% L-arabinose was added, to give a final concentration of 0.2% L-arabinose. This culture was then incubated for 4 h (37°C, 300 rpm). The cells (~20 ml wet volume) were then collected by centrifugation (6000 rpm × 30 min) and washed 3× with 100 mM HEPES, pH 7.1 (buffer A). The cells were then resuspended in buffer A and incubated at 4°C with 50 μg all-trans-retinal (added as a concentrated ethanol solution) for 3 h. The cells were collected by centrifugation (6000 rpm × 30 min), then resuspended in 60 mL buffer A containing 0.3 mg/mL lysozyme, and stirred for 4 h at room temperature. The cells were again collected by centrifugation (6000 rpm × 30 min), then lysed with 20 ml of 20% sodium cholate, pH 7.1 (30 min., 4°C). The cells were centrifuged again (6000 rpm × 30 min), and the supernatants collected. After extracting 3× more with the same cholate solution, the pooled supernatants were diluted 10× with buffer A and centrifuged at 180,000 g for 45 min to collect the membrane pellet. This cholate-washed membrane pellet was then further extracted 3× with 3.0% β-octyl-D-glucoside (OG) in buffer A (30 min, with stirring, 4°C). The pooled supernatants, containing OG-solubilized pR (~15 mg), were then diluted 6× with buffer A.
Column purification
The diluted OG-solubilized membrane extract (10 mg in 300 mL total volume of 0.5% OG) was loaded on a 25 × 1 cm column containing Phenylsepharose™ (6 fast flow high sub; Amersham Pharmacia Biotech). The column was eluted with a 0.5%-2.0% OG gradient in buffer A (300 mL total volume, 0.5 mL flow rate). The pR eluted at an OG concentration of 1.5–2.0%. Fractions having an A280/A546 ratio of 4.0 or lower were pooled (9.5 mg pR recovered in all) and concentrated using Vivaspin™ 20 concentrators having a 5000 MW cutoff (Vivascience, Westford, MA). A portion of the Phenylsepharose™-purified pR (1.5 mg) was diluted to an OG concentration of 0.5% with 0.5 M KCl, 100 mM acetate. It was then loaded on a 10 cm × 1 cm hydroxylapatite (BioGel HTP, BioRad) column and eluted under pressure with a 0–600 mM phosphate gradient (200 mL total volume, flow rate 0.5 ml/min). Fractions with an A280/A546 ratio of 2.5 or lower were pooled and concentrated for subsequent experiments (0.5 mg).
Mutagenesis
Methodology for the site-directed mutagenesis of pR is discussed in detail elsewhere (R. Parthasarathy, T. Caterino, R.A. Krebs, M.S. Braiman, manuscript in preparation). The triple cysteine mutant (pR-TCM) has all three of its native cysteines (Cys-107, Cys-156, and Cys-175) replaced with serines, and was prepared using the same E. coli expression system and purification methods as the wild type.
Polyacrylamide gel electrophoresis
A 12% discontinuous SDS/polyacrylamide gel was used for molecular weight and purity analysis [17].
Flash photolysis
Time-resolved UV/vis spectroscopy methods were as described previously [18]. A Phenylsepharose™-purified pR sample was reconstituted into mixed micelles containing 1,2-diheptanoyl-SN-glycero-3-phosphocholine (DHPC), by adding a 1% solution of the short-chain lipid and then removing most of the detergent on a Sephadex G-25 column equilibrated with 1% DHPC in 100 mM NaCl. Proton release and uptake in the aqueous bulk medium were detected from the pR-containing micelles suspended in 1% DHPC, 100 mM NaCl, with 45 μM Cresol Red pH indicator dye. Flash-induced absorbance changes at 580 nm of samples with and without the Cresol Red were subtracted to determine the transient signals due to proton concentration changes. Photoexcitations were performed with 10-ns laser pulses of 3–6 mJ at 500 nm. The time courses in Figs. 3,4 are an average of 40 cycles with the exception of the Cresol Red experiments averaging 100 cycles (Fig. 5, bottom trace) [19-21].
This work was supported by Syracuse University, and by a grant of the Deutsche Forschungsgemeinschaft (Sfb 449-TPA5 to U. Alexiev and M. P. Heyn). A.-M. DeVita was supported by an NSF-REU award to Syracuse University Chemistry Department.
Previous work by our group and others has implicated a role for kinins in the ovulatory process. The purpose of the present study was to elucidate whether endogenous progesterone, which is an intraovarian regulator of ovulation, might be responsible for induction of the kinin system in the ovary during ovulation. The gonadotropin-primed immature rat was used as the experimental model, and the role of endogenous progesterone was explored using the antiprogestin, RU486.
Results
The results of the study revealed that RU486 treatment, as expected, significantly attenuated ovulation. Activity of the kinin-generating enzyme, kallikrein, was elevated in the ovary in control animals prior to ovulation with peak values observed at 4 h post hCG, only to fall to low levels at 10 h, with a recovery at 20 h post hCG. RU486 treatment had no significant effect on ovarian kallikrein activity as compared to the control group. Total ovarian kininogen levels in control animals increased significantly at 12–14 h after hCG – coinciding with initiation of ovulation. Thereafter, ovarian kininogen levels fell to low levels at 20 h, only to show a rebound from 24–38 h post-hCG. RU486 treatment had no significant effect on the rise of total ovarian kininogen levels from 12–14 h after hCG; however, from 30–40 h post hCG, RU486-treated animals had significantly higher total ovarian kininogen levels versus control animals, suggesting that endogenous progesterone may act to restrain elevations of kininogens in the post-ovulatory ovary. This robust elevation of ovarian kininogen levels by RU486 was found to be primarily due to an increase in T-kininogen, which is a potent cysteine protease inhibitor.
Conclusions
Taken as a whole, these results suggest that endogenous progesterone does not regulate kallikrein activity or kininogens prior to ovulation, but may provide a restraining effect on T-kininogen levels in the post-ovulatory ovary.
Background
The kininogen-kallikrein-kinin system is well established to be important in inflammatory processes due to its actions to induce vasodilation, prostaglandin biosynthesis, and tissue remodeling through regulation of proteases [1-3]. There are two types of kinins, B-kinin and T-kinin, which are products of cleavage of the substrates B-kininogen and T-kininogen, respectively [[1,2], for review]. The enzyme kallikrein cleaves B-kininogen to B-kinin, while the enzyme T-kininogenase cleaves T-kininogen to T-kinin [3,4]. B-kinin and T-kinin have similar vasodilation and prostaglandin regulatory actions, and are metabolized by the enzyme, angiotensin-converting enzyme [1,2]. T-kininogen, in addition to being a substrate for T-kinin production, is a potent cysteine protease inhibitor and has a role in tissue remodeling [5,6].
Due to its role in inflammatory processes and tissue remodeling functions, a number of investigators have focused on the kinin system as a possible important mediator in the ovulatory process. Along these lines, our group and others have demonstrated that kallikrein activity and kininogen levels in the ovary increase preceding and at the time of ovulation in the gonadotropin-primed immature rat [7-9]. A functional role for this activation of the kinin system was suggested by the fact that kinin agonists have been shown to induce ovulation, while conversely; kinin antagonists inhibit ovulation [10,11]. Additionally, work by Holland et. al. has shown that kallikrein (rKLK-1) gene expression increases in the ovary preceding ovulation [9]. While evidence is mounting supporting a role for the kinin system in ovulation, the precise regulators of this system in the ovary remains unclear. In our previous study, we noticed that the elevation of kallikrein activity and kininogens in the ovary preceding and during ovulation was paralleled by an increase in ovarian progesterone concentrations [7]. Since progesterone has been demonstrated to be an important intraovarian regulator of the ovulatory process [12,13], the aim of the present study was to determine whether endogenous progesterone is responsible for activation of the kinin system in the ovary during ovulation. The potent antiprogestin compound, RU486 was used to accomplish this aim.
Results
As shown in Figure 1, PMSG-hCG induced ovulation in control rats with peak ova released from the ovary from 24 h to 34 h after hCG administration. RU486 treatment significantly inhibited ovulation at 20 h and 30 h (p < 0.03) versus controls. Ovulation also appeared lower in RU486 -treated rats at 24 and 34 h after hCG as compared to controls; however, due to variability this effect was not significant at these time points. Ovarian estradiol levels in control rats were high at 0 h only to fall to basal levels from 10 h to 38 h (Fig. 2). RU486 treatment caused a slight but significant elevation in ovarian estradiol levels at 30 h compared to control rats (p < 0.05). Nevertheless, even at this time point, ovarian estradiol levels were close to basal levels. Ovarian progesterone levels rose rapidly to reach a peak at 12 h in control animals, followed by a fall at 20 h to 38 h to low levels which, however, were still higher than the initial 0 h levels (Fig 2). RU486-treated rats had significantly attenuated peak progesterone levels at 12 h compared to control animals (p < 0.01). Figure 3 demonstrates that ovarian kallikrein activity in control rats was elevated prior to ovulation with peak values at 4 h, only to fall to low levels at 10 h, with a recovery at 20 h. RU486 treatment had no significant effect on ovarian kallikrein activity as compared to controls. As shown in Figure 4, ovarian total kininogen levels in control animals rose in a gradual manner from 0 h to 12 h to reach peak levels at 12 h, a time that coincides with the beginning of ovulation. Kininogen levels then fell to low levels at 20 h only to show a rebound from 24 to 38 h. RU486 treatment had no effect on the rise of ovarian total kininogen levels from 0 to 10 h; however, from 30–38 h RU486-treated rats had significantly elevated ovarian total kininogen levels compared to control rats (Fig 4). As illustrated in Figure 5, T-kininogen is the major type of kininogen in the postovulatory ovary and the elevation of total kininogen levels by RU486 from 30–38 h was primarily due to an elevation of T-kininogen levels at all time points. B-kininogen levels, on the other hand, did show a small but significant decrease at 30 h, followed by a significant increase at 34 h in RU486-treated rats versus controls.
Effect of the antiprogestin RU486 upon ovulation rate in PMSG-hCG-primed immature rats. Twenty-three-day-old female rats were primed with 10 IU of PMSG and 48 h later ovulation was induced with hCG (10 IU). Either vehicle or RU486 (10 mg/kg) was administered 30 minutes prior to hCG. n = 6 rats per group. ** p < 0.03 vs. vehicle.
Effect of RU486 on ovarian steroid levels in the PMSG-hCG-primed immature rat. The model is the same as described in Figure 1. * p < 0.05 vs. Vehicle; ** P < 0.01 vs. Vehicle.
Effect of RU486 on ovarian kallikrein activity in the PMSG-hCG-primed immature rat. The model is the same as described in Figure 1. There is no statistical difference between the groups.
Effect of RU486 on total ovarian kininogen levels in the PMSG-hCG-primed immature rat. The model is the same as described in Figure 1. * p < 0.05 vs. Vehicle; ** p < 0.01 vs. Vehicle.
Species of kininogens in the ovary after RU486 treatment in the PMSG-hCG-primed immature rat. The model is the same as described in Figure 1. *p < 0.05 vs. Vehicle; ** p < 0.01 vs. Vehicle.
Discussion
The present study demonstrates that treatment with the antiprogestin compound, RU486, significantly attenuates ovulation in PMSG-hCG-treated immature rats. This finding is in agreement with previous reports in the literature and is consistent with the hypothesis that progesterone has a direct role in ovulation [12,14]. Along these lines, the progesterone receptor knockout mouse has been shown to develop follicles to the ovulatory stage, but follicle rupture does not occur in these animals despite administration of a superovulatory dose of exogenous gonadotropins [12]. The precise genes/proteins regulated by progesterone in the ovary to modulate ovulation remains unclear. This study investigated the possible role that endogenous progesterone may have in regulation of the kinin system, which has been implicated as a mediator of the ovulatory process. Several interesting observations were yielded by the study. First, even though ovulation was lowered by RU486 treatment, activation of the kinin system as reflected by the elevation in ovarian kallikrein activity and kininogen levels prior to and during ovulation was not significantly affected. This suggests that progesterone may not be responsible for the elevation in kinin system activity in the ovary observed prior to and at the time of ovulation, and that progesterone's modulatory influence in the ovulatory process may be mediated through some system other than the kinin system. Our finding of a lack of significant effect of RU486 on ovarian kallikrein activity during ovulation is somewhat at variance with the report of Tanaka et al. [15]. These investigators suggested a role for progesterone in the regulation of ovarian kallikrein activity based on studies using epostane, a progesterone synthesis inhibitor. In these studies, epostane treatment inhibited ovulation and ovarian kallikrein activity. It should be pointed out however that epostane, in addition to inhibiting progesterone synthesis, has also been shown to strongly inhibit synthesis of 17β-estradiol, testosterone and 4-androstene-3,17-dione in the gonadotropin-stimulated immature rat ovary, so it is not truly specific for inhibition of progesterone synthesis [16]. It should also be pointed out that the temporal pattern for elevation of kallikrein activity differs between the two groups. We see an early peak elevation from 0–4 h followed by a fall to low levels at 10 h and another smaller increase at 10–12 h. Tanaka et. al. see no early increase and their peak levels occur at 12–14 h post-hCG. These differences may reflect different assay specificities as the chromogenic substrates used in their study (D-Val-Leu-Arg-paranitroanilide) and ours (Pro-Phe-Arg-methyl-coumarylamide) are different. The substrate differences could be important because different kallikreins have been shown to have different substrate specificity [17]. The rat kallikrein gene family consists of 13 genes, of which six are expressed in the ovary (rKLK1, rKLK3, rKLK7, rKLK8, rKLK9 and rKLK12) [9]. In an attempt to explain the disparate temporal pattern between the two groups, Holland et. al. examined the gene expression pattern of all six kallikrein genes in the ovary in the same gonadotropin-primed immature rat model used by both groups. They found that only one gene is increased prior to ovulation, rKLK1 or true kallikrein [9]. They also noted that the pattern of gene expression for rKLK1 mirrored the temporal pattern for kallikrein activity observed in our study rather than that reported by Tanaka et. al. (e.g. it was elevated from 0–2 h after hCG) [9]. This led the authors to suggest that our substrate, Pro-Phe-Arg-methyl-coumarylamide is a more specific substrate and is detecting true rKLK1 enzyme activity. We cannot rule out the possibility that the activity of another kallikrein isoform, which is more specifically recognized by the substrate used by Tanaka et. al., changes after hCG, which could explain the temporal pattern observed by Tanaka et. al. and the apparent progesterone regulation they observed. However, Holland et. al. did not find any other kallikrein gene elevation prior to ovulation which would support this possibility [9]. Finally, we did not assess for changes in one other component of the kinin system, the bradykinin receptor. The bradykinin B2 receptor has been demonstrated to be localized in the ovary in theca and granulosa cells [18], and thus regulation by progesterone at this level of the kinin signaling system could be possible and cannot be excluded.
A second interesting observation from our study was that in the post-ovulatory ovary, endogenous progesterone might actually be acting to restrain kininogen elevations. This suggestion is based on the finding that after ovulation, RU486-treated rats exhibited a significant robust elevation of total ovarian kininogen levels as compared to control animals. This elevation was principally due to a RU486-induced increase in T-kininogen levels in the ovary at 30–38 h post hCG. While the function of T-kininogen in the post-ovulatory ovary is not known, T-kininogen can be cleaved to yield the vasodilatory compound, T-kinin, and in its uncleaved state, it is a potent cysteine protease inhibitor [1,2,5,6]. An important class of cysteine proteases in the ovary that T-kininogen may regulate is cathepsins. Cathepsins degrade type I and IV collagen, fibronectin and laminin, and have been implicated to play a role in stimulation of steroidogenesis and degradation of extracellular matrix – events that would be occurring in the post-ovulatory ovary as the corpus luteum is forming [13,17-20]. In support of a possible cathepsin cysteine protease regulatory function for T-kininogen in the ovary, we previously reported that a strong inverse correlation exists between T-kininogen levels and cathepsin activity in the gonadotropin-stimulated immature rat ovary (e.g. as T-kininogen levels increased, cathepsin activity decreased) [7]. We believe that T-kininogen is primarily serving a cysteine protease regulatory function in the ovary, as we have been unable to detect T-kininogenase, the enzyme that cleaves T-kininogen to T-kinin, in the ovary [7].
Conclusions
In conclusion, the current results do not provide support for a role for endogenous progesterone in activation of the kallikrein-kinin system in the ovary preceding ovulation. Thus, progesterone modulation of ovulation most likely is mediated by a system other than the kinin system. Intriguingly, the results suggest that endogenous progesterone may actually be more important in regulating the kinin system in the post-ovulatory ovary. Of significant interest, this post-ovulatory regulation was not observed on kallikrein activity and only minor effects were observed on B-kininogen, which together represent the classical kinin pathway. Rather, the effect was observed specifically on the newest member of the kinin/kininogen family, T-kininogen. Since we have been unable to detect T-kininogenase activity in the ovary, we believe that T-kininogen is functioning primarily as a cysteine protease inhibitor in the ovary. Thus, endogenous progesterone restrainment of T-kininogen in the post-ovulatory ovary could be a mechanism for modulation of cysteine protease activity in the post-ovulatory ovary, thereby facilitating the extensive tissue remodeling that is known to occur after ovulation. Additional studies are underway in our laboratory to further explore this possibility.
Materials and MethodsAnimals
Immature 23-day-old female Sprague Dawley rats (Holtzman, Madison, WI) were given PMSG at 0800 h. Forty-eight hours later, the animals were given hCG (10 IU, s.c. in saline; Sigma, St. Louis, MO) to induce ovulation. Thirty minutes prior to hCG injection, either vehicle or the antiprogestin RU486 (10 mg/kg in ethylene glycol ip; Roussel UCLAF, Romanville, France) was administered. In the PMSG-hCG induction model, mature follicles first begin to ovulate at 12–15 h, with peak number of ova appearing in the oviducts at 24–30 h [7]. To access the role of the kinin system in the ovulatory process, groups of animals were killed at various time points after hCG treatment; the ovaries were removed to determine ovulation and to measure the various components of the kininogen-kallikrein-kinin system. After removal, the ovaries were cleaned of fat, snap frozen in liquid nitrogen, and stored at -70°C until the various assays were performed. All protocols involving animal use in this study were approved by our Institutional committee for the care and use of animals in research (CAURE).
Determination of ovulation
To determine ovulation, the animals were killed by decapitation and the oviducts were removed, pressed between two microscope slides, and examined under a microscope for the presence of ova as described previously [7].
Preparation of Tissue Homogenate
Ovaries excised from each animal were homogenized using a Dyna-Mix (Fisher Scientific, Pittsburg, PA) in 400 μl of 50 mM ice-cold PBS, pH 7.4. The homogenates were centrifuged at 4°C at 20,000 × g for 20 min. The supernatants were then assayed as described below.
Determination of Kininogens
Measurements of kininogen levels were performed as described previously by our laboratory using the protocols described by Barlas et al. [7,21]. Total kininogens. Plasma (100 μl) or tissue homogenate (100 μl) was incubated with 900 μl of 0.03 N HCl at 37°C for 15 min to destroy kininase activity. After neutralization with 25 μl of 1 M NaOH, 25 μl of 5 mg/ml trypsin was added (final concentration of trypsin was 1 mg/ml) and incubated at 37°C for 1 h to liberate kinins from the kininogens. The reaction was stopped by heating at 100°C for 10 min. The liberated kinins were quantified according to the RIA method of Greenbaum and Okamoto [2]. An aliquot of the sample (100 μl) was incubated at 4°C with 400 μl of 125I-labeled bradykinin (BK) and 100 μl of BK polycolonal antibodies, which were raised from the rabbit and able to recognize all kinins. After 2–24 h, 1% of bovine gamma globulin and 25% polyethylene glycol were added, respectively, to separate bound and free kinins. Radioactivity was counted by a gamma counter. Total kininogen was expressed as μg/mg protein of BK equivalents.
T-kininogen
After trypsin digestion, plasma was treated with an equal volume of 30 % trifluoroacetic acid (TFA). The supernatant, after centrifugation for 10 min at 3,000 rpm, was applied to a C-18 extraction column previously primed with 5 ml methanol and 5 ml of 1% TFA. After an initial wash with 1% TFA, the kinins were eluted with 1.5 ml of 50% acetonitrile in 1 % TFA. The elute was evaporated to dryness on a Speed Vac centrifuge, redissolved in 40 μl of distilled water, and applied to an HPLC column equilibrated with a mixture of 18% acetonitrile and 82% triethylammonium formate (v/v), pH 4.0. Fractions (14 ml/2 min) were collected and subjected to RIA. T-kinin (μg kinin equiv./mg protein) = [Total kinins released by trypsin] × % T-kinin in total kinin eluted. BK-kininogen was estimated by subtracting the amounts of T-kinin released from T-kininogen by trypsin from the amount of kinins released for total kininogen.
Assay of Kallikrein
Kallikrein activity was measured as described previously by our laboratory [7,22]. Briefly, tissue homogenate was incubated with a substrate, Pro-Phe-Arg-methylcoumarylamide (MCA), in 0.1 M Tris-HCl buffer containing 0.15 M NaCl., pH 8.0. The initial hydrolysis of the substrate for the first 5 min is recorded. One unit of enzyme activity of kallikrein releases amino methylcoumarin (AMC) from the substrate at a rate of 1 nM/5 min. The hydrolysis of AMC was measured using a fluorescence spectrophotometer with excitation at 370 nm and emission at 460 nm.
Statistical Analysis
The results given in the text are expressed as means ± SEM. Six rats were used per group and the experiments were repeated three times. The differences between experimental groups were analyzed using the student t-test; p < 0.05 was considered significant.
Authors' Contributions
DWB conceived of the study and participated in its design, coordination, and analysis, and in the treatment of animals with hCG and measurement of ovulation. LMG participated in the measurement of kallikrein activity and kininogens and the interpretation of data. VBM participated in the conceptualization and interpretation of data and in the measurement of steroid levels. XXG participated in the design of the study, interpretation and analysis of data, measurement of kallikrein and T-kininogenase activity and kininogen levels.
Acknowledgments
We are grateful to Mrs. Jannie J. Jones for skillful technical assistance. This work was supported by a research grant (11-16-04-3013-66) from the Medical College of Georgia Research Institute.
With the invention of the ion-selective electrode (ISE), ionic magnesium (iMg) is a common blood assay. This could be advantageous, as iMg is the biologically active form of Mg. There is some evidence that iMg has considerable within subject variability.
Results
Individual ranges averaged .08 mmol/L (range .05 to .14). Coefficients of variation (CV) ranged from 3% to 7% (mean 4%) while analytical variation was determined to be 2.3%. Biological variability thus accounts for almost half of the variability, which is clinically significant, as 9 of the 13 subjects recorded at least one value below a reference range of .46 – .60 mmol/L. A significant within-day variation (p < .001) was noted, with differences between 7:00 and 10:00 as well as 10:00 and 22:00. Between day variations were not significant (p = .56).
Conclusions
A plausible explanation of this data is that iMg has a circadian rhythm. Thus, cautious interpretation of single iMg values is warranted until future research determines the nature of iMg variability.
Background
Advances in ionic sensitive electrode (ISE) technology has allowed ionic magnesium (iMg) to become a common measure of Mg status. However, some evidence has shown that iMg may be quite variable and thus not a reliable measure. The primary purpose of this study was to analyze the variability of iMg to determine if it is a reliable measure. If iMg is variable, a secondary purpose was to determine if a diurnal pattern exists over the course of three days.
The rationale for studying the variability of iMg lies in the physiological importance of magnesium, the prevalence of Mg deficiencies and the emergence of iMg as a potentially sensitive measure of Mg status. The usefulness of iMg in both clinical or research settings hinges on the establishment of the assay's variability.
Magnesium is the second most abundant cation found in the intracellular fluid and the fourth most abundant cation in the extracellular fluid [1]. It is a cofactor for more than 325 enzymatic reactions including adenosine triphosphate (ATP) metabolism, glucose utilization, muscle contraction and synthesis of fat, protein, and nucleic acids [2]. It is also involved in intermediary metabolism, neuromuscular activity, secretion, excitation-secretion coupling, cardiovascular health and bone metabolism [3].
A number of disease states have been associated with magnesium imbalances and include: cardiovascular diseases, neuromuscular disorders, higher mortality rates [4], renal diseases, drug toxicities, asthma [5], migraines, premenstrual syndrome, pre-eclampsia, eclampsia, menopausal bone problems [3], atherosclerosis, diabetes mellitus, obesity [6], and hypertension [7]. As well, Lukaski [8] has noted a decrease in athletic performance as a result of magnesium deficiency/imbalances.
Low dietary intakes of Mg may account for low Mg status in many individuals. Altura [9] found that dietary intakes of magnesium in the United States have been declining since the turn of the century from about 500 mg/day to 175–225 mg/day. According to the National Research Council of Canada, [10], this is due to the increasing use of fertilizers (lacking Mg) and food processing (removing Mg). Some investigators believe that the current RDA of 350 mg/day for men and 300 mg/day for women as recommended by the US National Academy of Sciences is too low and should be 450–500 mg/day [11].
According to Djurhuus et al., [6] there is no consensus regarding measurement of magnesium. Although muscle Mg, obtained through a needle biopsy, is thought to be reliable, it is time consuming to perform, very invasive and causes discomfort to the patient. Magnesium status can also be measured in the serum, erythrocytes, and lymphocytes or through a magnesium load test with urinary excretion [6]. However, Djurhuus et al., [6] have found that urinary Mg is quite variable, so generally it cannot be used to evaluate Mg status. The total amount of Mg in serum (TMg) is the most common means for measuring magnesium status. Although TMg has been the most common measure for magnesium status, since the introduction of the ISE for Mg, ionic Mg has been widely used and may become the new standard.
There is approximately 1000 mmol of magnesium in the human body [12], with muscle and bone comprising approximately 80% of total body Mg [6]. The serum portion of blood contains less than 1 percent, yet is the most accessible source for Mg measurement. Serum Mg can further be subdivided into its component parts: ionic, complex-bound, and protein-bound. It is the free (ionic) portion, however, that is most important because it is physiologically active [13]. Ionic magnesium levels have been found by Altura & Altura, [5] to be altered in some disease states and should prove to be of importance in disease management.
Despite the clinical advantages that iMg has to offer, its usefulness may be compromised by physiological variability. Studies have shown a circadian rhythm associated with TMg as well as iMg. In 1978, Touitou et al.[12] found a significant circadian rhythm in TMg. As well, Willimzig, Latz, Vierling & Mutschler [14] found a noticeable circadian fluctuation of TMg with a peak in evening hours and strong fluctuations in the morning. Ising, Bertschat, Gunther, Jeremias & Jeremias [15] were the first to discover a significant circadian rhythm in iMg and observed the highest concentrations around 9:00 and the lowest concentrations around 15:00. A further study relating to the discoveries by Ising et al., [15] was conducted by Jacomella et al., [16] to see if glucose loading affected the circadian rhythm of iMg and found no significant results.
Although assessment of iMg would appear to have obvious clinical and research implications in the screening, monitoring, diagnosis and treatment of individuals, determination of an intra-individual variability would confound appropriate interpretation of iMg values. This variability has only been hinted at in previous research [12,14-17]. Knowledge of these changes over time is vital to the collection of specimens at appropriate times, selection of relevant reference values, and in diagnosis, because the absence of the expected rhythm may indicate the presence of disease [18]. This research thus represents a critical step in the validation of iMg measurements. If iMg proves to be variable, the nature and physiological mechanisms underlying the variance would need to be determined.
Results
Participants' self-report journals revealed that no unforeseen circumstances occurred during the three days of testing. All subjects maintained normal routines of exercise and lifestyle (work, diet, and sleep), with no unexpected changes in their daily routines (other than the six blood samples per day).
The mean dietary magnesium intakes were 378.2 +/- 157.7 mg/day. The dietary analyses showed that four subjects (two males and two females) had intakes less than the RDA of 350 mg/day for men and 300 mg/day for women, with both above mentioned male subjects having less than 70% of the RDA. Values over 70% of the RDA are in the acceptable range for adequate nutrition [8].
Mean iMg values and ranges are presented in Figure 1. Nine of the thirteen subjects (69.2%) recorded at least one value below the reference range (0.46 – 0.60 mmol/L) suggested by Ising et al., [15]. The highest value recorded (0.56 mmol/L) was reached by two subjects.
Means and ranges of iMg in 13 healthy individuals
The CVt, CVa, CVi and index of individuality ratios are indicated in Table 2. The total variability had a mean of 4.3% and ranged from 3.1 to 7.2%. The estimated analytical variability was 2.3%. The mean physiologic or "true" biological variance of iMg was thus a CVi of 2.0%. The average index of individuality was .88.
Partitioning of the variability and index of individuality
Subject
CVt
CVa
CVi
CVt/CVg
1
7.2
2.3
4.9
1.47
2
3.8
2.3
1.5
.77
3
4.1
2.3
1.8
.84
4
3.2
2.3
0.9
.65
5
4.1
2.3
1.8
.83
6
4.8
2.3
2.5
.97
7
3.1
2.3
0.8
.63
8
4.8
2.3
2.5
.97
9
4.4
2.3
2.1
.90
10
3.7
2.3
1.4
.76
11
3.8
2.3
1.5
.78
12
5.6
2.3
3.3
1.14
13
3.6
2.3
1.3
.73
MEAN
4.3
2.3
2.0
.88
*Coefficient of Variability (CV) for total (t), analytical (a) and individual (i) variability expressed in percentages. An index of individuality (CVt/CVg) where CVg is the between-subject variability is also shown. CVg was determined to be 4.9%.
A representative subject's iMg values over the three days are shown in Figure 2. Initial plotting of all of the subject's iMg data revealed the possibility of an inherent rhythm to the data. Mean values for the six time periods of data collection as well as the corresponding coefficients of variation (CV) are shown in Table 3.
Typical participant data
Mean iMg Values and Coefficients of Variation for Six Time Periods
T1 (7:00)
T2 (10:00)
T3 (13:00)
T4 (16:00)
T5 (19:00)
T6 (22:00)
iMg (mmol/L)
0.50
0.48
0.48
0.49
0.49
0.50
CV (%)
3.9
5.2
5.0
5.3
4.8
5.2
* All time periods are within a range of +/- 40 minutes.
Within-day and between-day variability
A 3 (Day: 1,2,3) by 6 (Time: 7:00, 10:00, 13:00, 16:00, 19:00, 22:00) repeated measures analysis of variance (ANOVA) was calculated on the iMg values. There was a significant main effect for time (F(5,2) = 6.71, p < 0.001). The main effect for day was not significant (F(2,5) = 2.07, p = 0.142). However, there was a significant interaction (F = 6.71, p < 0.001).
As noted, plotting the data across time points in a given day for a given subject indicated that there was an inherent rhythm for the measured variable. Therefore data was transformed and slope scores were used in the subsequent ANOVA. A non-significant result was obtained for the between-day variation (F(2,12) = 0.915, p = 0.557).
Discussion
The key finding of this study is that iMg is not variable from day to day, yet is variable in healthy subjects over the course of one day. The within-day variability is illuminated in the descriptive statistics, a partitioning of variability components (Table 2) and the repeated-measures ANOVA statistic. In figure #1 for example, participant #1 exhibited a range in iMg from .42 to .56 mmol/L. The .42 value, recorded at 10:00 AM on day 3 would put her in a hypomagnesemic state, whereas the .56 mmol/L value, recorded at 10:00 PM on day 1 places her in the upper normal range. Less drastic ranges were observed in the other subjects but nevertheless, the clinical implications in basing a diagnosis or treatment on a single measurement are significant. Another problem is that the diurnal rhythms noted are somewhat variable from one individual to another, so that no particular time represents a universal maximum or minimum.
Partitioning of variability
It is essential to obtain separate estimates of the individual (CVi) and analytical (CVa) components of variability that make up the total (CVt). Analytical variability was determined to be a CV of 2.3%. This value exceeds the criterion noted by Fraser and Harris [18], which states that the maximum allowable analytical variation should be less than or equal to half the average within-subject variation. The average within subject variability was 2.0%. On the other hand a CV of 2.3% is less than a 3% value considered acceptable by Nova Biomedical®. While analytical variability may be pushing or beyond acceptable limits, 47% (or 2.0/4.3) of the total variability of iMg is due to "true" physiological variance.
Within-day and between-day variability
Observing a within-day variability provides essential information for clinical interpretation of iMg. The data indicate that there are specific peaks of iMg concentration as well as troughs, and that these fluctuations follow a diurnal rhythm. The discrepancy between consecutive blood samples (7:00 and 10:00) indicates that iMg may not be a suitable means of diagnosing Mg status. A patient in a clinic may show a normal ionic magnesium level if tested at 7:00 but could be deficient if tested again at 10:00. More research in this area is needed for a better understanding of within-day iMg variability. The findings of the present study are consistent with Ising et al., [15] who found a maximum value in the morning and a minimum value later in the day.
Coefficients of variation (CV) were calculated for the combined set of blood samples taken during each time period which showed that the lowest variation (3.9%) occurred during the first blood sample (7:00), while the second, fourth and sixth blood samples (10:00, 16:00, and 22:00) had the highest CV at 5.2%, which suggests that the best time for blood collection is first thing in the morning. These results are in agreement with Fraser & Harris [18] who reported that the ideal method for specimen sampling is to collect the specimen from fasting, non-exercised subjects between 7:00 and 9:00.
A non-significant between-day variability is important to the utility of iMg as a reliable means of measuring Mg status, as it shows that multiple samples when converted to sinusoidal data for each day for three consecutive days will not significantly vary. This will add confidence to the clinician when measuring iMg in hospitalised patients, as stability can be monitored successfully from day to day, or more importantly when values are changing daily and are no longer stable. It should be kept in mind though that the total duration of this study was only three days. Many analytes show cyclical rhythms, which can be circadian, monthly, or seasonal in nature [17].
Circadian rhythm
There are a number of reasons that could explain the discovered within-day variability, including analytical error or biological variability perhaps due to exercise, diet, stress or circadian rhythm. Speculation on causal factors that could explain iMg variability must be guarded as neither the purpose nor design of this study was geared toward causality. The possibility of a circadian rhythm has been found in previous research [15] and could explain the biological variability found in the subjects in this study. A preliminary examination of plotted individual data showed an inherent rhythm to the data. However, when all data was pooled together and smoothed to a sinusoidal curve of best fit, the data no longer showed a rhythm to it. Therefore, a circadian rhythm may be evident in each individual, which may explain the significant within-day variability. Also, as mentioned above, exercise and diet were not controlled throughout the study, which may influence the results. The fact that biological variability of iMg does exist but many biological causes are not controlled for in a clinical setting, points to the pressing need for more research along these lines.
Index of Individuality
In order to assess the utility of iMg values, an index of individuality was calculated. When the index of individuality is expressed as CVt/CVg, is less than 0.6, conventional population-based reference values are of very limited diagnostic value [18]. On the other hand though, a low index of individuality means that the indice being measured could find value in the tracking of a disease progression or the effectiveness of the treatment. In contrast, when CVt/CVg is more than 1.4, observed values can be compared usefully with reference values [18]. With an average ratio of .88, neither conclusion can be drawn. In other words caution is warranted in comparing iMg values to reference ranges and in tracking an individual's values.
The means and ranges of iMg (Figure 1) indicate that the reference range suggested by Ising et al., [15] is not useful with respect to each individual. Each subject's range falls on the low end of the spectrum, yet compared to other population-based reference ranges, this is one of the ranges with lower values. The individual ranges would fall completely out of some other population-based ranges. Therefore, the need for subject-based reference intervals may be necessary.
Magnesium status
Although measurement of the variability of iMg was the focus of the study, Mg status is also very important, as iMg is a means for determining total body magnesium status. Using the reference range by Ising et al. [15] (0.46 – 0.60 mmol/L) and the values obtained from the present study, one is able to determine the status of each subject. The mean values for each subject (18 samples) revealed that one subject rests on the border of Mg deficiency (O.46 mmoI/L) while all other subjects have normal mean iMg values but are on the low end of the range (highest individual mean is O.52 mmol/L). Throughout the three days of testing, however, nine of the subjects recorded one or more value below the reference range and would be considered hypomagnesemic. The results of the diet analyses revealed that only two subjects had intakes of dietary Mg below 70% of the RDA, thus suggesting that diet may not be the cause of the low iMg values. One explanation could be that the NOVA® instrument gives low values (compared to other instruments). This suggestion can be discounted though for testing of control samples of known iMg concentrations were always within acceptable ranges. A more likely explanation is that most published reference ranges for iMg are inappropriate, since all subjects were healthy individuals.
Greenway et al. [4] proposed that to use iMg as a valid biochemical marker, it is essential to establish a reliable reference interval in a healthy population. Close examination of previous studies that determined reference ranges for iMg has shown that wide spectrums of ranges have been developed. This may be due to the lack of partitioning that has been reported (for example for age, blood group, gender, race, time of day or posture when sampled) and possibly due to the differences among the three different companies and the ISE that they use.
Conclusions
This study concluded that there is a significant within-day variation in iMg (six blood samples per day taken every three hours from 7:00 till 22:00) and a non-significant between-day variation (three consecutive days of sampling). The significant within day variability of iMg should be accounted for in clinical and research settings and caution should be used in interpreting single measurements. A plausible explanation of the data is that iMg has a circadian rhythm.
Materials and MethodsDescription of Participants
The characteristics of the 13 participants (nine males and four females) that volunteered for the study are listed in table 1. Participant selection for this study was based largely on volunteers who were willing to donate 18 blood samples over three days, but participants were also screened for smoking (non-smokers), blood pressure (resting blood pressure not higher than 144/94), drug use (prescription and non-prescription), vitamin and/or mineral use. As well, alcohol consumption was not permitted throughout the duration of the study.
Characteristics of participants
Parameter
Mean +/- S.D.
Range
Age (yrs)
24.8 +/- 5.7
21 – 41
Height (cm)
173.5 +/- 8.7
165.5 – 184.0
Weight (kg)
67.1 +/- 12.5
47.7 – 89.3
Systolic Blood Pressure (mmHg)
115.3 +/- 8.2
103 – 125
Diastolic Blood Pressure (mmHg)
72.9 +/- 8.2
58 – 80
Dietary Mg Status (mg/day)
378.2 +/- 157.7
132 – 533
The Lakehead University ethics committee approved the study. After informed consent was obtained from each subject, three consecutive days of testing occurred with six blood samples taken each day for a total of eighteen blood samples. The first blood sample was taken at 7:00 a.m. and every three hours thereafter (7:00, 10:00, 13:00, 16:00, 19:00, 22:00). Blood samples were collected throughout the study in 7 ml green topped Vacutainer® tubes (lithium-heparin added) by antecubital venipuncture. Prior to blood withdrawal, the subjects were seated for ten minutes. When blood was collected, a tourniquet was applied gently to the upper arm and released prior to actual blood flow.
The Nova 8 stat analyzer® (Nova Biomedical Canada Ltd., Mississauga, Ontario) housed in the same laboratory that testing occurred, was used for immediate analysis of [iMg] and Hct, from whole blood samples. Throughout the study, the same technician did all the testing. Within-run variation or CVa (analytical error) was calculated by replicate analysis of the specimens (i.e., ten samples in a row of the same blood). Between run variability was not recorded, although control samples in the normal physiological range were consistently within +/- 1 mmol/L of their known concentration.
During the three days of testing, subjects recorded diet, sleep, exercise and stress in a self-monitored logbook. Subjects received information that explained the procedure and importance of accuracy for the self-reported logs and were instructed to continue to follow their normal routine as close as possible.
Dietary intakes were analysed using computerised diet software (Diet Analysis Plus®, 1996, West Publishing Co, St. Paul, MN). Subjects recorded three consecutive days of testing (the same days as the blood samples – two weekdays and one weekend day) and all data was entered into the computer by the same technician.
Daily Log
The variables included in the daily log include: (1) Quality of sleep, (2) Number of hours of sleep, (3) length of exercise session, (4) intensity of exercise session, (5) minor illnesses, (6) minor injuries, (7) menstruation, and (8) major stressful events.
Dependent Variables
(1) Ionic magnesium (mmol/L) – corrected for hematocrit
(2) Hematocrit (%)
Independent Variables
(1) Time
(2) Day
Statistical Analysis
The primary purpose of this study was to determine the variability of iMg in order to determine its utility. Descriptive statistics (mean, standard deviation, range and coefficient of variation (CV)) and graphic presentations of the data were used to gain an initial appreciation of the variability of iMg. CVs were computed to assess analytical error (CVa) from ten repeated measures of subject samples (subjects as their own control). Total within-subject variability (CVt) was computed for each subject's 18 samples with the difference between CVa and CVt attributable to biological variability (CVi). To calculate the CV for the six time periods, all blood values obtained at each time period were pooled together (13 subjects × 3 days = 39 values). The average CV for the six time periods constituted the CVg or between-subject variability. An index of individuality was calculated using a ratio of the total within-subject variability and the between-subject variability (CVt/CVg), in order to assess the utility of iMg in either monitoring a patient or classifying them based on reference ranges.
A 13 (subjects) × 6 (time periods) repeated measures analysis of variance (ANOVA) was used to calculate the within-day variability. Calculation of the between-day analysis requires a data transformation to provide a score for each subject that represents the events of each day across each time measurement. It is important to recognise that the score must consider the events within the measurement interval as a function of events in the preceding interval but while impacting events in each subsequent interval. Plotting the data across time points in a given day for a given subject indicated that there was an inherent rhythm for the measured variable. Therefore, a linear (summative) transformation was inappropriate for the data. A summative scalar such as the coefficient of variation (standard deviation/mean) only considers the range of the variance distributed over the group mean, thus, a transformation that considers range as well as rhythmicity of responses over the entire collection period was needed. Considering the above, the raw data was transformed using a cosine function (transformed score = ½ sin (raw score)). The slope of the transformed data curve was then calculated for each subject, for each day, and the slope scores used in a subsequent ANOVA across the three days.
The first two authors, I.J. Newhouse and K.P. Johnson contributed equally to the preparation of this manuscript.
Acknowledgements
The authors would like to acknowledge the participants who volunteered for this study and also Nova Biomedical® Canada Ltd.
Vascular endothelial growth factor (VEGF) mRNA levels increase in rat skeletal muscle after a single bout of acute exercise. We assessed regional differences in VEGF165 mRNA levels in rat gastrocnemius muscle using in situ hybridization after inducing upregulation of VEGF by treadmill running (1 hr) or electrical stimulation (1 hr). Muscle functional regions were defined as oxidative (primarily oxidative fibers, I and IIa), or glycolytic (entirely IIb or IId/x fibers). Functional regions were visualized on muscle cross sections that were matched in series to slides processed through in situ hybridization with a VEGF165 probe. A greater upregulation in oxidative regions was hypothesized.
Results
Total muscle VEGF mRNA (via Northern blot) was upregulated 3.5-fold with both exercise and with electrical stimulation (P = 0.015). Quantitative densitometry of the VEGF mRNA signal via in situ hybridization reveals significant regional differences (P ≤ 0.01) and protocol differences (treadmill, electrical stimulation, and control, P ≤ 0.05). Mean VEGF mRNA signal was higher in the oxidative region in both treadmill run (~7%, N = 4 muscles, P ≤ 0.05) and electrically stimulated muscles (~60%, N = 4, P ≤ 0.05). These regional differences were not significantly different from control muscle (non-exercised, non-stimulated, N = 2 muscles), although nearly so for electrically stimulated muscle (P = 0.056).
Conclusions
Moderately higher VEGF mRNA signal in oxidative muscle regions is consistent with regional differences in capillary density. However, it is not possible to determine if the VEGF mRNA signal difference is important in either the maintenance of regional capillarity differences or exercise induced angiogenesis.
Background
The adaptations of skeletal muscle to endurance-type training are well characterized. They include increases in mitochondrial volume density and increases in the activity of enzymes involved in oxidative phosphorylation to produce ATP [1]. The increased metabolic capacity of trained muscle is accompanied by an angiogenic response which increases capillary density and/or capillary to fiber ratio [2,3], preserving the functional match between oxygen delivery and metabolic demand within the muscle. The angiogenic response in skeletal muscle is thought to be mediated by a number of angiogenic factors including, most importantly, vascular endothelial growth factor (VEGF). VEGF is a 45 kDA heparin-binding homodimeric glycoprotein with a predominant specificity to vascular endothelial cells [4-7]. Recent investigations demonstrate that VEGF increases vascular permeability [4], endothelial cell proliferation in vitro[8], and angiogenesis in vivo[9]. We have previously demonstrated that VEGF mRNA is upregulated in rat gastrocnemius muscle following 1 hour of acute submaximal treadmill exercise [10], in dog gastrocnemius muscle following 1 hour of electrical stimulation [11], and in human vastus lateralis following 30 min of one-legged knee extension exercise at 50% maximal capacity [12]. Other studies show similar upregulation of VEGF mRNA in chronically electrically stimulated rat skeletal muscle [13,14], and in human vastus lateralis following 45 min of one-legged knee extension exercise [15].
VEGF mRNA transcripts are produced within skeletal muscle fibers and may localize in the sub-sarcolemmal region [10], but it is unknown whether differences in VEGF mRNA expression exist between functionally diverse muscle regions with markedly different fiber type compositions. Because of capillary density and metabolic differences between oxidative and glycolytic muscle regions, we hypothesize a greater up-regulation of VEGF mRNA in oxidative regions of a mixed skeletal muscle. The rat gastrocnemius is an ideal muscle to investigate these questions as it is comprised of approximately 28% fast oxidative glycolytic (FOG), 65% fast glycolytic (FG), and 7% slow (S) muscle fibers [[16], according to a metabolic enzyme-based fiber type classification scheme]. Importantly, this muscle also shows significant regionalization i.e., deep versus superficial regions of both medial and lateral heads of the rat gastrocnemius can be characterized as oxidative (predominantly S and FOG fibers) or glycolytic (predominantly FG fibers), respectively [16].
In the present study we used a 1 hr. treadmill exercise protocol or a 1 hr. electrical stimulation protocol to upregulate total VEGF mRNA production in rat gastrocnemius muscle. Both protocols have previously been demonstrated to upregulate VEGF mRNA levels by 3–4 fold [10,13,14]. In situ hybridization with a VEGF165 probe was then applied to muscle samples to detect the fiber specific localization of VEGF mRNA transcripts. In order to observe the pattern of regional VEGF production, serial cross sections were used to identify fiber types as I, IIa, IIb, and IId/x by histochemistry and immunohistochemistry. Under this classification scheme, fibers are characterized as follows: type I fibers are slow-twitch with a high oxidative capacity; type IIa fibers are fast-twitch with a high oxidative and glycolytic capacity; type IIb fibers are fast-twitch with a low oxidative and a high glycolytic capacity; and type IId/x fibers (common in small mammals) are fast-twitch with a metabolic profile intermediate to that of IIa and IIb fibers [17].
ResultsDetermination of fiber types and VEGF mRNA signal on serial slide sections
Figure 1 shows serial sections of rat gastrocnemius after 1 hour of treadmill running. The VEGF mRNA signal is apparent as a dark stain (panel A), and may be compared to the sense control (panel B) which shows no non-specific binding in the same muscle region. In panel C, type I (dark stain), IIa (light stain), and IIb-IId/x (intermediate stain) fibers can be identified based on the myofibrillar actomyosin ATPase histochemical staining procedure [18]. Immunohistochemistry was used to positively identify subpopulations of fibers expressing only the 2B MHC. Panel D shows positive binding (dark) of BF-F3 antibody against 2B MHC. In this manner subpopulations of IIb and IId/x fibers were identified. Panel E shows positive binding (dark) of the NCL-MHCs antibody against the slow MHC (type I), and panel F shows positive binding (dark) of the A4.74 antibody against 2A MHC. Identification of I and IIa fibers by immunohistochemistry was fully concordant with identification of these fiber types by the myofibrillar actomyosin ATPase histochemical staining procedure.
Serial sections of rat gastrocnemius muscle. This is muscle after a 1 hr treadmill run processed through in situ hybridization, histochemistry, and immunohistochemistry. (A) Intracellular location (dark stain) of VEGF165 mRNA transcripts from in situ hybridization with representative type I, IIa, IIb and IId/x fibers indicated. (B) Sense strand control from in situ hybridization which shows no significant non-specific background staining in this protocol. (C) Myofibrillar actomyosin ATPase staining procedure to identify type I (dark stain), type IIa (light stain), and the mixed population of type IIb-IId/x (medium stain) fibers. (D) Positive binding of 2B MHC antibodies (BF-F3) to type IIb fibers is shown as dark stain. (E) Positive binding of slow MHC antibodies (NCL-MHCs) to type I fiber shown as dark stain. (F) Positive binding of 2A MHC antibodies (A4-74) to type IIa fibers shown as dark stain. Magnification by 25× objective in all photos.
Northern analysis
Figure 2 shows representative Northern blots for total VEGF mRNA levels examined in (A) resting control, (B) exercised, and (C) electrically stimulated muscles. It is clear that VEGF mRNA levels are elevated over resting after both 1 hour of treadmill exercise and 1 hour of electrical stimulation. This is shown clearly in Figure 3 which gives the quantitative densitometry for VEGF mRNA, normalized to 18S ribosomal RNA. A significant increase in VEGF mRNA over resting control is observed after 1 hour of treadmill running (~3.5-fold, P = 0.015), and after 1 hour of electrical stimulation (~3.5-fold, P = 0.015). There were no significant differences in the VEGF mRNA levels between exercised and electrically stimulated muscles.
Representative Northern blot for total VEGF mRNA levels. (A) resting, (B) exercised, and (C) electrically stimulated gastrocnemius muscles. Blots were reprobed and normalized to 18S ribosomal RNA (18S).
Quantitative densitometry of total VEGF mRNA levels. (A) resting, (B) exercised, and (C) electrically stimulated rat gastrocnemius muscles. Lane loading was normalized by densitometry on 18S ribosomal RNA. *Significantly different from rest-control.
Regional VEGF mRNA signal
Oxidative regions were located in the deep part of the gastrocnemius, and were identified based on the expression of a majority of type I and IIa fibers (~77% of the total fiber number in the regions analyzed). Glycolytic regions were located superficially, and were identified based on the expression of only type IIb and IId/x fibers.
The heterogeneity of fiber phenotypes in oxidative regions affords the possibility of comparing VEGF mRNA signal between adjacent fibers differing in metabolic capacity. However, data collected in this regard were considered insufficient to test the hypothesis that fiber phenotype is related to VEGF mRNA signal strength. That is, a non-biased test of this hypothesis requires a random sampling of fibers from each phenotypic class in multiple muscle samples. Such fibers must then be clearly identified on serial slide preparations to determine fiber phenotype and VEGF mRNA signal. This was not possible in all muscle samples due to freezing artifact and tissue degradation during (particularly) the in situ hybridization process. Regional comparisons are not limited in this way as muscle regions were easily identified on serial sections for measurement of VEGF signal in a random sample of fibers within the region (see Methods).
Data presented in Table 1 show the normalized mean VEGF mRNA signal in oxidative versus glycolytic muscle regions from rats after 1 hour of treadmill exercise, rats after 1 hour of electrical stimulation, and control rat muscle (neither exercise or electrical stimulation). Mean VEGF signal was determined from 27 randomly sampled fibers in each defined region within a single muscle sample (see Methods). By ANOVA, there were significant differences in VEGF mRNA signal strength between functional regions (P = 0.002) and between protocols (treadmill, electrical stimulation, and control, P = 0.036). In treadmill run muscle the glycolytic regional mean was 93% of the oxidative regional mean (P ≤ 0.05), and in electrically stimulated muscle the glycolytic regional mean was 59% of the oxidative regional mean (P ≤ 0.05). Significant regional differences were not apparent in control muscle (P = 0.463). Post-hoc testing reveals a significantly larger regional difference in VEGF mRNA signal in electrically stimulated versus treadmill run muscle (P = 0.017), and no significant difference in this regard between electrically stimulated and control muscle, although the latter comparison reaches a P-value of 0.056. Regional differences were not significantly different between treadmill run and control muscle (P = 0.912).
Normalized1 mean VEGF165 signal from in situ hybridization in fibers randomly sampled from oxidative and glycolytic regions of the rat gastrocnemius. Values are given as mean ± SE.
Oxidative region
Glycolytic region
Number of fibers sampled
VEGF signal (Mean of fibers sampled)
Number of fibers sampled
VEGF signal (Mean of fibers sampled)
Treadmill
Muscle 1
27
1.00 ± 0.05
27
0.93 ± 0.05
Muscle 2
27
1.00 ± 0.03
27
0.95 ± 0.02
Muscle 3
27
1.00 ± 0.03
27
0.92 ± 0.03
Muscle 4
27
1.00 ± 0.02
27
0.91 ± 0.01
Mean2
1.00 ± 0.00
0.93 ± 0.01*
Electrical
Muscle 1
27
1.00 ± 0.04
27
0.76 ± 0.03
Muscle 2
27
1.00 ± 0.11
27
0.63 ± 0.14
Muscle 3
27
1.00 ± 0.09
27
0.65 ± 0.06
Muscle 4
27
1.00 ± 0.69
27
0.31 ± 0.25
Mean2
1.00 ± 0.00
0.59 ± 0.1*
Control
Muscle 1
27
1.00 ± 0.04
27
0.83 ± 0.04
Muscle 2
27
1.00 ± 0.03
27
0.99 ± 0.05
Mean2
1.00 ± 0.00
0.93 ± 0.02
1Normalized within each muscle sample to the mean of the oxidative region. 2ANOVA mean using individual muscles as the unit of analysis (P = 0.002 for regional differences and P = 0.036 for protocol differences). *Significantly different from mean of oxidative region by paired t-test, p ≤ 0.05.
Discussion
Previous studies demonstrate that an acute bout of exercise or electrical stimulation is sufficient to significantly upregulate total VEGF mRNA levels in rat and human skeletal muscle within 1 hour [10,12,15,19]. We have replicated these previous findings by showing (through Northern analysis) that 1 hr. of treadmill running or electrical stimulation induces an approximately 3.5-fold upregulation in VEGF mRNA in rat gastrocnemius muscle (Figure 2 and 3). More specifically, in the present descriptive study, non-radioactive in situ hybridization with a VEGF165 probe [20] was used to visualize the fiber specific localization of VEGF mRNA transcripts on muscle tissue cross sections after muscles were stimulated to upregulate VEGF through exercise or electrical stimulation. This technique revealed modest VEGF signal differences between functional regions of the rat gastrocnemius that differed in oxidative capacity, with oxidative muscle regions showing higher signal after both electrical stimulation and treadmill running.
In situ hybridization allows visualization of the intracellular location of mRNA transcripts, but it is not a technique designed to provide a quantitative measure of mRNA signal strength. This is problematic, even for quantitative comparison between muscle regions on the same slide preparation. For example, signal strength can vary across a slide preparation with probe binding success, which depends on the elimination of endogenous RNases, and on tissue permeability to the VEGF probe [21]. Additional error is introduced by freezing artifacts, which are difficult to avoid, and non-specific background signal. To address the latter limitation, we successfully minimized background signal as demonstrated in a control muscle cross-section probed with a sense strand transcript of the VEGF probe (Figure 1B). Error is not necessarily an insurmountable problem so long as it is randomly distributed across muscle slide preparations. On the assumption of random error distribution, our approach was to use the individual muscle as the unit of analysis and to analyze multiple muscle samples. Fibers were randomly sampled from muscle functional regions to determine a regional mean VEGF signal within each muscle sample, and this signal was normalized to an internal standard to allow for direct comparison between muscles (see Methods). Non random error (confounding) was also considered, including fiber size differences between regions and the number of nuclei visible in myofibrillar actomyosin ATPase slide preparations between regions (see Figure 1C). While fibers were larger in glycolytic regions, size differences did not relate to VEGF signal. Also, the VEGF probe did not co-localize with muscle nuclei within the fiber (see Figure 1A), another possible source of confounding.
Relevant to the main finding of this study, that VEGF mRNA signal is stronger in deep oxidative regions of the muscle expressing a high proportion of type I and IIa fibers, is the extensive literature describing skeletal muscle capillarity, blood flow, and the angiogenic response to exercise or electrical stimulation [22]. Fibers with an oxidative phenotype have a greater number of surrounding capillaries than glycolytic fibers [23]. This difference in vascular supply is consistent with the metabolic profiles of type I and IIa fibers versus IIb and IId/x fibers. The former have a high oxidative capacity and high demand for O2 delivery, while the latter have a lower oxidative capacity and a reduced demand for O2[17]. Studies of mixed muscle show greater capillarity and/or higher rates of blood flow during exercise in regions composed primarily of oxidative fibers compared to regions with a high proportion of glycolytic fibers [24-28]. Thus, the regional difference observed in VEGF signal may be related to differences in innate fiber phenotype. However, we cannot exclude other possibilities. For example, the difference may be related to differences in metabolic activity between regions during the stimulation protocols, rather than fiber phenotype per se (see below), or indeed to regional differences in other VEGF producing cell types e.g., endothelial cells which are well know to produce VEGF [29].
Whether the regional differences in VEGF mRNA demonstrated here underlie either 1) regional differences in capillarity or 2) regional differences in exercise induced angiogenesis cannot be determined from our study. Regarding angiogenesis, our findings are not consistent with previous studies which suggest that capillary growth in response to electrical stimulation is initiated in the vicinity of IIb fibers, not I or IIa fibers [14,22]. However, our findings are consistent with recent findings by Annex et al. [30]. These authors demonstrate that VEGF mRNA upregulation after chronic electrical stimulation is followed by an increase in VEGF protein in the extracellular matrix between fibers, and that the protein increase is highest in innate oxidative versus innate glycolytic muscle. The increase in protein takes place over a 21 day time frame which is much longer than the transient increase in VEGF message demonstrated here, but certainly noteworthy.
We cannot explain why electrical stimulation tends to produce larger differences in VEGF mRNA signal compared to treadmill running. Indeed, treadmill running did not produce a regional difference that was significantly different from the control muscle processed through in situ hybridization. This may be related to the small number of control muscles processed, or with fundamental differences between the exercise and electrical stimulation protocols. The running and electrical stimulation protocols differ in the way that motor units are involved during muscle contraction, and thus they may also differ in the pattern of VEGF gene activation between muscle regions. During exercise, slow twitch motor units serving oxidative fibers are recruited first during moderate exercise, and motor units serving glycolytic fibers are recruited later and during more severe exercise [1]. This is in contrast to electrical stimulation of sufficient intensity where all motor units are thought to be activated simultaneously [31]. The larger regional difference in VEGF mRNA signal after electrical stimulation may be due to the fact that this protocol was more metabolically demanding than the treadmill exercise protocol, or indeed because the rat muscle was hypoxic during electrical stimulation. In support of the first possibility, we observed a significant force decline in many of the electrically stimulated muscles i.e., some muscles showed a decline in peak tension development of more than 25% by the end of 1 hr. of stimulation. However, we have no corroborative evidence to support this possibility such as the glycogen content of muscle fibers after exercise or electrical stimulation. In support of the latter possibility, the gastrocnemius muscle may have been hypoxic, despite the fact that all rats were mechanically ventilated during the electrical stimulation protocol. Hypoxia has been demonstrated to upregulate VEGF independent of exercise [10], although in this case we would have expected higher total VEGF levels by Northern analysis in the electrically stimulated muscles, and this was not the case (Figures 2 and 3).
Conclusions
In summary, this descriptive study demonstrates modest regional differences in the mean VEGF mRNA signal in rat gastrocnemius muscles that were induced to upregulate VEGF production by 1 hr. of treadmill running or electrical stimulation. This regional patterning of VEGF mRNA is at least superficially consistent with differences in capillarity and oxidative capacity, although not consistent with the angiogenic process itself which may be preferentially initiated around IIb fibers. Despite some consistencies with morphologic and metabolic observations, it is not possible to conclude that the observed VEGF mRNA differences are physiologically important, or that they are involved in initiating and/or maintaining regional heterogeneity in muscle capillary density. However, the technique of in situ hybridization may prove useful to address these questions in the future.
Materials and MethodsAnimals
This study was approved by the University of California, San Diego, Animal Subjects Committee. Female Wistar rats were used (age 6–8 weeks, ~200 grams body weight). All animals were housed in cages and allowed standard rat food and water ad libitum prior to study.
Study protocols used to induce upregulation of VEGF mRNA
Both treadmill exercise and muscle electrical stimulation protocols were used to initiate the total VEGF mRNA response in rat skeletal muscle. For the exercise protocol, 6 rats were first familiarized with a rodent treadmill (Omnipacer model LC-4, Omnitech, Columbus, OH), and then required to run for 1 hour at 20–35 m/min on an incline of 10°. Running speed varied between rats but was maintained for an individual rat near the maximal running speed that could be sustained for a 1 hour period. For rats of this age and body size we have previously shown that a running speed of 20 m/min at 10° incline represents ~55% of maximal oxygen consumption [10].
For the electrical stimulation protocol 6 rats were anesthetized by i.p. injection of sodium pentobarbitol. Because anesthetization depresses ventilatory drive, and because hypoxia can also induce VEGF mRNA upregulation, great care was taken in the electrical stimulation protocol to ensure that animals were adequately mechanically ventilated. For each rat, the trachea was intubated and ventilation was maintained at a tidal volume of 2.5 ml, ventilatory frequency of 50 breaths/min, and 1 cm of positive end expiratory pressure. Preliminary work showed this level of ventilation to be sufficient to preserve normal arterial PO2 and PCO2. The left gastrocnemius, soleus, and plantaris muscles were surgically isolated and the sciatic nerve, which innervates this muscle complex, was cut and attached to an electrode. The hindlimb was fixed so that no movement of the limb was possible during muscle contraction and the Achilles tendon was attached to a force-displacement transducer (Grass Instruments, Co., Quincy, Mass) to monitor tension development during contraction. Contractions were initiated by train rate impulses, 4–8 volts, 200-ms duration, 1 ms-1 frequency, given at a rate of 2 stimulations (contractions) per second for 1 hour. Signs of muscle fatigue i.e., a decline in tension development below 75% of peak, were evident using this stimulation protocol after 30–60 minutes in approximately 50% of the rats tested. In the remaining rats, tension was well maintained over the course of the stimulation protocol. There was no detectable difference in VEGF mRNA levels by Northern analysis between muscles that fatigued versus muscles that maintained tension development. Six resting rats (non-exercised) were processed as a control for Northern Analyses and in situ hybridization. These were the age matched cage mates of those rats run on the treadmill (2 rats per cage). With the exception of the exercise protocol, these rats were handled in an identical manner to the exercised rats.
Tissue collection and processing
After the exercise protocol rats were allowed to rest for 1 hour and then were anesthetized by i.p. injection of sodium pentobarbitol. For the 6 rats run on the treadmill, the left gastrocnemius (both heads combined) was removed for RNA isolation, frozen in liquid nitrogen, and stored at -80°C. In these same rats the right gastrocnemius was removed for in situ hybridization and fiber type determinations. This tissue was mounted in TBS tissue freezing medium (Triangle Biomedical Sciences, Durham, NC), frozen in isopentane cooled in liquid nitrogen, and stored at -80°C. For the 6 electrically stimulated rats, skeletal muscle was also harvested 1 hour after the completion of the electrical stimulation protocol. Electrical stimulation can be applied to one leg at a time only. In our protocol the left leg is used. Thus, the left gastrocnemius was harvested for RNA isolation and Northern analyses in six rats. An additional four rats provided a left gastrocnemius for in situ hybridization, histochemistry, and immunohistochemistry. This minor inconsistency in study design is not expected to affect study results. Six resting rats provided a right gastrocnemius as a control for Northern analysis.
For Northern analysis, 18 muscle samples were initially processed (6 rest control, 6 electrically stimulated, and 6 treadmill run), but some samples were degraded during the RNA isolation leaving 3 rest-control, 4 exercised, and 4 electrically stimulated muscle samples. In order to ascertain regional and fiber type patterns of VEGF mRNA production in those muscles stimulated to upregulate VEGF, 4 each of treadmill exercised and electrically stimulated muscle samples, and 2 control muscles, were successfully processed through in situ hybridization.
RNA isolation and Northern analyses
Total cellular RNA was isolated from each gastrocnemius muscle by the method of Chomczynski and Sacchi [32]. RNA preparations were quantified by absorbance at 260 nm and RNA intactness and integrity assessed by ethidium bromide staining after separation by electrophoresis in a 6.6% formaldehyde-1% agarose gel. Fractionated RNA was transferred by Northern blot to Zeta-probe membrane (Bio-Rad, Hercules, CA). Following transfer, RNA was cross-linked to the membrane by ultraviolet irradiation for 1 min and stored at 4°C. The blots were then probed with oligolabeled [alpha-32P] deoxycytidine triphosphate cDNA probes which had a specific activity > 1 × 109 disintegrations.min-1.μgDNA-1(11). The rat VEGF probe is a 0.9 kb cDNA Pst I/Sma I insert cloned into a pBluescript II KS+ vector [33]. Prehybridization and hybridization were performed in 50% formamide, 5×SSC (20×SSC is 0.3 M sodium chloride, 0.3 < sodium citrate), 10× Denhardt's solution (50× Denhardt's solution is 2% Ficoll, 2% polyvinyl pyrrolidone, 2% Bovine Serum Albumin Factor V), 50 mM sodium phophate (pH 7.0), 1% sodium dodecyl sulfate (SDS), and 250 μg/ml salmon sperm DNA at 42°C. Blots were washed with 2×SSC and 0.1% SDS at room temperature and 0.1% SSC and 0.1% SDS at 65°C. Blots were exposed to XAR-5 X-ray film (Eastman Kodak, New Haven, CT) for 2–3 days by use of a Cronex Lighting Plus screen at -80°C. Autoradiographs were quantitated by densitometry within the linear range of signals and normalized to ribosomal 18S RNA levels.
In situ hybridization
Frozen gastrocnemius muscle samples were cryo-sectioned in a Reichert Jung Cryocut 1800 cryostat (Cambridge Instruments, Buffalo, NY) at -20°C to 10 μm and mounted in cross-section on Fisherbrand Superfrost/Plus microscope slides (Fisher Scientific, Pittsburgh, PA). Slides were immediately processed through the in situ hybridization protocol of Braissant and Wahli [20] to detect VEGF mRNA transcripts on tissue sections. After tissue fixation for 20 minutes in 4% paraformaldehyde in diethyl pyrocarbonate (DEPC) treated phosphate buffered saline (PBS), sections were incubed for 2 × 15 minutes in PBS containing 0.1% active DEPC, and equilibrated for 15 min in 5×SSC. The sections were then prehybridized for 2 hr at 58°C in the hybridization mix (50% formamide, 5×SSC, salmon sperm DNA 40 μg/μl; 500 μl on each section). The hybridization probes (see below) were denatured for 5 min at 80°C and added to the hybridization mix (400 ng/ml). The hybridization reaction was carried out overnight at 58°C with 50 μl of hybridization mix on each section. During hybridization, sections were kept in a box saturated with 5×SSC, were covered by a rectangle of Parafilm, and were sealed with DPX Mountant (Fluka Chemie, Neu-Ulm, Switzerland) to prevent drying. After hybridization, sections were washed for 30 min in 2×SSC (room temperature), 1 hour in 2×SSC (65°C), 1 hr in 0.1×SSC (65°C), and equilibrated for 5 min in Buffer 1 (Tris-HCl 100 mM and NaCl 150 mM, pH 7.5). Sections were then incubated for 2 hours at room temperature with alkaline phosphatase-coupled anti-digoxigenin antibody (Boehringer Mannheim, Indianapolis, IN) diluted 1:5000 in Buffer 1 containing 0.5% Blocking reagent (Boehringer Mannheim). Excess antibody was removed by two 15 min washes in Buffer 1, and the sections were equilibrated for 5 min in Buffer 2 (Tris-HCl 100 mM, NaCl 100 mM, and MgCl2 50 mM, pH 9.5). Color development was performed at room temperature overnight in Buffer 2 containing NBT and BCIP (Boehringer Mannheim). Staining was stopped by a 10 min wash in Tris/EDTA (10/1 mM, ph 8.0), and non-specific staining was removed by 1 hour in 95% EtOH. Sections were rehydrated in deionized water for 15 min to remove precipitated Tris and then dehydrated through successive ethanol baths (70, 95, 100%) and Hemo-De (Fisher Scientific) (2 × 15 min each) and mounted in Permount (Fisher Scientific).
Hybridization probes were derived from a ~600-bp cDNA that contains the entire coding region of the human VEGF165[8] subcloned into the pBluescript II KS+ vector (Stratagene, La Jolla, CA). The final riboprobe was the same size (~600-bp) as the cDNA. Constructs in this vector were linearized at appropriate restriction sites to allow the sythesis of digoxigenin-UTP labeled complementary RNA in the antisense or sense orientation (using T3 or T7 RNA polymerase, respectively; Boehringer, Indianapolis, IN). The riboprobes synthesized in the "sense" orientation served as background control.
Determination of fiber types
Fiber types were determined in serial cross sections. Two methods were used: (1) The histochemical method of Ogilvie and Feeback [18], based on staining for myofibrillar actomyosin ATPase, and (2) immunohistochemistry based on the reaction of specific antibodies for mysosin heavy chain (MHC) isoforms [34,35]. In the histochemical staining procedure, sections were pre-incubated in (0.025 M potassium acetate, 9 mM calcium chloride-dihydrate, pH to 4.5 with glacial acetic acid) for 8 minutes, rinsed 3× two min/each in (0.1 M Trizma base (Sigma), 0.018 M calcium chloride-dihdrate), and incubated for 25 minutes in (0.154 M ATP-disodium salt, 0.053 M Glycine, 0.029 M calcium chloride, 0.065 M sodium chloride, 0.048 M sodium hydroxide, pH 9.4). Following incubation sections were rinsed 3× in 0.014 M calcium chloride-dihydrate, stained for 1 min in Toluidine blue, rinsed for 5 sec in deionized water, and dehydrated with 5 dips in 95% ethanol, 5 dips in 2 changes of 100% ethanol, and 2 changes of Hemo-De (Fisher) for 5 min each. Sections were mounted with Permount (Fisher). This method allows for the clear determination of type I and IIa fibers, but distinguishing IIb from IId/x subpopulations is difficult and subjective as both stain with nearly the same intensity. For this reason immunohistochemistry was also used on one muscle sample to clearly identify IIb fibers based on specificity of the 2B MHC antibody. For immunohistochemistry, mouse polyclonal and monoclonal antibodies specific for slow (Novocastra, NCL-MHCs), 2A (Blau, A4.74), and 2B (DSM, BF-F3) MHC isoforms were used. Serial sections were incubated overnight with primary antibodies diluted (NCL-MHCs, 1:100; and BF-F3, A4.74, 1:1) in a 0.2 M potassium phosphate solution (PPS). After washing 3× with PPS, sections were incubated for 3 hr at room temperature with either peroxidase-conjugated goat anti-mouse immunoglobins (for NCL-MHCs, and A4.74) or biotinylated rabbit antimouse IgM (for BF-F3) (DAKO A/S, Glostrup, Denmark). The latter was also incubated for 1 hour with streptavidin-HRP (NEN Life Science Products, Boston, MA) diluted 1:100 in PPS. Sections were again washed 3× in PPS and color was evolved by incubating slides in the dark between 2–10 min in a 3,3'-diaminobenzidine (DAB) solution (Vector Laboratories Inc., Burlingame, CA). Sections were rinsed in deionized water, dehydrated in successive ethanol baths, cleared with Hemo-De (Fisher), and mounted with Permount (Fisher).
Quantification of the regional VEGF mRNA signal from in situ hybridization
Regional VEGF mRNA signal within an individual muscle fiber was quantified by densitometry from microscope images obtained by a Sony 3ccd color video camera (Sony Corporation, Japan) attached to a Jenalumar optical microscope (Jenoptik JENA, Germany), and processed by the Sigma Scan Pro imaging software (SPSS Inc., Chicago, IL). Optical density measurements were made on a Macintosh computer using the public domain NIH Image program (developed at the U.S. National Institutes of Health and available from the internet by anonymous FTP from ). This program can calculate mean grey-scale value within the cross sectional area of a single muscle fiber.
Unfortunately, measured grey-scale values give arbitrary units. Within a given microscopic field of view, an optical density measurement depends both on the strength of the VEGF mRNA signal, and also on the amount of light transmitted through the tissue sample (which in turn depends on the degree of freezing artifact in muscle preparation and many other factors). Uncontrolled factors affecting light transmission imply that optical density comparisons between different slide preparations are not valid. In order to aggregate data from different slides, it was necessary to normalize all VEGF mRNA optical density signals to an internal standard. For example, to describe the difference between an oxidative and glycolytic muscle region, the optical density value of each individual fiber was normalized to the mean of all fibers within the oxidative region (which was arbitrarily set to a value of 1.0) (see Table 1).
A random sampling strategy was used to obtain mean optical density values for different muscle regions identified on histochemically stained sections. Oxidative regions were located in the deep part of the gastrocnemius and expressed a majority of type I and IIa fibers (~77%), while glycolytic regions were located superficially, and expressed only type IIb and IId/x fibers. For each muscle, a microscopic field of view was located in a region which contained typically between 100–200 fibers in cross section at 125× magnification. The same field of view was then located on a serial section processed through in situ hybridization. Twenty seven fibers total (three fibers from each of 9 equal quadrats) were randomly selected from the region on the in situ slide preparation and densitometric measurements of the VEGF mRNA signal within fibers were made.
Analyses and statistics
Analysis of variance was used to test for differences in VEGF mRNA levels from Northern analysis between muscle samples from exercised (n = 4) and electrically stimulated rats (n = 4). ANOVA, using the individual muscle sample as the unit of analysis, was also used to test for the main effects of functional region (oxidative vs. glycolytic) and protocol (treadmill running, electrical stimulation, control) on VEGF mRNA signal. Post-hoc testing, using Fisher's Least Square Difference correction for multiple comparisons was used to test for differences between protocols in the regional difference of VEGF mRNA signal. Paired t-tests were used to compare regional means within a given protocol. All statistical analyses were performed using Systat Version 5.2 (Systat Inc., Evanston, IL). Statistical significance was set at p ≤ 0.05 for all tests.
Authors' contributions
Author 1 (TB) participated in the conception, design, and coordination of the study, carried out the animal potocols, tissue processing through in situ hybridization, histochemistry, and immunohistochemistry, microscopic quantification of VEGF signal, and drafted the manuscript. Author 2 (TG) did the Northern analysis. Author 3 (ZF) participated in the immunohistochemistry and in situ hybridization. Author 4 (EB) participated in the immunohistochemistry and in situ hybridization. Author 5 (5) participated in the in situ hybridization. Author 6 (OMC) participated in the design and coordination of tissue processing and histochemistry. Author 7 (PW) participated in the conception, design, and coordination of the study.
Acknowledgements
This study was supported by a National Heart, Lung, and Blood Institute Grant, HL-17731. T.P. Gavin was funded by a National Heart, Lung, and Blood Institute Fellowship HL-09624. The authors would like to thank the laboratory of Dr. Gary Sieck, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, for kindly providing the MHC antibodies used in this study.
The cockroach salivary gland consists of secretory acini with peripheral ion-transporting cells and central protein-producing cells, an extensive duct system, and a pair of reservoirs. Salivation is controled by serotonergic and dopaminergic innervation. Serotonin stimulates the secretion of a protein-rich saliva, dopamine causes the production of a saliva without proteins. These findings suggest a model in which serotonin acts on the central cells and possibly other cell types, and dopamine acts selectively on the ion-transporting cells. To examine this model, we have analyzed the spatial relationship of dopaminergic and serotonergic nerve fibers to the various cell types.
Results
The acinar tissue is entangled in a meshwork of serotonergic and dopaminergic varicose fibers. Dopaminergic fibers reside only at the surface of the acini next to the peripheral cells. Serotonergic fibers invade the acini and form a dense network between central cells. Salivary duct segments close to the acini are locally associated with dopaminergic and serotonergic fibers, whereas duct segments further downstream have only dopaminergic fibers on their surface and within the epithelium. In addition, the reservoirs have both a dopaminergic and a serotonergic innervation.
Conclusion
Our results suggest that dopamine is released on the acinar surface, close to peripheral cells, and along the entire duct system. Serotonin is probably released close to peripheral and central cells, and at initial segments of the duct system. Moreover, the presence of serotonergic and dopaminergic fiber terminals on the reservoir indicates that the functions of this structure are also regulated by dopamine and serotonin.
Background
Cockroaches have acinar salivary glands that consist of secretory acini and an extensive duct system [1,2](see Fig. 1a). In addition to the salivary glands proper, the salivary gland complex includes a pair of reservoirs with their ducts, and an extrinsic muscle associated with the orifice of each reservoir [3].
Morphology of the salivary glands in the cockroach Periplaneta americana a: Low-power micrograph of the salivary gland complex. The salivary glands are paired and consist of several lobes of secretory acini. The ducts (arrowheads) of each gland unite to a single efferent salivary duct (long arrows) that then fuses with the opposite duct to form the main salivary duct. Two reservoirs (asterisks) are associated with the secretory tissue. The reservoirs open into reservoir ducts (broad arrows) that accompany the efferent salivary ducts. b: Schematic representation of the structural organization of secretory acini. Each acinus consists of two peripheral cells with long microvilli and several central cells with numerous secretory granules. The apical surface of the central cells is covered by a sheath of flattened fenestrated centroacinar cells and by a thin discontinuous layer of cuticule. The central cells are stimulated only by serotonin, whereas the peripharal cells respond to dopamine and probably also to serotonin. The duct cells have basal and apical infoldings and are only responsive to dopamine. Scale bar = 2 mm
The physiology of the salivary gland complex and the neuronal and cellular control of salivation are poorly understood. The following picture emerges from currently available evidence. The salivary glands secrete saliva of two different qualities, either with or without proteins [4]. Salivation appears to be mainly controled by direct serotonergic and dopaminergic innervation from the subesophageal ganglion and the stomatogastric nervous system [5-8]. A pair of large dopaminergic neurons located within the subesophageal ganglion and termed SN1 (salivary neuron 1) send their axons via nerves extending along the salivary ducts toward the salivary glands where they ramify extensively [6,8]. This appears to be the only source of dopaminergic innervation of the salivary glands. Serotonergic innervation is achieved via several small axons within the salivary nerves and the esophageal nerve. The latter branches into several thin nerves that form a network over the acinar tissue [8,9]. Despite this general picture, the exact topography of the dopaminergic and serotonergic nerve fibers, their varicosities, and/or their terminals with respect to the various structures and cell types engaged in saliva production and modification is still insufficient for a stringent correlation of innervation and gland physiology.
The acini of cockroach salivary glands are grape-like structures and composed mainly of peripheral cells and central cells [2] (see Fig. 1b). Peripheral cells reside in pairs at the distal end of each acinus, possess long apical microvilli equipped with Na+,K+-ATPase, and are specialized for water and electrolyte transport [10]. Central cells are densely packed with secretory granules and produce the proteinaceous components of the saliva [2,4]. The saliva secreted within the acinar portions of the glands then passes through the salivary ducts composed of a simple epithelial layer. The duct cells have an extensive basal labyrinth carrying Na+,K+-ATPase molecules and apical infoldings studded with vacuolar H+-ATPase molecules [10], suggesting that this cell type modifies the ionic composition and/or the volume of the primary saliva.
Both serotonin and dopamine have been shown to stimulate salivation in isolated salivary glands; however, the quality of the saliva differs upon exposure to these substances [4]. Superfusion of salivary glands with serotonin leads to the exocytosis of secretory granules and the production of a protein-rich saliva, suggesting that at least the central cells are responsive to serotonin. Saliva produced upon dopamine application, in contrast, is completely free of proteins, indicating that this neurotransmitter acts selectively on the ion-transporting cells, viz., the peripheral cells and/or the duct cells. Electrophysiological studies on salivary duct cells have further shown that dopamine induces a slow depolarization, evokes an increase in the intracellular Ca2+ concentration, and elicits an intracellular Na+ elevation and K+ reduction in these cells [11,12]. Serotonin, in contrast, appears to have no effect on salivary duct cells [11].
The above results lead to a model in which salivary duct cells are stimulated exclusively by dopamine and central cells exclusively by serotonin. Peripheral cells may be responsive only to dopamine or to both neurotransmitter substances. In order to examine this model further, we have analyzed the exact spatial relationship of dopaminergic and serotonergic nerve fibers to these cells types by studying anti-dopamine and anti-serotonin immunofluorescence on whole-mount preparations of salivary glands in conjunction with high-resolution confocal microscopy. Close apposition of fiber terminals and/or varicosities to a distinct cell type provides evidence for a selective innervation of the respective cell type. We demonstrate that the innervation pattern is more complex than expected from the above model, but that it essentially supports this working hypothesis. In addition, we have examined the serotonergic and dopaminergic innervation of other structures associated with the salivary glands (see Fig. 1a), viz., the reservoirs, the reservoir ducts, and the muscles attached to the reservoir orifices.
ResultsSpecificity of antibody labeling
On cryostat sections of cockroach salivary glands, both anti-serotonin and anti-dopamine intensely stained fiber-like structures and individual punctae, the latter probably representing cross-sectioned fibers (Fig. 2a,2c). Specificity of labeling was tested by preabsorption of the primary antibodies with the corresponding antigens, serotonin or dopamine, respectively. Under these conditions, immunoreactivity was highly reduced or absent (Fig. 2b,2d), suggesting that these antibodies identify their appropriate antigens within cockroach salivary glands.
Specificity of anti-serotonin and anti-dopamine labeling
a-d: Fluorescence confocal images, representing the summarized view of 9-μm-thick image stacks. e-h: Nomarski contrast images of the same areas. a,b: Cryostat sections of salivary glands incubated with anti-serotonin in the absence or in the presence of 1 mg/ml serotonin. c,d: Sections reacted with anti-dopamine in the absence or in the presence of 1 mg/ml dopamine. Immunoreactivity of the tissue is highly reduced in the presence of the corresponding antigen. Scale bars = 100 μm
Further support for the specificity of anti-dopamine immunoreactivity was provided by colabeling experiments with affinity-purified antibody against tyrosine hydroxylase (TH), a common probe for dopaminergic neurons in insects [6,8]. TH is the first and rate-limiting enzyme in the synthesis of the catecholamines dopamine, norepinephrine and epinephrine, of which dopamine is the major amine found in insects [13]. When the anti-TH antibody was applied to whole-mounts of salivary glands, it produced a labeling pattern that precisely corresponded to the anti-dopamine-immunoreactive structures (Fig. 3).
Codistribution of anti-dopamine and anti-TH immunolabeling Whole-mounts of salivary glands were triple-labeled with anti-dopamine (green), anti-TH (red) and BODIPY FL phallacidin (blue), and imaged by confocal microscopy. The image shows a lobule of acinar tissue; the peripheral cells are arranged in pairs and their apical arrays of phallotoxin-stained microvilli appear as "bow ties". A sparse network of fibers resides on the tissue and is labeled by both anti-dopamine and anti-TH. Scale bar = 100 μm
It should be noted that colabeling experiments with anti-dopamine and anti-serotonin were not successful as these antibodies required different protocols for tissue fixation. The anti-dopamine provided specific labeling only in specimens fixed in the presence of at least 0.5% glutaraldehyde. The anti-serotonin, however, only exhibited specific immunoreactivity in tissue fixed without glutaraldehyde.
Distribution of serotonergic and dopaminergic nerve fibers over the secretory acini
The distribution of serotonergic and dopaminergic fibers within the salivary gland complex was probed by confocal fluorescence microscopy of whole-mount preparations stained with anti-serotonin or anti-dopamine. In order to locate the various acinar cells and to provide a spatial reference for the position of the immunoreactive fibers within the tissue, specimens were co-labeled with fluorochrome-tagged phallacidin, a probe for actin filaments [2]. Peripheral cells with their densely packed, long microvilli are arranged in pairs that are visualized as brightly fluorescent "bow ties" in phallotoxin-stained preparations (Figs. 3c, 4, 7, 8). The acinar lumen surrounded by central cells with their short microvilli is delimited by weak labeling with phallotoxin (Figs. 4d,4e,4f, 7d,7e,7f, 8a,8d).
Distribution of serotonergic nerve fibers on the salivary gland acini Whole-mounts of salivary glands were double-labeled with anti-serotonin (red) and BODIPY FL phallacidin (blue), and imaged by confocal microscopy. Each image shows the sum of 8 consecutive optical sections (inter-section distance 0.35 μm), representing a total thickness of 2.8 μm. Serotonergic fibers and fiber endings (white arrowhead) form a network on the acinar surface (a) over the peripheral cells (asterisks). The fibers extend deep into the acini (yellow arrowheads) between the central cells that are identified by short, phallotoxin-labeled microvilli (arrows) on their luminal surface. Scale bar = 50 μm
Serotonergic fibers formed a dense network on the surface of the acinar lobules (Fig. 4a). Fibers branched on the lobule surface and displayed either uniform staining over extended stretches or had an irregular beaded appearance. The former fibers appeared thicker in diameter than the varicose fibers and united into bundles that interlinked adjacent lobules (Fig. 5). Moreover, some of these fiber bundles extended away fom the acinar tissue (data not shown); these may represent branches of the esophageal nerve that innervate the acinar tissue and that may have been ruptured during dissection of the salivary gland complex. Other fiber bundles linked the serotonergic network on the acinar tissue to the salivary nerve, supporting the view that both the stomatogastric nervous system and the subesophageal ganglion contribute to the serotonergic innervation of the glandular tissue [8]. Serial confocal sections (Fig. 4a,4b,4c,4d,4e,4f) or cryostat sections (Fig. 2a) through the acinar lobules demonstrated further that serotonergic fibers were not confined to the tissue surface but extended throughout the acini, forming a dense three-dimensional meshwork. These invading fibers were mostly varicose in appearance and resided either below the peripheral cells, suggesting a location between peripheral and central cells, or were localized far deeper than the peripheral cells, suggesting a position among the central cells.
Serotonergic nerve fibers in nerves that interlink adjacent acini Whole-mounts of salivary glands were double-labeled with anti-serotonin (red) and BODIPY FL phallacidin (blue), and imaged by confocal microscopy. Nerves of large (broad arrows) or small (long arrows) diameter interlink the acinar lobules (asterisks) and contain serotonergic fibers. Scale bar = 50 μm
In addition to serotonergic fibers, the nerves interlinking the acinar lobules contained dopaminergic fibers with varicosities and fiber terminals (Fig. 6). In some regions of these nerves, the dopaminergic fibers ramified extensively and had numerous varicosities (Fig. 6b), suggesting that these structures represent neurohemal organs. Individual dopaminergic fibers of these nerves approached acinar lobules and formed a widely spaced network on the lobule surface (Fig. 7). These acinar-tissue-associated dopaminergic fibers had few varicosities irregularly distributed over their length and sidebranches with terminals on the tissue surface (Fig. 7a). Serial confocal sections through acinar lobules demonstrated that, in contrast to serotonergic fibers, dopaminergic fibers did not invade the acinar tissue but were confined to the surface (Fig. 7a,7b,7c,7d,7e,7f). However, extensive cross-linking by the use of glutaraldehyde as a fixative may have prevented penetration of the antibodies into the tissue, and thus the lack of anti-dopamine-immunoreactive structures within the acinar lobules might have been an artefact. Several lines of evidence indicated that this was not the case. First, immunoreactivity was also confined to the surface of the acinar lobules when anti-dopamine was applied to cryostat sections (Fig. 2c). Second, an identical staining pattern was obtained with anti-dopamine on whole-mounts fixed with a low concentration of glutaraldehyde (0.5%; data not shown), and with anti-TH on whole-mounts prepared by the same glutaraldehyde-free fixation protocol as that used for labeling with anti-serotonin (data not shown). Finally, anti-dopamine-positive fibers could be detected not only on the surface, but also within the tissue of other structures of the salivary gland complex (see below).
Dopaminergic nerve fibers within nerves that extend between acini Whole mounts of salivary glands were double-labeled with anti-dopamine (red) and BODIPY FL phallacidin (blue), and imaged by confocal microscopy. Acinar lobules (asterisks) are connected by nerves of large (broad arrows) and small (long arrows) diameter, containing dopaminergic fibers. In some of theses nerves, the dopaminergic fibers branch extensively and have numerous varicosities (b), suggesting that these sites represent neurohemal organs. Scale bar = 50 μm
Distribution of dopaminergic nerve fibers on the salivary gland acini Whole mounts of salivary glands were double-labeled with anti-dopamine (red) and BODIPY FL phallacidin (blue), and imaged by confocal microscopy. Parameters of image acquisition and data presentation are identical to those of Fig. 4. Dopaminergic fibers and their endings (arrowheads) reside on the acinar surface over and between the peripheral cells (asterisks). The inner part of the acinar lobule with the central cells and the acinar lumen (arrows in f) lacks dopaminergic fibers. Scale bar = 50 μm
Three-dimensional (red-green) views of serotonergic and dopaminergic fibers associated with acinar lobules Salivary glands were double-labeled with BODIPY FL phallacidin (a,d) and anti-serotonin (b) or anti-dopamine (e). Stacks of confocal images were recorded, and three-dimensional reconstructions were made by using Carl Zeiss LSM510 software. c,f: The corresponding images of staining with phallotoxin and with antibody were added (a+b or d+e; the phallotoxin image was multiplied with the factor 0.7 to reduce its intensity) in order to present both staining patterns together. The rectangle in a indicates the area that is presented at higher magnification in c. b,c: A dense network of serotonergic fibers extends throughout the entire acinar tissue. e,f: Dopaminergic fibers, in contrast, form a loose network only on the acinar surface. Scale bar = 50 μm
In conclusion, serotonergic and dopaminergic fibers had a dissimilar distribution over the acinar tissue. These differences between serotonergic and the dopaminergic innervation can be directly visualized in Figure 8, presenting three-dimensional views of the two fiber types associated with the acinar lobules. A striking feature of the serotonergic innervation was its richness not only at the lobule surface, but throughout the entire acini. Dopaminergic fibers, in contrast, were sparse and resided only on the surface of the lobules. Moreover, dopaminergic fibers appeared to form neurohemal organs between acinar lobules.
Serotonergic and dopaminergic nerve fibers along the efferent salivary ducts and the reservoir ducts
Each of the paired reservoir ducts was accompanied by a large salivary nerve with a 5-μm-thick dopaminergic axon residing at its center (Fig. 9g). The dopaminergic axon extended all the way toward the acinar tissue, supporting the conclusion that it provides the only source of dopaminergic innervation of the salivary gland complex [6,8]. Along the salivary nerves, thin dopaminergic fibers branched off the central axon. These varicose fibers either remained in a superficial position within the nerves, or they left the nerves and spread, either individually or in small bundles, over the outer surface of the reservoir duct (Fig. 9h). Some of these dopaminergic fibers extended from the salivary nerve toward the adjacent efferent salivary duct. Here, varicose fibers and fiber terminals formed a widely spaced network on the outer duct surface (Fig. 9e,9f) and also invaded the epithelium, as demonstrated by vertical optical sections through the ducts (Fig. 9f, inset). It must be noted that, although this dopaminergic innervation was found along almost the entire efferent salivary duct, only a minority of the epithelial cells had intimate contact to dopaminergic fibers.
Distribution of dopaminergic fibers on salivary ducts, the reservoir, and the reservoir muscle Summarized views of confocal image stacks through whole-mounts double-labeled with anti-dopamine (red) and BODIPY FL phallacidin (blue). The upper left inset presents a scheme of the various structures examined and outlines the areas shown in a-j. Asterisks in a,c,d,i indicate acinar tissue. a: Small salivary ducts (broad arrows) are mostly without dopaminergic fibers (long arrow). b: A small salivary duct without dopaminergic innervation at higher magnification. c: A dopaminergic fiber approaches a small salivary duct and terminates on the duct surface (arrowhead). d: A dopaminergic fiber (arrow) invades the epithelium of a small duct. A vertical section (inset) through the duct at the position indicated by the line in d demonstrates that the dopaminergic fiber (arrow) resides below the duct surface (broken line). e,f: Dopaminergic fibers (arrows) form a loose network on a large salivary duct and terminate on this structure (arrowheads). The inset in f shows a horizontal confocal section through the duct and visualizes a dopaminergic fiber within the duct epithelium, below the duct surface (broken line). g: The salivary nerve coming from the subesophageal ganglion and extending along the reservoir/salivary duct complex contains a single thick dopaminergic axon. h: On the reservoir duct, small dopaminergic varicose fibers reside superficially within the salivary nerve (arrows) or leave the nerve and extend over the duct surface. i: A loose network of dopaminergic fibers with fiber terminals (arrowheads) covers the reservoir. j: Dopaminergic fibers and terminals (arrowheads) within the reservoir muscle. White scale bars = 100 μm; yellow scale bars = 25 μm
In addition to the large dopaminergic axon, the salivary nerves contained several thin serotonergic fibers in a superficial position and with numerous varicosities (Fig. 10g, inset). Moreover, individual serotonergic fibers left the nerves, spread toward the reservoir ducts and terminated on the duct surface (Fig. 10g). In rare cases, serotonergic fibers could be traced to an efferent salivary duct and appeared to end on this structure (data not shown). The majority of efferent salivary ducts, however, was without serotonergic innervation (Fig. 10f).
Distribution of serotonergic fibers on salivary ducts, the reservoir, and the reservoir muscle The upper left inset indicates the structures shown in a-j. a-c, f-j: Summarized views of confocal image stacks through whole-mounts, double-labeled with anti-serotonin (red) and BODIPY FL phallacidin (blue). Asterisks in a,b,c indicate acinar tissue. a: A dense network of serotonergic fibers is associated with the acini (asterisks), whereas small salivary ducts (broad arrows) are mostly without serotonergic fibers. b: A small salivary duct without serotonergic innervation at higher magnification. c: A small salivary duct with a network of serotonergic fibers (arrows). d,e: Vertical sections through the salivary duct shown in c (planes indicated by white lines), demonstrating that the serotonergic fibers (arrows) reside below the duct surface (broken lines). f: A large salivary duct (broad arrow) without serotonergic innervation. g: The reservoir duct is accompanied by a nerve (arrows and inset) containing several serotonergic fibers. Fibers in a superficial position within the nerve have numerous varicosities (inset). Individual fibers also extend over the reservoir duct and have terminals (arrowheads) associated with this structure. h: A loose network of serotonergic fibers, with their terminals (arrowheads), covers the middle part of the reservoir. i: At the orifice, the reservoir has a relatively dense network of serotonergic fibers on its surface. Note that i is enlarged twofold compared with h. j: The reservoir muscle contains numerous serotonergic fiber terminals (arrowheads). White scale bars = 100 μm; yellow scale bars = 25 μm
Distribution of serotonergic and dopaminergic nerve fibers over small salivary ducts
Most of the salivary duct system upstream of the paired efferent salivary ducts was exclusively, but only locally, innervated by dopaminergic fibers (Figs. 9a,9b,9c,9d). The varicose fibers and fiber terminals formed a loose network on the outer duct surface and often invaded the epithelium (Fig. 9d, inset). On duct segments next to the acinar tissue, however, both dopaminergic and serotonergic varicose fibers extended from the acinar lobules to the duct surface and deep into the epithelium (Fig. 10c,10d,10e). Again, the innervation of these most proximal segments of the duct system was only local, and on the majority of these small salivary ducts close to the acinar tissue, no serotonergic or dopaminergic fibers could be detected at all (Figs. 9a,9b, 10a,10b).
Association of serotonergic and dopaminergic nerve fibers with the reservoir system
The paired reservoirs reside amidst the acinar tissue (Fig. 1a). Nerve fibers that entangled the acinar tissue extended toward the reservoirs, and both serotonergic and, as reported previously [6], dopaminergic fibers were detected on the surface of the reservoirs. However, we noticed differences in the distribution of serotonergic and dopaminergic fibers over this structure. Dopaminergic fibers branched and formed a loose network over the entire reservoir. These fibers had a varicose morphology and terminated on the reservoir (Fig. 9i). The serotonergic innervation pattern, in contrast, varied along the reservoir. On the distal half of the reservoir, serotonergic fibers appeared to be absent (data not shown). Its middle section had a loose network of varicose serotonergic fibers on the outer surface (Fig. 10h); these fibers were linked to acinar-tissue-associated serotonergic fibers via small nerves, indicating that they originated in the stomatogastric nervous system and/or the subesophageal ganglion. Finally, the basal part of the reservoir next to the orifice into the reservoir duct had a relatively dense network of varicose serotonergic fibers and fiber terminals on its surface (Fig. 10i). These serotonergic fibers could be traced back directly to the salivary nerve accompanying the reservoir duct, suggesting that they originated in the subesophageal ganglion.
The reservoir muscle is attached near the orifice of each reservoir [3]. Both serotonergic and dopaminergic fibers were detected within this muscle (Figs. 9j, 10j). The fibers branched extensively and had numerous varicosities and nerve terminals.
Discussion
In the present study, serotonergic and dopaminergic nerve fibers were identified by immunolabeling of the cockroach salivary gland complex with anti-serotonin/anti-dopamine antibodies and confocal fluorescence imaging. The results of these analyses are summarized in a schematic manner in Figure 11 and demonstrate that:
Schematic presentation of the distribution of serotoneric and dopaminergic fibers over the salivary gland complex The salivary gland is innervated by the salivary nerve (1) and via branches of the esophageal nerve (2). The salivary nerve accompanies the reservoir duct and contains one thick dopaminergic axon. Moreover, several serotonergic fibers ramify within the nerve and have numerous varicosities. Nerves containing numerous dopaminergic varicose fibers (3) link the acinar lobules and may function as neurohemal organs.
1. Serotonergic varicose fibers are associated with the lobule surface and invade each acinus to form a dense network over and within the entire acinar lobule. Thus, serotonergic varicosities and fiber terminals are found next to peripheral cells and central cells.
2. Dopaminergic varicose fibers form a loose network only on the surface of the acinar tissue, being closely positioned to peripheral cells.
3. In nerves interlinking adjacent acinar lobules, dopaminergic fibers ramify extensively and have numerous varicosities (Fig. 11, #3), suggesting that these structures represent sites for the neurohemal release of dopamine.
4. Segments of the salivary duct system immediately adjacent to the acini are sparsely innervated by both serotonergic and dopaminergic fibers. Segments of the duct system further downstream are exclusively associated with dopaminergic fibers. These fibers reside on the outer surface of the ducts and invade the epithelium where they terminate between the duct cells.
5. The entire reservoir system, composed of the reservoir, the reservoir duct, and the reservoir muscle, is innervated by dopaminergic and serotonergic fibers. Within the salivary nerve along the reservoir duct, these serotonergic fibers branch and form varicosities (Fig. 11, #1), suggesting that this portion of the nerve serves the neurohemal release of serotonin.
Innervation of the acinar tissue
The innervation of the cockroach salivary gland has been investigated previously by light microscopy of methylene-blue stained preparations and by electron microscopical techniques [5,7]. These studies have established that the salivary gland receives innervation via the salivary nerves emerging form the subesophageal ganglion and via the esophageal nerves of the stomatogastric nervous system. By labeling with anti-serotonin and anti-TH, evidence has been provided that each of the paired salivary nerves contains a single dopaminergic axon and several thin serotonergic axons, whereas the stomatogastric nervous system provides only serotonergic innervation of the salivary gland [6,8,9]. Although the focus of the present study has not been on the origin of the innervation, our results are generally in agreement with the conclusions of the aforementioned studies. The acinar tissue thus appears to have a dual innervation by serotonergic fibers, whereas dopaminergic innervation is provided by the salivary nerve only. This situation raises the question as to whether serotonergic fibers of stomatogastric and of subesophageal origin have a different distribution over the acinar tissue, or in other words, whether they innervate different cell types. However, because of the density and complexity of the serotonergic fiber network associated with the acinar tissue, individual fibers could not be traced back to their source, and therefore this question has to remain unanswered for now.
Over and within the acinar lobules, serotonergic fibers form a dense three-dimensional plexus with numerous varicosities. On the acinar surface, each peripheral cell appears to have a neighboring serotonergic fiber. Likewise, although we have no marker at hand that permits the identification of individual central cells, the density of the serotonergic fiber network within the acinar tissue suggests that every central cell has immediate contact to a serotonergic fiber. The bouton-like structures along these fibers possibly represent sites of neurotransmitter release, similar to the situation at the Drosophila neuromuscular junction [14]. Moreover, by transmission electron microscopy of cockroach acinar tissue, axonal profiles with numerous synaptic vesicles have been observed not only on the acinar surface but also embedded between central cells [5,15]. At these sites, the axonal profiles are without glial wrappings and occasionally have an electron-dense plaque on the axonal membrane, indicating an active zone. Finally, preliminary data suggest that the serotonin-positive varicosities as well as the dopamine-positive varicosities associated with the acinar tissue colocalize with a marker for synapses (O. Baumann, D. Kühnel, P. Dames and B. Walz, in preparation). It may be concluded that serotonin is liberated both on the surface of the acini, next to each peripheral cell, and deep within the acini, next to each central cell.
Physiological studies have demonstrated that central cells are responsive to serotonin, and that serotonin application stimulates the exocytosis of secretory granules [4]. For peripheral cells, in contrast, direct evidence for a physiological response to serotonin is lacking. The close spatial relationship of serotonergic varicosities to peripheral cells however indicates that serotonin also acts on this cell type. We suggest that serotonin stimulates electrolyte and water transport across peripheral cells in order to flush the secretory products of the central cells out of the acini.
Dopaminergic fibers are confined to the surface of the acini and form a relatively loose network. Thus, not every pair of peripheral cells has a dopaminergic varicose fiber in its immediate vicinity. Moreover, within the nerves interlinking adjacent acinar lobules, dopaminergic fibers ramify extensively and have numerous swellings, indicating that these structures serve the neurohemal release of dopamine. This confinement of dopaminergic fibers to the periphery of the acinar lobules is in agreement with the suggestion that only the peripheral cells are sensitive to dopamine [4]. The sparsity of dopaminergic fibers in association with the acinar tissue and the presence of putative sites of neurohemal release of dopamine suggest further that dopamine acts not as a neurotransmitter, but is released into the hemolyph to function as a paracrine substance or neurohormone.
Dopaminergic neurohemal organs have not been described in insects so far, whereas serotonergic, octopaminergic, histaminergic and peptidergic neurohemal organs seem to be quite common in the peripheral nervous system [e.g., [9,16-18]]. It must be admitted, however, that the presence of varicose fibers detected by light microscopy can only be taken as an indication of neurosecretion, and thickened fiber sites could also result from an accumulation of cell organelles, such as mitochondria. Unequivocal identification of these structures as neurohemal organs requires confirmation by use of other techniques. Therefore, a detailed analysis of the distribution of a synapse-specific protein and of the ultrastructure of the fibers associated with salivary gland complex is in progress (O. Baumann, D. Kühnel, P. Dames and B. Walz, in preparation). Preliminary data suggest an enrichment of a synapse-specific protein within these varicosites, providing further support for the conclusion that these structures serve as neurohemal organs.
Innervation of the salivary duct system
Although the innervation of the cockroach salivary gland has been studied previously by various techniques, an association of nerve fibers with the salivary duct system has not been reported so far, except for the paired efferent salivary ducts [5]. The reason for this may be that smaller duct segments are embedded between the acinar lobules and are thus not immediately accessible to conventional light-microscopic techniques, and that nerve fibers are sparse along the duct system and therefore detectable by electon microscopy in serial sections only. By confocal fluorescence microscopy, however, it is possible to determine the exact spatial relationship between fluorescently labeled fibers and the duct epithelium.
Dopaminergic fibers are present over the entire length of the duct system but innervate only small areas. Thus, only a small number of duct epithelial cells resides in close apposition to dopaminergic fibers. Surprisingly, rather than remaining on the outer epithelial surface, varicose fibers invade the epithelium, suggesting that dopamine is released deep within the epithelial layer.
Physiological studies have demonstrated that the duct epithelial cells are responsive to dopamine [11,12]. However, how do all duct cells become stimulated when only a fraction of them has intimate contact to dopaminergic varicosities? One possibility is that the putative neurohemal structures at the acinar periphery represent the major source of dopamine acting on the salivary duct cells. An alternative, but not exclusive, possibility is that direct stimulation of only a few epithelial cells is sufficient to activate ion transport mechanisms in the entire epithelium, because the cells are extensively coupled by gap junctions [19], and second messengers may diffuse through gap junctions from the dopamine-activated cells to their neighbors. This suggestion is directly supported by ratiometric imaging of dopamine-induced spatiotemporal intracellular Ca2+ changes in the salivary duct epithelial cells loaded with Fura-2. Dopamine stimulates a Ca2+ elevation in duct cells at several points along the ducts, and from there, the increase in intracellular Ca2+ spreads over the duct as a Ca2+ wave at a velocity of 3.7 μm s-1[11].
The presence of serotonergic varicose fibers on some duct segments may seem to contradict the results of previous physiological studies, showing that duct cells are unresponsive to serotonin [11]. However, serotonergic innervation is restricted to segments immediately adjacent to the acini and to the efferent salivary duct. Our physiological studies on the cockroach salivary duct, in contrast, have been performed on areas in between these segments [11] and, thus, on areas that are only associated with dopaminergic fibers. The identification of serotonergic varicose fibers only on distinct segments of the duct system indicates that the various segments differ in their properties and functions. This hypothesis is in line with results on the morphological characteristics of the duct segments. Whereas secretory granules have been detected in the duct cells next to the secretory acini, cells in the major portion of the duct system lack granules but have an extensive basal labyrinth and numerous mitochondria [1,20].
The reservoir complex – innervation and possible functions
The functions and the physiology of the reservoir system are still enigmatic. It has been demonstrated that ligation of the salivary ducts prevents the filling of the reservoirs [3], suggesting that the acinar tissue is the source of at least part of the reservoir content, and that the reservoirs may become filled by back pressure of the secreted fluid when the hypopharynx is closed. The contracted reservoir muscle may serve as an occlusor of the reservoir orifice, and when the muscle relaxes, pressure of the hemolymph on the reservoir walls may cause emptying of the reservoirs [3]. In this scenario, the reservoir would play a primarily passive role and serve as a storage compartment for watery saliva. The reservoir content may be released during ingestion in order to moisten and digest the food [3]. Moreover, the reservoir may have some osmoregulatory function and satisfy the water requirements of the animal in times of water shortage.
The present study demonstrates that both serotonergic and dopaminergic varicose fibers are associated with the reservoir wall and the reservoir duct, and that the pattern of serotonergic innervation varies over the length of these structures. These findings indicate that the reservoir and the adjoining duct serve not only as a passive storage compartment or passageway, respectively, but have some active functions that may be regulated by dopamine and serotonin. For example, the epithelium of the reservoir wall may modify the composition of the primary fluid made within the glandular tissue. In agreement with this hypothesis is the finding that creatinine and urea have been detected in the contents of the reservoir but not in homogenized glandular tissue, suggestive of an excretory function for the reservoir [20]. Moreover, the epithelial cells of the reservoir wall are intensely stained for Na+,K+-ATPase, indicating that these cells are active in ion transport across the reservoir wall (W. Blenau and O. Baumann, unpublished results). Preliminary results suggest further that not only the serotonergic innervation but also the cellular architecture varies along the length of the reservoir (W. Blenau and O. Baumann, unpublished results), supporting the view that the various regions of the reservoir differ in their physiological properties.
Innervation of the salivary gland complex by other sources
Electron microscopy [5] and immunofluorescence staining with a neuron-specific marker (our unpublished data) visualized that the salivary nerve contains the axons of the giant neurons SN1 and SN2, and several small axons. The present study confirms that one of the large axons (SN1) is dopaminergic, and that most, if not all, small axons are serotonergic [6,8]. The second large axon (SN2) must thus contain a different, yet unidentified neurotransmitter or neurohormone, and serotonergic and dopaminergic neurons do not provide the only innervation of the salivary gland complex. Moreover, we should not dismiss the possibility that the salivary gland complex is innervated by neurons that are located in other parts of the nervous system than the subesophageal ganglion and that contain neither dopamine nor serotonin. In locusts, evidence has been presented that neuronal processes with FMRFamide-related peptides extend from the prothoracic and mesothoracic ganglia via transverse nerves to the salivary glands and ramify over the acinar tissue [8,21]. The physiological roles of FMRFamide-related peptides in this system are unknown; it has been proposed that these neurotransmitters may modulate rather than activate salivation in locust salivary glands [21]. In order to obtain a complete view on the innervation pattern and the neuronal control of salivation in the cockroach, several issues are yet to be solved: (1) the neurotransmitter content of the SN2 neuron, (2) the spatial relationship of the SN2 axon terminals to the various cell types, (3) the functional role of SN2 in salivation, and (4) the possibility of innervation by other sources.
Conclusions
Earlier research in our laboratory established the importance of serotonin and dopamine in salivation by the cockroach salivary gland. The present data extend these findings by determining the exact spatial relationship of the serotonergic and dopaminergic fiber endings and varicosities to the various structures and cell types composing the salivary gland complex. Close apposition of fiber terminals and/or varicosities to a distinct cell type provides evidence for a selective innervation of the respective cell type.
The distribution pattern of serotonergic and dopaminergic varicose nerve fibers over and within the acinar tissue supports the concept that central cells are stimulated by serotonin only, whereas peripheral cells are responsive to both serotonin and dopamine. The salivary duct system, previously thought to be regulated by dopamine only, may vary in functions along its length, as the initial acinar-close segments have a dopaminergic and a serotonergic innervation. Finally, the finding of a complex serotonergic and dopaminergic innervation pattern of the reservoirs, the adjoining reservoir ducts and the reservoir muscles warrants further investigations of the physiology of these structures.
Materials and MethodsAnimals and preparation
A colony of the American cockroach (Periplaneta americana) was maintained at 25°C under a 12-h light:12-h dark regime and with free access to food and water. Young male and female imagines were sacrificed, and the salivary glands were dissected under physiological saline (160 mM NaCl, 10 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 10 mM glucose, 10 mM TRIS, pH 7.4), as described previously [2].
Antibodies
Anti-serotonin was obtained from Sigma (Taufkirchen, Germany; Product No. S5545); this antiserum was made in rabbits against serotonin conjugated to bovine serum albumin. Anti-dopamine, raised in goats against glutaraldehyde-conjugated dopamine, was provided by H.W.M. Steinbusch (Maastricht University, Maastricht, The Netherlands). Affinity-purified rabbit antibody against rat TH was purchased from Chemicon (Temecula, CA; Product No. AB152). This antibody has been reported previously to cross-react with TH of an invertebrate, Aplysia[22]. Secondary antibodies conjugated to Cy3 or Cy5 were obtained from Rockland (Gilbertsville, PA) and Dianova (Hamburg, Germany).
Fixation protocols
For labeling with anti-serotonin, salivary glands were fixed for 2 hours at room temperature with 2% paraformaldehyde, 0.075% lysine-HCl, 10 mM Na-periodate in 0.1 M phosphate buffer (PB), pH 7.0 [10,23]. Specimens were washed for 10 minutes in PB and treated further as described below.
For labeling with anti-dopamine, salivary glands were fixed for 30 minutes on ice with 5% glutaraldehyde in PB supplemented with 10 mM ascorbic acid (PB/AA). For colabeling with anti-dopamine and anti-TH, 0.5% glutaraldehyde, 3% paraformaldehyde in PB/AA was used as a fixative. After fixation, specimens were washed for 10 min on ice in PB/AA, treated for 30 min with 0.5% sodium borohydride in PB/AA to reduce free aldehyde groups, and washed again for 10 min in PB/AA.
Immunofluorescence labeling
Fixed salivary glands were either directly used for immunolabeling or processed for cryostat sectioning. In the latter case, preparations were incubated with 10% sucrose in PB or PB/AA for 30 minutes on ice, infiltrated with 25% sucrose in PB or PB/AA overnight at 4°C, and then shock-frozen in melting isopentane. Sections (8–10 μm thick) were cut at -30°C in a cryostat, collected on poly-L-lysine-coated coverslips, air-dried, and stored at 4°C until use.
For labeling with anti-serotonin, salivary glands or cryosections were permeabilized with 0.01% Tween 20 in PBS, reacted with 50 mM NH4Cl in phosphate-buffered saline (PBS), washed in PBS, and blocked with 1% normal goat serum, 0.8% bovine serum albumin, 0.1% fish gelatine, and 0.5% Triton X-100 in PBS. After being labeled overnight at 4°C with anti-serotonin (diluted 1:10,000 in the above blocking solution), specimens were washed in PBS and reacted for 1 hour (cryostat sections) or 3 hours (whole-mounts) with Cy3-conjugated goat anti-rabbit-IgG. In case of whole-mount preparations, the F-actin probe BODIPY FL phallacidin (Molecular Probes, Eugene, OR) was added to the secondary antibody solution. After a final extensive wash in PBS, specimens were mounted in Mowiol 4.88 (Farbwerke Hoechst, Frankfurt, Germany), containing 2% n-propyl-gallate as an anti-fading reagent.
For labeling with anti-dopamine, entire salivary glands or cryosections were permeabilized and blocked with a solution consisting of 1% normal donkey serum, 0.8% bovine serum albumin, 0.1% fish gelatine, and 0.5% Triton X-100 in PBS supplemented with 10 mM ascorbic acid (PBS/AA). Preparations were then labeled overnight at 4°C with anti-dopamine (diluted 1:8,000 in blocking solution), washed extensively with PBS/AA, and reacted with Cy3-conjugated donkey anti-goat-IgG and (in the case of whole-mounts) BODIPY FL phallacidin in PBS/AA. For double-labeling of whole-mount preparations with anti-dopamine and anti-TH, both primary antibodies were applied together (anti-TH diluted 1:200); the tissue was then washed, incubated with Cy3-conjugated donkey anti-goat-IgG, washed again, and reacted with Cy5-conjugated goat anti-rabbit IgG and BODIPY FL phallacidin.
Confocal microscopy
Specimens were examined with a Zeiss LSM 510 confocal laser-scanning microscope (Carl Zeiss, Jena, Germany) equipped with a 488-nm Argon laser, a 543-nm Helium-Neon laser, a 633-nm Helium-Neon laser, and differential interference contrast optics. Low-magnification images were recorded with a Fluar 10x/0.5, images at higher magnification either with a Plan-Neofluar 40x/1.4 or with a C-Apochromat 40x/1.2 W. In the case of double-labeled specimens, BODIPY FL and Cy3 were excited sequentially with the 488-nm and the 543-nm laser lines, by using the multitracking function of the LSM 510 software, and detected through 505–530-nm bandpass and 560-nm longpass filters. In the case of triple-labeled specimens, BODIPY FL and Cy5 were excited simultaneously at 488 nm and 633 nm and detected through 505–550-nm bandpass and 650-nm longpass filters; subsequently, Cy3 fluorescence was imaged through a 560–615-nm bandpass filter by using the 543-nm laser line for illumination. Specimens labeled with only one fluorochome and viewed with the instrument settings as used for double- or triple-labeled preparations demonstrated that there was no bleed-through between the detector channels under these recording conditions.
Controls for labeling specificity
Specificity of antibody binding was checked by treating cryostat sections in the described fashion except that primary antibodies were omitted from the procedure. No fluorescence was detected when these control specimens were viewed under the same instrumental settings as used for imaging sections that had been reacted with primary antibody. As a further control, primary antibody solutions were supplemented with 1 mg/ml dopamine or 1 mg/ml serotonin, preincubated for 30 minutes, and then used for immunofluorescence labeling of cryostat sections.
Abbreviations used
PB – phosphate buffer
PB/AA – phosphate buffer with ascorbic acid
PBS – phosphate-buffered saline
PBS/AA phosphate-buffered saline with ascorbic acid
SN1 – salivary neuron 1
SN2 – salivary neuron 2
TH – tyrosine hydroxylase
Authors' contributions
O.B. participitated in the design and coordination of the study, carried out part of the image analysis, and drafted the manuscript. P.D. and D.K. carried out the immunolabeling experiments and did the confocal imaging of the specimens. B.W. proposed the project and participated in the design and coordination of the study.
Acknowledgements
We should like to thank Dr. Wolfgang Blenau and Mrs. Katja Rietdorf for the maintenance of the cockroach culture and for stimulating discussions, and Mrs. Bärbel Wuntke for technical assistance.
These studies investigate the role of mitoKATP channels, protein kinase C (PKC) and Mitogen activated protein kinase (p38MAPK) on the cardioprotection of ischemic (IP) and pharmacological preconditioning (PP) of the human myocardium and their sequence of activation.
Results
Right atrial appendages from patients undergoing elective cardiac surgery were equilibrated for 30 min and then subjected to 90 min of simulated ischemia followed by 120 min reoxygenation. At the end of each protocol creatinine kinase leakage (CK U/g wet wt) and the reduction of MTT to formazan dye (mM/g wet wt) were measured. Similar protection was obtained with α1 agonist phenylephrine, adenosine and IP and their combination did not afford additional cardioprotection. Blockade of mitoKATP channels with 5-hydroxydecanoate, PKC with chelerythrine, or p38MAPK with SB203580 abolished the protection of IP and of PP. In additional studies, the stimulation of mitoKATP channels with diazoxide or activation of PKC with PMA or p38MAPK with anisomycin induced identical protection to that of IP and PP. The protection induced by diazoxide was abolished by blockade of PKC and by blockade of p38MAPK. Furthermore, the protection induced by PMA was abolished by SB203580 but not by 5-hydroxydecanoate, whereas the protection induced by anisomycin was unaffected by either 5-hydroxydecanoate or chelerythrine.
Conclusions
Opening of mitoKATP channels and activation of PKC and p38MAPK are obligatory steps in the signal transduction cascade of IP and PP of the human myocardium with PKC activation being downstream of the opening of mitoKATP channels and upstream of p38MAPK activation.
Background
Ischemic preconditioning (IP) is a powerful protective endogenous adaptive response of the heart against a prolonged ischemic insult. [1,2] However, the application of IP requires a physical cut of the blood supply which can be difficult or impractical in many clinical situations. A way to circumvent these potential problems associated with the clinical application of IP may be preconditioning by pharmacological (PP) means or manipulation of the signalling pathway involved in the protection. A number of membrane receptors are involved in the phenomenon of IP including α1[3-5] and β-adrenoceptors, [6] opioid [7] and adenosine A1 and A3 receptors. [8,9] Other factors such as heat shock proteins, [10] bradykinin, [11] calcium [12] and nitric oxide synthase activity [13] have also been shown to participate in the protection of IP, however, whether the various forms of PP share the same molecular mechanism with IP is not fully elucidated.
The intracellular sequence of events that translate the binding of the various agonists to their membrane receptors into the protection of preconditioning remains under intense investigation. It has been reported that α1-adrenoceptors are coupled with protein kinase C (PKC) through phospholipase activity, [14] that in turn activate p38 Mitogen Activated Protein Kinase (p38MAPK) in some cardiac preparations. [15] ATP sensitive potassium channels have also been implicated in the signal transduction mechanism of IP, [16,17] and recent evidence from several investigators [18-20] including ourselves [21] has shown that the mitochondrial and not the sarcolemmal KATP channels are involved. The order of involvement of the above mediators remains controversial although recently it has been suggested that mitochondrial KATP channels are the triggers in the signal transduction mechanism rather than the end effectors. [18]
The aims of the present series of studies were to investigate the efficacy of pharmacological preconditioning of the human myocardium with α1-adrenoceptor and adenosine receptor agonists as compared to ischemic preconditioning and to elucidate the contribution and sequence of activation of PKC, p38MAPK and mitoKATP channels.
Results
All samples entering the studies completed the applied protocol and were included in the analysis.
Ischemic versus pharmacological preconditioning (Study 1)
Figures 4A and 4B demonstrate that SI/R alone resulted in a significant increase in CK leakage and decrease in MTT reduction when compared to the aerobic controls. They also show that PP with phenylephrine or adenosine administered prior to SI/R is as protective as IP and that their use in combination does not result in additive protection.
Protocol for Study 1 to investigate the efficacy of preconditioning via α1-adrenoceptors or adenosine (A) receptors alone and in combination with IP (n = 6 specimens/group): (i) time-matched aerobic control, (ii) SI/R alone, (iii) IP induced with 5 min of SI/ 5 min R before SI, (iv) phenylephrine (P) for 5 min and 5 min washout (W) before SI, (v) adenosine (A) for 5 min and 5 min washout (W) before SI, (vi) phenylephrine (P) for 5 min and 5 min washout (W) before IP, (vii) adenosine (A) for 5 min and 5 min washout (W) before IP and (viii) adenosine (A) for 5 min and 5 min washout (W) followed by phenylephrine (P) for 5 min and 5 min washout (W) prior to SI.
Protocol for Study 2 to investigate the role of PKC, p38MAPK and mitoKATP channels in the cardioprotective effect of IP and PP with phenylephrine (P) or adenosine (A) (n = 6 specimens/group): (i) time-matched aerobic control, (ii) SI/R alone, (iii) IP alone prior to SI, (iv) chelerythrine (CHE) for 10 min prior to IP, (v) SB203580 (SB) for 10 min prior to IP, (vi) 5-hydroxydecanoate (5-HD) for 10 min prior to IP, (vii) phenylephrine (P) for 5 min and 5 min washout (W) before SI, (viii) chelerythrine (CHE) for 10 min prior to phenylephrine (P) for 5 min and 5 min washout (W), (ix) SB203580 (SB) for 10 min prior to phenylephrine (P) for 5 min and 5 min washout (W), (x) 5-hydroxydecanoate (5-HD) for 10 min prior to phenylephrine (P) for 5 min and 5 min washout (W), (xi) adenosine (A) for 5 min and 5 min washout (W) before SI, (xii) chelerythrine (CHE) for 10 min prior to adenosine (A) for 5 min and 5 min washout (W), (xiii) SB203580 (SB) for 10 min prior to adenosine (A) for 5 min and 5 min washout (W) and (xiv) 5-hydroxydecanoate (5-HD) for 10 min prior to adenosine (A) for 5 min and 5 min washout (W).
Study 3 protocol designed to elucidate the sequence of the participation of involvement of mitoKATP channels, PKC and p38MAPK in the signal transduction cascade of cardioprotection. For this purpose, in addition to the (i) aerobic time-matched control and (ii) SI/R alone the following groups were studied (n = 6 specimens/group): (iii) diazoxide alone for 10 min prior to SI, (iv) chelerythrine (CHE) for 20 min with diazoxide added for the last 10 min prior to SI, (v) SB203580 (SB) for 20 min with diazoxide added for the last 10 min prior to SI, (vi) PMA alone for 10 min prior to SI, (vii) SB203580 (SB) for 20 min with PMA added for the last 10 min before SI, (viii) 5-hydroxydecanoate (5-HD) for 20 min with PMA added in the last 10 min before SI, (ix) anisomycin (ANS) alone for 10 min prior to SI, (x) chelerythrine (CHE) for 20 min with anisomycin (ANS) added for the last 10 min prior to SI and (xi) 5-hydroxydecanoate (5-HD) for 20 min with anisomycin (ANS) added for the last 10 min prior to SI.
Creatine Kinase (CK) leakage into the media (A) during the 120 min reoxygenation period and MTT reduction by the slices (B) at the end of the reoxygenation period in human atrial myocardium subjected to various protocols (see Figure 1) to investigate the efficacy of preconditioning via α1-adrenoreceptors or adenosine receptors alone and in combination with IP (Study 1). Data are expressed as mean ± SEM of six experiments. *p < 0.05 vs SI/R alone group.
Role of PKC, p38MAPK and mitoKATP channels in IP and PP (Study 2)
Figures 5A and 5B show that both IP and PP with phenylephrine or adenosine are equally abolished by the PKC antagonist chelerythrine, the p38MAPK inhibitor SB203580 or the mitoKATP blocker 5-hydroxydecanoate suggesting that these steps are necessary in the signal transduction cascade of preconditioning.
Creatine Kinase (CK) leakage into the media (A) during the 120 min reoxygenation period and MTT reduction by the slices (B) at the end of the reoxygenation period in human atrial myocardium subjected to various protocols (see Figure 2) to investigate the role of PKC, p38MAPK and mitoKATP channels in ischemic and pharmacological preconditioning with phenylephrine (P) or adenosine (Study 2). Data are expressed as mean ± SEM of six experiments. *p < 0.05 vs. SI/R alone group. (5-HD: 5-hydroxydecanoate, SB: SB203580).
Sequence of activation of PKC, p38MAPK and mitoKATP channels (Study 3)
The results on CK leakage and MTT reduction shown in Figures 6A and 6B demonstrate that identical protection is obtained with the mitoKATP channel opener diazoxide, the PKC activator PMA and the p38MAPK activator anisomycin. Importantly, they also show that whilst the protection of diazoxide is abolished by the PKC antagonist chelerythrine and the p38MAPK antagonist SB203580, the protective effect of PMA is abolished by SB203580 but not by the mitoKATP channel blocker 5-hydroxydecanoate, and the protective action of anisomycin is unaffected by chelerythrine and 5-hydroxydecanoate.
Creatine Kinase (CK) leakage into the media (A) during the 120 min reoxygenation period and MTT reduction by the slices (B) at the end of the reoxygenation period in human atrial myocardium subjected to various protocols (see Figure 3) to investigate the sequence of activation of the mediators of pharmacological and ischemic preconditioning (Study 3). Data are expressed as mean ± SEM of six experiments. *p < 0.05 vs SI/R alone group. (CHE: chelerythrine, 5-HD: 5-hydroxydecanoate, SB: SB203580, PMA: phorbol-12-myristate-13-acetate).
Discussion
The present studies have demonstrated that protection with PP by activation of α1-adrenoreceptors or adenosine receptors is identical to that of IP in the human myocardium. In addition, they have shown that mitoKATP channels, PKC and p38MAPK are an integral part of the cellular signal transduction involved in this cardioprotection in which mitoKATP channels are placed upstream and p38MAPK placed downstream of PKC. These studies provide novel information to understand the underlying mechanism of protection by preconditioning of the human myocardium and the results have obvious clinical importance that warrants further discussion.
Pharmacological versus ischemic preconditioning
The demonstration that activation of α1-adrenoreceptors with phenylephrine is as protective as IP of the human myocardium in these studies is supported by other results in the rat heart [3,5] but contrast with the results reported by Cleveland et al [4] in the human myocardium. The differing results may be due to the use of 50 μM phenylephrine by Cleveland et al, [4] a concentration 500 times the one used in our studies. Using an identical preparation, we have recently shown [22] that phenylephrine exerts maximal protection at 0.1 μM and that protection is lost at concentrations ≥ 10 μM. It should also be pointed out that the human trabeculae preparation used by Cleveland et al [4] differed from our preparation in that muscles were electrically stimulated and subjected to only 45 min of simulated ischemia which may be another potential explanation.
The protection induced by adenosine was also identical to that of IP and the use of the two in combination did not result in additional benefit up and above that seen with each intervention alone. These results in the human myocardium are supported by studies in other animal species [23] but contrast with the results reported by Leesar et al [24] in the vivo human heart and by McCully et al [25] in the isolated rabbit heart. In the study by Leesar et al [24] patients undergoing percutaneous transluminal coronary angioplasty were subjected to three 2-minute balloon inflations 5 minutes apart. Under these conditions the administration of adenosine was more effective in limiting ST-segment shift than the balloon inflation protocol. Although we have previously shown [2] that 4 to 5 min of ischemia are sufficient to precondition the human myocardium, independent of whether this time is attained with one or two cycles of ischemia, it is conceivable that in their study [24] the balloon inflation protocol may not have been sufficient to induce enough ischemia to trigger preconditioning. This is a strong possibility under clinical conditions where collateral flow may lessen the severity of ischemia. Furthermore, in that study [24] changes in ST-segment, that are modulated by sarcolemmal KATP channels, [26] were used as the main end-point and it is now recognized that the protection of preconditioning may be mediated by mitochondrial rather than by sarcolemmal KATP channels. [18-21] The results reported by McCully et al [25] that adenosine is more potent than and extends the cardioprotection of IP in the rabbit heart are difficult to explain because they used a protocol of 5 min ischemia/5 min reperfusion for preconditioning which is identical to the one used in our studies and shown to afford optimal protection in other preparations. [5,9] However, since adenosine and IP use identical cellular signal transduction mechanism to induce protection, the most likely explanation of their results may be the presence of some unknown factor that may have influenced the severity of the IP insult and therefore its protective efficacy.
Mechanism of preconditioning
The elucidation of the factors involved in the signal transduction pathway of preconditioning has been the subject of intense investigation and although the participation of factors such as PKC [17,19,27] and mitoKATP channels [16-21] is well established, their relevance and the sequence of activation remains controversial. The present studies are the first to demonstrate that PP and IP of the human myocardium share identical signal transduction cascade that involves mitoKATP channels, PKC and p38MAPK in that order. Wang et al [28] showed that PKC inhibition with chelerythrine or calphostin C completely abolished the beneficial effects of diazoxide in the isolated rat heart, thus providing further support that mitoKATP channels are upstream of PKC. However, Pain et al [18] and Miura et al [29] using chelerythrine and calphostin C to block PKC were unable to confirm this observation in the rabbit, suggesting that these differences could arise as a result of different animal species. In spite of this, it is interesting to note that Pain et al [18] reported that genistein, a tyrosine kinase antagonist, blocked the protection induced by diazoxide which indicates that activation of kinases also lies downstream of mitoKATP channels in the rabbit heart.
Our demonstration that activation of any one of the components of the transduction cascade investigated in our studies (mitoKATP channels, PKC, p38MAPK) can provide identical protection and that blockade of any of them individually completely abolishes protection indicates that in the human myocardium there is only one pathway of protection by preconditioning. The failure to obtain additional protection when more than one agent was used to induce PP or when these agents were used in combination with IP further support this thesis. But again, the mechanism of preconditioning the human myocardium may not be applicable to all species as suggested by the need to combine the inhibition of PKC and tyrosine kinase to abort the protection of preconditioning in pigs. [30]
The molecular interactions between the various components of the preconditioning pathway are not well understood. There is evidence that mitoKATP channel opening increases radical oxygen species production which in turn may activate PKC, [18] but it is well known that mitoKATP channels are also modulated by PKC [19]. The above and the realization of selective translocation of PKC with preconditioning, [27,28,31] suggest that PKC may play a dual role upstream and downstream of mitoKATP channels. Our observation that blockade of PKC abolishes the protection by diazoxide does not eliminate that possibility.
Our results have shown that activation of p38MAPK is a crucial step in the transduction pathway of preconditioning in the human myocardium. Activation of p38MAPK has also been connected to preconditioning in the rabbit heart, [32] however the relationship could not be established in rat [33] and pig [34] hearts. Therefore, it seems that once more the components of the signal transduction pathway of preconditioning are species-dependent.
It is still unknown whether the activation of p38MAPK is the last step of the transduction cascade that phosphorylates the end-effector and whether there is a simple or multiple effectors and their location. p38MAPK can phosphorylate a wide range of proteins, some of which may be potential candidates for end-effectors of preconditioning. Thus, for example, the low molecular weight heat shock protein HSP27 may be phosphorylated by p38MAPK via the intermediate MAPKAPK2 [35] and this may lead to polymerization of actin [36] and to increase tolerance of the cytoskeleton to stress. [37] Translocation of PKC isoforms to mitochondrial sites, intercalated discs and nucleus may suggest that p38MAPK activation in these places may activate enzymes involved with energy production, intercellular communication through cell junctions or gene transcriptions.
Our results suggest that mitoKATP channels are not the end-effectors of cardioprotection by preconditioning of the human myocardium. The concept that mitoKATP channels may be the end-effectors of preconditioning has been based on the efficacy of diazoxide, a highly selective mitoKATP channel opener, to mimic cardioprotection by preconditioning [20,21] and the blockade of this protection by 5-hydroxydecanoate, [20,21] a specific mitoKATP channel blocker; however, it was never clear whether the actions of these channels were limited to the mitochondria or were part of a more complex signal transduction cascade with effects on other cellular structures. The effect of opening the mitoKATP channels remains controversial and whereas some investigators have reported large decreases in mitochondrial membrane potential affecting respiration and resulting in a reduction in Ca2+ uptake into mitochondria, [38] others have observed little effect on membrane potential, bioenergetics or Ca2+ uptake but important changes on matrix and inter-membrane space volumes. [39] The reason given for these discrepancies was the use of high doses of mitoKATP channel openers in the former studies [39] but this does not clarify how changes in mitochondrial volume are connected to PKC activation, that is downstream in the signal transduction cascade.
The diagram in Figure 7 describes our proposal of the signal transduction cascade of preconditioning in the human myocardium. It is suggested that upon activation of sarcolemmal receptors (e.g. adenosine receptors, α1-adrenoreceptors) mitoKATP channels are opened via Gi proteins and PKC, possibly PKC-δ [28]. The opening of mitoKATP channels will activate PKC possibly via the production of radical oxygen species [18]. PKC may then translocate to various cellular sites including the mitochondria, sarcolemma and intercalated discs, and nucleus, a phenomena that may involve specific PKC isoforms, where p38MAPK will be activated. In turn, p38MAPK may activate a simple or multiple end-effectors directly or via MAPK intermediates. It is clear that more studies are required to fully elucidate the signal transduction pathway of preconditioning and the coupling between its components.
Proposed schematic representation of the signal transduction mechanism leading to cardioprotection by pharmacological and ischemic preconditioning of the human myocardium. Upon activation of sarcolemmal receptors mitoKATP channels are opened via G proteins and possibly PKC-δ. The opening of mitoKATP channels will activate PKC possibly via the production of radical oxygen species (ROS). PKC may then translocate to various cellular sites including the mitochondria, sarcolemma and intercalated discs, and nucleus, a phenomenon that may involve specific PKC isoforms, where p38MAPK will be activated. In turn, p38MAPK may activate a single or multiple end-effectors directly or via MAPK intermediates. PLC: phospholipase C, PLD: phospholipase D.
Clinical implications
The findings that the cardioprotection achieved by activation of α1-adrenoreceptors or adenosine receptors is as potent as the one obtained with IP and that their use in combination does not result in additional benefit have obvious major clinical implications because maximal protection can be attained without the need to occlude the coronary arteries to induce ischemia. The procurement of cardioprotection with stimuli of receptors as diverse as α1-adrenoreceptors and adenosine receptors can be clinically advantageous because some of these agents may be contraindicated in certain conditions (e.g. α1-adrenoreceptor agonists in hypertension, adenosine in the presence of alterations of the cardiac conduction system). These interventions can be useful to combat ischemic injury in different clinical conditions such as coronary angioplasty, cardiac surgery and heart transplantation; however, it is necessary to mention that our studies were performed in an in-vitro preparation and therefore any extrapolation to the clinical setting should be made with caution.
The realization that cardioprotection by PP and IP of the human myocardium is mediated by one obligatory signal transduction pathway also opens the therapeutical window for direct manipulation of its components. Recently, we have demonstrated that although the myocardium from patients with poor left ventricular function (ejection fraction < 30%) or with diabetes cannot be protected with IP, the mitoKATP channel opener diazoxide elicited protection in the former but not in the latter [40]. This suggests that if part of the signal transduction cascade is affected by disease states, cardioprotection can still be obtained by bypassing the defective components. It is also of clinical relevance that blockade at any stage of the signal transduction involved in preconditioning does not seem to exacerbate injury suggesting that this pathway is solely used for cardioprotection and not in tissue injury.
Possible limitations of the study
A potential limitation of our studies was the use of atrial tissue as opposed to ventricular myocardium and the fact that this was a necrosis in vitro model, and therefore any extrapolation must be conducted with caution. Another limitation might be that right atrial appendages were obtained from patients subjected to various medical treatments such as nitrates and beta blockers, which may have influenced the simulated ischemia/reoxygenation injury and the protection induced by preconditioning. However this effect is unlikely to be significant as all the specimens responded similarly to ischemic injury and preconditioning.
Conclusions
The current studies demonstrate that protection with pharmacological preconditioning by activation of α1-adrenoreceptors or adenosine receptors is identical to that of ischemic preconditioning in the human myocardium. In addition, they show that mitoKATP channels, PKC and p38MAPK are an integral part of the cellular signal transduction involved in this cardioprotection in which mitoKATP channels are placed upstream and p38MAPK placed downstream of PKC.
Materials and methodsExperimental preparation
Experiments were performed on trabeculae obtained from the right atrial appendage of patients undergoing elective coronary artery bypass graft surgery or aortic valve replacement in a cell necrosis model that was developed and characterised in our laboratory and that has been described previously [41]. The investigation conforms with the principles outlined in the Decleration of Helsinki. Donor patients were excluded if they had enlarged atriums, atrial arrhythmias, poor left ventricular function (ejection fraction <30%), right ventricular failure or were taking oral hypoglycaemic agents, opioid analgesia, KATP channel openers or catecholamines. Local ethical committee approval was obtained for the harvesting technique. Temperature was maintained at 37°C throughout the experiments and simulated ischemia was induced by bubbling the medium with 95% N2/5% CO2 (pH 6.8–7.0) instead of 95% O2/5% CO2 (pH 7.36 and 7.45) and replacing D-glucose with 2-deoxy-D-glucose.
Solutions and chemicals
The incubation medium was prepared daily with de-ionized distilled water and contained (in mM): NaCl2 (118), KCl (4.8), NaHCO3 (27.2), MgCl2 (1.2), KH2PO4 (1.0) CaCl2 (1.25), HEPES (20) and D-glucose (10) or 2-deoxy-D-glucose (10). All the chemicals were purchased from Sigma Chemicals.
Experimental time course
All the muscles were equilibrated for a 30 min period before being randomly assigned to serve as time-matched aerobic controls or subjected to a 90 min period of simulated ischemia (SI). The muscles were then reoxygenated (R) for another 120 min by incubation in 10 ml of oxygenated medium with added glucose. At the end of the experimental protocols, samples from the incubation media used during the reoxygenation period were collected for the assessment of creative kinase (CK) leakage and the tissue was taken for the assessment of viability (reduction of 3- [4,5-dimethylthiazol-2-yl]-2,5-diphenyltetiazolium bromide (MTT). All agents tested were added for 5, 10 or 20 min at the end of the equilibration period and before the induction of SI. The doses of the agent used in the present studies (phenylephrine at 0.1 μM, adenosine at 100 μM, chelerythrine at10 μM, 5-hydroxydecanoate at 1 mM, SB203580 at 10 μM, PMA at 1 μM diazoxide at 100 μM anisomycin at 1 nM) were selected following preliminary dose-response studies for each of the drugs.
Study groups
In Study 1, the efficacy of preconditioning via α1-adrenoceptors or adenosine receptors alone and in combination with IP was investigated following the protocol described in Figure 1.
Study 2 investigated the role of PKC, p38MAPK and mitoKATP channels in the cardioprotective effect of IP and of PP with phenylephrine and adenosine following the protocol described in Figure 2.
Study 3 was designed as described in Figure 3 to elucidate the sequence of the participation of involvement of mitoKATP channels, PKC and p38MAPK in the signal transduction cascade of cardioprotection.
In the three studies, n = 6 specimens from different subjects/group were used.
Assessment of tissue injury and viability
Tissue injury was determined by measuring the leakage of CK into the incubation medium during the 120 min reoxygenation period. This was assayed by a kinetic ultraviolet method based on the formation of NAD (Sigma Catalogue No. 1340-K) and the results expressed as U/g wet wt.
Tissue viability was assessed by the reduction of MTT to a blue formazan product at the end of the experimental time. The absorbance of the blue formazan formed was measured on a plate reader (Benchmark, Bio-Rad Laboratories, California, USA) at 550 nm and the results expressed as mM/g wet wt [41].
Statistical analysis
All data are presented as mean ± SEM. All values were compared by ANOVA with application of a post hoc Tukey's test. Statistical significance was taken as p < 0.05.
Authors' Contribution
ML participated in the design of the study, carried out all the studies, collated and analyzed the data and drafted the manuscript. MG participated in the design of the study, analysis and presentation of the data and revision of the manuscript. Both authors read and approved the final manuscript.
Acknowledgments
This work was supported in part by a grant from Mason Medical Research Foundation and by individual contribution from Professor M Galiñanes.
MurryCEJenningsRBReimerKAPreconditioning with ischemia: a delay of lethal injury in ischemic myocardium.198674112411363769170GhoshSStandenNBGaliñanesMPreconditioning the human myocardium by simulated ischemia: Studies on the early and delayed protection.20004533935010728354BanerjeeALocke-WinterCRogersKBPreconditioning against myocardial dysfunction after ischemia and reperfusion by an alpha-1 adrenergic mechanism.1993736566708396503ClevelandJCMeldrumDRRowlandRTIschemic preconditioning of human myocardium: protein kinase C mediates a permissive role for α1-adrenoceptors.1997273H902H9089277509MengXShamerBDPulidoEJAdrenergic induction of bimodal myocardial protection: signal transduction and cardiac gene reprogramming.1999276R1525R153310233047LochnerAGenadeSTrompEPodzuweitTMoolmanJAIschemic preconditioning and the β-adrenergic signal transduction pathway.199910095896610468527SchultzJJHsuAKGrossGJIschemic preconditioning is mediated by a peripheral opioid receptor mechanism in the intact rat heart.19972913551362920162110.1006/jmcc.1996.0369DowneyJMLiuGSThorntonJDAdenosine and the anti-infarct effects of preconditioning.199327388458028CarrCSHillRJMasamuneHEvidence for a role for both A1 and A3 receptors in protection of isolated human atrial muscle against simulated ischemia.19973652599415272JoyeuxMGodin-RibuotDRibuotCResistance to myocardial infarction induced by heat stress and the effect of ATP-sensitive potassium channel blockade in the rat isolated heart.1998123108510889559890BrewEBMitchellMBRehringTFThe role of bradykinin in cardiac functional protection after global ischemia-reperfusion in the rat heart.1995269H1370H13787485570MeldrumDRClevelandJCSheridanBCRowlandRTBanerjeeAHarkenAHCardiac preconditioning with calcium: clinically accessible myocardial protection.19961127787868800168JonesWKFlahertyMPTangXLIschemic preconditioning increases iNOS transcript levels in conscious rabbits via a nitric oxide-dependent mechanism.199931146914811042334510.1006/jmcc.1999.0983StrasserRHBraun-DullaeusRBWalendzikHMarquetantRα1-receptor-dependent activation of protein kinase C in acute myocardial ischemia: mechanisms of sensitization of the adenylyl cyclase system.199270130413121315640LazouASugdenPHClerkAActivation of mitogen-activated protein kinases (p38-MAPKs, SAPKs/JNKs and ERKs) by the G-protein coupled receptor agonist phenylephrine in the perfused rat heart.19983324594659601075GrossGJAuchampachJABlockade of ATP-dependent potassium channels prevents myocardial preconditioning.1992702232331310443Speechly-DickMEGroverGJYellonDMDoes ischemic preconditioning in the human involve protein kinase C and the ATP-dependent K channel? Studies in an atrial in vitro model.199577103010357554138PainTYangXCritzSDOpening of mitochondrial KATP channels triggers the preconditioned state by generating free radicals.20008746046610988237SatoTO'RourkeBMarbanEModulation of mitochondrial ATP-dependent K+ channels by protein kinase C.1998831101149670924GarlidKDPaucekPYarov-YarovoyVCardioprotective effect of diazoxide and its interaction with mitochondrial ATP-sensitive K+ channels: Possible mechanism of cardioprotection.199781107210829400389GhoshSStandenNBGaliñanesMEvidence for mitochondrial KATP channels as effectors of human myocardial preconditioning.20004593494010728419LoubaniMGaliñanesMAlpha 1 adrenoceptors during simulated ischemia and reoxygenation of the human myocardium: effect of the dose and time of administration.200112210311211436042LiuGSThorntonJVan WinkleDMStanleyAWOlssonRADowneyJMProtection against infarction afforded by preconditioning is mediated by A1 adenosine receptors in rabbit heart.1991843503562060105LeesarMAStoddardMAhmedMBroadbentJBolliRPreconditioning of human myocardium with adenosine during coronary angioplasty.199795250025079184580McCullyJDToyodaYUematsuMStewartRDLevitskySAdenosine-enhanced ischemic preconditioning adenosine receptor involvement during ischemia and reperfusion.2001280H591H602LiRALeppoMMikiTSeinoSMarbanEMolecular basis of electrocardiographic ST-segment elevation.20008783783911073877MitchellMBMengXAoLBrownJMHarkenAHBanerjeeAPreconditioning of isolated rat heart is mediated by protein kinase C.19957673818001280WangYHiraiKAshrafMActivation of mitochondrial ATP-sensitive K(+) channel for cardiac protection against ischemic injury is dependent on protein kinase C activity.19998573174110521247MiuraTLiuYKitaHOgawaTShimamotoKRoles of mitochondrial ATP-sensitive K channels and PKC in anti-infarct tolerance afforded by adenosine A1 receptor activation.20003523824510636286VahlhausCSchulzRPostHRoseJHeuschGPrevention of ischemic preconditioning only by combination of protein kinase C and protein tyrosine kinase in pigs.1998381972099514996PingPZhangJQiuYIschemic preconditioning induces selective translocation of protein kinase C isoforms epsilon and eta in the heart of conscious rabbits without subcellular redistribution of total protein kinase C activity.1997814044149285643NakanoABainesCPKimSOIschemic preconditioning activates MAPKAPK2 in the isolated rabbit heart: evidence for involvement of p38 MAPK.20008414415110666409SchneiderSChenWHouJSteenbergenCMurphyEInhibition of p38 MAPK α/β reduces ischemic injury and does not block protective effects of preconditioning.2000280H449H508BehrendsMSchulzRPostHInconsistent relation of MAPK activation to infarct size reduction by ischemic preconditioning in pigs.2000279H1111H111910993774StokoeDEngelKCampbellDGCohenPGaestelMIdentification of MAPKAP kinase 2 as a major enzyme responsible for the phosphorylation of the small mammalian heat shock proteins.19923133073131332886GuayJLambertHGingras-BretonGLavoieJNHuotJLandryJRegulation of actin filament dynamics by p38 map kinase-mediated phosphorylation of heat shock protein 27.19971103573689057088HuotJHouleFSpitzDRLandryJHSP 27 phosphorylation-mediated resistance against actin fragmentation and cell death induced by oxidative stress.1996562732798542580LiuYSatoTSeharaseyonJSzewczkyAO'RourkeBMarbanEMitochondrial ATP-dependent potassium channels. Viable candidate effectors of ischemic preconditioning.1999874273710415518KowaltowskiAJSeetharamanSPaucekPGarlidKDBioenergetic consequences of opening the ATP-sensitive K(+) channel of heart mitochondria.2001280H649H65711158963GhoshSStandenNBGaliñanesMFailure to precondition pathological human myocardium.20013771171811693741ZhangJGGhoshSOckelfordCGaliñanesMCharacterization of an in vitro model for the study of the short and prolonged effects of myocardial ischaemia and reperfusion in man.20009944345311052925oai%3Apubmedcentral.nih.gov%3A122066!!!pmc!bmcphys