A better understanding of dementia, including its causes, underlying pathophysiological processes and earliest possible identification, has become a major public health priority. Changes in cognition associated with age are complex, especially with regard to distinguishing usual from pathological brain ageing. Multiple and often intertwined pathological factors, including atrophy, neurodegeneration, inflammation, stroke and genetic-related factors, cause dementia [1]. Here, we explore the link between vascular disease, cognitive decline and dementia risk. Given the relatively high proportion of dementia attributable to possibly reversible midlife vascular causes [2], it has been suggested that vascular risk manipulation may result in up to a 50% reduction in the forecasted dementia prevalence rate in individuals who are 65 years old or older [3,4]. Vascular risk factors for dementia may also contribute to impairments observed in the pre-clinical stage of cognitive decline. This has raised questions regarding (a) whether vascular disease can predict cognitive change and dementia risk in otherwise non-impaired individuals and (b) the duration and possible reversal of cognitive symptoms and dementia depending on vascular disease manipulation and treatment. The aim of this review is to describe the current understanding of the division between pre-clinical cognitive impairment in the context of vascular disease versus the absence of vascular disease. The focus will be on the term 'vascular cognitive impairment, no dementia' (VCIND), an umbrella term that broadly encompasses cognitive deficits associated with vascular disease which fall short of a dementia diagnosis, in order to determine whether within the context of this condition there is a pre-clinical state linked to a high risk of dementia progression.

Ageing in the developed world is associated with changes in the vascular system which result in atherogenesis, increased pulse pressure and increased risk of developing vascular disease as a consequence of a direct effect on the vascular system (for example, arterial hypertension and vasculitis) or indirect metabolic and haemodynamic effects secondary to other disorders (for example, diabetes, congestive heart failure and obesity) [5]. Vascular disease can reduce cerebral perfusion, causing oxidative stress and neurodegeneration [6]. Vascular disease has also been reported to accelerate atrophy and result in white matter abnormalities, asymptomatic infarct, inflammation and reduced glucose metabolism, cerebral blood flow and vascular density [5,7]. Such pathological changes have been associated with not only dementia of the vascular type but also Alzheimer disease (AD) [8,9]. The mechanisms by which vascular disease and its risk factors cause pathology and how such changes impact cognitive function are incompletely understood but are thought to depend on age, disease co-morbidity, lifestyle and genetic susceptibility and predisposition. An important question is whether different vascular disease factors can be separated from each other and from the effect of ageing itself in order to identify their unique impact on cognitive function. A more complete understanding of the relationship between vascular disease, cognitive decline and dementia risk will have important implications in identifying vulnerable population subgroups and a potential treatment target.

Age-related cognitive change can be placed along a continuum from normal to severely demented, with intermediate stages of cognitive decline. To date, there is no consensus on where boundaries between disease and non-disease lie, if such a definite boundary exists. Rather than a strict dichotomisation, the determination of impairment may instead be based on the likelihood or probability that ageing is not occurring in accordance with normative expectations. Furthermore, classification systems of cognitive change themselves raise questions regarding the level of cognitive and functional dysfunction used to reliably categorise individuals and those risk factors apart from ageing itself that contribute to mild and severe cognitive change.

Some MCI classifications do not consider possible vascular contributing factors to cognitive impairment, and therefore the distinction between MCI and VCIND becomes blurred [36]. This raises questions of the benefit in distinguishing MCI from VCIND and whether formulating a new diagnostic category such as VCIND is even necessary. Whether vascular disease is considered in the diagnosis of MCI may influence general population prevalence estimates and longitudinal patterns of progression. Empirical evidence as to the degree of co-morbidity across different pathologies (including, for example, AD and VaD pathologies) will help new classifications, particularly as biomarkers of specific pathologies emerge.

Generally, the differential diagnosis between MCI and VCIND is clinical and based on the distinction between AD and VaD. AD is characterised by a steady and progressive loss of memory and cognitive faculties, including language deterioration, impaired visuospatial skills and poor judgement [37]. AD has a distinct neuropathological pattern of beta-amyloid plaques and neurofibrillary tangles [38,39]. Other significant correlates of cognitive decline include synaptic loss, neuronal death and disruption to the cholinergic pathway [40]. In contrast, the disease course of VaD is highly variable, generally following a stepwise pattern of decline and fluctuating course [41,42]. For a diagnosis of VaD, it is recommended that radiographic features of vascular disease, including evidence of an ischaemic lesion, white matter hyperintensities and/or hypometabolism, be confirmed [37]. However, in some cases, AD and VaD have been found to result in similar cognitive, functional and behavioural disturbances, and frequently both pathologies co-occur [43-45]. AD and VaD may share associated risk factors (stroke, arterial hypertension, increasing age and low educational attainment), structural changes, neuropathological profiles (white matter lesions and apoptosis) and neurochemical changes (that is, in the cholinergic system) [46]. Overlap may also result from the large degree of silent risk pathology in older people.

In older people, multi-morbidity is common and a strict dichotomisation between degenerative and vascular dementing disorders at both pre-clinical and dementia stages is difficult to undertake, possibly artificial and perhaps not the most useful approach. Below, we explore the cognitive, neuro-imaging and neuropathological profiles of MCI and VCIND to determine whether evidence exists for the separation of both conditions.

Current pharmacological and non-pharmacological modifications of vascular disease and its risks have been found to have only a marginal effect on reducing dementia prevalence in the general population [113]. Indeed, most strategies, whether pharmacological or non-pharmacological, are ineffective in the prevention of dementia and are potentially harmful. With regard to MCI, while pharmacological and lifestyle modifications have been found to be effective in ameliorating cognitive impairment in selected older cohorts [114,115], no consistently positive results have emerged from randomised controlled trials for such manipulation in the prevention of MCI or future dementia progression in older people in the general population [116,117]. Where the primary prevention of VCIND has been considered, physical activity has been found to reduce the risk of VCIND in women, but not men [118]. Why gender differences emerged in this study is unclear but is thought to be linked to gender reporting bias in physical activity levels or to statistical error (type 2 error) due to the small number of male cases. However, before recommendations based on this result can be made, it must be replicated in population-representative samples using objective measures of physical activity.

Overall, the results of intervention trials of vascular risk reduction on the prevention of cognitive decline and dementia have not been encouraging so far. There are, however, other reasons why cardiovascular and cerebrovascular disease should be prevented and treated, especially for the prevention of recurrent stroke and hypertension, which themselves are strong risk factors for functional impairment and mortality [119]. Further trials are needed to determine whether the manipulation of different vascular diseases and vascular risk factors prevents VCIND and dementia progression. However, it is unlikely that a single strategy will cure or prevent all dementia; rather, early treatment may require a combination of therapies with different targets and time frames for instigation (that is, early, mid- and later life).

A more complete understanding of the relationship of vascular disease to cognitive decline and dementia risk is needed. Even when vascular pathology appears to be the main underlying process, the effect can be heterogeneous, with diverse neuropsychological, clinical, radiological and morphological profiles, often in the presence of other pathologies. To date, the risk of dementia progression cannot yet be accurately predicted from pre-dementia states captured in the concepts of MCI and VCIND. Cognitive decline can also occur prior to vascular insult (for example, pre-stroke dementia), raising the question of causality [1]. How vascular disease relates to dementia and its influence across an individual's life span and the unique and interactive mechanisms of action on neurodegeneration must be investigated further to identify the best treatment and preventative target.

Before case screening for individuals at high risk of dementia secondary to vascular disease can be undertaken, the concept of VCIND will require evidence-based consensus criteria and validation as a pre-clinical state that confers high dementia risk in both clinical and population-based studies. Many questions remain open, particularly with regard to identifying where the state of VCIND begins and ends. This review suggests that cognitive change would be expected to be influenced by vascular disease type and severity, disease onset, co-occurring factors, underlying vulnerability (for example, age, education and genetics) and whether pathology occurs secondary to another process (for example, AD). Accurate early detection of the general population at increased risk of dementia is central for the implementation of interventions to prevent dementia and other memory-related problems in older individuals.

AD: Alzheimer disease; A-MCI: anmestic subtype of mild cognitive impairment; MCI: mild cognitive impairment; VaD: vascular dementia; VCI: vascular cognitive impairment; VCIND: vascular cognitive impairment, no dementia.

The authors declare that they have no competing interests.

AD and FM are employees of BioMed Central and receive a fixed salary. DG, TG and GW are Editors-in-Chief of Alzheimer's Research & Therapy and receive an annual honorarium.

Tramiprosate (also referred to as homotaurine and 3-amino-1-propanesulfonic acid, or 3APS) is an Aβ-binding compound that was developed using in vitro and in vivo model systems [9] that left some uncertainty regarding the brain concentration necessary for a pharmacodynamic effect in human AD. While a phase II study did suggest a reduction in cerebrospinal fluid Aβ in AD subjects treated with tramiprosate [10], it was unknown whether the degree of reduction would be sufficient to translate into clinical benefit. The small and brief phase II program was not designed to demonstrate clinically a disease-slowing effect; as expected, subjects in the 12 week treatment trial treated with placebo showed no decline, and, therefore, there was no possibility of showing reduced decline with treatment. The development of tramiprosate as a pharmaceutical treatment for AD was halted when the first phase III trial failed to demonstrate significant beneficial effects on the primary analysis of cognitive and clinical outcomes.

Similar problems were faced in the tarenflurbil development program. In vitro and in vivo, the drug clearly modulates gamma secretase activity, reducing generation of Aβ [11,12]. In early human studies, however, there was no strong biomarker evidence of amyloid reduction in cerebrospinal fluid (CSF), and concerns about inadequate brain concentrations in humans were voiced. The phase II trial, though relatively large, was underpowered to show a disease-modifying effect, and the primary analysis of the impact of treatment on cognitive and functional outcomes was negative [13]. However, post hoc analyses appeared to be consistent with a beneficial drug effect, leading to the launch of large phase III trials. The program was terminated when the first phase III trial showed no evidence of beneficial effect.

A somewhat similar situation has arisen in the development program of bapineuzumab, a monoclonal amino terminus-specific anti-amyloid antibody [14]. Perhaps misled by encouraging cognitive data from a small phase I trial hinting at a symptomatic effect, the sponsors sought evidence of efficacy in the modestly sized phase II program. Though there was evidence of benefit in a number of secondary analyses, particularly in the apolipoprotein E ε4 negative subgroup, the primary cognitive efficacy analysis was negative. The sponsors, Elan and Wyeth, have nonetheless proceeded with a very large phase III program.

Why have so many programs yielded discouraging efficacy data in phase II and III clinical trials? In phase II, the problem may be primarily one of statistical power. Most programs seek to be able to demonstrate a 25% to 33% slowing of progression of mild or mild to moderate AD. But in view of substantial inter-subject and inter-site variance issues, a large and long trial is necessary. Estimates using preliminary data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) [15] suggest that to demonstrate a 33% reduction in progression rate (as measured by change in Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-cog) or Clinical Dementia Rating 'sum of boxes' (CDR-SB)) in an 18 month trial in mild AD, approximately 300 subjects per group are required (M Donohue et al., unpublished). No phase II program has approached this size. It should be noted that the ADNI experience probably underestimates the sample size required, in that it may be expected that variance will be greater in commercial trials, particularly those that are international, than among the academic North American sites participating in ADNI.

As with the anti-amyloid agents tramiprosate, tarenflurbil and bapineuzumab, the phase II trial of the anti-tangle compound Rember (methylene blue) did not meet its primary efficacy objectives [16]. In view of the modest group sizes and short duration, this too is not surprising, regardless of whether the drug ultimately proves effective in pivotal trials. But as with the anti-amyloid programs, caution must be exercised in the interpretation of post hoc analyses of the Rember trial data.

The development of dimebon represents the one recent AD program with strikingly positive results. At the primary 6 month analysis, strongly significant beneficial effects were seen on all outcome measures [17]. With continuation of the blind through 12 months of treatment, the effects on outcome measures increased, consistent with (though not definitive evidence of) a disease-modifying effect [17]. The success of this modestly sized trial is indicative of the immediate symptomatic benefit associated with the treatment. If a putative disease modifying drug yields short-term benefits, short (6 month) trials may be sufficient for regulatory approval. Symptomatic effects are plausible with neuroprotective and anti-amyloid drugs. But in the absence of such effects, a modest, phase II-type trial will be insufficient; little or no cognitive and clinical decline can be observed in 6 months, so no slowing of progression can be demonstrated. A consensus has arisen that 18 months or longer is an appropriate duration of treatment for studies aiming to show slowing of decline in AD.

But what about the two negative phase III trials of plausible anti-amyloid agents? There was certainly a 'phase II problem' – that is, phase III proceeded without evidence of efficacy in phase II. As expected, the small phase II tramiprosate study did not show any efficacy signal, but the modest reduction in CSF Aβ42 was considered encouraging. But it is unknown what the size of this biomarker signal must be to predict clinical efficacy with prolonged treatment. Further, there were questions about the magnitude and consistency of central nervous system drug penetration and concentration. But in addition to these uncertainties, the power of the phase III North American tramiprosate trial was lower than expected. The placebo group decline was smaller, and the standard deviation of the change score was higher, than expected; the power to demonstrate the target effect size of 25% slowing with the group sizes of 350 was limited. The tarenflurbil phase III program followed a phase II study that (not surprisingly) failed to achieve its primary efficacy objectives, so the risk of a negative phase III program had to be considered substantial; only the post hoc phase II analyses were encouraging. In addition, there was no convincing evidence of pharmacodynamic effect; specifically, no reduction in CSF Aβ in humans had been demonstrated. The negative trial results may reflect inadequate brain penetration in humans to yield a sufficient reduction in the generation of Aβ.

On the basis of these plausible explanations, the negative results of the phase II and III anti-amyloid trials cannot be considered to be strong evidence against the amyloid cascade hypothesis. The scientific basis for the hypothesis remains quite compelling. Aβ42 is highly toxic to neuronal cells and synaptic function, particularly in its oligomeric states. Each of the known genetic causes of AD is closely linked to Aβ generation: Down syndrome to APP over-expression, and familial autosomal dominant AD to mutations of APP and presinilins 1 and 2 that increase amyloidogenic cleavage of APP. Occam's Razor points strongly to amyloid as the pivotal molecule. Recent reports of a small number of AD patients with progressive dementia despite apparent amyloid plaque clearance resulting from active vaccination [18] does not disprove the hypothesis; clinical data on these individuals are limited, plaques are probably not the most important form of Aβ, the course of disease had these patients not been treated is unknown, and perhaps earlier treatment is necessary for a profound effect on outcome.

This last point may be key. There is strong evidence that the pathiobiology of AD precedes dementia by many years. In Down syndrome, amyloid deposition in brain precedes dementia by years or decades [19,20]. There is a high prevalence of brain amyloid in non-demented elderly individuals at autopsy [21], perhaps an indication of a long pre-symptomatic stage. Similarly, neuroimaging evidence of brain amyloid deposition is common [22]. Subtle memory symptoms and cognitive decline have been documented more than a decade before dementia onset [23]. If, as suggested by the recent active vaccine study autopsy report, dementia can progress despite elimination of amyloid plaques in AD patients [18], perhaps it is necessary to intervene earlier in the disease process.

At present, the diagnosis of AD requires the presence of dementia. What is the relationship of pre-dementia cognitive dysfunction to AD? What is the significance of amyloid brain deposition in the absence of cognitive impairment? With two plausible (if not yet proven) methods for identifying brain amyloid deposition, positron emission tomography (PET) scanning and CSF measurement of Aβ42, identification of such individuals is quite feasible. If either subtle cognitive impairment or amyloid deposition in brain consistently predicts AD dementia, it should be considered an early stage of AD. That is, we should revise the standard NINCDS-ADRDA criteria for AD [24].

Dubois and colleagues [25] have proposed one possible revision. They suggest that a 'research diagnosis' of AD be based on the presence of gradually progressive episodic memory impairment with evidence of AD neurobiology documented by the presence of one or more among several characteristic biomarker signals. The biomarker signals include medial temporal lobe atrophy by volumetric magnetic resonance imaging (MRI), temporal parietal hypoperfusion by [18F]fluorodeoxyglucose (FDG)-PET, amyloid deposition by PET, or CSF findings (elevated tau or phospho-tau, and/or low Aβ42) characteristic of AD. The proposed criteria can be applied in the pre-dementia or dementia stages.

It is plausible that effective disease-modifying interventions might be only minimally effective or even futile at the dementia stage; neuroprotection or favorable effects on the inciting amyloid dysregulation might be overwhelmed by extensive neuronal/synaptic degeneration and plaque pathology. Extending the diagnosis of AD to include individuals with mild cognitive impairment and even normal cognition when there is biomarker evidence of AD-type pathophysiology might facilitate the development of disease-modifying drugs for the treatment of individuals most likely to respond. The earlier the disease-modifying intervention, the greater the expected impact on the disease course. This idea is supported by a number of therapeutic studies in transgenic animal models [26,27].

Altering the definition of AD may not be necessary for the development of early interventions. A possible development strategy for early intervention using the current diagnostic criteria would be to conduct studies aiming to demonstrate that an intervention increases the time to dementia diagnosis. Several completed studies have enrolled subjects with amnestic mild cognitive impairment (MCI), with the primary analysis of survival to consensus diagnosis of AD [28]. This design has the advantage of clear clinical validity, a desirable feature in consideration of the uncertain regulatory status of the MCI designation. At least one set of MCI criteria seems to predict a high likelihood of AD diagnosis (approximately 15% per year), so that such a trial can have a reasonable size with adequate power to demonstrate a treatment effect [29]. But progression from MCI to AD is not a discrete event; the loss of function necessary to meet criteria for dementia occurs gradually, and it is challenging to assign a specific date to dementia onset. This subjectivity may be aggravated in large international trials. The progression of cognitive and functional impairment caused by AD pathobiology is insidious; defining a discrete disease onset seems arbitrary.

If the diagnostic criteria for AD are modified to encompass individuals prior to the onset of dementia, it would be straightforward to design trials with standard AD co-primary outcome measures. The ADNI longitudinal data demonstrate acceptable decline rate and variance for the ADAS-cog and CDR-SB in amnestic MCI; the size of trials adequately powered to demonstrate slowing of cognitive/clinical progression would be large but perhaps manageable. Adding selection criteria, and perhaps covariates to adjust for disease state, will reduce sample sizes substantially. In particular, for the development of anti-amyloid programs such as secretase inhibitor, anti-aggregation agents and anti-amyloid immunotherapy, trials can select MCI patients with biomarker evidence of brain amyloid deposition. Two options are feasible, though each presents challenges. The advent of F18 amyloid binding radiotracers has established the feasibility of amyloid brain imaging at most sites with PET scanners, but this is an expensive undertaking that has not yet been fully validated. CSF Aβ42 is strongly associated with neuroimaging evidence of amyloid deposition [30] and is essentially universally available, though requiring lumbar puncture during study screening may not be welcomed by investigators and especially subjects. The addition of covariates such as MRI volumetric measures to analysis plans will reduce unexplained variance and further increase statistical power to demonstrate slowing of progression. If the community of AD investigators, clinicians and regulators were to adopt early AD diagnostic criteria, feasible early AD trials could be launched immediately (Table 1).

While disease-modifying treatment of MCI is expected to yield more dramatic benefits than treatment of mild AD, perhaps the most appropriate population for intervention is the earlier, pre-symptomatic (or very mildly symptomatic) subjects with biomarker evidence suggestive of AD. The ultimate goal of disease-modification programs, prevention of AD dementia, is conceivable if treatment is started before appreciable neuronal damage and synaptic dysfunction have occurred. Preliminary evidence from some studies suggest that markers of amyloid accumulation predict dementia even in asymptomatic individuals [31].

But in the absence of symptoms, it will not be possible to fulfill the conventional US Food and Drug Administration requirement for AD drug development: demonstration of efficacy on co-primary measures, specifically a cognitive performance test and a functional/global measure. It would require huge and lengthy studies to show slowing of cognitive and clinical progression or delay to diagnosis of dementia in subjects not yet showing any symptoms. To study interventions in this population, we will require validated surrogate markers.

A biomarker is any objectively measured characteristic that reflects normal or pathological processes, or responses to therapeutic intervention. As discussed above, biomarkers can be valuable in selecting subjects for clinical trials and for therapeutic interventions, for reducing unexplained variance and thus improving statistical power, and for establishing proof of concept in early phase drug development.

In rare cases, a biomarker can take the place of a clinical endpoint for establishing efficacy in a phase III clinical trial; that is, a biomarker can be validated as a surrogate endpoint. Examples of such surrogate markers include blood glucose and hgA1c in diabetes, blood pressure and cholesterol in cardiovascular disease, intraocular pressure in glaucoma, and lymphocyte subset ratios and viral load in HIV disease. To validate a biomarker as a surrogate endpoint, several issues must be addressed. There must be a well-accepted scientific framework connecting the biomarker to disease mechanisms and the prediction of clinical outcomes. Further, drug effects on the biomarker must be related to drug effects on clinical outcome; ideally, the biomarker should fully capture treatment effects, as confirmed by clinical trials of multiple interventions.

It is unlikely that an ideal surrogate for disease-modifying intervention in AD will become available in the foreseeable future. However, in consideration of the enormous clinical need, and the likelihood that the development of highly effective disease-modifying treatments will require the use of surrogate endpoints, it is reasonable to assume that regulatory agencies will consider acceptance of surrogates that are less than ideal.

A validated surrogate marker is essential for the study of AD interventions in asymptomatic or very mildly symptomatic individuals. It may be feasible to gain acceptance of a surrogate AD biomarker with a small number of trials demonstrating concordant treatment effects on the biomarkers and clinical symptoms. Even if the benefits of disease-modifying treatments in mild AD dementia are limited, they may well be sufficient to establish this concordance. Indeed, ongoing anti-amyloid trials that have incorporated biomarkers could provide this evidence. Consensus among clinical experts, based on robust data, that candidate biomarkers track disease progression at various stages of disease will strengthen the case for validation. A leading candidate surrogate marker is brain atrophy rate as measured by volumetric MRI; a huge body of evidence supports a link between regional brain atrophy and progression of AD pathobiology [32-34].

Building the consensus necessary to shift regulatory guidelines, clinical trial design and clinical practice will require large-scale cooperation among pharmaceutical and biotech companies, academic leaders, advocacy groups, funders and regulators. It is fortuitous that such cooperative efforts have been steadily gaining traction in recent years. Regular meetings involving all of the stakeholder groups have been productive; the semiannual Alzheimer's Association Research Roundtable, the annual Leon Thal Symposium sponsored by the Lou Ruvo Brain Institute, and the meetings of the Task Force on Use of Biomarkers in Alzheimer's Trials are leading examples that have advanced the field. They demonstrate the eagerness of many companies to share experience and ideas in pursuit of solutions to problems in AD therapeutic research.

Perhaps the best example of a cooperative effort among many to overcome the hurdles in drug development is ADNI. Led by Michael Weiner at the University of California San Francisco, ADNI is jointly funded by the National Institute on Aging, the Alzheimer's Association and other foundations, and contributions from pharmaceutical companies. It is a long-term effort to collect longitudinal cognitive, clinical, CSF and neuroimaging data on cohorts of individuals with mild AD, mild cognitive impairment and normal cognitive aging to allow optimal use of biomarkers in trial design. ADNI brings together leaders from academia, industry, government agencies and advocacy groups on at least a biweekly basis to jointly assess the study progress, and to discuss roadblocks and paths forward. To maximize scientific advance, all ADNI data are publicly posted in real-time; a huge number of presentations and publications from ADNI as well as outside investigators bears evidence of its success. ADNI has also spawned or supported similar collaborative efforts in Europe, Japan, Australia and China.

Data and ideas arising from ADNI and the various collaborative meetings have provided the ideas and data behind the discussion in this article. With continuation of these efforts, the common goal of optimal trial design is readily achievable. The challenges of determining populations for study, cognitive and clinical outcome measures, validation of biomarkers and analytic plans can be met within a few years. Consensus will lead to practical regulatory pathways, and the successful introduction of disease-modifying interventions that will blunt the AD epidemic that is growing with the aging world populations.

AD: Alzheimer's disease; ADAS-cog: Alzheimer's Disease Assessment Scale-cognitive subscale; ADNI: Alzheimer's Disease Neuroimaging Initiative; APP: amyloid precursor protein; CDR-SB: Clinical Dementia Rating 'sum of boxes'; CSF: cerebrospinal fluid; MCI: mild cognitive impairment; MRI: magnetic resonance imaging; PET: positron emission tomography.

PSA is the recipient of grant awards from Pfizer, Baxter, Neuro-Hitech, Abbott and Martek, and is a consultant to Elan, Wyeth, Eisai, Neurochem, Schering-Plough, Bristol Myers Squibb, Lilly, Neurophage, Merck, Roche, Amgen, Genentech, Abbott, Pfizer, Novartis and Medivation. He holds stock options in Medivation.

Alzheimer's disease (AD) [MIM 104300], as the most common form of progressive dementia, is characterised neuropathologically by the presence of intracellular neurofibrillary tangles and features resulting from the deposition of amyloid β-peptide (Aβ) extracellularly in the form of senile plaques and within blood vessels in the brain in the form of cerebral amyloid angiopathy. The pathogenesis of AD is understood to be partly explained by mechanisms involved with (but not restricted to) two of the most prevailing hypotheses: the Aβ cascade hypothesis and the cholinergic hypothesis. The Aβ cascade hypothesis, which developed in the early 1980s, suggests that the commonly observed neurodegenerative abnormalities of AD, particularly senile plaques, develop following the accumulation of the 39 to 42 amino acid peptide Aβ in the brain [1]. This accumulation of Aβ is likely a consequence of imbalance between production of Aβ from amyloid precursor protein (APP; Figure 1) and its removal. Aβ removal can be mediated via drainage through interstitial fluid in the space surrounding blood vessels in the brain, by receptor-mediated transport of Aβ from the brain to the peripheral circulation, by enzymatic degradation or various combinations of all the these [1]. Over the years the Aβ hypothesis has not been universally accepted due to its perceived failing that brain Aβ deposition correlated poorly with cell death or disease severity in AD whereas neurofibrillary tangle pathology, another established neuropathological hallmark for AD, correlated better and, as such, was suggested to be more relevant, with Aβ representative of a secondary phenomenon [2]. Yet this failing was arguably addressed in the past decade with the identification of soluble and diffusible oligomeric forms of Aβ that are now thought to be the more harmful forms and which have been correlated with AD pathogenesis [3]. Furthermore, questions have now arisen as to whether neurofibrillary tangles are merely a marker of disease progression and not, in fact, a mediator of cell death as has been proposed (see [3,4] for reviews).

The cholinergic hypothesis posits that loss of cholinergic function (derived from the action of the neurotransmitter acetylcholine (ACh)) in the central nervous system is a significant part of the cognitive decline associated with AD (see [5] for a review; Figure 1). This hypothesis pre-dates the seeds of the Aβ cascade hypothesis but is supported by considerable evidence that there is reduced synthesis and altered transport of ACh, selective loss of cholinergic neurons, disruption of ACh receptor signalling, as well as reductions in the levels of these receptors in AD brains (see [5] for a review). Indeed, the majority of licensed 'cholinesterase' therapeutics currently used to treat and partially delay some of the progressive symptoms of AD are drugs that target acetylcholinesterase-mediated breakdown of ACh, thereby increasing the amount and prolonging the life of ACh in the brain [5]. Importantly, neither of the properties of these hypotheses exist in isolation and there is now a body of evidence supporting complex levels of interaction between the two and, more than likely, with other systems that fall beyond the scope of this review (see [5] for a review; Figure 1).

The renin angiotensin system (RAS) is best known for its role in the kidney in controlling blood pressure and bodily fluid homeostasis. The 'classical' RAS is known by the role of renin in cleaving inactive angiotensinogen to produce angiotensin (Ang)I, which is, in turn, converted by angiotensin-1 converting enzyme (ACE) to the (vaso-)active AngII [6]. AngII exerts its well known hypertensive effects following binding to its two receptors (AT₁R and AT₂R) [7]. However, the past two decades have revealed that the classical RAS was only the tip of the iceberg of what is now known to be a very complex system involving new AngI and AngII metabolites, additional receptors and regulating mechanisms (see [7] for a review; Figure 2). This added complexity has also offered new potential therapeutic targets for hypertension in addition to those already targeting the RAS through inhibition of ACE (ACE inhibitors (ACE-Is)) or by preventing AngII from binding its receptors (angiotensin receptor blockers (ARBs)) [8] (Figure 2).

Studies over the past two decades have added to our knowledge of the pathways involved in the pathogenesis of AD, including alterations to ACE and other angiotensin-related components of the RAS. Whether these are primary mediators or secondary consequences of the disease still remains to be shown, however, and further studies are certainly needed. There may be insufficient evidence at this time with which to make a final judgement as to what benefit or consequences ACE-Is or ARBs have on the development or progression of AD, but there is sufficient evidence to justify more measured consideration when prescribing ARBs or ACE-Is for hypertension as well as active future study of how pharmacological targeting of AngII in this pathway could offer therapeutic value in mitigating against secondary pathogenic Aβ-mediated changes in AD.

Aβ: amyloid β peptide; ACE: angiotensin-1 converting enzyme; ACE-I: ACE inhibitor; ACh: acetylcholine; AD: Alzheimer's disease; Ang: angiotensin; APP: amyloid precurser protein; ARB: angiotensin receptor blocker; AT₁R/AT₂R: angiotensin II receptors; BBB: blood brain barrier; CSF: cerebrospinal fluid; Indel: insertion/deletion polymorphism; RAS: renin angiotensin system.

The authors declare that they have no competing interests.

Aβ, tau and other protein aggregates in neurodegenerative diseases are almost always found in an abnormal structural conformation compared to the non-aggregated protein [1-4]. Many aggregates show the characteristic features of amyloid and accumulate in an 'abnormal' fibrillar β-pleated sheet structure [11]. A common theme of genetic alterations that cause AD is that they increase the likelihood that Aβ will aggregate into amyloid [6]. Mutations in tau that cause frontal temporal dementia with parkinsonism linked to chromosome 17 (FTDP-17 MAPT) also alter tau in a way that increases its likelihood to aggregate into amyloid-like structures [12-14]. Furthermore, there is evidence that mutations in, or over-expression of, other proteins linked to neurodegeneration enhance the likelihood that they are assembled into misfolded aggregates, and many of these aggregates also have characteristic features of amyloid [1-4].

When a normal protein misfolds and aggregates, it no longer resembles a self protein; thus, it is subject to recognition by the immune system. Misfolded self protein aggregates resemble pathogen-associated molecular patterns (PAMPs) and thymus independent type 2 (TI-2) antigens [15-18]. PAMPs are a group of molecules that are capable of activating a wide array of innate immune defenses. They can be proteins, polysaccharides, or nucleotides and are characterized by a repetitive molecular motif that is recognized as non-self and can bind and activate evolutionarily conserved pattern recognition receptors (PRRs) that initiate innate immune signaling. Classic PAMPs are bacterial lipopolysaccharide, flagellin, peptidoglycan, some viruses and virus-like particles, and double-stranded RNA. TI-2 antigens are similar polymeric molecules that directly stimulate B cells to secrete IgM by crosslinking of plasma membrane immunoglobulin. Despite their chemical diversity, the unifying features of both PAMPs and TI-2 antigens are that they are large in size (typically greater then 100 kDa and often much larger), and have repetitive epitopes (and at least for TI-2 antigens require rigid presentation of the epitope with a two-dimensional spacing of 5 to 10 nm), poor in vivo degradability, and the ability to activate complement [18]. Notably, these features of PAMPS and TI-2 antigens are quite reminiscent of the features of amyloid deposits [11].

Misfolded protein aggregates can theoretically activate the adaptive immune system. However, the key step in activation of adaptive immunity, major histocompatibility complex (MHC) presentation of non-self peptides, is likely to limit such activation [19]. MHC binds small peptides that are typically cleaved from a larger protein [20]. Unless a small peptide derived from the protein aggregate retains an abnormal 'non-self' configuration following disaggregation, proteolytic cleavage, and binding to the MHC, it will not be a strong activator of the adaptive immune system [21]. Moreover, classic MHC molecules are typically expressed at low levels in the CNS and though there clearly is continuous surveillance of the CNS by T cells, the low levels of T cells and limited MHC expression are likely to limit adaptive immune responses to misfolded self proteins in the CNS. Thus, misfolded self protein aggregates are not likely to strongly activate the adaptive immune response. Instead, the proteinopathy will largely activate the innate immune system.

The recent description of proteins that can exist as functional amyloid-like structures within select organelles challenges, to some degree, the notion that all aggregated misfolded proteins are PAMPs. Examples of physiologically 'functional' mammalian amyloids are currently limited to amyloid formation by select peptide hormones in secretory granules of the pituitary, amyloid present in semen, and Pmel17 in melanocytes [22-24]. Notably, both the secretory granule amyloid and Pmel17 amyloid are contained within intracellular vesicles that most likely sequester them from interaction with PRRs and other forms of innate immune surveillance. Furthermore, because the peptide hormone amyloids must dissociate in order for them to be active, they are distinct from many pathological amyloids and PAMPs, which are highly stable structures.

Amyloid or amyloid-like protein aggregates are highly resistant to degradation [11]. Perhaps the clearest illustration of this stability is seen in both in vivo imaging studies in Aβ protein precursor transgenic mice and cross-sectional pathology studies in inducible Aβ protein precursor transgenic mice [25-28]. These studies demonstrate that the amyloid deposits, once formed, are incredibly stable even in the absence of ongoing Aβ production. Though intracellular aggregates, such as those found in the polyglutamine diseases, can in certain circumstances be cleared in mice in which the transgene is turned off entirely, this situation is not replicated in the human disease and attempts to clear the aggregates may result in direct impairment of the protein quality control machinery [29-32]. Significantly, amyloid or amyloid-like protein aggregates catalyze the structural conversion of the normally folded protein into additional aggregates via a seeded nucleation-dependent process. Thus, following nucleation, the ongoing production of a 'normal' precursor drives additional amyloid formation [33,34]. In contrast to amyloid, the stability and 'seeding' or nucleating potential of other potentially pathogenic oligomeric structures formed by misfolded proteins has not been studied in detail [35,36].

We postulate that extracellular and intracellular protein aggregates act like PAMPs and result in chronic activation of the innate immune systems through PRRs. The concept that misfolded protein aggregates are PAMPs and activate innate immunity has been previously suggested by a number of groups and is supported by a plethora of experimental data [37-39]. Amyloid and amyloid like aggregates regardless of their peptide or protein subunit can be shown to bind and activate a whole array of PRRs, including Toll-like receptors, formyl peptide receptors, receptor for advanced glycation end products, scavenger receptors, complement and pentraxins [37-39]. Oligomeric assemblies of amyloidogenic proteins have not been studied as intensively with respect to PRR activation, but in the cases that they have they can be shown to elicit effects similar to amyloid fibrils (reviewed in [37]). Structurally, it is likely that oligomeric proteins resemble viruses or virus-like particles, which are known to function as PAMPs [40].

Most of the experiments that have established amyloid, amyloid like structures, and oligomers as PAMPs have involved direct application of these aggregates to cells in culture [37]. Such studies are complemented by histopathological studies that show co-localization of inflammatory cells and mediators with amyloid plaques and in vivo studies using multiphoton imaging that show the rapid mobilization of microglia to newly formed plaques [38,39,41-43]. These studies strongly support the concept that extracellular proteinopathies activate PRRs. Less well supported by direct experimental data is the notion that an intracellular protein aggregate acts like a PAMP, resulting in activation of PRRs and mobilization of the innate immune defenses. Nevertheless, a number of pathological features seen in intracellular proteinopathy-induced neurodegeneration suggest that these intracellular aggregates are activating the innate immune system. For example, binding to heat shock proteins, induction of autophagy and binding to intracellular PRRs can all activate the innate immune system [37,44-46].

The notion that intracellular proteinopathies activate innate immunity incorporates and extends some aspects of the danger theory of immune activation. This theory postulates that an intracellular stress or pathogen results in the cell generating a 'danger' signal that activates the immune system [47]. In this case we postulate that an intracellular protein aggregate causes the neuron or other CNS cell to send out 'danger signals' that activate the innate immune system. Such 'danger signaling' might explain the observation that inflammatory markers are often the earliest sign of pathology in experimental models of neurodegenerative proteinopathies [38,48-51]. Ultimately, intracellular or extracellular, stable protein aggregates acting as PAMPs will produce a chronic inflammatory condition.

A major ongoing debate in the AD and the larger neurodegenerative disease field is whether small soluble aggregates or larger, less soluble aggregates are the principle toxic species [52-55]. In the context of this hypothesis, both small soluble aggregates - oligomers - and larger aggregates - fibrils - will function as PAMPs. Depending on their concentration, location, and degradability, their ability to activate PRRs will likely vary. Significantly, the activation of PRRs by PAMPs elicits a response that is designed to result in clearance or sequestration and inactivation of the PAMP. In addition to a whole host of other variables, differential activation of the innate immune system by different protein aggregates and variable clearance of the aggregates following immune activation probably contribute to the imperfect correlations between the amounts and regional distribution of protein aggregates with clinical and neuropathological phenotypes [56]. Indeed, an aggregate that elicits the strongest innate immune response may be cleared more effectively. If this is the case, then it will always be challenging to link the aggregate to downstream pathology through cross-sectional analyses.

Histopathological, biochemical and molecular studies unambiguously show that the AD brain is subject to a chronic inflammatory condition [38]. The widespread gliosis and increased levels of numerous inflammatory factors, including, but not-limited to, chemokines, cytokines, and acute phase reactants, in the absence of overt lymphocytic or mononuclear infiltrates is consistent with inflammation resulting from innate immune activation and not adaptive immune responses. In AD, the inflammatory changes are noted in the earliest stages of the disease process and have also been shown to be early events in some of the AD mouse models of amyloid and tau pathology. As noted above, in some cases the earliest pathology noted is microglial activation and increased levels of select cytokines [38,48-51]. Though often more focal in nature, inflammation is a hallmark of other CNS proteinopathies and is often seen as an early change in mouse models of these diseases [57].

Recent studies have revealed a remarkable connection between inflammatory mediators and replicative senescence [58-64]. These studies demonstrate that a hallmark of replicatively senescent cells is a massive increase in the secretion of multiple pro-inflammatory proteins, including IL-6, IL-8 (CXCL8), IL-1α and β and monocyte chemoattractant peptide-1 (MCP-1, CCL2) [60,63]. In certain cases it has been shown that these secreted proteins can act in an autocrine manner to further maintain the senescent state, drive senescence of neighboring cells in a paracrine fashion, and promote degenerative or proliferative changes in neighboring cells [58,60]. It has also been shown that key inflammatory pathways, including those mediated by IL-6 and CXCR2 ligands, are not only upregulated by senescence but may play a critical role in inducing and maintaining senescence [59,62-64]. Notably IL-6 and MCP-1 are markedly upregulated in AD, as are CXCR2 receptors [38,65-71]. Finally, it is well-established that oxidative stress, which almost invariably accompanies chronic inflammation, can also induce senescence [72-74]. Oxidative stress can also arise independently of inflammatory pathways in CNS proteinopathies [75]. Extracellular Aβ has been reported to directly cause oxidative stress through production of reactive oxygen species [76,77]. Oxidative stress arising from mitochondrial dysfunction has also been reported to be associated with numerous neurodegenerative proteinopathies [75]. Thus, senescence appears to be associated with an induction of a pro-inflammatory state, but can also result from an inflammatory state. Moreover, inflammatory mediators and oxidative stress can synergistically act to drive senescence.

Senescence has largely been studied in the context of dividing cells, a phenomenon more specifically referred to as replicative senescence. Replicative senescence was first described by Hayflick, and the 'Hayflick limit' refers to the limited replicative capacity of primary human fibroblasts or other diploid cell lines to prolonged passaging in tissue culture [78]. Typically, such senescence results in cells with altered morphology (large, flattened cells with high cytoplasm to nucleus ratios), telomere shortening or telomerase malfunction, distinct senescence-associated hetrochromatic foci, and increased expression of a panel of senescence-associated biomarkers (for example, senescence-associated β-galactosidase activity, INK4A, IL-6) [58,60,79]. A functional definition is that a senescent cell has lost its replicative capacity and is no longer able to respond to growth factors. Though replicative senescence has been postulated to be a key driver of human aging, it is more likely that replicative senescence is simply one of many factors that contribute to the aging process.

To date there has been very little study of senescence of neurons or even glia cells. As neurons are terminally differentiated, it is immediately obvious that one of the critical hallmarks of replicative senescence, inability to divide, does not apply. Some would claim that because they are terminally differentiated, neurons cannot senesce [80]. However, we postulate that neurons do undergo physiological senescence and that this senescence is accelerated in AD and other CNS proteinopathies by inflammatory and oxidative stimuli. Moreover, we postulate that a senescent neuron will be defined functionally by its inability to respond appropriately to growth factors and its expression of senescence-associated proteins. In this scenario, other CNS cells, including glia, neuroglial stem cells, and endothelial and smooth muscle cells that form the cerebrovasculature, are also likely to undergo senescent changes in response to the chronic inflammatory environment. Senescent astrocytes might switch from a neuroprotective phenotype to one that is less suitable for supporting neuronal homeostasis. Senescence of microglia cells has been proposed as a mechanism for 'aging' microglia, less efficient scavenger cells with diminished phagocytic capacity and enhanced neurotoxic potential [81,82]. Put simply, we postulate that CNS cells will display a senescent phenotype that is physiologically similar to cells that have undergone replicative senescence, and be functionally impaired in a way that leads to neuronal dysfunction and degeneration. Ultimately, a growing number of senescent cells will lead to either widespread brain 'organ failure' as exhibited in AD, or more or less regional brain organ failure as seen in other neurodegenerative proteinopathies.

One of the key features of this hypothesis is that once triggered by the proteinopathy, senescent changes are likely to be both self-reinforcing and irreversible. In an autocrine fashion, inflammatory mediators secreted by the senescent neurons and glia would help to maintain the senescent state [60]. In a paracrine fashion, senescent cells induce additional inflammation and senescence of neighboring cells. Over long periods of time senescent neurons become increasingly dysfunctional and die due to a combination of diminished response to growth factors and possible pathophysiological effects of chronic exposure to an altered milieu of signaling factors as well as the direct signaling effects of the protein aggregates. Senescent changes can also affect neuronal stem cells, leading to diminished potential for renewal of neurons.

It is likely that the senescence response is not an 'all or none' phenomenon. There may be a graded continuum of responses to a proteinopathy-induced stress that depends both on the strength and acuteness of the stress as well as the preprogrammed response of the cell. Experimental data demonstrate that cells with high levels of anti-apoptotic proteins often undergo senescence whereas cells with lower levels of anti-apoptotic factors seem prone to undergo apoptosis [83-85]. Mature neurons, which are known to express high levels of anti-apoptotic proteins, may respond to potentially apoptotic stresses by senescing. At least in culture, a lower level of oxidative stress can drive senescence whereas a higher level can drive apoptosis. The notion that cellular 'stress', depending on the context, can result in two different endpoints, either apoptosis or senescence, may help to explain the disparate endpoints observed in various neurodegenerative proteinopathy models. In models where the cells are 'primed' to undergo apoptosis, a proteinopathy-driven stress will more likely drive apoptosis. In models where the proteinopathy is overwhelming, apoptosis may also be the primary endpoint. If the proteinopathy develops more insidiously, senescence may result.

Of course there is extensive neuronal loss in AD. So how would a senescent cell die? Senescent cells are stable for some period of time, and little information has been published on how they die. As noted above, in vitro studies suggest that senescence is an alternative pathway to apoptosis and that senescent cells are resistant to apoptosis. Of note, a recent study has shown that senescent keratinocytes die by autophagic cell death, a cell death pathway characterized by an increase in macroautophagic activity [86]. In many neurodegenerative diseases auto-phagic cell death has been implicated as an alternative to apoptotic or necrotic mechanisms [45,87].

A more speculative extension of this hypothesis in AD is that the senescent phenotype could be the key link between Aβ proteinopathy and secondary proteinopathies that are seen in the AD brain, including those involving tau, α-synuclein and TDP-43. At least for tau and α-synuclein there is experimental evidence that an Aβ proteinopathy can enhance, if not trigger, a tauopathy or synucleinopathy. Despite intense investigation, there is no consensus regarding the pathways that relay the signals between Aβ and tau, synuclein, or TDP-43. Although it is possible that we simply have not identified the single critical factor, it is perhaps more likely that Aβ proteinopathy induces a plethora of changes that drive the secondary proteinopathies. Given that senescence triggers gross changes in the transcriptome and secretome, perhaps senescent changes mediate the secondary proteinopathy?

Recent studies of the senescence-associated secretory phenotype (SASP), also termed the senescence-messaging secretome (SMS), have identified a number of secreted biomarkers associated with replicative senescence [58,60,63]. Depending on the cell type examined, the method of induction of senescence, and the methodology used to identify the secreted proteins, the secretome and transcriptome that defines the SASP/SMS can be variable. However, certain proteins are invariably identified in these studies. Many of the proteins consistently upregulated during senescence are inflammatory mediators, including IL-6, IL-8 and other chemokines and cytokines [58,60,63]. Though not a perfect correlation, there is extensive overlap between the biomarkers that comprise the core SASP/SMS phenotype and secreted biomarkers of AD (Table 1). This overlap provides evidence for a SASP/SMS in the AD brain, with many of these features seen in mouse models of AD. At least in AD, it has been challenging to define the nature of the 'immune-system dysregulation' based on pathway-type analyses; the overlap between the SAPS/SMS and biomarkers of AD suggest that the 'immune-system dysregulation' may, in fact, reflect senescence [71].

Given the large number of inflammatory proteins that have been implicated in the SASP/SMS, one of the challenges in moving forward is to discriminate between 'classic' reactive neuroinflammation and senescence in the diseased brain. If one uses replicative senescence as a guide, the key observations that would distinguish a senescence from an inflammatory phenotype are: that inflammatory mediators are expressed by cells such as neurons that do not normally produce them; that neurons and other cells in the AD brain exhibit cytoplasmic and nuclear markers of senescence; and that the cells displaying these markers of senescence are both functionally impaired and exhibit a SASP/SMS phenotype. The links between senescence and inflammation could also be evaluated through a number of experimental paradigms. For example, does overexpression of inflammatory factors, such as IL-6 in the brain, drive senescent markers in CNS cells? Does forced expression of classic inducers of senescence, such as P53 or INK4A, in adult neurons drive senescence and inflammation?

Though clearly there is much work to be done to move this hypothesis forward, there is sufficient evidence in the literature to make the case for further study. In the AD brain neurons strongly stain for MCP-1 and IL-6, suggesting that these inflammatory mediators are being 'ectopically expressed', or at least dramatically upregulated, in cells that are not professional immune cells [65]. In addition, neurons stain and can be shown to actually express the mRNA for a number of other secreted inflammatory proteins [37,38]. Notably, some of the neuronal expression of inflammatory markers can be seen in mouse models of AD and other neurodegenerative disorders [48,88]. Cytoplasmic or nuclear protein markers associated with senescent cells are often cell-cycle proteins, tumor suppressors, or cell-cycle regulators [80]. In AD and other neurodegenerative diseases there is often marked upregulation of cyclins, p53 and related proteins, and cyclin-dependent kinase inhibitors that have also been implicated in the senescent phenotype (Table 2) [89-91]. Often these markers are upregulated in tangle-bearing neurons. Notably, there are no reports about the presence of a widely used 'biomarker' of senescence, senescence associated β-galactosidase activity, in AD or any other neurodegenerative condition [92].

Changes in telomere length, telomere activity, and the presence of distinct heterochromatic nuclear bodies (called senescence-associated heterochromatic foci) are also well-established markers of replicative senescence [80,93]. However, given the terminally differentiated state of neurons, it is unclear whether these markers would be expected to be seen in physiological senescence of neurons in the AD brain, or even in any of the CNS cells. For example, even senescence-associated heterochromatic foci are typically only seen in human cells and have been much more difficult to demonstrate in mouse cells [94,95]. Furthermore, mouse cells can undergo replicative senescence without shortened telomeres [80]. Thus, it is clear that the nuclear biomarkers of human replicative senescence may not apply to studies of CNS senescence.

Autophagy represents an additional potential link between senescence and neurodegeneration. Altered autophagy has been implicated in AD and many other neurodegenerative conditions [45,96]. Genetic removal of genes involved in autophagy results in neurodegeneration [97,98]. More generally, autophagy plays a key role in organismal aging and lifespan [99,100]. Genetic alterations that induce premature aging phenotypes are associated with autophagy induction. Autophagosomes accumulate in senescent fibroblasts [101]. In addition, it also has been shown that autophagy is an effector mechanism of replicative senescence [102]. It is activated during senescence, as are a subset of autophagy-related genes, and inhibition of autophagy delays oncogene-induced replicative senescence [103,104]. Finally, it has recently been postulated that autophagy is a key mechanism in immune responses to intracellular pathogens [105].

The main working hypothesis in AD and other CNS proteinopathies has been that some species of the aggregated misfolded protein are directly neurotoxic [1-4,106]. There have been many variations on the theme of the presumptive neurotoxic protein aggregate. In primary neuronal culture systems it has been possible to reproducibly demonstrate that a variety of protein aggregates cause some form of 'neurotoxicity'. For example, Aβ aggregates ranging from oligomers (dimers, trimers, tetramers, dodecamers to 50-to 100-mers), soluble protofibrils, actively growing fibrils, to mature fibrils have been implicated as potential pathological entities in AD [3,107-114]. Despite this intense focus on finding the exact assembly that is the real 'neurotoxin', there is little, if any, consensus in the field regarding this issue, in large measure because it has been difficult to unequivocally demonstrate that some of the proposed misfolded neurotoxins exist in vivo. Given the considerable body of data showing that protein aggregates can be directly neurotoxic by causing calcium influx, altering synaptic plasticity, impairing axonal transport and mitochondrial function, or altering other homeostatic functions in the cell, we believe that direct neurotoxic effects of protein aggregates probably do play some role in AD and other neurodegenerative diseases. However, we would argue that a slow degenerative phenotype is hard to reconcile with a direct toxic mechanism and that if indeed there were a 'smoking gun' aggregate that was directly neurotoxic, that it would likely show a much better and more consistent correlation with disease or disease progression than do any of the current aggregates.

Organisms have developed many ways to adapt to stressful stimuli. In the mature nervous system a key adaptive mechanism is to try to keep largely irreplaceable neurons alive. One of the most remarkable examples of this is that pathogenic viruses can be cleared from neurons by activation of the innate and adaptive immune systems in a non-cytolytic fashion [115]. In the periphery this immune activation would typically result in significant collateral damage and killing of the infected cells. As noted previously, there is some evidence that, in response to stress, a cell can either undergo apoptosis or senescence [104]. Given that mature neurons have many mechanisms to protect them from apoptosis, we might speculate that the same stress that induces apoptosis in a primary embryonic neuronal culture may induce senescence in the intact mature CNS.

Replicative senescence does not appear to be easily reversible [80]. In the few examples where replicative senescence has been reversed in culture, the reversal typically requires inactivation or downregulation of tumor suppressors [116,117]. Thus, from a therapeutic point of view, reversal of a replicatively senescent phenotype poses serious problems as it increases the likelihood for tumorigenesis. Indeed, it is generally thought that replicative senescence is a mechanism designed to suppress tumorigenesis. If cells in the AD or other neurodegenerative disease brain are physiologically senescent, it may be very challenging to reverse the senescent phenotype. Of course, an enhanced understanding of physiological neuronal or glial senescence may reveal distinct differences between replicative senescence and the physiological senescence of these specialized cells. Understanding whether such differences are present may reveal new therapeutic approaches to treat many neurodegenerative diseases.

Recognizing that proteinopathy-induced inflammation may drive senescence, and thereby induce neurodegeneration, reinforces therapeutic efforts designed to prevent the formation of or clear the proteinopathy and also potentially reveals new pathways that could be the focus of therapeutic efforts. In the former case, the rationale is obvious - prevent the proteinopathy and the downstream cascade is prevented. Of course, as discussed in recent reviews, this type of therapy targeting the trigger of the disease is likely to be much more effective as primary prevention and may have little therapeutic benefit once degeneration is entrenched [5,118]. If some aspects of the degenerative cascade downstream of the proteinopathy are self-reinforcing and difficult to reverse, then therapeutics aimed at the initiator of the cascade will almost certainly have limited benefit when administered once these downstream cascades have begun. In the latter case, the notion that the inflammatory and senescent phenotypes may be mutually reinforcing responses that are capable of inducing pathological changes in a paracrine fashion establishes a new framework for understanding the interplay between chronic neuroinflammation and neuronal dysfunction. Further elucidation of an inflammatory senescence network may reveal multiple new targets for intervention. In particular, novel anti-inflammatory approaches designed to reduce the paracrine effects of the SASP/SMS may limit the spread of a neurodegenerative process, and thereby limit the collateral damage caused by a proteinopathy.

The major risk factor for developing AD and other neurodegenerative diseases is age. Because of this association many in the field have proposed that aging contributes to the risk for developing AD and other neurodegenerative diseases. Often the semantic distinction between age and aging is not well-defined even by those who use the terms. We use the term aging to refer to distinct biological processes that are altered as an organism grows older, and age will simply be used to denote time. Though it remains possible, and even likely, that aging does contribute to the risk of developing AD or other neurodegenerative conditions, genetic studies indicate that aging effects can be overcome. Essentially, the same neurodegenerative disease can be driven in a relatively young person by a genetic alteration that accelerates the induction of the proteinopathy. For example, when the polyglutamine expansion is large enough, Huntington's disease can occur in children [119].

As alluded to previously, replicative senescence has been implicated as a key component of the aging process. In the context of a neurodegenerative cascade we would propose that senescence of CNS cells is a reinforcing pathway downstream of a proteinopathy that can, in an autocrine and paracrine fashion, create an environment that results in aging of the brain. Indeed, senescence appears to reinforce both chronic inflammation and oxidative stress, two factors that are thought to play a key role in the aging process. This might explain why genetically driven early onset forms of AD and other neurodegenerative diseases mimic late onset sporadic forms of the disease.

The proteinopathy-induced senescent cell hypothesis of AD and neurodegenerative disease that we describe here provides a novel integrative intellectual framework for future studies of pathological cascades in AD and other neurodegenerative diseases. Such studies may broaden our understanding of the phenotype of senescing cells and also identify novel therapeutic targets for the treatment or prevention of AD and other neurodegenerative disorders.

Aβ: amyloid β; AD: Alzheimer's disease; CNS: central nervous system; IL: interleukin; MAPT: microtubule associated protein tau; MCP: monocyte chemoattractant peptide; MHC: major histocompatibility complex; PAMP: pathogen-associated molecular pattern; PRR: pattern recognition receptor; SASP: senescence-associated secretory phenotype; SMS: senescence-messaging secretome; TI-2: thymus independent type 2.

TEG is a co-editor in chief of this journal, for which he receives an honorarium. He serves on the SAB for Alzheimer's disease for Élan Pharmaceuticals, and has received sponsored research support from Lundbeck and Myriad Genetics. He is an inventor on several patents related to gamma-secretase modulators and anti-Aβ immune therapy. VMM has no conflicts.