1

The Therapeutics of Alzheimer’s Disease where we stand and where we are heading

Dennis J. Selkoe, MD

Center for Neurologic Diseases Department of Neurology Brigham and Women’s Hospital Harvard Medical School Boston, MA 02115

Running head: AD therapeutics [email protected] 617-525-5200 One figure

Abstract

Few diagnoses in modern medicine evoke more apprehension in patient and family than Alzheimer’s disease. Defined as a clinical and pathological entity a century ago, the disorder only came under intense molecular scrutiny in the mid 1980s. Genetic, histopathological, biochemical and animal modeling studies have combined to provide evidence that the disease may begin with an imbalance between the production and clearance of the self-aggregating This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an ‘Accepted Article’, doi: 10.1002/ana.24001

2

amyloid β-protein (Aβ

) in brain regions serving memory and cognition. This concept has

been furthered by recent analyses in humans of cerebrospinal fluid and neuroimaging biomarkers that suggest an approximate sequence of AD-type brain alterations beginning more than two decades before onset of dementia. Although the Aβ hypothesis of Alzheimer causation does not explain all features of this multifactorial syndrome, experimental agents that lower or neutralize Aβ have become the major focus of therapeutic research. Several clinical trials in mild-tomoderate AD have not met standard cognitive and functional endpoints, but there were important shortcomings in the agent and/or the trial design in each case. Based on the lessons learned, the field has moved on to test potentially disease-modifying agents in mild AD patients or via secondary prevention in pre-symptomatic subjects bearing amyloid plaques. Immunotherapeutic agents are receiving the most study, but other anti-amyloid strategies and, importantly, nonamyloid targets such as tau and neuroinflammation are of great interest. The pace of recent developments augurs well for one or more experimental agents being shown to slow cognitive decline without major side effects. However, research funding from all sources will need to increase dramatically and soon to stave off the approaching tsunami of Alzheimer’s disease.

A remarkable rise in life expectancy in the last century and the declining prevalence of certain causes of late-life mortality have allowed dementia to become a common occurrence with age. Epidemiological studies suggest that Alzheimer’s disease (AD) and other neurodegenerative dementias will become even more prevalent by mid-century, constituting a major personal and societal tragedy. These sobering projections heighten concern about the inability to date to slow the progression of AD and related dementias. Yet that this should be so

3

is not surprising, as research on the molecular basis of AD began less than 30 years ago, and the development of life-prolonging treatments for cardiovascular disease such as statins followed several decades of intensive study of atherosclerosis and lipoprotein biology.

As in other chronic diseases, treatments for AD can be roughly divided into those that are symptomatic (inducing relatively rapid but temporary amelioration of symptoms without apparent slowing of disease progression over years) and those that are potentially diseasemodifying (not necessarily improving symptoms acutely but slowing or halting the decline of cognitive function over years). In the former category, AD patients in the United States have long been offered one of three FDA-approved acetyl cholinesterase inhibitors (donepezil, rivastigmine or galantamine) that have similar modes of action and effects, and also memantine, a glutamate antagonist that is believed to help ameliorate behavioral symptoms, particularly in the moderate phase of the disease and beyond. Treatment with one or both of these two classes of symptomatic drugs has become the standard of care for patients with AD, and most clinical trials of potentially disease-modifying agents in symptomatic AD patients are conducted additively. Many AD patients also receive drugs for psychological and behavioral symptoms such as anxiety, agitation, aggression or depression, although these agents should be prescribed with caution and for relatively brief periods, as they are often fraught with side effects and can temporarily worsen cognition.

New symptomatic compounds and various methods of behavioral modification intended to better manage AD are being evaluated, and these are clearly very important areas for more clinical research. But the central quest of the AD therapeutic field is to identify compounds which can

4

significantly slow progression of the cardinal cognitive symptoms of the disorder without major side effects, and the rest of this article will review approaches to disease modification.

Assessing the reasons for recent therapeutic failures

Current pessimism about progress in finding therapies for AD must be considered in light of the specific reasons for the failures of individual agents in advanced clinical trials. For example, semagacestat, a non-selective inhibitor of the γ-secretase enzyme that generates the amyloid βprotein (Aβ), had a therapeutic index of about 3; that is, the concentration of the drug that inhibited by 50% the cleavage of the Notch protein by γ-secretase was only 3-fold higher than the concentration which inhibited by 50% the γ-cleavage of Amyloid Precursor Protein (APP). Accordingly, some subjects developed serious side effects apparently attributable to decreasing the processing of Notch and perhaps other γ-substrates besides APP 1. The worsening of cognitive performance seen in some patients in this aborted trial may have been due to these side effects; there is no current evidence that it was due to excessive lowering of brain Aβ levels, certainly not below the physiological range. In the case of bapineuzumab, a monoclonal antibody to the free N-terminus of Aβ that can bind monomers, soluble oligomers and amyloid plaques, the unanticipated emergence of amyloid-related imaging abnormalities with edema (ARIA-E; formerly called vasogenic edema) markedly limited dosing 2, leading to a lack of cognitive benefit in Phase 3 despite small but significant decreases in amyloid PET scan signals and CSF phospho-tau levels. In contrast, solanezumab, a monoclonal to the Aβ mid-region that binds principally monomers, evoked no ARIA-E, could thus be dosed 15-30 times higher than bapineuzumab in two Phase 3 trials, and produced a significant ~35% decrease (p200 uM) and penetrated the brain poorly.

The salient lesson from such trial specifics is that failures in Phase 2 and 3 can ensue when the agent in question lacks potency, does not have a sufficient safety margin, and/or is given too late to expect significant disease slowing. It is also true that some agents have not received sufficiently extensive evaluation in animal testing to accurately predict their efficacy vs. safety margins. Collectively, these concerns must now lead to more carefully designed clinical trials of more rigorously vetted and efficacious compounds.

Another potential reason for the recent failures of certain anti-amyloid agents could be that lowering or neutralizing Aβ is not a useful therapeutic approach for AD, regardless of disease stage. But this conclusion would not be consistent with the genetics of the human disease. Inherited missense mutations solely within and immediately flanking the Aβ region of APP or else in the catalytic subunit of γ-secretase, presenilin 4, can cause clinically and neuropathologically typical AD. Conversely, an inherited missense mutation at the second amino

6

acid of the Aβ region sharply reduces the β-secretase cleavage of APP lifelong and thus prevents the development of AD 5. Moreover, direct support for the clinical feasibility of anti-amyloid approaches comes from the independent analyses of the solanezumab data in Lilly’s Expedition 1 and 2 trials by the Alzheimer’s Disease Cooperative Study (ADCS), which documented a significant lessening of cognitive decline in mild AD patients by an agent that is entirely specific for Aβ. As a result of this finding, solanezumab has been chosen for Phase 3 trials in asymptomatic Presenilin and APP mutation carriers by the Dominantly Inherited Alzheimer’s Network (DIAN) and also by the ADCS for its “A4” (Anti-Amyloid in Asymptomatic Alzheimer’s disease) secondary prevention trial in 65-85 year old cognitively normal subjects prone to developing ‘sporadic’ AD. Furthermore, Lilly has announced it will begin another large (~2100 subject) Phase 3 trail of solanezumab confined to mild AD. Finally, the successes in slowing other human amyloidoses by targeting the respective amyloid subunit protein (e.g., transthyretin) predicts that anti-Aβ agents have the potential to show clinical efficacy in AD.

Targeting the right patients in AD clinical trials

A key lesson underscored by the generally disappointing Phase 3 trial results to date is the relative imprecision of a clinical diagnosis of AD, even when physicians follow rigorous, preestablished clinical criteria. In both the bapineuzumab and solanezumab trials, as many as ~30% of the trial patients did not show abnormal amyloid burdens by PET scanning; this was much more common among those not carrying an Apolipoprotein ε4 allele, which is known to strongly predispose to amyloid deposition. Therefore, those trial subjects did not have evidence of the pathobiological process of AD (amyloid deposition is a sine qua non of the diagnosis) and could

7

not be benefitted by an anti-amyloid agent. Subsequent trials will need to consistently use biomarkers which document the presence of amyloid pathology (i.e., amyloid PET scans and/or lowered CSF Aβ42 levels) to ensure that only subjects with neuropathological AD are enrolled.

A related issue is the frequent co-occurrence of cerebrovascular disease in subjects with AD neuropathology. In one recent study, 226 cognitively normal subjects at entry were followed for a mean of ~10 years and developed clinically diagnosed AD sometime during follow-up. Of these 226 clinical AD patients, 143 died during the study, and 126 had complete neuropathological exams. Among the latter, 101 (80%) met the neuropathological criteria for AD, but of these, 67 (66%) also had at least one other neuropathological condition, most commonly one or more cerebral infarcts 6. This finding is consistent with earlier evidence that many clinically diagnosed AD subjects have multiple pathological processes underway (e.g., 7). Some of these ‘other’ processes (e.g., cortical Lewy bodies; neurons positive for the TDP-43 protein) may actually be downstream cellular changes of the AD cascade induced by progressive Aβ accumulation, so the burden of these lesions might theoretically respond to anti-Aβ therapeutics. However, the mixed nature of many late-onset AD cases, particularly the presence of cerebrovascular lesions, means that highly specific therapeutic agents (e.g., anti-amyloid or anti-tau antibodies) should be tested in subjects with as “pure” a form of AD as confirmable with clinical and biomarker evidence, so that one can judge whether the AD process per se can be ameliorated. Such agents could prove helpful to humans who have pathology largely restricted to the classical AD cerebral phenotype, but they could also be tried later in those with apparent mixed dementias to see whether some partial benefit ensues.

8

Selecting therapeutic approaches with promise based on understanding the AD pathogenic sequence

Despite ongoing interest in further testing of anti-amyloid agents, it is vitally important that compounds for alternative therapeutic targets which are well-validated in AD pathogenesis are brought forward. In this context, we will briefly review the latest information about the multifactorial events that appear to contribute to the progression of the disease. Attempts to e the patho en



in AD have relied principally on four approaches:

1) examining the development of AD-type neuropathological changes in subjects with Down’s syndrome (Trisomy 21) of increasing age; 2) recognizing that AD-causing mutations in APP, presenilin 1 or presenilin 2 lead to clinical and neuropathological phenotypes largely indistinguishable from those of ‘sporadic’ late-onset AD; 3) studying temporal changes during the aging of transgenic mice expressing mutant human APP with or without mutant human presenilin (although such models are perforce biased to an amyloidogenic mechanism); and most recently, 4) examining the time-dependent appearance of fluid and neuroimaging biomarkers in both familial and ‘sporadic’ AD subjects.

Collectively, these four approaches have suggested that AD can involve the accumulation of Aß42 in association and limbic cortices many years -- probably two decades or more -- before the onset of abnormalities on standard cognitive tests for dementia.

The most recent and

compelling evidence for this conclusion comes, for example, from the biomarker analyses performed on presymptomatic subjects carrying presenilin or APP mutations in the Dominantly Inherited Alzheimer Network (DIAN) Study 8. Numerous families carrying such deterministic

9

mutations have been studied collectively to ascertain the time course of fluid biomarker changes, neuroimaging changes and subtle cognitive changes prior to the expected onset of AD symptoms (based on the age of symptom onset in a parent with the same mutation). The initial analyses of this cohort suggest that Aß42 levels in CSF begin to decline as early as 25 years before expected symptom onset 8. This is followed by the appearance of fibrillar amyloid deposits in the brain (as detected by amyloid PET scans), increased levels of tau in CSF and progressive brain atrophy roughly 15 years before expected symptom onset 8.

Cerebral hypometabolism and subtly

impaired performance on challenging measures of memory seem to begin some 10 years or so before expected symptom onset 8. If this time course is generally similar to that of ‘sporadic’ AD, and there is growing evidence from cross sectional studies that it may be

9-11

, then humans

destined to develop AD have detectable biochemical and histopathological abnormalities two decades or more before overt clinical symptoms. Two key lessons that emerge from such studies of pre-symptomatic AD are: (i) profound brain alterations occur long before the dementia can be diagnosed

12, 13

; and (ii) therapeutic interventions directed only at the mild-to-moderate clinical

stage may be too late to ameliorate symptoms (Figure 1).

In light of these emerging data, the time-honored principle of targeting prevention to the earliest steps in a disease cascade would encourage more and better designed trials of amyloid-lowering or -neutralizing agents. The imperative to overcome major disappointments in the recent past requires us to identify Aß/amyloid-directed compounds that have rigorously demonstrated benefits in animal models (e.g., in blinded, publically registered mouse trials) and very few adverse effects upon prolonged animal dosing. More careful testing in multiple mouse models and larger mammals and then trying the cleanest compounds in humans at early stages in the AD

10

process, especially before the time of major cognitive symptoms, is likely to lead to progress (reviewed in 14). (For a complete list of current clinical trials, go to http://www.clinicaltrials.gov/ and enter ‘Alzheimer’s disease’ in the search field.)

A final issue in pinpointing the temporal sequence of AD pathogenic changes: the above biomarker analyses as well as studies of Down’s syndrome brains highlight very early accumulation of Aβ42, and yet some neuropathological surveys of humans of increasing age have suggested that altered tau proteins and neurofibrillary tangle formation may precede amyloid build-up 15-17. However, the latter approach focuses on detection of occasional taupositive neurons and tangles in young and middle-aged humans who died of various causes without other evidence of an AD-type process. Because tangles have long been known to occur in diverse brain disorders that are etiologically distinct from AD, including some in early life (e.g., subacute sclerosing panencephalitis, a rare childhood complication of measles), one can conclude that tau alteration/tangle formation can be a non-specific response on the part of some neurons to a variety of insults. The observation of occasional tau-positive tangles or dystrophic axons in young humans dying of unrelated causes does not mean that these individuals would have developed pathologically and clinically typical AD had they lived into late life. So the detection of incidental tangles in some younger humans does not necessarily define the first neuronal alteration of AD per se, particularly as such brains have not been probed for accumulation of soluble Aβ oligomers. Indeed, the fact that profound, ultimately fatal tangle formation in frontotemporal lobar dementia due to tau mutations does not lead secondarily to Aβ deposition and an AD phenotype argues strongly that tangle-related neurodegeneration does not induce the Aβ phenotype, while there is much evidence for the converse sequence. The

11

hypothesis that these two processes can arise independently in some humans and that early Aβ accumulation can accelerate subcortical and cortical tau alteration is certainly reasonable 11, 18.

Finding anti-amyloid approaches with better therapeutic indexes

Aβ immunotherapy -- At this writing, immunotherapeutic approaches using either passive antibody infusion or active Aβ vaccination are the furthest along in clinical testing among the various anti-amyloid agents. Much debate on this topic centers on which Aβ epitope and which form of the peptide is the most desirable to target immunologically. Active vaccination with synthetic Aβ immunogens yields a range of endogenous, principally N-terminally directed Aβ antibodies. Passively administered anti-Aβ antibodies have mostly been chosen to target the N-terminal region, although not necessarily the aspartate at the extreme N-terminus. Some of the Aβ N-terminal antibodies studied in human trials have been strongly associated with ARIA-E, including bapineuzumab and gantenerumab, while the mid-region antibody solanezumab has not. Some N-terminal antibodies with a weakened effector function, for example of the IgG4 subclass, are also being evaluated, because they are believed to have a lower propensity to induce ARIA-E and thus might be dosed higher and longer. On the other hand, a head-to-head comparison of a particular Aβ antibody having either an IgG2a (strong) or IgG1 (weak) effector subtype suggested that the latter was not as efficacious in lowering Aβ burden in an APP transgenic mouse 19.

Antibodies to the C-terminus of Aβ have not progressed very far in clinical trials. A few antibodies have been designed to bind selectively to a low-abundance form of Aβ in AD cortex,

12

namely, that having a modified N-terminus which lacks the first two residues and has its third residue (glutamate) cyclized to a pyroglutamate 19, 20. Despite its low abundance, there is preclinical evidence that this pyroGlu Aβ peptide can strongly seed the aggregation of conventional Aβ beginning at aspartate 1 21. An antibody selectively binding this form was recently shown in APP transgenic mice to decrease total Aβ load in the brain, as if targeting this low-abundance species (60%) decreases in all Aβ species can be achieved in human plasma and CSF. However, proteomic analyses have revealed that BACE1 has well over 100 physiological substrates besides APP (e.g., 26, 27), and interference with their processing via BACE1 gene deletion has lead to adverse phenotypes in mice. For example, both decreased peripheral and central myelination 28, 29 and decreased muscle-spindle formation30 can be caused by genetically reducing the processing of just one of these substrates, neuregulin-1, in mice. Such reports predict the likely recognition of adverse neurochemical effects from chronically inhibiting BACE1 and 2, particularly if done at high (>50%) levels of inhibition. Although this likelihood will presumably temper the enthusiasm for the recent progress of BACE inhibitor trails, selective β-secretase regulation

14

remains an attractive approach based on our understanding of the early role of Aβ accumulation in AD pathogenesis.

γ-Secretase inhibitors and modulators -- Inhibition of γ-secretase remains theoretically attractive, but the risk of interfering with the processing of the Notch receptors and many other substrates means that clinical γ-secretase inhibitors must be highly selective for APP over Notch. A lack of such selectivity may have contributed to the cessation of the phase 3 trial of semagacestat 1, which apparently had a therapeutic index for inhibiting the cleavage of APP versus Notch of less than 3. However, another partially Notch-sparing γ-secretase inhibitor, avagacestat, was also discontinued after phase 2 testing due to adverse gastrointestinal and dermatological effects suggestive of Notch inhibition at higher doses, despite it having a selectivity ratio of ~190-fold for APP over Notch cleavage inhibition in culture 31. As a result, research is underway to identify Notch-sparing γ-secretase inhibitors with even higher substrate selectivity ratios.

A principal approach to this problem has been the development of γ-secretase modulators (GSMs), which can shift the γ-secretase cleavage site N-terminally by approximately one helical turn (~3.6 amino acids) within the APP transmembrane domain (and presumably those of all other γ-secretase substrates), resulting in less Aβ42 and more Aβ38 and Aβ37 production, while leaving the major γ-secretase cleavage event at Aβ40 unchanged or only slightly decreased 32-34. Therefore, γ-secretase processing is “modulated” (partially shifted) but not fully inhibited, apparently by changing presenilin conformation 35. This allows the major γ-cleavages of APP, Notch and other substrates to occur and also does not affect the initial ε-cleavage that releases

15

the APP and Notch intracellular signaling domains into the cytoplasm. The precise mechanisms of action of the various GSMs are unclear, but it has been reported that recent (“second generation”) GSMs, both NSAID and non-NSAID type compounds, have these effects by binding to presenilin (the catalytic subunit of γ-secretase) 34, although it has also been postulated that some early (“first generation”) NSAID-derived GSMs may bind to a protein site within the APP transmembrane domain itself 36 or to a common site that includes sequences of both the APP substrate and the presenilin protease. Increasingly potent second generation GSMs have now been synthesized (e.g.,37, 38), and the question of whether this Notch-sparing modulatory approach will lower Aβ42 levels sufficiently to slow AD progression will be answered in current and future GSM trials.

Aß aggregation inhibitors – Because there is no evidence that the physiologically secreted Aβ monomer present in human biological fluids in picomolar to nanomolar levels is pathogenic, a major goal of anti-amyloid approaches is to prevent the monomer’s progressive aggregation into potentially cytotoxic oligomers and then amyloid fibrils. Despite the theoretical attractiveness of accomplishing this, it has been difficult in practice to obtain efficacious and safe anti-aggregation compounds. One reason is the general challenge of designing small molecules that can interfere effectively with protein-protein interactions, in contrast to inhibiting specific enzymes like β- or γ-secretase. The few Aβ aggregation inhibitors in the literature have not provided clear enough benefit in preclinical models or have failed in early clinical testing. However, one much-studied member of this therapeutic class is the naturally occurring glycolipid sugar, scyllo-inositol, which was discovered in in vitro assays to inhibit synthetic Aβ peptide aggregation 39. These investigators went on to show that oral administration of scyllo-

16

inositol in substantial pharmacological doses could lower amyloid plaque burden and prevent behavioral decline in APP transgenic mice, while its close stereoisomers had much less or no benefit 40. Moreover, scyllo-inositol prevented Aβ-mediated inhibition of LTP in mouse hippocampus, and oral administration to normal rats prevented memory deficits caused by acute cerebroventricular injections of soluble Aβ oligomers 41. Based on such evidence, scyllo-inositol was advanced into a Phase 2 trial in typical mild-to-moderate AD patients. The primary endpoints on a cognitive battery and an activities of daily living scale were not met, but CSF Aβ42 levels were significantly lowered, and post-hoc analyses suggested a trend toward some cognitive benefits and lower CSF phospho-tau levels in the mild patients 42. This agent is now being tried against secondary symptoms of agitation and aggressiveness in moderate AD patients. It should also be noted that some monoclonal antibodies which avidly bind Aβ monomers or soluble oligomers are likely to have anti-aggregation properties.

Therapeutic targets beyond Aβ

As these and other anti-amyloid approaches proceed, the field must design and execute more trials of agents that target the tau alterations which lead to profound neurofibrillary degeneration in AD. Indeed, it has been shown that the cytoskeletal neurotoxicity and behavioral deficits induced by Aβ oligomers in experimental models require the expression of tau, in that genetic deletion of tau strongly prevents these adverse effects 25, 43. Therapeutic approaches against neurofibrillary degeneration could include active and passive tau immunotherapy, tau-lowering agents including anti-sense and microRNAs, microtubule-stabilizing drugs, and perhaps selective inhibitors of certain kinases implicated early in the process of tau hyperphosphorylation. The

17

recent development of positron-emitting small molecules that cross the blood brain barrier and can image at least fibrillar deposits of tau in humans will be a real boon to therapeutic development around tau. And the long-standing ability to quantify tau monomers in human CSF, arguably the best available biomarker for neurodegeneration in AD, needs to be accompanied by measuring soluble oligomers of tau (and of Aβ) that are likely to exist in that fluid but have not yet been definitively identified.

Clinical trials targeting tau and the cytoskeleton in AD are underway. The most advanced is the small molecule methylthioninium (TRx0237), a derivative of the dye methylene blue, which may function to inhibit the aggregation of tau in AD and other human tauopathies and is now in a Phase 3 trial in AD 44. In addition, a microtubule-stabilizing drug, BMS 241027, is intended to decrease cytoskeletal disruption and related neurodegeneration and is in Phase 1 testing [http://clinicaltrials.gov/ct2/results?term=NCT01492374&Search=Search].

Another attractive topic for vigorous therapeutic study is the microglial-mediated inflammation that characterizes the AD cortex and that likely occurs quite early in the process of progressive Aß accumulation and fibrilization 45. The long-debated question of whether innate immune responses are “good” or “bad” in the AD brain will almost certainly be answered “both”. This aspect of AD cytopathology has drawn mechanistic study for many years, with an increasingly granular picture of the many microglial and monocyte surface receptors, chemokines, cytokines and acute phase proteins that appear to create complex feedback loops during the development of AD neuropathology. Abnormalities of the innate immune system comprise both peripheral macrophage/monocyte cells that survey the brain and respond to local alterations (including Aβ

18

deposition) and resident microglia that have been in the CNS since early embryogenesis (reviewed in 46). A major challenge to date in correctly analyzing the roles of these components has been the identification of highly specific protein markers that can distinguish these two “arms” of innate immunity in the CNS, but newly emerging data suggest that such specific markers exist 47.

Progress in the cell biology of immune cells in the brain has been accompanied by the identification of genetic risk factors for late-onset AD that are polymorphisms in certain key inflammatory proteins, notably complement receptor 148 and the microglial surface receptor TREM 2 49, 50. The classical complement cascade has been found to help mediate synaptic pruning during brain development 51, and adverse activation or repression of this process in AD cortex could play a role in the neurodegeneration of the disease. Polymorphisms in TREM2 are hypothesized to potentially alter the phagocytic properties of resident microglia, perhaps allowing decreased clearance of fibrillar Aß deposits. Many other inflammatory molecules likely play roles in the peri-plaque microgliosis and astrocytosis that are invariant features of AD. This is a highly complex but potentially very compelling area for discovering small molecules, microRNAs and antibodies which can regulate these cells to a lowered inflammatory state or help clear Aβ species.

Yet another area for therapeutic development revolves around changes in lipids involved in cell signaling and/or membrane dynamics. There has long been evidence that the levels of certain lipids are altered in postmortem AD brain, but like other changes documented at the end of the disease, it has been difficult to know whether these arose as an early part of the pathogenic

19

process and thus are attractive therapeutically. As just one example of a specific lipid change potentially relevant to AD and models thereof, a lipidomic analysis of APP transgenic mice revealed increases in arachidonic acid and its metabolites 52. Accordingly, the levels of the activated Group IV isoform of phospholipase A2 (GIVA-PLA2) were increased in both AD and transgenic mouse hippocampi, and genetic ablation of this enzyme in the mice protected them against Aβ-dependent deficits in memory 52. These preclinical observations suggest that inhibiting GIVA-PLA2 might prove beneficial in AD. An inhibitor of lipoprotein-associated PLA2 (rilapladib) is currently in a Phase 2 trial in AD subjects 53.

Of course, there exists a wealth of preclinical reports of specific molecular changes and manipulations in various AD model systems, a few of which could ultimately turn out to have therapeutic relevance in humans. On the other hand, many tantalizing findings have been observed by only the originating laboratory, and one should not consider advancing preclinical candidates into human testing until there is widespread and rigorous confirmation that a target is directly relevant to modulating AD-like phenotypes (Figure 1). Biomedical research in this and other fields has a rather weak record of reproducing academic publications about potential therapeutic targets when the preclinical findings are moved into biopharmaceutical companies for confirmation (e.g., 54, 55)

Light at the end of the tunnel

20

The many disparate observations and daunting complexity of the Alzheimer syndrome have occasionally given rise to a sense of therapeutic nihilism, at least in some quarters. But a careful analysis of the current state of preclinical and clinical research suggests that we have gained a rather detailed view of some of the fundamental mechanisms of the disorder and have achieved considerable, albeit not universal, consensus about the approximate sequence of key steps in molecular pathogenesis. This palpable progress -- coupled with the absolute imperative for society to devote its best scientific efforts to identifying disease-modifying treatments for the looming AD epidemic -- makes Alzheimer therapeutic research a burgeoning area with true promise, particularly if we attempt to prevent the cardinal abnormalities of the disease as early as possible.

21

Figure 1 Aligning potential disease‐modifying agents for AD with the course of the disease. Red  boxes, sequence of steps in the discovery of compounds or biologics as investigational new  drugs (INDs) for AD. Blue boxes, speculative stages in the long presymptomatic and  symptomatic phases of AD in a hypothetical individual who undergoes Aß buildup for one  of several possible reasons (e.g., a Presenilin or APP mutation; ApoE4 inheritance;  increased ß‐secretase activity, etc.) and develops very early symptoms by around age 70.  Green boxes, clinical trial categories dependent on the stage of AD. Red X, trials in moderate  AD not recommended. Yellow X, trials in mild AD recommended with caution. [Figure adapted from 14 and reprinted with permission from AAAS.…(to be obtained).]

22

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Doody RS, Raman R, Farlow M et al. A phase 3 trial of semagacestat for treatment of Alzheimer's disease. N Engl J Med. 2013;369:341-350 Sperling R, Salloway S, Brooks DJ et al. Amyloid-related imaging abnormalities in patients with Alzheimer's disease treated with bapineuzumab: a retrospective analysis. Lancet Neurol. 2012;11:241-249 Galasko DR, Graff-Radford N, May S et al. Safety, tolerability, pharmacokinetics, and Abeta levels after short-term administration of R-flurbiprofen in healthy elderly individuals. Alzheimer Dis Assoc Disord. 2007;21:292-299 De Strooper B, Iwatsubo T, Wolfe MS. Presenilins and γ-Secretases: Structure, function and role in Alzheimer's disesae. In: Selkoe DJ, Mandelkow E, Holtzman DM, eds. The Biology of Alzheimer's Disease. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 2012:259-277 Jonsson T, Atwal JK, Steinberg S et al. A mutation in APP protects against Alzheimer's disease and agerelated cognitive decline. Nature. 2012;488:96-99 Wilson RS, Segawa E, Boyle PA et al. The natural history of cognitive decline in Alzheimer's disease. Psychol Aging. 2012;27:1008-1017 Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. The neuropathology of probable Alzheimer disease and mild cognitive impairment. Ann Neurol. 2009;66:200-208 Bateman RJ, Xiong C, Benzinger TL et al. Clinical and biomarker changes in dominantly inherited Alzheimer's disease. N Engl J Med. 2012;367:795-804 Jack CR, Jr., Knopman DS, Jagust WJ et al. Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010;9:119-128 Morris JC, Roe CM, Grant EA et al. Pittsburgh compound B imaging and prediction of progression from cognitive normality to symptomatic Alzheimer disease. Arch Neurol. 2009;66:1469-1475 Jack CR, Jr., Knopman DS, Jagust WJ et al. Tracking pathophysiological processes in Alzheimer's disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol. 2013;12:207-216 Dubois B, Feldman HH, Jacova C et al. Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria. Lancet Neurol. 2007;6:734-746 Sperling RA, Aisen PS, Beckett LA et al. Toward defining the preclinical stages of Alzheimer's disease: Recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011;7:280-292 Selkoe DJ. Preventing Alzheimer's disease. Science. 2012;337:1488-1492 Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol (Berl). 1991;82:239-259. Braak H, Del Tredici K, Rub U et al. Staging of brain pathology related to sporadic Parkinson's disease. Neurobiol Aging. 2003;24:197-211 Braak H, Del Tredici K. The pathological process underlying Alzheimer's disease in individuals under thirty. Acta Neuropathol. 2011;121:171-181 Price JL, Morris JC. Tangles and plaques in nondemented aging and "preclinical" Alzheimer's disease. Ann Neurol. 1999;45:358-368 Demattos RB, Lu J, Tang Y et al. A plaque-specific antibody clears existing beta-amyloid plaques in Alzheimer's disease mice. Neuron. 2012;76:908-920 Frost JL, Liu B, Kleinschmidt M et al. Passive immunization against pyroglutamate-3 amyloid-beta reduces plaque burden in Alzheimer-like transgenic mice: a pilot study. Neurodegener Dis. 2012;10:265-270 Nussbaum JM, Schilling S, Cynis H et al. Prion-like behaviour and tau-dependent cytotoxicity of pyroglutamylated amyloid-beta. Nature. 2012;485:651-655 Walsh DM, Teplow DB. Alzheimer's disease and the amyloid beta-protein. Prog Mol Biol Transl Sci. 2012;107:101-124 Lambracht-Washington D, Rosenberg RN. Advances in the development of vaccines for Alzheimer's disease. Discov Med. 2013;15:319-326 Shankar GM, Li S, Mehta TH et al. Amyloid-beta protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nat Med. 2008;14:837-842 Jin M, Shepardson N, Yang T et al. Soluble amyloid beta-protein dimers isolated from Alzheimer cortex directly induce Tau hyperphosphorylation and neuritic degeneration. Proc Natl Acad Sci U S A. 2011;108:5819-5824

23 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

Hemming ML, Elias JE, Gygi SP, Selkoe DJ. Identification of beta-secretase (BACE1) substrates using quantitative proteomics. PLoS ONE. 2009;4:e8477 Kuhn PH, Koroniak K, Hogl S et al. Secretome protein enrichment identifies physiological BACE1 protease substrates in neurons. Embo J. 2012;31:3157-3168 Willem M, Garratt AN, Novak B et al. Control of peripheral nerve myelination by the beta-secretase BACE1. Science. 2006;314:664-666 Hu X, Hicks CW, He W et al. Bace1 modulates myelination in the central and peripheral nervous system. Nat Neurosci. 2006;9:1520-1525 Cheret C, Willem M, Fricker FR et al. Bace1 and Neuregulin-1 cooperate to control formation and maintenance of muscle spindles. Embo J. 2013;32:2015-2028 Coric V, van Dyck CH, Salloway S et al. Safety and tolerability of the gamma-secretase inhibitor avagacestat in a phase 2 study of mild to moderate Alzheimer disease. Arch Neurol. 2012;69:1430-1440 Weggen S, Eriksen JL, Das P et al. A subset of NSAIDs lower amyloidogenic Aβ42 independently of cyclooxygenase activity. Nature. 2001;414:212-216. Haass C, Kaether C, Thinakaran G, Sisodia S. Trafficking and Proteolytic Processing of APP. In: Selkoe DJ, Mandelkow E, Holtzman DM, eds. The Biology of Alzheimer's Disease. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 2012:205-229 Crump CJ, Johnson DS, Li YM. Development and mechanism of gamma-secretase modulators for Alzheimer's disease. Biochemistry. 2013 Lleo A, Berezovska O, Herl L et al. Nonsteroidal anti-inflammatory drugs lower Abeta42 and change presenilin 1 conformation. Nat Med. 2004;10:1065-1066 Kukar TL, Ladd TB, Bann MA et al. Substrate-targeting gamma-secretase modulators. Nature. 2008;453:925-929 Kounnas MZ, Danks AM, Cheng S et al. Modulation of gamma-secretase reduces beta-amyloid deposition in a transgenic mouse model of Alzheimer's disease. Neuron. 2010;67:769-780 Rogers K, Felsenstein KM, Hrdlicka L et al. Modulation of gamma-secretase by EVP-0015962 reduces amyloid deposition and behavioral deficits in Tg2576 mice. Mol Neurodegener. 2012;7:61 McLaurin J, Franklin T, Chakrabartty A, Fraser PE. Phosphatidylinositol and inositol involvement in Alzheimer amyloid-beta fibril growth and arrest. J Mol Biol. 1998;278:183-194 McLaurin J, Kierstead ME, Brown ME et al. Cyclohexanehexol inhibitors of Abeta aggregation prevent and reverse Alzheimer phenotype in a mouse model. Nat Med. 2006;12:801-808 Townsend M, Cleary JP, Mehta T et al. Orally available compound prevents deficits in memory caused by the Alzheimer amyloid-beta oligomers. Ann Neurol. 2006;60:668-676 Salloway S, Sperling R, Keren R et al. A phase 2 randomized trial of ELND005, scyllo-inositol, in mild to moderate Alzheimer disease. Neurology. 2011;77:1253-1262 Roberson ED, Scearce-Levie K, Palop JJ et al. Reducing endogenous tau ameliorates amyloid beta-induced deficits in an Alzheimer's disease mouse model. Science. 2007;316:750-754 Wischik C, Staff R. Challenges in the conduct of disease-modifying trials in AD: practical experience from a phase 2 trial of Tau-aggregation inhibitor therapy. J Nutr Health Aging. 2009;13:367-369 Wyss-Coray T, Rogers J. Inflammation in Alzheimer's Disease-A brief review of the basic science and clinical literature. In: Selkoe DJ, Mandelkow E, Holtzman DM, eds. The Biology of Alzheimer's Disease. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 2012:335-357 Weiner HL, Frenkel D. Immunology and immunotherapy of Alzheimer's disease. Nat Rev Immunol. 2006;6:404-416 Butovsky O, Siddiqui S, Gabriely G et al. Modulating inflammatory monocytes with a unique microRNA gene signature ameliorates murine ALS. J Clin Invest. 2012;122:3063-3087 Brouwers N, Van Cauwenberghe C, Engelborghs S et al. Alzheimer risk associated with a copy number variation in the complement receptor 1 increasing C3b/C4b binding sites. Mol Psychiatry. 2012;17:223-233 Jonsson T, Stefansson H, Steinberg S et al. Variant of TREM2 associated with the risk of Alzheimer's disease. N Engl J Med. 2013;368:107-116 Guerreiro R, Wojtas A, Bras J et al. TREM2 variants in Alzheimer's disease. N Engl J Med. 2013;368:117127 Stevens B, Allen NJ, Vazquez LE et al. The classical complement cascade mediates CNS synapse elimination. Cell. 2007;131:1164-1178 Sanchez-Mejia RO, Newman JW, Toh S et al. Phospholipase A2 reduction ameliorates cognitive deficits in a mouse model of Alzheimer's disease. Nat Neurosci. 2008;11:1311-1318

24 53. 54. 55.

Tselepis AF, Rizzo M, Goudevenos IA. Therapeutic modulation of lipoprotein-associated phospholipase A2 (Lp-PLA2). Curr Pharm Des. 2011;17:3656-3661 Landis SC, Amara SG, Asadullah K et al. A call for transparent reporting to optimize the predictive value of preclinical research. Nature. 2012;490:187-191 Prinz F, Schlange T, Asadullah K. Believe it or not: how much can we rely on published data on potential drug targets? Nat Rev Drug Discov. 2011;10:712

Page 25 of 25

Annals of Neurology

Aligning potential disease-modifying agents for AD with the course of the disease. Red boxes, sequence of steps in the discovery of compounds or biologics as investigational new drugs (INDs) for AD. Blue boxes, speculative stages in the long presymptomatic and symptomatic phases of AD in a hypothetical individual who undergoes Aß buildup for one of several possible reasons (e.g., a Presenilin or APP mutation; ApoE4 inheritance; increased ß-secretase activity, etc.) and develops very early symptoms by around age 70. Green boxes, clinical trial categories dependent on the stage of AD. Red X, trials in moderate AD not recommended. Yellow X, trials in mild AD recommended with caution. [Figure adapted from 14 and reprinted with permission from AAAS.…(to be obtained).] 190x254mm (72 x 72 DPI)

John Wiley & Sons