Translational Research in Neurology: Dementia

NEUROLOGICAL REVIEW

SECTION EDITOR: DAVID E. PLEASURE, MD Translational Research in Neurology Dementia

Lawrence S. Honig, MD, PhD

ementia disorders are characterized by clinicopathological criteria. Molecular un- derstandings of these disorders, based on immunohistochemical studies, biochemical investigations, genetic approaches, and animal models, have resulted in advances in diagnosis. Likewise, translational research has allowed us to apply our increasingD basic scientific knowledge of neurodegeneration to the rational development of new investigational therapies based on our current understanding of disease pathogenesis. This review discusses the application of translational research to both diagnosis and treatment of dementia disorders. The development of biomarkers has yielded imaging and biochemical methods that assist the physician more than ever in the diagnosis of neurodegenerative dementias, especially Alzhei- mer disease. New diagnostic criteria for disease are based on these molecular-based techniques. And these biomarkers are of potential use in monitoring disease activity during therapeutic trials. Translational investigations likewise have led toward new avenues in targeted dementia research. This is particularly so in the development and testing of disease-modifying treatments that might slow or deter progressive deterioration. Recent clinical trials have not been based on empirical trials of established drugs but, rather, on trials of drugs shown, through experiments in biochemical, cell culture, and animal models, to interfere with known elements of the pathogenetic cascade of Alzheimer disease. Arch Neurol. 2012;69(8):969-977. Published online April 2, 2012. doi:10.1001/archneurol.2011.2883

Dementia is defined as a disorder mani- tigation, experimental animal science, or fest by loss of mental capacity affecting a neuropathological studies, to be applied person’s ability to function. Dementia af- to the development of diagnostic tech- fects more than 6 million Americans to- niques or disease-directed specific thera- day, most of whom are elderly. Dement- pies. Translational research is bidirec- ing disorders were at one time viewed as tional, with clinical data often informing of psychiatric origin in the younger popu- or influencing basic studies, and allow- lation and of “senile” derivation (ie, a con- ing in turn further development based sequence of aging) in the elderly. There on basic research of clinically applied was little hope in treating either of these. methods. Translational research has emerged as a Categorization of the dementias as clini- dominant driving force in both diagnos- copathological disorders started in ear- tic and therapeutic advances in demen- nest in the early 20th century with mod- tia. Multidisciplinary research efforts allow for basic knowledge, whether obtained ern neuropathology, in conjunction with in the laboratory through in vitro inves- careful clinical studies. Neurodegenera- tive diseases include Alzheimer disease Author Affiliations: Taub Institute for Research on Alzheimer’s Disease and the (AD), Lewy body dementia, frontotempo- Aging Brain, Gertrude H. Sergievsky Center, and Department of Neurology, ral dementia (FTD), corticobasal degen- Columbia University Medical Center, New York, New York. eration, progressive supranuclear palsy,

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 969

Clinical Disorder Protein Term Genes Alzheimer disease ␤-Amyloid ␤-Amyloidopathy PSEN1, PSEN2, APP, APOE, others Progressive supranuclear palsy Tau Tauopathy MAPT Corticobasal degeneration Tau Tauopathy MAPT Frontotemporal dementia Tau Tauopathy MAPT TDP-43 TDP-43-opathy TARDP, PGN, C9ORF72 Lewy body dementia ␣-Synuclein ␣-Synucleinopathy GBA Parkinson disease dementia ␣-Synuclein ␣-Synucleinopathy SNCA, LRRK2, GBA Creutzfeldt-Jakob disease Prion protein Prionopathy PRNP Huntington disease Huntingtin HTT

Abbreviation: TDP-43, TAR DNA-binding protein 43. a Neurodegenerative dementing disorders are characterized by derangements in particular proteins, with characteristic deposition of abnormal proteins in neurons, glia, or brain extracellular space. In some diseases, there are co-pathologies. For example, Alzheimer disease is marked by abnormal deposits of both ␤-amyloid and tau. Various genes have been shown to be involved in these disorders, either through mutations with autosomal dominant or recessive inheritance, or through risk factors of polymorphisms or mutations.

vascular dementia, Huntington disease, and Creutzfeldt- may be at risk of developing MCI and dementia. These Jakob disease. Clinical neuropathology, with classical asymptomatic persons have preclinical disease and can staining techniques, did provide the original basis for iden- be called pre-MCI. They would only be identifiable by bio- tifying these disorders. However, our increasing molecu- markers suggesting that, even without any symptoms, they lar understanding has sharpened the differentiation of show the earliest pathological molecular, cellular, or ra- these diseases. The use of immunochemical assays has diological changes suggesting the beginning stages of a allowed us to categorize dementias by molecular typol- dementing disorder that has not become symptomatic. ogy (Table 1). This has been supplemented by the iden- Recent consensus work groups, sponsored jointly by the tification of genes leading to specific dysfunctions. Alz- National Institute on Aging of the US National Insti- heimer disease is most uniquely characterized tutes of Health and the Alzheimer Association, have uti- biochemically by the accumulation of A␤ in plaques and lized the increasing body of knowledge on molecular dis- vessels, with concomitant intracellular accretions of tau. ease markers to formulate important new criteria, not only Lewy body dementia shares with Parkinson disease the for AD3 but also for MCI4 and presymptomatic or pre- accumulation of abnormal ␣-synuclein aggregates. About clinical disease.5 These criteria should allow for better half of the cases of FTD, together with cases of progres- definition of individuals who might be in the MCI or pre- sive supranuclear palsy and corticobasal degeneration, clinical phases of dementia. The use of amyloid imaging may be grouped as disorders marked by deposits of ab- to ascertain amyloid deposition in living persons and the normal tau. Other degenerative dementias are charac- use of a cerebrospinal examination to ascertain abnor- terized by abnormalities of TAR DNA-binding protein 43 malities in amyloid, tau, and phosphorylated tau are key (TDP-43), fused in sarcoma protein (FUS), huntingtin techniques in the process. It is possible that the earlier protein, prion protein, and other proteins (Table 1). The identification of disease during prodromal periods4-6 may identification of the specific cellular proteins promi- allow for more effective therapeutic intervention be- nently involved in these neurodegenerative disorders has cause the disease process could be prevented, inter- led to in vitro and rodent models, which have been used rupted earlier, or attacked before the onset of more ir- for probing disease pathways, and to the promise of di- reversible changes such as neuronal cell loss. This review rected, rather than empirical, therapeutic trials. will principally focus on AD dementia and its precursor The dementing disorders of adults are age- stage MCI, because AD is responsible for more than 80% dependent, with generally increasing incidence in the later of cases of dementia in the elderly. decades of life. The 3 most frequent pathologies are AD, Major advances in dementia over the past decade have Lewy body dementia, and vascular disease abnormali- transpired involving both diagnostic and therapeutic areas. ties. Two or more of these pathologies commonly coex- These advances spring from the increasingly diverse num- ist in the same individual.1,2 It is unclear whether this is ber of biochemical and microscopic techniques, genetic simply coincidental, evidence of a shared diathesis, or in- analyses, and ensuing cell culture, mice, or other model dicative of an interaction between the pathoetiologic cas- systems. Diagnostic accuracy has increased through the cades of these diseases. In addition to those millions of translation of knowledge of molecular pathogenesis to Americans with dementia, there are larger numbers of the development of biomarkers and genetic markers as- individuals who have cognitive impairment of insuffi- sisting in diagnosis, even prior to symptoms. The num- cient severity, extent, or functional consequence to meet ber of therapeutic trials have increased rapidly because criteria for dementia. This state is most often called mild the elucidation of the biochemical pathways, together with cognitive impairment (MCI). Most of these persons are ac- the ability to test interventions in model systems, has led tually experiencing the earliest stages of one of the de- to rational drug trials, rather than nonspecific or more mentia disorders already mentioned, most commonly AD, arbitrarily chosen empirical trials of established sub- and thus this is also called “prodromal AD.” There are stances, that have had an effect on accumulating disease also those persons who have no symptoms at all but who burden. Herein, we will discuss the advances in transla-

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 970

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 tional research with regard to diagnosis and treatment of dementia. Table 2. Translating Biology Into Biomarkers of Possible Diagnostic Utility in Dementia TRANSLATIONAL RESEARCH IN DIAGNOSIS OF AD Measures and Methods Magnetic resonance imaging structural measures Advances in translational research have spearheaded our Global volume measures (whole-brain, ventricular, gray matter, white matter, and cortical thickness measures) diagnostic understanding of dementia for the past cen- Regional measures (including hippocampal volumes) tury. The pioneering correlative clinicopathologic inves- Magnetic resonance imaging functional measures tigations of Alois Alzheimer and Frederick Lewy, among Arterial spin labeling others, provided the basis for the definitions of the main BOLD imaging degenerative dementias, based on clinical symptoms and Spectroscopy measures (eg, NAA) neuropathological findings. Because neuropathological Nuclear medicine imaging functional measures HMPAO-SPECT (regional perfusion) descriptions typically are made only at autopsy, after the 18F-FDG–PET regional metabolism patient has died, a diagnosis of dementia while the pa- Nuclear medicine molecular imaging tient is still alive has traditionally required much clini- PET amyloid imaging (eg, 11C-PiB, 18F-florbetapir, 18F-florbetaben, cal effort. Neurological testing and neuropsychological 18F-flutemetamol) testing provide measures of an individual’s cognitive per- PET transporter imaging (eg, 18F-AV-133) 123 formance, which can be judged normatively. However, SPECT transporter imaging (eg, I-ioflupane) Cerebrospinal fluid analyses the limitations of such testing to establish diagnosis in- A␤42 clude (1) the large variability in human performances and Tau (2) the susceptibility of testing to cultural, educational, Phosphorylated tau and achievement factors. Particular problems include in- 14-3-3 Protein accurate early diagnosis of disease, incorrect antemor- ␣-Synuclein tem diagnosis of established clinical disease (eg, Alzhei- Blood, plasma, and serum analyses DNA markers (genetic analyses: PS1, PS2, APP, APOE, MAPT, mer dementia vs Lewy body dementia), and the TARDP, PGN, C9ORF72, SNCA, GBA, LRRK2) overlapping of dementing disorders (eg, combined Alz- A␤42, A␤40 heimer and Lewy body dementia disorders). Thus, there Proteomics is a recognized need for tests that can provide a “pathological” diagnosis during life. Traditional tissue testing Abbreviations: BOLD, blood oxygenation level–dependent; FDG, methods, such as biopsy, are not likely to be widely ap- fluorodeoxyglucose; HMPAO, 99mTc-hexamethylpropyleneamine oxime; NAA, N-acetylaspartate; PET, positron emission tomography; PiB, Pittsburgh plicable to disorders of brain function, both because of Compound B; SPECT, single-photon emission computed tomography; 11C, the surgical invasiveness of a brain biopsy and because carbon 11; 18F, fluorine 18; 123I, iodine 123. of the large numbers of persons with a disorder of brain function. Thus, translational research involving biomark- aging, the conversion of MCI to AD, and the progres- ers (Table 2), either through magnetic resonance imaging sion of AD.10,11 It is now commonplace for local radiolo- (MRI), single-photon emission computed tomography gists to include the diagnostic possibility of AD after re- (SPECT), or positron emission tomography (PET) or viewing images with marked mesiotemporal atrophy. through biochemical analysis of cerebrospinal fluid (CSF), For clinical trials, longitudinal measurements of brain DNA, or blood serum or plasma, is now being used to volumes are now frequently incorporated because of develop newer improved diagnostic methods. their potential use in monitoring the effects of drug treatment. Brain Atrophy as a Feature of Dementia Regional Variations in Clinical and Neuropathological studies have shown that the losses of Neuropathological Changes neurons, synapses, and white matter are accompani- ments of dementia. These observations have translated Neuropathological and clinical changes are not diffuse, into in vivo structural neuroimaging studies for the di- but are variably distributed in different dementias. Func- agnosis and follow-up of dementia. Structural imaging, tional measures of brain activity have been valuable in originally performed using computed tomography and the differential diagnosis. These include the nuclear medi- then supplanted by MRI, has shown the value of mea- cine techniques of SPECT and PET. These methods have suring brain loss in dementia.7 Brains incur detectable provided imaging of brain function through the mea- volume losses due to synaptic and neuronal degenera- surement of functional surrogate agents mapping blood tive losses. These amount to a 1% to 2% change per year perfusion, and oxygen and glucose metabolism. These using various global measures of brain volume or ven- nuclear medicine studies provide “pattern” biomarkers tricular volume, with somewhat greater volumetric allowing for a better distinction between disorders pri- changes in specific regional measures of hippocampal vol- marily affecting temporoparietal cortices (AD), parieto- ume.8,9 Large-scale detailed longitudinal data from many occipital cortices (Lewy body dementia), and frontotem- hundreds of subjects have emerged as a result of the Alz- poral regions (FTD).12 Alzheimer disease prominently heimer’s Disease Neuroimaging Initiative. These data have affects temporal cortices but also causes decreased ac- provided us with a much better understanding of the rates tivity in association cortices in frontal and parietal re- of structural atrophy and of how these relate to normal gions. Frontotemporal dementias typically are marked

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 971

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 by decreased brain activity in frontal and anterior tem- of brain amyloid levels in human research subjects.16 poral regions. Lewy body dementia is often accompa- Nuclear medicine imaging techniques using PET or nied by significant decreases in brain activity in parieto- SPECT can also detect changes in nigrostriatal function occipital regions, with lesser affliction of frontal and seen in Parkinson disease and Lewy body dementia. Posi- temporal regions. These patterns may correlate with the tron emission tomographic imaging with fluorine 18– clinical symptomatology, but have added biomarker value radiolabeled fluorodopa can reveal presynaptic dopa- for assisting in the differential diagnosis of dementing dis- minergic insufficiency. Likewise, PET imaging using the orders. However, there remains uncertainty regarding the investigational agent 18F-AV-133, or SPECT imaging using interpretation and specificity of these imaging pat- radiolabeled tropanes (eg, 123I-ioflupane [DaTscan], which terns.13 Functional MRI (fMRI) may provide even less in- was recently approved in the United States) that are taken vasive measures of brain function than nuclear medi- up by the vesicular monoamine transporters, can con- cine imaging. firm clinically suspected dopaminergic deficits in Par- kinson disease or Lewy body dementia. Imaging of Specific Molecular and Neurochemical Changes Molecular Brain Changes Reflected in CSF

Specific molecular and neurochemical changes mark each Molecular brain changes are also discernible through dementia, and, in a revolutionary fashion, neuroimag- analysis of CSF. Obtained through generally painless lum- ing can identify such changes through use of specific li- bar puncture, CSF contains brain proteins that are shed gands. In the research sphere, the use of A␤-binding agents into the surrounding fluid and that are increasingly the has provided the means to visualize early evidence of subject of analysis. Alzheimer disease is associated with plaques in the living brain. The first agent to be widely certain hallmark changes in the CSF, including reduc- used, the carbon 11–radiolabeled agent BTA or Pitts- tion of A␤42. It has been proposed that there is a lower burgh Compound B, has been handicapped by the need concentration of A␤42 in the CSF of patients with AD for a cyclotron near the PET imaging center, but fluo- owing to deposition in the brain parenchyma17; how- rine 18–radiolabeled agents, including florbetapir (AV- ever, there is a decreased level of A␤42 in various other 45), florbetaben (BAY-949172 or AV-1), and non-AD disorders.18 An alternative explanation of low CSF flutemetamol, have now achieved wide use in research levels is that a decreased level of A␤42 relates to de- studies. The Alzheimer’s Disease Neuroimaging Initia- creased brain synaptic activity. Tau protein is elevated tive studies, funded through a partnership between the in the CSF of patients with AD, but this finding is non- National Institutes of Health and industry, have shown specific. An increased CSF tau level presumably reflects the ability, even in different centers using different imaging increased neurodegeneration, with the release of this in- instruments, to reliably detect amyloid binding. A very traneuronal cytoskeletal protein. An increased CSF phos- high percentage of cases of clinical AD, and essentially phorylated tau (P-tau; particularly, tau phosphorylated 100% of cases of pathologically proven AD, show amy- at the 181 position) is a more specific hallmark of AD, loid binding through PET imaging techniques using these related to increased phosphorylation of tau in AD. Be- agents. Similarly, persons with MCI who progress to AD cause each of these biomarkers in isolation is not highly generally show evidence of amyloid deposition by amy- specific, combinations of biomarkers, such as ratios of loid PET imaging. However, amyloid binding seen on PET A␤42 to tau or other indices (such as the amyloid tau scans also occurs in a moderate percentage (20%-40%) index, ATI = A␤42/[240 ϩ (1.18 ϫ tau)]), have had par- of persons with normal cognition, depending on age.14,15 ticular utility in diagnosis. Cerebrospinal fluid biomark- The favored implication of this finding is that persons ers have now been incorporated into the new clinical cri- with such ligand binding are at the beginning, asymp- teria for AD, MCI, and presymptomatic AD.3-6 Blood tomatic, stage of AD, which might become manifest at plasma or serum tests are increasingly sought and might some later age. Imaging findings have now been incor- ultimately provide the least invasive measure, but, to date, porated into the new clinical criteria for AD, MCI, and there has been less clear success in measurements of blood, presymptomatic AD.3-6 Because clinical drug trials are whether by measures of specific markers, such as A␤40 aimed at interrupting AD at the earliest stage, these trials and A␤42, or by changes in clusters of proteins, as re- may increasingly demand evidence of amyloid binding vealed by proteomic analysis.19 Like amyloid-imaging, CSF by PET as an eligibility criteria to increase the homoge- biomarker profiles can be used to select individuals with neity and accuracy of diagnosis. In one recent European molecular characteristics of AD for clinical trials and, po- trial16 that did not use amyloid binding as an eligibility tentially, may be used to monitor therapeutic efficacy. factor, nearly 20% of subjects entering the trial with clini- Decreased CSF levels of tau might be evidence of a de- cally probable AD had no significant amyloid binding, creased degree of neurodegeneration, and preliminary data which suggests a likely erroneous diagnosis of AD in those in some drug trials (eg, bapineuzumab) have shown such cases. Similar evidence for imperfect clinical diagnosis change.20 is present in the observational Alzheimer’s Disease Neu- roimaging Initiative study as well as in neuropathologi- Genetic Factors cal series. Ligand imaging may assist in diagnosis but is also potentially useful in monitoring treatment. Drug trials Genetic factors were discovered in the epidemiologic with an intravenously administered antibody specific for search for risk factors for dementia. Early-onset AD can A␤42 (ie, bapineuzumab) have shown in vivo reduction be caused by genetic mutations in APP, the gene coding

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 972

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 for the ␤-amyloid precursor protein, or in PSEN1 or cally important and include agitation, aggression, hallu- PSEN2, the genes involved in the ␥-secretase–mediated cinations, delusions, depression, and incontinence. Al- cleavage of APP to A␤42 (Table 1). Late-onset AD is as- though these symptoms can be refractory to present sociated with the presence of the APOE gene ε4 allele, treatments, the neurotransmitter-based medications, in- which conveys significant risk for AD: about a 2-fold risk cluding neuroleptics, antidepressants, anxiolytics, and an- for 1 copy and a more than 4-fold risk for 2 copies of this tispasmodic drugs, have been valuable; however, the ad- allele compared with individuals with no ε4 alleles. Over- verse effects are sometimes limiting. all, for persons aged 55 to 75 years in the US population, only about one-third of the general population has Symptomatic Therapies for AD an ε4 allele, but about two-thirds of those with AD have 1 or more ε4 alleles. Although the presence of an ε4 al- These were first developed after the recognition of the lele has not been proven to alter disease course or prog- particular cholinergic deficit in AD. Drugs causing in- nosis, it is increasingly clear that there may be differen- creased brain acetylcholine, through acetylcholinester- tial sensitivity to drug therapy efficacy or side effects in ase inhibition, were studied. The first such drug was ta- those with vs without an ε4 allele. Several studies have crine hydrochloride, which became available in 1993. suggested an effect of ε4 allele on efficacy (eg, on do- Three additional cholinesterase inhibitors have been de- nepezil in MCI),21 although this may relate to diagnosis, veloped successfully for clinical use. Donepezil hydro- on rosiglitazone in AD22 (although subsequent studies chloride, galantamine hydrobromide, and rivastigmine have not confirmed any effect of this drug), and on bap- tartrate have been shown to be of modest symptomatic ineuzumab in AD,20 although this may relate more to side benefic in AD. Following the development of these drugs, effects. Although definite effects of APOE are not proven memantine, an activity-dependent glutamatergic N- in efficacy data, it is reasonably clear that having the APOE methyl-D-aspartate receptor antagonist, was demon- ε4 allele is a risk factor for the adverse effect of amyloid- strated to have modest symptomatic benefit in persons reducing therapy; this adverse effect was originally called with moderate to severe AD.31 However, none of these 5 vasogenic edema but is now called ARIA, standing for amy- neurotransmitter-based drugs, labeled by the US Food loid-related imaging abnormality.23,24 There is a mark- and Drug Administration for use in AD (four acetylcho- edly increased risk of this radiologic brain change, which linesterase inhibitors and one N-methyl-D-aspartate glu- may or may not be symptomatic in persons with APOE tamatergic receptor antagonist), provide more than a mod- ε4 alleles. This observation is responsible for the proto- est symptomatic benefit for patients with AD. And these col in which treatment is stratified by genetic back- drugs do not appear to modify the molecular pathology ground in the bapineuzumab phase 3 studies, in which or clinical course of patients with AD. Investigations of individuals without ε4 alleles are eligible for higher doses other drugs that involve different neurotransmitter sys- of drug than those with ε4 alleles. tems, as well the use of intranasal insulin, are ongoing.32 The results of recent genetic analytic studies have shown that there is an increasing number of gene families for which Disease-Modifying Therapies polymorphisms might increase the genetic risk of late- onset “sporadic” AD. These genes include members of pro- Because established therapies do not actually modify dis- tein processing, membrane or cholesterol production, and ease course, the strongest focus of drug discovery for de- immune/inflammatory pathways (Sorl1, CR1, PICALM, menting disorders is investigations into agents that might CLU, MS4A4, CD2AP, CD33, and EPHA1).25-27 Although alter disease progression through affecting basic disease variations in these genes appear to be responsible for only pathophysiology. A variety of empirical trials of exist- small fractions of the risk of AD, the genetic findings make ing medications or natural substances, including vari- it likely that newly identified pathways could be ad- ous vitamins, fish oils, botanical compounds (eg, Gingko dressed by drug treatment strategies. biloba), hormones (eg, estrogen), HMG-COA reductase For Parkinson disease, there are likewise genes that inhibitors, and anti-inflammatory agents, have sug- are relatively determinative of inherited disease, whether gested possible utility, but subsequent studies have proven autosomal dominant (eg, SNCA and LRRK2) or reces- inefficacy in the treatment of AD. Hence, it is transla- sive (eg, PARK2, PARK7, and PINK1), and those that may tional medical science that most likely will provide ad- convey risk (UCHL1, SNCAIP, and GBA). In particular, vances in treatment of AD and the other dementing dis- GBA may provide risk of dementia in Parkinson disease orders (Table 3). or Lewy body dementia.28 There are now several auto- Pathological studies show that AD is marked by ac- somal dominant genes known to convey risk of FTD, in- cumulation of ␤-amyloid protein in the brain in the form cluding MAPT (tau), GRN, CHMP2B, VCP, and C9ORF72 of plaques and by accumulation of hyperphosphory- (Table 1 and Table 2).29,30 lated tau in the brain in the form of tangles. In vitro, animal model, and human genetic evidence all point to a TRANSLATIONAL RESEARCH IN THERAPY centrality of ␤-amyloid in the pathophysiological cas- FOR AD cade causing AD.33 Thus, many drugs in development have mechanisms of action based on either decreasing ␤-amy- Therapies for dementia may be divided into sympto- loid production (through inhibition of ␤-secretase or matic therapies for secondary symptoms, disease- ␥-secretase), increasing ␤-amyloid clearance (through ac- specific symptomatic therapies, and disease-modifying tive or passive immunization), or decreasing ␤-amyloid therapies. Secondary symptoms of dementia are clini- aggregation/fibrillization (Table 3).

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 973

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 ment, notably including avagacestat (BMS-708163), for Table 3. Clinical Investigational Drug Studies in Alzheimer which a phase 2 trial (http://www.dementiatoday.com Disease /?p=3565) has been completed, and phase 3 development is reportedly planned. Inhibition of ␤-secretase also Objectives and Drugs results in decreased A␤42 production in animal mod- Increase ␤-amyloid clearance els,38 and inhibitors of this enzymatic activity are under Active immunization development. Enhancement of ␣-secretase activity by AN-1792, ACC-001 (vanutide cridificar),a CAD-106, UB-311, V-950 etretinate has been reported, and limited clinical trials Passive immunization of this agent have started. Bapineuzumab (AAB-001),a solanezumab (LY2062430),a crenezumab (MABT-5102A),a gantenerumab (R-1450),a THERAPIES TO INCREASE ␤-AMYLOID ponezumab (PF-04360365), GSK-933776 CLEARANCE IN AD Immunoglobulin therapy Human immunoglobulin intravenous therapya Decrease ␤-amyloid production Active Immunization ␥-Secretase modulators (tarenflurbil [R-flurbiprofen]) ␥-Secretase inhibitors (semagacestat [LY450139], begacestat Injection with ␤-amyloid peptide is efficacious in mouse [GSI-953], avagacestat [BMS-708163],a GSI-136, PF-3084014, models of AD: A␤ deposits in the brain are reduced,39 and MK0752) memory function is improved in maze tasks.40 The first ␣-Secretase enhancers (acitretin [etretinate, varenicline]) ␤-Secretase inhibitors (CTS-21166) drug with this mechanism of action tried in humans was 41 Decrease ␤-amyloid fibril formation or aggregation AN-1792, for which a phase 2 trial was halted prema- Tramiprosate turely owing to the occurrence of meningoencephalitis, Scyllo-inositol (D-005)a which was symptomatic in about 5% of treated patients. Deter tau aggregation Limited clinical data obtained from this study41 did not Methylthioninium (methylene blue) show efficacy, although a follow-up study42 did seem to Inhibit tau phosphorylation show possible evidence suggestive of efficacy. Postmor- GSK-3 inhibitors (lithium, valproic acid) 43,44 Deter neurodegeneration tem studies of some immunized individuals showed Davunetidea (NAP, AL-108) apparent evidence that immunization may have had the Nerve growth factor (NGF) intended effect in those treated. Individuals who had a NGF-adeno–associated virus, AAV2-NGF (CERE-110) serological antibody response did show successful clear- Brain-derived neurotrophic factor (BDNF) ance of amyloid,43 and immunized individuals may have Exenatide shown beneficial changes in abnormal neurites.44 The possibility that amyloid clearance was successful but was still a Currently under development in phase 2 or phase 3 trials. accompanied by lack of clinical improvement or clinical deterioration has raised the issue of whether A␤- INHIBITORS OF ␤-AMYLOID PRODUCTION IN AD removing strategies will indeed be successful.43 How- ever, with such small numbers, with the lack of clarity A␤42 production can be inhibited by reduction of ␥-secre- as to the influence of subclinical adverse effects in treated tase or ␤-secretase activities or by enhancement of ␣-secre- patients, and with the lag from treatment to autopsy, it tase activity (Figure). Tarenflurbil (Flurizan; Myriad is not possible to reach any conclusions at this time. It is Pharmaceuticals, Inc), otherwise known as R- likely that the severe inflammatory symptoms from AN- flurbiprofen, acts in vitro as a selective inhibitor of A␤42 1792 were related significantly to the immunological ad- production both in vitro and in mouse models.34 This drug juvant. Experimental laboratory studies have led to the recently failed to show any efficacy or harm in a large introduction of other forms of A␤ immunization, includ- phase 3 double-blind randomized placebo-controlled trial ing different haptens and adjuvants, such as vanutide cridi- involving more than 2000 subjects.35 It is unclear whether ficar (ACC-001) and CAD-106, which are now being used this was because of a differing action in humans than mice, in phase 2 trials. Such trials, as well as passive immuni- whether there was simply an insufficient dose or inad- zation trials, should result in data that may clarify equate penetration into the central nervous system, or whether amyloid reduction strategies will be effective in whether the putative mechanism of action did not oc- stabilizing, slowing, or, less likely, reversing AD clinical cur in humans. Semagacesat (LY450139) is an inhibitor symptoms. of ␥-secretase that showed evidence of having an effect on amyloid production in vitro, in mice, and, demon- Passive Immunization strably, in humans.36 However, a large phase 3 double- blind randomized placebo-controlled trial was recently Passive immunization with antibodies to ␤-amyloid pep- stopped upon interim review because of evidence that tide has demonstrated efficacy in mouse models of AD. subjects treated with the drug were doing less well cog- Bapineuzumab (AAB-001) is a humanized mouse mono- nitively than those treated with placebo.37 The reasons clonal antibody to the N-terminal portion of ␤-amyloid. for these unexpected findings are unclear but might pos- It has undergone a phase 2 trial20 and is now undergo- sibly relate to the changed balance of amyloid products, ing 2 large multicenter phase 3 trials, with more than 4000 to insufficient overall inhibition of ␥-secretase, or to “off- patients being studied. Although the phase 2 trial20 did target” effects of this inhibitor on other cellular pro- not meet the efficacy end points, there was clear evi- cesses. Other ␥-secretase inhibitors are under develop- dence of radiological adverse effects, including vaso-

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 974

Decreased Positive AD pattern MRI- Cognitive Activities Increased Aβ42 production Decreased Aβ42 clearance CSF Aβ42 Aβ imaging in detected symptoms and level FDG PET atrophy function

Excessive Aβ42

Excessive oligomieric and/or polymeric Aβ42

Neurofibrillary tangles and Neuritic plaques neuritic threads (Aβ42 and Aβ40 deposits) (phosphorylated tau)

Synaptic losses Neuronal losses

Clinical AD

Figure. Hypothetical working model of molecular and clinical features of Alzheimer disease (AD). The pathogenetic cascade of AD disease is hypothesized to initially involve either overproduction or impaired clearance of A␤42 (A). This peptide aggregates into oligomers and then polymeric deposits. Development of diffuse and neuritic plaques ensues, with concomitant depositions of hyperphosphorylated tau as neurofibrillary tangles and neuritic threads. Ultimately, there are synaptic losses and neuronal losses. In parallel to the cascade depicted (A), indicators of the degree to which clinical biomarkers may be present at different stages of the pathogenetic cascade are shown (B); the darker blue color indicates a higher likelihood of the listed biomarker being positive. Thus, early aspects of ␤-amyloid dysregulation can be detected through decreased A␤42 in the cerebrospinal fluid (CSF) and through nuclear medicine–based amyloid imaging. Later stages of the cascade show detectable early brain functional changes through positron emission tomography (PET) with fluorodeoxyglucose (FDG), subsequent brain atrophy, cognitive symptoms, and functional change.

genic edema of the brain in apolipoprotein ε4 carri- the immune response. A phase 3 trial of human immu- ers,20,23 and suggestive evidence of a benefit from active noglobulin intravenous therapy is ongoing in mild to treatment, particularly in noncarriers of the apolipopro- moderate AD. An alternative to immunization with tein ε4 allele.20 The phase 3 trial is ongoing in both car- protein or passive immunoglobulin administration has riers (at lower dose) and noncarriers of the ε4 allele. A been the concept of DNA vaccination,47 in which different humanized mouse monoclonal antibody to the injection of specific DNA in mice has led to decreased midportion of ␤-amyloid, solanezumab (LY2062430), is ␤-amyloid in the brain. also undergoing phase 3 trials, after phase 2 trials and cerebrospinal fluid testing showed encouraging bio- THERAPIES TO INHIBIT ␤-AMYLOID marker changes.36,45 This antibody may act more to clear AGGREGATION IN AD soluble ␤-amyloid, although, secondarily, it may cause plaque clearance. Additional antibodies including cren- Antifibrillation/Antiaggregation Agents ezumab and gantenerumab are also undergoing phase 2 trials (Table 3). In vitro and animal model experiments have suggested Nonspecific Immunoglobulin Therapy and Other that treatment with agents that might prevent the aggre- Immunization Strategies gation of ␤-amyloid might be effective against AD. The first such agent tested in a phase 3 trial, tramiprosate (also The success in mice of specific anti-A␤ passive immu- known as homotaurine), failed to show any efficacy in nization prompted small human phase 1 and phase 2 cognitive tests or first-order biomarker tests. The devel- trials of nonspecific human immunoglobulin intrave- oper of that drug has discontinued the development of nous therapy.46 Biochemical experiments have sug- the molecule as a drug but has recharacterized it as a gested that human intravenous immunoglobulin con- “medical food” (Vivimind; Bellus Health) marketed in tains some polyclonal anti–␤-amyloid antibody at low Canada. However, evidence that it might be effective levels,46 and this would be one putative mechanism of against AD is lacking. A different agent that appears to action of human immunoglobulin preparations, if they show promising preclinical results is ELN D005, also indeed were shown to have beneficial effects in AD. known as scyllo-inositol, which has now completed phase Alternative mechanisms of action include modifying 2 testing and for which phase 3 testing is being planned.

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 975

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 THERAPIES TO PREVENT NEURONAL Accepted for Publication: October 25, 2011. DEGENERATION IN AD Published Online: April 2, 2012. doi:10.1001 /archneurol.2011.2883 Agents That Might Inhibit Tau Dysfunction or Correspondence: Lawrence S. Honig, MD, PhD, Taub In- Support Neuronal Integrity stitute for Research on Alzheimer’s Disease and the Ag- ing Brain, Gertrude H. Sergievsky Center, and Depart- ment of Neurology, Columbia University Medical Given the prominence of neurofibrillary degeneration Center,630 West 168th St (P&S Unit 16), New York, NY in AD, it is reasonable to speculate about agents that 10032 ([email protected]). might prevent the formation of the phosphorylated tau Financial Disclosure: Dr Honig has received personal that aggregates into neurofibrillary tangles in the brains compensation for consulting activities with Bayer, Bio- of persons with AD. In vitro and animal experiments gen Idec, Dainippon, Johnson & Johnson, and Pfizer. He have suggested that a variety of agents might deter the has also received personal compensation for editorial ac- formation of tangles. Methylene blue (methylthio- tivities for Archives of Neurology. He has received fund- ninium) was tried in a small phase 2 study but showed ing for research activities from Bayer, Elan, Genentech, no clear efficacy. Taxol has been suggested to be of po- Johnson and Johnson, Lilly, and Pfizer. tential utility on the basis of animal studies, but its tox- Funding/Support: This work was supported by funding icity will likely preclude human trials. Davunetide, also from the National Institutes of Health/National Insti- known as AL-108, has been proposed, based on animal tute on Aging (grant P50AG08702), the Alzheimer’s As- experimentation, to be of potential utility in AD, and a sociation (grant IIRG 08-92010), the Alzheimer’s Dis- phase 1 trial has been performed. Growth factors might ease Discovery Foundation, the Panasci Fund for Lewy improve neuronal survival. The injection of the nerve Body Research, and the Taub Institute for Research. growth factor (NGF) gene into the brain48 has been tried in a limited number of persons in a phase 1, proof- of-concept experiment. Since then, injections of DNA REFERENCES coding for brain-derived neurotrophic factor has been 49 the subject of phase 1 experiments. Neuronal degen- 1. White L. Brain lesions at autopsy in older Japanese-American men as related eration is marked by a number of final processes, in- to cognitive impairment and dementia in the final years of life: a summary cluding cell membrane breakdown, calcium influxes, report from the Honolulu-Asia aging study. J Alzheimers Dis. 2009;18(3):713- caspase activations, and other cell-death pathways. Re- 725. 2. Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. The neuropathology of search continues on developing agents that might inter- probable Alzheimer disease and mild cognitive impairment. Ann Neurol. 2009; vene beneficially at these more terminal steps of ner- 66(2):200-208. vous system injury that occur in common in ischemic 3. McKhann GM, Knopman DS, Chertkow H, et al. The diagnosis of dementia due and neurodegenerative disease. to 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(3):263-269. NON-AD DEMENTIAS 4. Albert MS, DeKosky ST, Dickson D, et al. The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Insti- tute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for The biological underpinnings of dementia with Lewy bod- Alzheimer’s disease. Alzheimers Dement. 2011;7(3):270-279. ies, FTD, progressive supranuclear palsy, corticobasal de- 5. Sperling RA, Aisen PS, Beckett LA, et al. Toward defining the preclinical stages generation, and Creutzfeldt-Jakob disease are reason- of Alzheimer’s disease: recommendations from the National Institute on Aging- ably described, but the pathogenetic cascade of these Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 2011;7(3):280-292. non-AD dementias is less well understood than it is for 6. Dubois B, Feldman HH, Jacova C, et al. Revising the definition of Alzheimer’s AD. Thus, advances in diagnosis and therapy are in less disease: a new lexicon. Lancet Neurol. 2010;9(11):1118-1127. advanced stages for these non-AD dementias. For FTD 7. Sabuncu MR, Desikan RS, Sepulcre J, et al; Alzheimer’s Disease Neuroimaging and related disorders, possible medications (including tau- Initiative. The dynamics of cortical and hippocampal atrophy in Alzheimer disease. disruptive or neuronal supportive medications for the tau Arch Neurol. 2011;68(8):1040-1048. 8. Barnes J, Bartlett JW, van de Pol LA, et al. A meta-analysis of hippocampal at- forms of this disorder) are under consideration, includ- rophy rates in Alzheimer’s disease. Neurobiol Aging. 2009;30(11):1711-1723. ing AL-108, which is being tried in a phase 2 study of 9. Scher AI, Xu Y, Korf ES, et al. Hippocampal morphometry in population-based progressive supranuclear palsy. Likewise, memantine, incident Alzheimer’s disease and vascular dementia: the HAAS. J Neurol Neuro- which has neuroprotective effects in vitro, is being tried surg Psychiatry. 2011;82(4):373-376. 10. Ewers M, Sperling RA, Klunk WE, Weiner MW, Hampel H. Neuroimaging mark- in a phase 2 study of FTD. Quinacrine hydrochloride has ers for the prediction and early diagnosis of Alzheimer’s disease dementia. Trends shown efficacy in vitro in deterring prion protein aggre- Neurosci. 2011;34(8):430-442. gation, leading to trials of this agent in Creutzfeldt- 11. Vemuri P, Wiste HJ, Weigand SD, et al; Alzheimer’s Disease Neuroimaging Ini- Jakob disease; to date, these limited trials have been with- tiative. Serial MRI and CSF biomarkers in normal aging, MCI, and AD. Neurology. out success.50,51 Even though our understanding of the 2010;75(2):143-151. 12. Chen K, Ayutyanont N, Langbaum JB, et al; Alzheimer’s Disease Neuroimaging molecular cascades of AD is limited, it is nonetheless more Initiative. Characterizing Alzheimer’s disease using a hypometabolic conver- advanced than our understanding of the molecular patho- gence index. Neuroimage. 2011;56(1):52-60. logical underpinnings of these other neurodegenerative 13. Womack KB, Diaz-Arrastia R, Aizenstein HJ, et al. Temporoparietal hypometabo- disorders. Thus, translational investigations leading to lism in frontotemporal lobar degeneration and associated imaging diagnostic errors. Arch Neurol. 2011;68(3):329-337. potential diagnostic or therapeutic methods for the 14. Sojkova J, Driscoll I, Iacono D, et al. In vivo fibrillar beta-amyloid detected using non-AD degenerative disorders are lagging behind those [11C]PiB positron emission tomography and neuropathologic assessment in older for AD. adults. Arch Neurol. 2011;68(2):232-240.

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 976

©2012 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 15. Aizenstein HJ, Nebes RD, Saxton JA, et al. Frequent amyloid deposition without 34. Kukar TL, Ladd TB, Bann MA, et al. Substrate-targeting gamma-secretase significant cognitive impairment among the elderly. Arch Neurol. 2008;65(11): modulators. Nature. 2008;453(7197):925-929. 1509-1517. 35. Green RC, Schneider LS, Amato DA, et al; Tarenflurbil Phase 3 Study Group. 16. Rinne JO, Brooks DJ, Rossor MN, et al. 11C-PiB PET assessment of change in Effect of tarenflurbil on cognitive decline and activities of daily living in patients fibrillar amyloid-beta load in patients with Alzheimer’s disease treated with bap- with mild Alzheimer disease: a randomized controlled trial. JAMA. 2009;302 ineuzumab: a phase 2, double-blind, placebo-controlled, ascending-dose study. (23):2557-2564. Lancet Neurol. 2010;9(4):363-372. 36. Siemers ER, Friedrich S, Dean RA, et al. Safety and changes in plasma and ce- 17. Fagan AM, Mintun MA, Mach RH, et al. Inverse relation between in vivo amyloid rebrospinal fluid amyloid beta after a single administration of an amyloid beta imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol. 2006; monoclonal antibody in subjects with Alzheimer disease. Clin Neuropharmacol. 59(3):512-519. 2010;33(2):67-73. 18. Blennow K, Hampel H, Weiner M, Zetterberg H. Cerebrospinal fluid and plasma 37. Siemers E, Henley D, Sundell K, et al. Evaluating semagacestat, a gamma- biomarkers in Alzheimer disease. Nat Rev Neurol. 2010;6(3):131-144. secretase inhibitor, in a phase III trial. Alzheimer’s & Dementia. 2011;7(4): 19. Ray S, Britschgi M, Herbert C, et al. Classification and prediction of clinical Alz- S484-S485. doi:10.1016/j.jalz.2011.05.2355. heimer’s diagnosis based on plasma signaling proteins. Nat Med. 2007;13 38. Citron M. Beta-secretase inhibition for the treatment of Alzheimer’s disease— (11):1359-1362. promise and challenge. Trends Pharmacol Sci. 2004;25(2):92-97. 20. Salloway S, Sperling R, Gilman S, et al; Bapineuzumab 201 Clinical Trial Inves- 39. Schenk D, Barbour R, Dunn W, et al. Immunization with amyloid-beta attenu- tigators. A phase 2 multiple ascending dose trial of bapineuzumab in mild to mod- ates Alzheimer-disease-like pathology in the PDAPP mouse. Nature. 1999; erate Alzheimer disease. Neurology. 2009;73(24):2061-2070. 400(6740):173-177. 21. Petersen RC, Thomas RG, Grundman M, et al; Alzheimer’s Disease Cooperative 40. Wilcock DM, Rojiani A, Rosenthal A, et al. Passive immunotherapy against Abeta Study Group. Vitamin E and donepezil for the treatment of mild cognitive in aged APP-transgenic mice reverses cognitive deficits and depletes parenchy- impairment. N Engl J Med. 2005;352(23):2379-2388. mal amyloid deposits in spite of increased vascular amyloid and microhemorrhage. 22. Risner ME, Saunders AM, Altman JF, et al; Rosiglitazone in Alzheimer’s Disease Study J Neuroinflammation. 2004;1(1):24. Group. Efficacy of rosiglitazone in a genetically defined population with mild-to- 41. Gilman S, Koller M, Black RS, et al; AN1792(QS-21)-201 Study Team. Clinical moderate Alzheimer’s disease. Pharmacogenomics J. 2006;6(4):246-254. effects of Abeta immunization (AN1792) in patients with AD in an interrupted 23. Sperling RA, Jack CR Jr, Black SE, et al. Amyloid-related imaging abnormalities trial. Neurology. 2005;64(9):1553-1562. in amyloid-modifying therapeutic trials: recommendations from the Alzheimer’s 42. Vellas B, Black R, Thal LJ, et al; AN1792 (QS-21)-251 Study Team. Long-term Association Research Roundtable Workgroup. Alzheimers Dement. 2011;7 follow-up of patients immunized with AN1792: reduced functional decline in an- (4):367-385. tibody responders. Curr Alzheimer Res. 2009;6(2):144-151. 24. Sperling R, Salloway S, Brooks DJ, et al. Amyloid-related imaging abnormalities in patients with Alzheimer’s disease treated with bapineuzumab: a retrospective 43. Holmes C, Boche D, Wilkinson D, et al. Long-term effects of Abeta42 immuni- analysis. Lancet Neurol. 2012;11(3):241-249. sation in Alzheimer’s disease: follow-up of a randomised, placebo-controlled phase 25. Reitz C, Cheng R, Rogaeva E, et al; Genetic and Environmental Risk in Alzheimer I trial. Lancet. 2008;372(9634):216-223. Disease 1 Consortium. Meta-analysis of the association between variants in SORL1 44. Serrano-Pozo A, William CM, Ferrer I, et al. Beneficial effect of human anti- and Alzheimer disease. Arch Neurol. 2011;68(1):99-106. amyloid-beta active immunization on neurite morphology and tau pathology. Brain. 26. Naj AC, Jun G, Beecham GW, et al. Common variants at MS4A4/MS4A6E, CD2AP, 2010;133(pt 5):1312-1327. CD33 and EPHA1 are associated with late-onset Alzheimer’s disease. Nat Genet. 45. Siemers ER, Quinn JF, Kaye J, et al. Effects of a gamma-secretase inhibitor in a 2011;43(5):436-441. randomized study of patients with Alzheimer disease. Neurology. 2006;66(4): 27. Jun G, Naj AC, Beecham GW, et al; Alzheimer’s Disease Genetics Consortium. 602-604. Meta-analysis confirms CR1, CLU, and PICALM as Alzheimer disease risk loci and 46. Relkin NR, Szabo P, Adamiak B, et al. 18-Month study of intravenous immuno- reveals interactions with APOE genotypes. Arch Neurol. 2010;67(12):1473-1484. globulin for treatment of mild Alzheimer disease. Neurobiol Aging. 2009;30 28. Shulman JM, De Jager PL, Feany MB. Parkinson’s disease: genetics and (11):1728-1736. pathogenesis. Annu Rev Pathol. 2011;6:193-222. 47. Lambracht-Washington D, Qu BX, Fu M, et al. DNA immunization against amy- 29. Rohrer JD, Warren JD. Phenotypic signatures of genetic frontotemporal dementia. loid beta 42 has high potential as safe therapy for Alzheimer’s disease as it di- Curr Opin Neurol. 2011;24(6):542-549. minishes antigen-specific Th1 and Th17 cell proliferation. Cell Mol Neurobiol. 30. Borroni B, Pilotto A, Bianchi M, Gilberti N, Padovani A. Genetic contributors to 2011;31(6):867-874. frontotemporal lobar degeneration: beyond monogenic disease. Mini Rev Med 48. Tuszynski MH. Nerve growth factor gene delivery: animal models to clinical trials. Chem. 2011;11(11):988-1001. Dev Neurobiol. 2007;67(9):1204-1215. 31. Herrmann N, Li A, Lanctoˆt K. Memantine in dementia: a review of the current 49. Nagahara AH, Tuszynski MH. Potential therapeutic uses of BDNF in neurological evidence. Expert Opin Pharmacother. 2011;12(5):787-800. and psychiatric disorders. Nat Rev Drug Discov. 2011;10(3):209-219. 32. Craft S, Baker LD, Montine TJ, et al. Intranasal insulin therapy for Alzheimer dis- 50. Collinge J, Gorham M, Hudson F, et al. Safety and efficacy of quinacrine in hu- ease and amnestic mild cognitive impairment: a pilot clinical trial. Arch Neurol. man prion disease (PRION-1 study): a patient-preference trial. Lancet Neurol. 2012;69(1):29-38. 2009;8(4):334-344. 33. Selkoe DJ. Biochemistry and molecular biology of amyloid beta-protein and the 51. Geschwind MD. Clinical trials for prion disease: difficult challenges, but hope for mechanism of Alzheimer’s disease. Handb Clin Neurol. 2008;89:245-260. the future. Lancet Neurol. 2009;8(4):304-306.

ARCH NEUROL / VOL 69 (NO. 8), AUG 2012 WWW.ARCHNEUROL.COM 977