part 1: abstracts of presentations

Brain and Cognition 49, 185–193 (2002)doi:10.1006/brcg.2002.1462

Part 1: Abstracts of Presentations

Neuroimaging Biomarkers in Dementia

Sandra E. Black

Program in Aging and Behavioral Neurology Unit, Sunnybrook & Women’s College Health SciencesCentre; University of Toronto; and Rotman Research Institute, Toronto, Ontario, Canada

Diagnosis. The search for diagnostic brain imaging biomarkers is rationally basedon topographical selectivity of the major dementing pathologies. For example, thefrontal–temporal regions are affected in primary tauopathy and limbic structures aretargeted, followed by isocortical association areas seven in Alzheimer’s disease (AD),a primary amyloidopathy. Overlap with age-related atrophy in regions such as thehippocampus and heterogeneity in topographical predilections, however, reducesspecificity and sensitivity of imaging biomarkers. The search for such biomarkers inAD has focussed on the medial temporal structures and hippocampal volume hasbeen 30 and 40% smaller in AD subjects in many studies. Sensitivity and specificity,however, remain problematic when this is the only structure measured. Functionalimaging techniques such as PET and SPECT in the resting state reliably show de-creases in parietal–temporal blood flow or metabolism in about 80% of AD subjects.To be clinically useful, diagnostic techniques must exceed the 80–90% accuracyachievable by using standardized clinical criteria. A combination of volumetric andfunctional measures can achieve higher diagnostic accuracy but with increased im-aging costs.

Early detection of AD. Neuroimaging biomarkers may also assist in early detectionof preclinical AD. Functional alterations in the posterior cingulate and medial tempo-ral regions have been noted in individuals with memory complaints who later developdementia, but such findings are difficult to apply in individual cases. A new approachmay be functional MRI in which individual cognitive activation patterns can be deter-mined and may be abnormal in those destined to develop AD. MRI spectroscopymay also be useful in revealing compromise of neuronal tissue in persons at risk fordeveloping dementia. Recent studies emphasizing comorbidity of Alzheimer’s andVaD also suggest that analysis of the degree of vascular injury in brain white matterand deep nuclei related to small vessel disease will also be important, along withatrophy measures, for assessing dementia risk. If neuroprotective therapies emerge,the need to identify patients at risk will become an urgent priority to rationally selectthose most likely to benefit from such therapies.

Monitoring progression and response to therapy. Quantitative structural brainimaging will likely be increasingly used as surrogate outcome measure to determineif new treatments are slowing down the rate of progression of atrophy. How bestto measure longitudinal change in the brain remains problematic. However, medialtemporal lobe thickness has been proposed as a simple measure to monitor longitudi-nal change, but so far has been difficult to reproduce. Brain loss can be calculatedby subtracting annual coregistered T1-weighted MRI scans from each other, but againthere are problems doing this reliably. Segmentation techniques as well as planimetric

1850278-2626/02 $35.00

2002 Elsevier Science (USA)All rights reserved.

186 ROTMAN INSTITUTE ABSTRACTS

methods can also be used. Although current methods have not yet reached the re-quired accuracy and ease of use to be applied on a daily basis in the clinic, it is likelythat they will need to be incorporated into routine clinical practice in the near future.

High- and Low-Tech Approaches to the Early Diagnosis of Alzheimer’s Disease

Howard Chertkow

Department of Neurology and Neurosurgery, McGill University, Montreal, Quebec, Canada

Family members of an individual with Alzheimer’s disease (AD) have a higherrate of developing the illness than others in the population. Several studies havedocumented subtle cognitive and functional neuroimaging abnormalities in these in-dividuals which seem predictive of later development of AD. In fact, there may beindicators of risk present years prior to the clinical development of the disease. Fur-thermore, Alzheimer’s disease usually begins slowly with mild memory loss. Mildmemory problems in older people occur in about 16% of the elderly population, andnot all of these individuals go on to deteriorate. At the same time, individuals withsuch ‘‘mild cognitive impairment’’ have a high rate of progression to Alzheimer’sdisease, up to 15% per year. Developing tools for predicting who will progress toAlzheimer’s disease is an urgent problem, since preventative medications are cur-rently in development. In this talk the current and future approaches to early diagnosisof Alzheimer’s disease in subjects who are totally asymptomatic, or at the stage ofmild cognitive impairment, will be reviewed. These approaches, largely at the experi-mental stage, range from ‘‘low tech’’ to ‘‘high tech.’’ The long list of suggestedpredictive tools includes combinations of sample clinical markers, neuropsychologi-cal test batteries, cognitive tests, genotyping (i.e., apolipoprotein E genotype), CSF,blood, and urine markers for amyloid and tau components; heme oxygenase 1, P97,and neuronal thread protein; sensory testing of smell; neuroendocrine profiles; andforms of anatomic and functional neuroimaging. The various advantages and disad-vantages of these different tests will be discussed, along with our own approach,which involves a number of markers assessed longitudinally. In our own longitudinalstudy, a combination ‘‘algorithm’’ appeared necessary to stratify risk in any individ-ual subject with MCI. There were no clear clinical ‘‘risk’’ markers other than ageand severity of the (mild) memory loss. Apo-E genotype also failed to delineate ahigher risk group for progression. It is likely that one can design an initial practical‘‘low-tech’’ approach for physicians to use with memory-impaired patients to stratifyrisk of progression, along with a complimentary ‘‘high-tech’’ assessment valid forstratifying ‘‘uncertain’’ cases.

Cognitive Impairment Due to Subcortical Ischemic Vascular Disease

Helena C. Chui

University of Southern California, Los Angeles, California; and Rancho Los Amigos NationalRehabilitation Center, Downey, California

Ischemic vascular disease (IVD) is the second most common cause of dementiain the Western World. This article focuses on subcortical ischemic vascular disease(SIVD), one subtype of IVD that is more challenging to diagnose, but eminently

ROTMAN INSTITUTE ABSTRACTS 187

preventable. Hypertension and diabetes mellitus are the leading causes of small-arterydisease, subcortical brain ischemic brain injury, and stepwise or slowly progressivedecline in cognitive function. Compared to Alzheimer’s disease, the pattern of cogni-tive impairment in SIVD features greater impairment of executive function, but betterpreservation of recognition memory. Pure motor or sensory deficits, sparing of thevisual fields, and parkinsonian gait disturbance are characteristic. Neuroimaging stud-ies, especially MRI, are more sensitive for detecting SIVD than the clinical examina-tion—often showing subcortical lacunes and deep white matter changes that are clini-cally ‘‘silent.’’ The brain can be protected against SIVD by early diagnosis andmanagement of risk factors. Once end-organ damage has occurred, treatment outcomeis less satisfactory.

Frontotemporal Dementia

Morris Freedman

Behavioral Neurology and Rotman Research Institute, Baycrest Centre for Geriatric Care;and University of Toronto, Toronto, Ontario, Canada

The concept of frontotemporal dementia (FTD) was introduced in the 1980s bygroups from England and Sweden, respectively. FTD accounts for as many as 20%of degenerative dementias and presents with profound early changes in character,social conduct and insight. Clinical–pathological data suggest at least three subtypes:disinhibited, apathetic, and stereotypic. The disinhibited subtype is characterized byrestlessness, purposeless overactivity, jocularity, callous unconcern, and profoundsocial breakdown. The apathetic subtype is characterized by inertia, aspontaneity,loss of volition, unconcern, mental rigidity and perseveration, and early incontinence.The stereotypic subtype is characterized by pronounced behavioral stereotypes andcompulsive and ritualistic traits. The disinhibited form has been associated with orbi-tofrontal and temporal lobe damage, the apathetic type with dorsolateral frontal con-vexity lesions, and the stereotypic subtype with striatal involvement. However, thereis substantial overlap in symptomatology among subtypes, especially with progres-sion of disease.

There is a strong hereditary component with about 50% of patients having a posi-tive family history. Genetic studies suggest that mutations in the tau gene on chromo-some 17 are related to FTD.

Two main types of pathology have been described in FTD: (i) prominent microva-cuolar change without specific histological features (frontal lobe degeneration type)and (ii) severe astrocytic gliosis with or without ballooned cells and inclusion bodies(Pick type). There is controversy whether these two pathologies represent differentparts on a spectrum of a single disorder or whether they represent different diseases.

The clinical profile of FTD contrasts with Alzheimer’s disease in which cognitivechanges are most prominent in the early phases and the ability to interact well atan interpersonal level and preservation of social graces, manners, and courtesy ismaintained until late in the disease.

FTD is part of a broader entity, i.e., frontotemporal lobar degeneration, which alsoincludes nonfluent progressive aphasia and semantic dementia. Nonfluent progressiveaphasia is characterised by a progressive decline in language with nonfluent effortfulspontaneous speech, phonological and grammatical errors, word retrieval difficulties,impaired repetition, and difficulty reading and writing. Understanding word meaning


is relatively preserved. In semantic dementia, there is loss of meaning for both verbaland nonverbal concepts (semantics). Affected individuals have severe naming andword comprehension deficits in the setting of fluent, effortless, and grammaticalspeech output; relatively preserved repetition; and preserved ability to read aloud andwrite orthographically regular words. Inability to recognise the meaning of visualpercepts, i.e., associative agnosia, also occurs.

The clinical presentations of frontotemporal lobar degeneration are determined bythe distribution of the pathology and not by histopathology. In FTD, there is promi-nent involvement of the frontal lobes bilaterally. In nonfluent progressive aphasiathe left frontotemporal lobes are chiefly affected. The lesions in semantic dementiaare most marked in the anterior temporal neocortex, predominantly in the middleand inferior temporal gyri.

The mechanisms of FTD are not well understood. In addition, there is currentlyno treatment. This will hopefully change as research into this recently recognisedform of dementia continues to advance.

Current and Future Therapy of Alzheimer’s Disease and Related Conditions

Serge Gauthier

McGill Centre for Studies in Aging; Professor of Medicine (Neurology), McGill University,Montreal, Quebec, Canada

The current strategy in the treatment of AD includes accurate diagnosis, treatmentof concomitant illnesses, patient and caregiver education, and support through thedifferent stages of the disease. Cholinesterase inhibitors (CIs) have been demon-strated to have a larger effect on cognition in the earlier stages of AD, but also toexert significant benefit in moderate to severe stages. Noncognitive effects includeslower decline in ADLs and improvement in some behavioral and psychologicalsymptoms of dementia (BPSD). It is possible that CIs delay the emergence of BPSDand the need for atypical neuroleptics. Most patients on stable doses of CIs show animprovement above pretreatment levels peaking at 3 months, followed by a declineparallel to natural history. Additive strategies in patients on stable doses of CIs arebeing considered, including neurotrophic and amyloid modifying agents, lookingfor a combination of disease modification and symptomatic control. The larger per-spective on AD is to delay onset of symptoms by 5 to 10 years using knowledge onpreventive and risk factors for the aging population. Individual risk assessment inthe presymptomatic stage may be possible using a combination of genetic markersand cognitive performance, leading to advice proportionate to risk.

Lewy Body Disease

Ian McKeith

Institute for the Health of the Elderly, University of Newcastle,Newcastle Upon Tyne, United Kingdom

Dementia with Lewy bodies (DLB) accounts for 15–20% of autopsy-confirmeddementia cases in old age. Lewy bodies (LB) are found in subcortical nuclei, limbic


cortex, and neocortex and are associated with significant neuronal loss including ex-tensive disruption of ascending cholinergic and dopaminergic projections. Most casesalso have additional pathologies, predominantly diffuse and neuritic amyloid plaques,similar to those seen in Alzheimer’s disease (AD). Neurofibrillary tangle density andbiochemical markers of tau pathology fall into the Alzheimer’s range in only 10–15% however. Vascular changes, visualized on MRI in vivo, are present in 30–50%of DLB cases at autopsy, in the form of both cortical infarcts and white matter, smallvessel disease. LB density and distribution are only weakly correlated with severityof cognitive impairment suggesting that LB per se do not constitute the primaryneurobiological substrate for dementia. Widespread neuritic degeneration associatedwith aggregates of α-synuclein has recently been reported in DLB and future clinico-pathological correlative studies will be focussed upon this.

Turning to clinical diagnostic issues, a series of autopsy validation studies showthe 1996 clinical Consensus criteria for probable DLB to have high specificity (0.9–1.0) but lower and variable specificity (0.22–0.83). These criteria are dependent uponthe accurate identification of fluctuating cognitive impairment, visual hallucinations,and parkinsonism, as core features of DLB. Repeated unexplained falls, syncope,depression, and REM sleep behavior disorder are features supportive of a diagnosisof DLB. Fluctuation, in particular, has been difficult to detect and quantify, withreports of low interrater reliability leading to diagnostic difficulties. Three indepen-dent and recently published methods of reliably assessing fluctuating cognition, withpotential to improve diagnostic sensitivity, will be presented. Other advances in diag-nosis are emerging from brain imaging studies, including regional volumetric mea-sures using MRI and in vivo assessment of nigrostriatal dopaminergic integrity withFP-CIT SPECT.

A four-step approach to the clinical management of DLB will be presented withan emphasis on the evidence base for the use of pharmacological agents includingatypical antipsychotics, cholinesterase inhibitors, and benzodiazepines. Treatment ofthe motor features of DLB will also be discussed.

Clinical Trials in Mild Cognitive Impairment

Ronald C. Petersen

Mayo Alzheimer’s Disease Research Center and Department of Neurology,Mayo Clinic Rochester, Minnesota

Research in aging and dementia is moving toward identifying persons at the earlieststage of cognitive impairment. The ultimate goal of this work is to characterize per-sons at their earliest signs of impairment to be able to intervene with treatmentsdesigned to prevent or slow the progression of the disease process. In recent yearsthe term ‘‘mild cognitive impairment’’ has been used to characterize subjects in thistransitional zone between normal aging and very mild Alzheimer’s disease (AD).Mild cognitive impairment refers to individuals who meet the following criteria: (i)subjective memory complaint preferably corroborated by an informant, (ii) objectivememory impairment for age and education, (iii) normal general cognitive function,(iv) intact activities of daily living, and (v) not demented. Several longitudinal studiesof aging are under way using variations of these criteria, and when subjects are char-acterized in this fashion, they tend to convert to clinically probable AD at a rate of10–15% per year. In some studies certain clinical/biological factors of the subjects


predict a more rapid progression, e.g., qualitative features of memory performance,apolipoprotein E4 carrier status, and atrophic hippocampal formations on MRI.

Mild cognitive impairment subjects have also become the focus of severaltreatment trials. Compounds such as antioxidants, cholinesterase inhibitors, anti-inflammatory agents, and nootropics are all being studied. One trial sponsored bythe Alzheimer’s Disease Cooperative Study will be discussed in detail. This studyis a double-blind placebo-controlled parallel arm trial evaluating the effects of high-dose vitamin E, donepezil, and placebo. Thus far 770 subjects have been enrolledand will be followed for 3 years. The entry criteria were similar to those stated above,and the primary outcome measure is conversion to clinically probable AD. Quantita-tive MRI scans are being acquired on a subset of the subjects and will be assessedlongitudinally.

In addition, newer therapies designed to prevent or slow the progression of theunderlying degenerative process, e.g., immunization therapies or secretase inhibitors,are being studied, and subjects with a mild cognitive impairment might be suitabletreatment candidates for these trials.

Genetic Variability and Pathobiology of Alzheimer’s Disease

Ekaterina Rogaeva

Centre for Research in Neurodegenerative Diseases, University of Toronto,Toronto, Ontario, Canada

Alzheimer’s disease (AD) is the most common neurodegenerative disorder ofaging, accounting for an estimated two-thirds of all cases of senile dementia. Thebrain pathology of AD is characterized by neuronal loss, neurofibrillary tangles andamyloid plaques, which mainly consist of Aβ peptide originating from β-amyloidprecursor protein (βAPP). Epidemiological studies suggest a complex etiology forAD with environmental and genetic factors influencing pathogenesis. Although themajority of cases are sporadic, a small number display family clustering. Geneticanalyses of these pedigrees have identified four genes that are involved in the devel-opment of AD. The ε4 allele of apolipoprotein E (APOE) is a well-established geneticrisk factor for both familial and sporadic late-onset AD. Three other genes, whenmutated, cause autosomal dominant early-onset AD: βAPP, presenilin 1 (PS1), andpresenilin 2 (PS2). Mutations in PS1 account for a majority of early-onset familialcases. The clinical implications and algorithms for genetic testing in dementia, how-ever, are still evolving. Our recent results of a screening for PS1 mutations in a largereferral-based series of AD cases revealed that 11% of cases (47/413) can be ex-plained by a mutation in the coding region of PS1, which includes 21 novel mutations.As would be expected 90% of cases with a PS1 mutation were affected by age 60.

Many AD cases, however, cannot be attributed to one of the known genetic causes,suggesting that other genes might be involved. The results of two genome scans haverevealed a few possible regions to which late-onset AD could be linked, with thehighest evidence for linkage being on chromosomes 1, 9, 10, and 12. In addition, anindependent study using plasma Aβ42 as a surrogate trait provided strong evidencefor linkage to the same locus on chromosome 10. It is remarkable that the four knowngenetic markers all converge upon a common biological pathway that involves the Aβmolecule, which indicates that one of the initiating events for AD is an abnormality inthe processing of βAPP and/or the clearance of Aβ peptide. Understanding the rea-


sons why changes in Aβ production should lead to progressive neurodegenerationstill remains an important challenge.

Pathobiology—Tauopathies

Kirk C. Wilhelmsen

Gallo Clinic and Research Center, University of California, San Francisco, California

The biology of the tau gene was implicated in dementia almost a century agowhen Lois Alzheimer observed what is now known to be aggregated tau protein inneurofibrillary tangles in demented patients. Subsequently, neurofibrillary tangles andaggregated tau proteins have been seen in many neurodegenerative diseases, and itwas widely believed that this was a marker of ongoing neurodegeneration. The detec-tion of linkage to chromosome 17, and the identification of mutations in the tau genein several familial neurodegenerative syndromes that are now collectively referredto as frontotemporal dementia and parkinsonism linked to chromosome 17, clearlyindicates that disregulation of tau biology is sufficient to cause neurodegenerativedisease. The most common syndrome associated with tau mutations is frontotemporaldementia. All individuals with tau mutations have aggregated neuronal tau protein,whereas most individuals with frontotemporal dementia do not have tau mutationsor aggregated tau protein.

The tau gene is a large gene with complex pattern of alternative transcriptionalsplicing. The protein has domains that have been shown to bind tubules, stabilizingmicrotubules. The differential splicing of exon 10 results in mRNA that codes foreither three or four repeats of the microtubule binding domains. Disregulation ofthis alternative splicing appears to be sufficient to cause disease. In addition, codingsequences of tau that presumably affect susceptibility to aggregation and the splicingof exon 10 lead to disease.

In animal models, overexpression of the human tau gene also appears to be suffi-cient to cause neurodegenerative disease, in particular causing a syndrome reminis-cent of amyotrophic lateral sclerosis. The tau gene is implicated in other neurodegen-erative diseases by mechanisms that have not elucidated including progressivesupranuclear palsy and the sporadic form of frontotemporal dementia. The etiologyof most cases of frontotemporal dementia remains to be discovered.

Role of Elevated Amyloid β Protein in Patients with Typical LateOnset Alzheimer’s Disease

Steven G. Younkin

Mayo Clinic Jacksonville, Jacksonville, Florida

Plasma Aβ (Aβ42 or both Aβ40 and Aβ42) is increased in subjects who have anyof the βAPP, PS1, or PS2 mutations linked to early onset familial Alzheimer’s diseaseand in subjects with trisomy 21, who invariably develop AD pathology by age 40.In 223 nondemented subjects between the ages of 20 and 96, there was a highlysignificant age-dependent increase in plasma Aβ40 (p , .0001) and Aβ42 (p 5.0009) beginning at approximately 65 years. Comparison of 180 first-degree AD rela-tives under age 65 with 129 controls showed significant increases in both Aβ40 (p 5


.0004) and Aβ42 (p , .0001) in the AD relatives. Plasma Aβ40 and Aβ42 were alsosignificantly elevated through three generations of five families where the probandwas a first-degree AD relative with elevated plasma Aβ. Importantly, the probandshad no FAD-like mutations in the coding region of PS1, PS2, or the APP exonsencoding Aβ. In an expanded set of AD families, segregation analysis (REGC,SAGE) was uninformative, suggesting a complex, polygenic mode of inheritance.Heritability analysis (‘‘polygenic’’ analysis, SOLAR) showed the heritability ofplasma Aβ42 and Aβ40 to be 73 and 47%, respectively. When the ApoE, ε4 alleledosage was included as a covariate, it was found to have an insignificant effect. Thehigh frequency of elevated plasma Aβ in the elderly and the high heritability ofplasma Aβ in AD families prompted us to reexamine plasma Aβ in a series of 159patients with typical late onset AD compared 163 controls in the same age range.Both Aβ40 and Aβ42 showed significant (p , .0001) increases in the AD patients.Remarkably, 71% of AD patients had plasma Aβ40 or Aβ42 in the range observedin subjects with trisomy 21 or mutations linked to early onset FAD. Of the 129 ADpatients with disease of 0–4 years duration, 19% had Aβ42 over 30 or Aβ40 over360 pM whereas only 3% with disease of long duration (over 5 years) were in thisrange (one subject). Thus, there appears to be a decrease in plasma Aβ in many ADpatients as disease progresses. Collectively, these results suggest that age-related,genetically driven increases in Aβ are likely to play an important role in typical lateonset AD and that elevated plasma Aβ is likely to be an excellent premorbid bio-marker for AD. In an effort to detect novel LOAD loci that increase Aβ42, plasmaAβ42 was used as a surrogate trait and linkage analysis was performed on extendedAD pedigrees identified through a late-onset patient with extremely high plasma Aβ.This analysis showed linkage to chromosome 10 with a maximal lod score of 3.93at 81 cM close to D10S1225. Remarkably, linkage to the same region was obtainedindependently by a corsortium (A. Goate, J. Hardy, and M. Owen) that performeda genome-wide screen of LOAD sib-pairs. These results provide strong evidence fora novel LOAD locus on chromosome 10 that acts to increase Aβ.

Mailing Addresses For Speakers

BLACK, Sandra, E., Division of Neurology, Room A-4-21, Sunnybrook and Wom-en’s College Health Sciences Centre, 2075 Bayview Avenue, Toronto, Ontario,M6A 2E1, Canada.Chertkow, Howard, Lady Davis Institute, Jewish General Hospital, 3755 CheminCote-Ste-Catherine, Montreal, Quebec, H3T 1E2, Canada.CHUI, Helena, C., Geriatric Neurobehavior & Alzheimer Centre, 800 Annex West,7601 Imperial Highway, Downey, California 90242, U.S.A.FREEDMAN, Morris, Behavioral Neurology Unit, Baycrest Centre for GeriatricCare, 3560 Bathurst Street, Toronto, Ontario, M6A 2E1, Canada.GAUTHIER, Serge, Alzheimer Disease Research Unit, McGill Centre for Studiesin Aging, 6825 LaSalle Boulevard, Verdun, Quebec, H4H 1R3, Canada.McKeith, Ian, Institute for the Health of the Elderly, Newcastle General Hospital,Westgate Road, Newcastle Upon Tyne E4 6BE, England.PETERSEN, Ronald, C., Mayo Alzheimer’s Disease Research Center, Mayo Clinic,200 First Street SW, Rochester, Minnesota 55905, U.S.A.ROGAEVA, Ekaterina, Centre for Research in Neurodegenerative Diseases, Univer-sity of Toronto, 6 Queen’s Park Crescent West, Toronto, Ontario, M5S 3H2, Canada.WESTAWAY, David A., Centre for Research in Neurodegenerative Diseases, Uni-versity of Toronto, 6 Queen’s Park Crescent West, Toronto, Ontario, M5S 3H2,Canada.


WILHELMSEN, Kirk C., Department of Neurology, University of California, GalloClinic and Research Centre, 5858 Horton Street, Suite 200, Emeryville, California94608, U.S.A.YOUNKIN, Steven G., Mayo Medical School, Mayo Clinic, Jacksonville, 4500 SanPablo Road, Jacksonville, Florida 32224, U.S.A.

part 1: abstracts of presentations

Documents