Association studies of novel polymorphisms in BACE and BACE2

s102

p%J

Poster Presentation: THE ASSOCIATION BETWEEN SECONDARY AND FUNCTIONAL DISABILITY IN DEMENTED

PATHOLOGY SUBJECTS

Vince S Thomas, Univ ofNebraska Med Ctr, Omaha, NE Dementia has been recognized as the strongest determinant for developing functional disability. But dementia patients typically present with concomitant illness. The objective of the current study was to examme the associations between specific diagnosed pathologies and specific functional disability in individual domains necessary for daily autonomous function. Using data on nearly 2900 subjects from the clinical examination of the 1991 Canadian Study of Health and Aging, multiple logistic regression models accounting for specific conditions and impairments that have demonstrated consistent associations with the development of physical disability in older people were estimated for disability in each of several domains (bathing, dressing, grooming, toileting, urinary and stool incontinence) assessed using the CAMDEX. The presence of a gait disorder, a history of falls, and current depression were significantly associated with disability in bathing. Among these, only the presence of a gait disorder was significantly associated with disability in any other domains (in dressing, grooming, toileting, and urinary incontinence). Diabetes was significantly associated with disability m dressing. Cardiac symptoms were significantly associated with disability in dressing, grooming, and toileting. In each case, it appeared to be protective against disability in that particular domain. Respiratory complaints were significantly associated with disability in dressing, urinary incontinence, and stool incontinence, while malignancies were significantly associated with disability in dressing and grooming. Chronic conditions are increasingly frequent in old age, many of which are associated with a likelihood of disability. When older persons present with possible cognitive problems, it is actually the disability that families are complaining of. Estimating the specificity and strength of secondary disease-disability associations in demented people is of practical importance because coexisting medical conditions may cause “excess disability” in patients with dementia that may be reducible by treatment and rehabilitation.

pi%J

COGNITIVE FUNCTION IN THE OLDEST FORM BETTER THAN MEN

OLD: WOMEN PER-

Eric van Exel, Jucobijn Gussekloo, Atlton Jwr de Crurn. Annetje Boofsnu-van der Wiel. Leiden Univ Med Ctr, Leiden Nether1and.y; Peter Houx, P.yychiatT and Neurop~ychology. Maastricht Netherlands; Ili(,k L Knook. Rudi Gj Westendorp, Lridm Univ Med Ctr, Leiden Nether1and.s Objective:The incidence of dementia is higher in women than in men. Limited formal education has been associated with poor cognitive function and susceptibility of dementia. If limited formal education is associated with dementia, we would expect women to have a poorer cognitive function than men. Our aim was to explore whether limited formal education can explain the higher incidence of dementia in older women. Design:The L&den 85 plus Study is a population~baaed study investigating all 85.year old inhabitants of the city of Leiden. From September 1997 until September 1999 a total of 599 participants were vislted at their place of residence. The response rate was 86%. Main outcome measures:The Mini-Mental State Examination was completed by all participants. Cognitive speed and memory were determined with four neuro-psychological tests in participants with a Mini-Mental State Examination score of 19 points or higher. Results:The proportion of women with limited formal education was significantly higher than that of men (70% versus 538, p=O.OOl). Despite this difference in formal education, women had better scores for cognitive speed and memory than men (p
MILD MEMORY IMPAIRMENT AND MEDICAL SULTS FROM A POPULATION-BASED STUDY

CARE:

RE-

Steven M Albert, Pam Nadash, Yaakov Stern, Columbia Univ, University, NY Mild memory impairment may be associated with an increased risk of medical care and greater health care costs. To investigate this question, a multi-ethnic cohort of I I90 older adults without dementia was identified from a population-based study (Washington Heights-Inwood Columbia Aging Project). All subjects completed a full battery of cognitive evaluations in 1996 or earlier and survived through 1996. Subjects with Medicare (1996, n=81 I) and NYC Medicaid claims (1996.1998, n=560) were then identified and examined relative to performance on tests of memory (a factor-analytic derived composite of seven terta). In multivariate models

Genetic Studies II

that adjusted for age, gender, and education, mean [median] Medicare charges in 1996 were $7713 [$7505], $7930 [$7726), and $5848 [$5677] for subjects with the lowest, middle, and highest memory test performance, respectively (p < .OOl by KroskalWallis [K-S] test of median scores). Mean [median] annual Medicaid costs by memory tenile were $16,115 [$15,794], $10.651 [$10,313], and $7552 [$7197] (p < .OOl by K-S test). Memory tertile groups did not differ in number of comorbid conditions. Low memory scores were associated with risk of incident dementia after 1996 (half of 59 subjects diagnosed with dementia during follow-up were in the lowest memory score tertile). Removing these subjects from analyses eliminated cost differences in Medicare but not Medicaid claims, suggesting that low memory performance in non-demented elders may be an independent risk factor for medical care costs that are captured in the Medicaid system.

THE ASSOCIATION RISK FACTORS

BETWEEN

DEMENTIA

AND

MID-LIFE

Michiko Yamada, Radiation Effects Research Fdn, Hiroshima Japan; Yusuyo Mimori, Hiroshima Univ Sch $Medicine, Hiroshima Japan; Hideo Susaki, Fumiyoshi Kasqi. Naomi Musurwri, Rudiution Effects Research Fdn, Hiroshima Jupun; Junko Ikeda. Sigenobu Nakamuru, Hiroshima Univ Sch r$ Medicine. Hm~shima Japm; Sarko F+vara, Radiation Effects Research Fdn, Hiroshima Jupm Objective: Previous studies of Alzheimer’s disease (AD) and other dementia have identified an association between these conditions and education, smoking, alcohol use, and co-morbid disorders, such as hypertension and diabetes mellitus (DM). We investigated the association of mid-life risk factors with the development of dementia, as detected 25 years later. Method: Study ubjects were the 1,774 Japanese men and women participating in the Adult Health Study (AHS) of the Radiation Effects Research Foundation. The AHS consists of atomic bomb survivors and their controls. Based on DSM-IV, 114 objects were diagnosed with dementia in examinations conducted between 1992.97. The number of cases of AD and vascular dementia (VaD) were 51 and 38, respectively. Information regarding education, physical activity, dietary habits, systohc blood pressure (SBP), total~cholesterol, body mass index, and history of DM were obtained in 1968.70. Sex, age, and radiation dose were also considered as explanatory variables. We evaluated the risk factors of dementia, AD, and VaD using logistic regression analysis. Results: The prevalence of dementia increased with increasing age, lower attained education, higher SBP in middle age, and a higher intake of salt in middle age. Odds ratios of AD for age (in 5.year increments), attained education (in 3.year increments), hi&tory of DM, and a lower intake of salt were 2.54,0.42, 3.09, and 0.53, respectively. Odds ratios of VaD for age (in 5-year increments), sex (women to men), SBP (in IO mmHg increments), and a greater Intake of milk were 1.33, 0.61, 1.33, and 0.35, respectively. Conclusion: In addition to age, sex, and education, dietary habits and a history of diseases in middle age were associated with the development of dementia.

Poster Presentation:

Genetic Studies II

ASSOCIATION STUDIES 1463] BACE AND BACEZ

OF

NOVEL

POLYMORPHISMS

IN

Petru Nowotny, Sumi Chakraverti, Jennifer M Kwon, Wushingtorr Urliv, St. Loui%, MO; Alison M Goate, Washington Univ Sch of Medicine, St. Louis, MO The release of amyloidogenic amylold-P peptide (AP42) from amyloid-P precursor protem (APP) requires cleavage by p- and y-secretasea. Mutations in these enzyme\ that increase AP42 production would be expected to increase the risk for Alzheimer‘s disease (AD). Four groups have identified a candidate for the “p-site APP-cleaving enzyme ” (BACE), also known aa “Asp2” because the enzyme is a tranxnembrane aspartic protease. This enzyme maps to chromosome llq23.24. Its homologue, BACE2 (Aspl), is mapped to chromosome 21q22 which includes the Down Syndrome region. Thus mutations or other polymorphic variants that alter the level of expression or activity of BACE and/or BACE2 may be risk factors for AD. We sequenced BACE cDNA from brain tissue of those with pathologically confirmed AD and elderly controls to look for polymorphisms in the coding sequence. We found one polymorphism at nucleotide 1239 where G is changed into C (from sequence GenbankAFl90725). The polymorphism was confirmed by PCR amplitication and restriction enzyme digestion (Hph I). In BACEZ, two polymorphisms have been found by sequencing 8 exons using genomic DNA from IS individuals, including 7 cases of familial dementia. We identified a Cm substitution in exon 6 at nucleotide 1200 (from sequence GenbankzAF050171) which can be screened with a BspM I digestion. An intronic C/T substitution was found at IO bp 5 of exon 4 which was confirmed with by EcoN I digestion. No pathogenic mutation< were observed in BACEZ in 5 indiwduala from FAD kindreds. Preliminary data in a small number of individuals failed to show differences in allele or genotype frequency between AD

Poster Presentation:

Genetic Studies II

s103

cases and controls. However, further genotyping experiments are undenvay in a large cohort of AD cases and controls.

ANGIOTENSIN I CONVERTING ENZYME (4661 THE CEPTIBILITY FACTOR FOR DEMENTIA Philippe

pii?iJ

SUSCEPTIBH,ITY DISEASE

LOCUS

FOR

ON CHROMOSOME

LATE

ONSET

ALZHEIMER’S

U508,

Brevannes

10

Pierre Amando

Jennings

WuvhinRton St. Louis, W&s:

Mwrs.

Washington

Umv Sch ofMedicine,

MO;

Julie

Richard

Jucksorwillr.

Williams.

Crook.

l/nil:

St.

St. Louis,

MO:

Mike J Owen,

Univ

Fabienne

Wowant-De

Amouyel. Lille

Louis,

Peter

MO;

Alison

Ho/mans,

M

Washington

of Wales Co11 of Medicine,

Vriezr,

John

A Hardy.

Goare. Univ.

talier

Clin.

Only four genes have been linked to Al/-heimer’s dtseaae (AD). The genes for P-amyloid precursor protein. and presenilins I and 2 have been linked to early onbet familial AD. Apolipoprotein E (APOE) is a risk factor for late onset AD (LOAD). Daw et al. have suggested that there may be between four to seven additional genes mvolved I” LOAD with one locus potentially contributing more of an effect than APOE. We have performed a genome wide screen in sibpair\ affected with LOAD to detect additional loci. Our initial results (292 ASPS) found 16 peaks with a multipoint LOD score >I. To follow up on this data we have genotyped additional markers within each peak region to reduce the interval between each pair of markers to lets than 5 cM and added additional samples to make a total of 567 affected sib pairs (ASPS). Several peaks have replicated. but the peak on chromosome 10 has given the most consistently high LOD scores. The maximum MLS on chromosome 10 in the initial data set was 1.87. This score increased to 3. I I with the additional markera and samples. The peak shifted slightly towards the pter, from DlOSl21 I (70.35 CM) to between DlOSl227 and DIOSl22.5 (peak at 60.42 CM). We have also stratified OUI sample by APOE genotype to examine any interactive effects between APOE and our putative loci. With the additional ASPS. the initial peak found in the E4-ve group disappeared. However, our E4+ve peak replicated giving a final MLS of 2.77 at 64.27 CM near marker DlOSl220 (initial results: MLS=2.21 at 70.35 CM). Thus, it appears that most of the effect on chromosome 10 is due to the E4+ve subgroup. suggesting a possible interaction between this putative locus and APOE4. Other groups have also reported linkage on chromosome IO using both sibpair analysis in LOAD families and case control association studies. We are currently analyzing several interecting candidate genes that map within this region on chromosome 10.

Emilr

Stephen

P Mcllro~.

The Queen’v Brlfavt

United

Univ,

Kewn Be/first

Kingdom;

ASSOCIATION

ALZHEIMER’S

B Dynan. United Pew

Bronn~h

OF EITHER

AZM

OR LRP

DISEASE

M McGleenon,

Kingdom:

John

A Povsmorr.

7hr

7 Lmvson, Queen’s

M Vuhidas.\r.

Belfust Univ.

City

Bdfast

Rico&,

INSERM

Frmce;

Villejuij France;

Florence

Ctr Hospitalirr

Ctr Hospitalier

U258,

Roux. LimrilBrevannrs

GENETIC (4671 AALZHEIMER’ S DENCE AND

Bosron.

MA:

Hosp. United

Kin,&m

The role of APOE as a major wsceptibihty gene for Alrheimer’q disease (AD) has been supported by a “umber of studies, however the precise mechanism by which risk for AD is modulated remains unclear. Within the last two years a number of studies have implicated polymorphislns in the low density lipoprotein receptor-related protein gene (LRP). a major receptor for APOE in the brain, with increased risk for AD. Genetic polymorphisms in alpha2-macroglobulin (A2M), a ligand of LRP which hmds P-amyloid and mediates its clearance through the LRP receptor, have also been associated with mcreased risk for AD. In addition, a recent study has linked a region of chromosome 12, between the LRP and A2M loci, with late-onset AD. In a” effort to replicate the these associations we decided to type the LRP exon 3 and tetranucleotide repeat polymorphisms and the A2M VlOOOl and exon 2 deletion polymorphisms in clinically well-defined group5 of sporadic AD case5 and controls from the relatively genetically homogeneous population of Northern Ireland. The dt&bution of all genotypes in this study did not differ from that predicted by Hardy-Weinberg equilibrium. We detected no difference in the distribution of genotypes or alleles from either LRP polymorphisms (Exon 3 genotypes: x2=2. 13, df=2, p=O.345: alleles: x2= 1.35, df= 1, p=O.24; Tetranucleotide repeat genotypes: x2=0.44, df=3, p=O.932; alleles: x2=0.45, df=2. p=O.799) or either A2M polymorphisms (A10001 genotypes: x*=0.48, df=2, p=O.786; alleles: x2=0.26, df=l. p=O.607; Deletion genotypes: x2= 1.03, df=2, p=O.S97: alleles~~=O.OO, df= 1, ~~0.956). Although this study had 95% power to detect a 15% difference between cases and controls with 95% confidence we conclude that the polymorphisms in the LRP and A2M genes contribute very little. if any, to risk for AD in Northern Ireland.

LINKAGE

Henri

France; Niwle

CHROMSOME

ON

Christine

STUDY

DISEASE

FOR LINKAGE

A Ro~mwz, Toronto,

Lind.su,v A Furrrr, Djumil

Lille

Richard,

Emile

Mondor,

Claude

Helbecque,

Inserm

Roux.

Limeil-

G-et&l

France;

Di Menza,

Ctr Hospi-

INSERM

USO8, Lille

EpidemiologIcal evidence suggests that vascular disorders and dementia may share common determinants. As a major player of vascular homeostasis, the renin angiotensin system (RAS) proteins constitute interesting sources of candidate genes. Recent reports suggested that the deletion allele(D) of the angiotensin l-converting enzyme gene (ACE), a key enzyme of the RAS, was associated with lower cognitive functions, but did not characterize any type of dementia. In a population study of 434 patients (age 7526 years) with late-onset Alrheimer direase (age at onset 7325 years) and of 475 controls of similar age (75-+X years), gender and origin, we characterized the frequencies of the ACED allele. The relative risk, adjusted for age. gender and the presence of at least one APOE t4 allele, to develop a dementia for ACE D allele carriers, was 2.3 195% CI:1.3-3.91 for subjects over 74 years of age (higher two quartiles of age). This result was confirmed in a case-control study of 127 demented and of 221 controls, in which two groups of cases were identified: Alzheimer disease affected patients (AD), and non-AD patients. The mea” ages (SD) of controls (n=221). AD (n=S6) and non-AD (n=71) were X4(5). 85(6), X5(6) years, respectively. The relative risks, adjusted for age, gender and the presence of at least one APOE $4 allele, to develop a dementia for ACE D allele carriers, were 2.5 195% CI:l.4-4.51, 2.3 [95% CI:l.O-5.21 and 2.6 [95% Cl:l.2-5.71 for all dementia, AD and non-AD respectively. Alltogether, these results suggest that the ACE D allele may constitute a genetic susceptibility factor for dementia. This observation reinforces the hypothesis of a major implication of vascular risk factors in the occurrence of all types of dementia.

Ekaterinu OF GENETIC

Frederic

de Lilly.

David-Fromentin,

FXI7l(?

Toronto. SPORADIC

France;

Ducimehrw.

Pasteur

I.sabrlle

IS A SUS-

Card@

Mayo

FL

(4651 LACK WITH

Inst

France:

GENE

Sate,

Boston

l/nil,

ON NOT

LATE-ONSET PROVIDE

TO CHROMOSOME

FAMILIAL STRONG

EVI-

lQ22,9QZl.l,

lOQ23,

Edo Richard,

Univ

llQ23.3

You Qiang Canada;

DOES

Song,

Michael Yun L&R,

Andrew

Patrrsorz.

Nicolauu. Unrv

Sch of Medicine,

Boston

of Toronto, Boston,

Univ

of

Sch of Medicine.

Toronto,

ON

Cunadu;

MA

Alsheimer’\ disease (AD) is a complex disorder which can be caused by several different genetlc defects. Many AD cases however, can not he attributed to one of the four known genetic causes, suggesting that other genes might he mvolved. A recent genome scan for AD revealed a few possible regions to which AD could be linked. I” linkage ctudies, complex disorders independent replication of positive findings i\ crucial in order to distinguish between true positives and false positives. We have begun a genome-wide survey in our pedigree dataset. We initially have examined the segregation of three clusters of anonymous markers from chromosomes 1 (D I S 1588. 75cM-DlS518), chromosome 9 (D9Sl52-36cM-D9S279). and chromosome IO (DIOSl423.7lcM-DIOS57l), which have shown suggestive evidence of linkage to AD in a” independent study by Kehoe and co-workers. In our dataset, these marker\ have generated negative evidence for linkage. As another priority of our genetic linkage study we chose chromosome I lq23.3, the region to which p-site APP cleaving enryme (BACE) has been recently mapped. Analysis of these locus in our pedigrees have generated no evidence for linkage, arguing that this locus is not prevalent cause of AD in our dataset. Furthermore, we have found no mutations or polymorphistn? in the open reading frame of BACE in 10 sporadic AD brains as well as in affected members of 20 FAD pedigrees in which no obligate recombinants were detected between AD and microsatellite markers flanking the BACE gene. This resuh wggests that the causative gene& within our dataset are located elsewhere m the genome.