Poster Presentations: P4P662

P4-081 GENETIC VARIABILITY IN THE CHOLESTERYL

ESTER TRANSFER PROTEIN (CETP) GENE MAY

BUFFERTHEEFFECTOFAPOE-ε4ON INCIDENCE

OFAMNESTIC MILD COGNITIVE IMPAIRMENT

(AMCI): RESULTS FOR THE EINSTEIN AGING

STUDY (EAS)

Richard Lipton1, Cuiling Wang1, Mindy Katz1, Molly Zimmerman1,

Carol Derby1, Joe Verghese1, Amy Sanders1, Laurie Ozelius2, Nir Barzilai1,1Albert Einstein College of Medicine, Bronx, New York, United States;2Mount Sinai School of Medicine, New York, New York, United States.

Background: Some carriers of the APOE4 allele do not develop AD, per-

haps because of gene-gene interactions. A single-nucleotide polymorphism

(SNP) at CETp codon 405 (isoleucine to valine V405; rs5882) has been as-

sociated with exceptional longevity and reduced risk of aMCI and AD in

some cohort studies. Based on the observation that dementia-free centenar-

ians with an APOE4 allele tend to carry the CETp V405 variant, we hypoth-

esized that CETPV405 may buffer the influence of APOE4 on the onset of

aMCI and AD. Methods: Analyses were based on follow-up data from the

EAS, a longitudinal study of an ethnically diverse community-based Bronx

cohort. Participants free of dementia and aMCI at baselinewith serial neuro-

psychological and neurological exams were eligible.We examined the influ-

ence of carrying at least one APOE 4 allele on the onset of aMCI, first in the

entire sample and then in subgroups stratified by CETPV405; V+ denotes at

least 1 copy of the CETPV405 variant; V- denotes the CETPI405I genotype.

The effect of genotype on time to onset of aMCI was assessed using Cox

models adjusting for age, gender, education, race and medical comorbid-

ities. Results: The eligible sample included 655 individuals free of both

aMCI and dementia at baseline.121 (18.5%) participants developed incident

aMCI over a mean 4.5 years of follow-up. In the entire sample, carriers of at

least one APOE 4 allele had a non-significant increased risk of incident

aMCI (HR 1.7; 95% CI: 0.7-3.8). In the V- group, the HR for incident

aMCI between APOE 4 carriers and non-carriers was 4.2 (95% CI: 1.2 -

14.9) in the combined ethnic sample and 8.5 (95% CI: 2.4-30.5) in Cauca-

sians. In the V+ group, the HR was 1.0 (95% CI: 0.32 - 3.3) in the combined

ethnic sample and 0.7 (95% CI: 0.1-5.4) in Caucasians. There was a differ-

ential effect of APOE4 on aMCI onset depending upon CETP genotype (P¼0.017). Conclusions: T his stratified analysis suggests a gene-gene interac-

tion between CETp V405 and APOE; the CETp V405 appears to buffer the

effect of APOE4 on aMCI. These findings need replication in other samples

and extension to other endpoints including AD.

P4-082 THE EFFECTS OFAPOLIPOPROTEIN GENOTYPE

ON THE DIAGNOSTIC ACCURACY OF

CEREBROISPINAL FLUID BIOMARKERS FOR

ALZHEIMER’S DISEASE

Ronald Lautner1, Henrik Zetterberg2, Kaj Blennow2, Anders Wallin2,

Sanna-Kaisa Herukka3, Harald Hampel4, Dan Rujescu5, Michael Ewers6,

Niklas Mattsson2, Oskar Hansson7, 1Institute of Neuroscience and

Physiology, University of Gothenburg, M€olndal, Sweden; 2University of

Gothenburg, M€olndal, Sweden; 3University of Eastern Finland, Kuopio,

Finland; 4University of Frankfurt, Frankfurt/Main, Germany;5Ludwig-Maximilians-University Munich, Munich, Germany; 6University

of California, San Francisco, San Francisco, California, United States;7Lund University, Lund, Sweden.

Background:APOEmight modulate the levels of cerebrospinal fluid (CSF)

biomarkers associated with Alzheimer’s disease (AD), especially Ab42.

Here, we aimed to evaluate the effects of the APOE E2/E3/E4 polymor-

phism on the diagnostic accuracy of CSF biomarkers for AD. Methods:

APOE genotype and the CSF levels of Ab42, T-tau and P-tau were deter-

mined in a cohort comprising of 1345 individuals, including 251 controls,

399 patients with stable mild cognitive impairment (MCI), 287 patients

with prodromal AD, 309 demented patients with AD, and 99 patients

with other dementias than AD. Results: CSF levels of Ab42, but not T-

tau or P-tau, were lower in APOE E4 carriers than in non-carriers in

a gene dose-dependent manner irrespective of diagnostic group. However,

the levels of Ab42, T-tau and P-tau differed significantly between subjects

with AD when compared to controls, cases with stable MCI, and patients

with other dementias than AD - even when analyzing subgroups according

to APOE E4 carrier status separately. ROC analysis revealed that Ab42 and

Tau proteins can be used to detect ADwith adequate accuracy (AUCs>0.82)

in individuals with zero or one E4 allele. However, the diagnostic accuracy

of Ab42 in individuals with two alleles was lower (AUC¼0.73). Multivar-

iate binary logistic regression revealed that Ab42 and the Ab42/tau

ratio are more strongly associated with AD than the APOE genotype.

Conclusions: APOE genotype modulates CSF levels of Ab42. However,

the CSF biomarkers, including Ab42, are strongly associated with AD inde-

pendently of APOE genotype.

P4-083 GENOME-WIDE ASSOCIATION ANALYSES OF

ONSETAGE IN LATE-ONSETALZHEIMER’S

DISEASE (LOAD) DEMONSTRATE NO STRONG

EFFECT OUTSIDE OF THE APOLIPOPROTEIN

REGION

Adam Naj1, Yo Park1, Ruchita Rajbhandary1, Kara Hamilton-Nelson1,

Gary Beecham1, Eden Martin1, Richard Mayeux2, Jonathan Haines3,

Lindsay Farrer4, Gerard Schellenberg5, Margaret Pericak-Vance6,

Alzheimer’s Disease Genetics Consortium7, 1University of Miami Miller

School of Medicine, Miami, Florida, United States; 2Columbia University

College of Physicians and Surgeons, New York, New York, United States;3Vanderbilt University Medical Center, Nashville, Tennessee, United States;4Boston University School of Medicine, Boston, Massachusetts, United

States; 5Perelman School of Medicine, University of Pennsylvania,

Philadelphia, Pennsylvania, United States; 6John P Hussman Institute for

Human Genomics, Miami, Florida, United States; 7ADGC, Philadelphia,

Pennsylvania, United States.

Background: LOAD risk loci may also contribute to variation in age of on-

set (AAO) of LOAD, as do the allelic variants in APOE. However, roles in

AAO for the nine newly identified risk loci have not been explored. Daw et

al. (2000) estimated 4 additional loci with effects as great as or greater than

APOE contribute to AAO variation. We examined variants at ten confirmed

LOAD risk loci (APOE, CLU, PICALM, CR1, BIN1, CD2Ap, EPHA1,

ARID5B, the MS4A region, ABCA7, and CD33) to determine if they con-

tribute to variation in AAO among 9,160 LOAD cases from the Alzheimer’s

disease Genetics Consortium (ADGC). Methods: Examining the variants

most significantly associated with LOAD risk at each locus, we tested asso-

ciation with AAO using linear modeling assuming additive effects, adjusted

for population substructure, and performed a random-effects meta-analysis

across datasets.We also performed a genetic burden analysis using genotype

scores weighted by effect sizes from association testing to examine the

aggregate contribution of these loci to variation in the AAO phenotype.

Results: Preliminary analyses confirmed association of APOE regional var-

iation with AAO (rs6857, P¼ 3.30x10�96). Variants at several other LOAD

risk loci also demonstrated statistically significant associations with AAO

(P<0.005), including rs6701713 in CR1 (P ¼ 0.00717), rs7561528 in

BIN1 (P ¼ 0.00478), rs561655 in PICALM (P ¼ 0.00223). APOE contrib-

utes to 3.1% of variation in AAO (R 2¼0.220) whereas the other nine genes

contribute to 1.1% of variation (R 2¼0.200) over baseline (R 2¼0.189).We

performed a GWAS to see if any other genomic variants with weak or no

association to LOAD risk were associated with AAO, and found several in-

teresting but not genome-wide significant associations in a chromosome

8q24.11 region including RAD21 (P ¼ 2.38 x10�6), chromosome 6q16.1

(P ¼ 6.1710 -6), CIB4 (P ¼ 7.70 x10�6), and ABCC4 (P ¼ 9.25x10�6).

Conclusions: In AAO analyses among LOAD cases, we confirmed that as-

sociations of APOE variants with AAO are by far the strongest in the ge-

nome, and in contrast to earlier hypothetical modeling, we show that the

combined effects of other loci do not exceed the effect of APOE on AAO.

If additional genetic contributors to AAO exist, they are individually very

small or are hidden in gene-gene interactions.