at SciVerse ScienceDirect

Neurobiology of Aging 34 (2013) 2457e2464

Contents lists available

Neurobiology of Aging

journal homepage: www.elsevier .com/locate/neuaging

BDNF Val66Met, Ab amyloid, and cognitive decline in preclinical Alzheimer’sdisease

Yen Ying Lim a,b,*, Victor L. Villemagne a,c,d, Simon M. Laws e,f,g, David Ames h,i, Robert H. Pietrzak j,Kathryn A. Ellis a,h, i, Karra D. Harrington a, Pierrick Bourgeat k, Olivier Salvado k, David Darby a,Peter J. Snyder l, Ashley I. Bush a, Ralph N. Martins e,f, Colin L. Masters a, Christopher C. Rowe c,d,Pradeep J. Nathanm,n,o, Paul Maruff a,p, for the Australian Imaging, Biomarkers and Lifestyle (AIBL)Research GroupaMental Health Research Institute, The University of Melbourne, Parkville, Victoria, AustraliabDepartment of Psychiatry, The University of Melbourne, Parkville, Victoria, AustraliacDepartment of Nuclear Medicine and Centre for PET, Austin Health, Heidelberg, Victoria, AustraliadDepartment of Medicine, Austin Health, The University of Melbourne, Heidelberg, Victoria, AustraliaeCentre of Excellence for Alzheimer’s Disease Research and Care, School of Medical Sciences, Edith Cowan University, Joondalup, Western Australia, Australiaf Sir James McCusker Alzheimer’s Disease Research Unit, Hollywood Private Hospital, Perth, Western Australia, AustraliagCo-operative Research Centre (CRC) for Mental Health, http://www.mentalhealthcrc.com, AustraliahAcademic Unit for Psychiatry of Old Age, Department of Psychiatry, The University of Melbourne, Kew, Victoria, AustraliaiNational Ageing Research Institute, Parkville, Victoria, AustraliajDepartment of Psychiatry, Yale University School of Medicine, New Haven, Conn, USAkCommonwealth Scientific Industrial Research Organization Preventative Health National Research Flagship, Australian e-Health Research Centre-BioMedIA, Brisbane, Queensland,Australial Lifespan Hospital System and Department of Neurology, Warren Alpert Medical School of Brown University, Providence, RI, USAmMedicines Discovery and Development, GlaxoSmithKline, Clinical Unit, Cambridge, UKnBrain Mapping Unit, Department of Psychiatry, University of Cambridge, Cambridge, UKo School of Psychology and Psychiatry, Monash University, Clayton, Victoria, AustraliapCogState Ltd., Melbourne, Victoria, Australia

a r t i c l e i n f o

Article history:Received 10 December 2012Received in revised form 18 March 2013Accepted 6 May 2013Available online 12 June 2013

Keywords:BDNF Val66Metbeta-AmyloidCognitive declineHippocampal atrophyAb accumulationPreclinical Alzheimer’s diseaseHealthy older adults

* Corresponding author at: Mental Health ResearParkville, VIC 3052, Australia. Tel.: þ61 3 9389 2932;

E-mail address: y.lim@mhri.edu.au (Y.Y. Lim).

0197-4580/$ e see front matter � 2013 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.neurobiolaging.2013.05.006

a b s t r a c t

Brain-derived neurotrophic factor (BDNF) Val66Met polymorphism has previously been implicated inAlzheimer’s disease (AD)erelated cognitive impairment. We aimed to determine the relationship be-tween BDNF Val66Met and beta-amyloid (Ab) on cognitive decline, hippocampal atrophy, and Abaccumulation over 36 months in 165 healthy adults enrolled in the Australian Imaging, Biomarkers andLifestyle study. In healthy adults with high Ab, Met carriers showed significant and moderate-to-largedeclines in episodic memory, executive function, and language, and greater hippocampal atrophy over36 months, compared with Val/Val homozygotes. BDNF Val66Met was not found to be related to rates ofchange in cognition or hippocampal volume in healthy adults with low Ab. BDNF Val66Met did not relateto the amount of Ab or to the rate of Ab accumulation in either group. High Ab levels coupled with Metcarriage may be useful prognostic markers of accelerated cognitive decline and hippocampal degener-ation in individuals in the preclinical stage of AD.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Current theories of Alzheimer’s disease (AD) place beta-amyloid(Ab) at the center of a series of events that result in synaptic lossand memory impairment (Masters and Selkoe, 2012). Although Ab-

ch Institute, 155 Oak Street,fax: þ61 3 9035 8642.

ll rights reserved.

related disruption to normal synaptic function is not well under-stood, growing evidence suggests that brain-derived neurotrophicfactor (BDNF) is 1 downstreammediator of Ab toxicity (Fahnestock,2011; Garzon and Fahnestock, 2007; Peng et al., 2005). For example,BDNF and its main receptor, tropomyosin-related kinase B (TrkB),are necessary for the synaptic excitation and neuronal plasticitythat subserve memory function (Hariri et al., 2003; Lee et al., 2012),impairment of which is the earliest and most frequent AD symptom(Sperling et al., 2011). Second, in AD andmild cognitive impairment,

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e24642458

BDNF mRNA is reduced substantially in the hippocampus andtemporal lobe (Garzon and Fahnestock, 2007; Lee et al., 2012; Penget al., 2005), with the extent of BDNF loss associated with themagnitude of cognitive impairment (Garzon and Fahnestock, 2007;Peng et al., 2005). Third, in AD murine models, pharmacologicallyincreasing BDNF can ameliorate synaptic dysfunction and improvememory (Caccamo et al., 2010; Lee et al., 2012). Likewise, changes inBDNF secretion occur with improved memory and cognitioninduced by aerobic exercise in humans at risk for AD (Baker et al.,2010; Intlekofer and Cotman, in press), and in AD murine models(Cotman et al., 2007; Intlekofer and Cotman, in press). However,there is currently no direct evidence that BDNF moderates any Ab-related changes in cognition or neurodegeneration in AD.

A polymorphism (rs6265, Val66Met) that affects secretion ofmature BDNF has been implicated in learning and memory inhealthy humans, whereby carriage of the Met allele is associatedwith poor memory, reduced hippocampal volume, and lower hip-pocampal activation on functional imaging (Egan et al., 2003; Haririet al., 2003; Peng et al., 2005; Sambataro et al., 2010). In AD, cross-sectional genetic association studies show that Met carriers havegreater impairment in memory and executive function (Ventrigliaet al., 2002), reduced regional cerebral blood flow in hippocampaland medial temporal lobe regions (Xu et al., 2010), and reducedhippocampal and prefrontal cortical volumes (Connor et al., 1997).In contrast, other studies have observed greater cognitive impair-ment (Harris et al., 2006; Nagata et al., 2012; Voineskos et al., 2011),reduced brain structure volumes (Dennis et al., 2011, Voineskoset al., 2011), and increased risk of AD (Voineskos et al., 2011) inVal/Val homozygotes. The absence of a clear relationship betweenBDNF Val66Met and AD could be due to the small samples, or tocross-sectional study designs, or could indicate that other factors(e.g., age, gender) may moderate relations between BDNF Val66Metand AD (Fahnestock, 2011). As yet, no BDNF Val66Met associationstudy has taken into account levels of Ab.

Because AD has a long preclinical stage in which the rate ofcognitive decline over periods of 3 years or less is agreed to be linear(Albert et al., 2007; Bateman et al., 2012; Villemagne et al., 2011),genetic influences may become clearer from prospective study ofthe preclinical stage. Furthermore, as the complexity of AD patho-logical changes increasewith disease severity, associations betweenBDNFVal66Met and ADmay bemore obvious in preclinical stages inwhich the disease presents almost exclusively as subtle abnormal-ities inmemory (Fahnestock, 2011; Lim et al., 2012). Thus, the aim ofthis study was to determine the extent to which BDNF Val66Met

Table 1Baseline demographic, clinical, and lifestyle characteristics of the overall group, and for e

Overall (n ¼ 165) PiB� (n ¼ 116)

Female, n (%) 82 (49.70%) 58 (50.00%)APOE ε4, n (%) 70 (42.42%) 38 (32.76%)Met carrier, n (%) 58 (35.15%) 44 (37.93%)Age (y) 71.36 (7.15) 70.03 (6.90)SUVR Neocortex 1.40 (0.40) 1.17 (0.09)MMSE 28.75 (1.20) 28.80 (1.20)CDR-SB 0.04 (0.19) 0.05 (0.21)Premorbid IQ 108.81 (7.31) 108.14 (7.86)HADS-D 2.78 (2.29) 2.73 (2.10)HADS-A 4.17 (2.88) 4.11 (2.68)Body mass index 26.52 (3.99) 26.46 (3.72)MAC-Q (at 36 mo) 25.08 (4.41) 25.10 (4.31)IPAQ (at 36 mo) 4056.49 (3537.80) 3910.72 (3271.87)

Note: One-way analysis of variance indicated that age was significantly different betweenand Met carriers, p < 0.05. The c2 test also indicated that number of APOE ε4 carriers wvariables shown in the table differed by PiB status or BDNF Val66Met polymorphism, allKey: BDNF, brain-derived neurotrophic factor; CDR-SB, Clinical Dementia Rating Scale, suHADS-D, Hospital Anxiety and Depression Scale, Depression Subscale; IPAQ, InternationaMini-Mental State Examination; PiB, Pittsburgh compound B; SUVR, standardized uptak

influences Ab-related memory decline (Lim et al., 2012; Villemagneet al., 2011) and reduction in hippocampal volume (HV) (Chételatet al., 2012) in healthy older adults. We hypothesized that Ab-related memory decline, and reduction in HV in healthy adults,would be greater in carriers of the Met allele. As BDNF may be adownstream mediator of Ab toxicity (Fahnestock, 2011), we alsohypothesized that BDNF Val66Met would not affect the rate of Abaccumulation. Finally, we explored whether changes in other areasof cognition were associated with BDNF Val66Met and Ab.

2. Methods

2.1. Participants

Participants were recruited from the group of healthy controls(hereafter termed healthy adults [HA]) enrolled in the AustralianImaging, Biomarkers and Lifestyle (AIBL) Flagship Study of Ageing(Ellis et al., 2009; Rowe et al., 2010). The process of recruitmentof HAs in AIBL has been described in detail (Ellis et al., 2009).Exclusion criteria included the following: schizophrenia; depres-sion (15-item Geriatric Depression Score [GDS] of 6 or greater);Parkinson’s disease; cancer (except basal cell skin carcinoma)within the last 2 years; symptomatic stroke; uncontrolled diabetes;or current regular alcohol use exceeding 2 standard drinks per dayfor women or 4 per day for men. A clinical review panel consideredall available medical, psychiatric, and neuropsychological infor-mation to confirm the cognitive health of each participant. OnlyHAs who had undergone Pittsburgh compound B (PiB)epositronemission tomography (PET) and neuroimaging with structuralmagnetic resonance imaging (MRI), BDNF genotyping, and cogni-tive testing at baseline, 18 months, and 36 months were included inthis study (demographic and clinical characteristics are shown inTable 1). The study was approved by and complied with the regu-lations of 3 institutional research and ethics committees (Ellis et al.,2009). All participants provided written informed consent beforestudy participation. Selection of healthy participants to undergoPiB-PET neuroimaging in the AIBL study was biased to ensure thatapproximately 50% were APOE ε4 carriers (Rowe et al., 2010).

2.2. Measures

2.2.1. NeuroimagingPiB-PET imaging methodology has been described in detail else-

where (Rowe et al., 2010; Villemagne et al., 2011). PET standardized

ach SUVR and BDNF Val66Met group

PiBþ (n ¼ 49) Val/Val (n ¼ 107) Met carrier (n ¼ 58)

24 (48.98%) 54 (50.47%) 28 (48.28%)32 (65.31%) 43 (40.19%) 27 (46.55%)14 (28.57%) d d

74.51 (6.81) 72.21 (6.96) 69.79 (7.30)1.97 (0.27) 1.42 (0.40) 1.37 (0.41)

28.61 (1.22) 28.81 (1.15) 28.71 (1.24)0.02 (0.10) 0.04 (0.20) 0.04 (0.17)

110.39 (5.57) 109.39 (6.90) 107.72 (7.96)2.90 (2.70) 2.83 (2.31) 2.68 (2.26)4.31 (3.32) 4.31 (2.98) 3.91 (2.69)

26.65 (4.35) 26.50 (3.58) 26.55 (4.47)25.03 (4.74) 25.06 (4.25) 25.13 (4.74)

4489.11 (4262.37) 4039.07 (3319.71) 4092.63 (3997.44)

PiB� and PiBþ groups, and premorbid IQ was significantly different between Val/Valere higher in the PiBþ group compared with the PiB� group, p < 0.001. No otherp’s > 0.05.m of boxes score; HADS-A, Hospital Anxiety and Depression Scale, Anxiety Subscale;l Physical Activity Questionnaire; MAC-Q, Memory Complaint Questionnaire; MMSE,e value ratio.

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e2464 2459

uptake value (SUV) data acquired 40 to 70minutes after PiB injectionwere summed andnormalized to the cerebellar cortex SUV, resultingin a region-to-cerebellar ratio termed SUV ratio (SUVR).

The methodology for MRI used for clinical screening and co-registration with PET images has also been described previously(Rowe et al., 2010; Villemagne et al., 2011). Hippocampal volume(HV) was normalized for head size using total intracranial volume(sumof graymatter, whitematter, and cerebrospinal fluid volumes).

2.2.2. GenotypingAn 80-mL blood sample was taken from each participant, and 10

mL was used for large-scale DNA extraction for AIBL bio-banking.The BDNF Val66Met polymorphism (rs6265) was included in acustom Illumina GoldenGate assay, which included 1536 SNPs, andwas performed by the Beijing Genomics Institute. The Val66Metpolymorphism had a call rate of greater than 99% and did not departfromHardyeWeinberg equilibrium in the AIBL HA group (p¼ 0.57).

2.2.3. Clinical and cognitive assessmentsThe AIBL clinical rating scales and neuropsychological battery

have been described in detail elsewhere and were administeredaccording to standard protocols by trained research assistants (Elliset al., 2009; Rowe et al., 2010). Premorbid intelligence was esti-mated using the Wechsler Test of Adult Reading (WTAR); depres-sive and anxiety symptoms were assessed using the HospitalAnxiety and Depression Scale (HADS); subjective memory impair-ment was measured using the Memory Complaint Questionnaire(MAC-Q); and levels of physical activity were measured using theInternational Physical Activity Questionnaire (IPAQ). All partici-pants underwent extensive medical, psychiatric, and neuropsy-chological assessments and neuroimaging after enrollment, and at18 and 36 months after baseline (Ellis et al., 2009).

2.3. Data analysis

Composite scores were computed by standardizing outcomemeasures on individual tests against the baseline mean and SD forthe entire group and then averaging these to compute a compositescore for episodic memory (Rey Complex Figure Test [RCFT,3-minute and 30-minute] delay; California Verbal Learning Test,Second Edition [CVLT-II] total, short delay, and long delay recall;Logical Memory 1 and 2; and the CogState One Card Learning task);executive function (Letter Fluency, Category Fluency Accuracy,Category Fluency Switching, Stroop Colours/Dots, and the CogStateOne Back task); language (Boston Naming Task, and CategoryFluency Total); and attention (Digit Symbol, CogState Detection andIdentification tasks).

Consistentwith earlier studies (Jacket al., 2008;Villemagneet al.,2011), the distribution of PiB SUVR data was skewed negatively and

Table 2Mean (SD) slopes for neuropsychological composites and neuroimagingmeasures, andma

Composite Mean (SD) slope

PiB� Val/Val(n ¼ 72)

PiB� Met carrier(n ¼ 45)

PiBþ Val/Val(n ¼ 35)

PiBþ(n ¼

NeuropsychologicalEpisodic memory �0.067 (0.362) �0.043 (0.378) �0.230 (0.357) �0.4Executive function �0.098 (0.387) �0.105 (0.406) 0.007 (0.386) �0.2Language 0.023 (0.149) 0.057 (0.155) 0.082 (0.144) �0.0Attention �0.098 (0.340) �0.025 (0.357) �0.113 (0.340) �0.0

NeuroimagingHippocampal volume �0.034 (0.049) �0.027 (0.052) �0.047 (0.049) �0.0Ab accumulation 0.016 (0.045) 0.012 (0.045) 0.058 (0.046) 0.0

Note: Values in boldface type indicate significant differences between groups.Key: PiB, Pittsburgh compound B.

could not be normalized with data transformations. Thus, baselinePiB SUVR was classified as negative (PIB�, SUVR < 1.5) or positive(PiBþ, SUVR�1.5) in accordancewith established criteria (Jack et al.,2008; Villemagne et al., 2011).

Data for composite cognitive test scores, HV, and SUVR for eachsubject across baseline,18-month, and 36-month assessments weresubmitted to a series of repeated-measures linear mixed model(LMM) analyses of covariance. We examined the relation betweenBDNF Val66Met (Met carrier, Val/Val homozygote), baseline PiBclassification (PiBþ, PiB�), and time (baseline, 18 months, and 36months) on composite cognitive test scores, HV, and Ab accumula-tion. BDNF, baseline PiB, time, BDNF � PiB interaction, BDNF � timeinteraction, PiB � time interaction, and BDNF � PiB � time inter-actionwere entered as fixed factors; participant as a random factor;age and APOE ε4 status as covariates, and composite cognitive testscore, HV, and Ab as dependent variables in separate analyses. Theseanalyses generated an estimate of the rate of change (i.e., slope) overthe 3 assessments. Consequently, for each of the outcome measure,3 planned comparisons of slopeswere conductedwithin the LMM inaccordance with our study hypotheses. Specifically, slopes werecompared between the following: PiBþ Val/Val homozygotes andPiBþMet carriers; PiB�Val/Val homozygotes and PiB�Met carriers;and PiB� Met carriers and PiBþ Met carriers. The magnitude ofdifference in the rates of change (slopes) of each group wasexpressed using Cohen’s d. Because premorbid intelligence anddepressive symptoms have been linked to BDNF Val66Met (Bicalhoet al., 2012), all LMMs were recomputed with premorbid IQ anddepression included as additional covariates. Although APOE ε4status was included as a covariate, it was not included as an inter-action term with BDNF, PiB or time in these analyses because thenumber of PiBþMet carriers whowere ε4 carriers (n¼ 11) and non-carriers (n¼ 3)was very small and the proportion ofAPOE ε4 carrierswas equivalent in Met carriers and Val/Val homozygotes (Table 1).

3. Results

3.1. Demographic and clinical characteristics

Demographicandclinical characteristics of the full sampleandthe4groupsclassifiedon thebasisof baselinePiBandBDNFVal66Metareshown in Table 1. Study groups did not differ on most demographicand clinical characteristics, although PiBþ HAs were 4 years olderthan PiB� HAs, and Val/Val homozygotes had higher premorbid IQthanMet carriers. The PiBþ grouphadmoreAPOE ε4 carriers than thePiB� group (c2¼ 14.94, p< 0.001), although therewas no differencein the number of APOE ε4 carriers between Met carriers and Val/Valhomozygotes (c2¼ 0.62, p¼ 0.43). At the 36-month assessment, thestudy groups did not differ in levels of subjective memory complaintor physical activity (Table 1).

gnitude of difference (Cohen’s d) in slopes, with age and APOE genotype as covariates

Cohen’s d (95% CIs)

Met carrier14)

PiB� Val/Val vs. PiB�Met carrier

PiBþ Val/Val vs.PiBþ Met carrier

PiB� Met carrier vs.PiBþ Met carrier

71 (0.308) �0.07 (�0.44, 0.31) 0.70 (0.05, 1.32) 1.18 (0.53, 1.80)91 (0.332) 0.02 (�0.35, 0.39) 0.80 (0.15, 1.43) 0.48 (�0.14, 1.08)15 (0.124) �0.22 (�0.60, 0.15) 0.70 (0.05, 1.32) 0.48 (�0.13, 1.08)34 (0.296) �0.21 (�0.58, 0.16) �0.24 (�0.86, 0.38) 0.03 (�0.57, 0.63)

81 (0.044) �0.14 (�0.51, 0.23) 0.71 (0.07, 1.34) 1.07 (0.43, 1.69)38 (0.040) 0.09 (�0.28, 0.46) 0.45 (�0.18, 1.07) L0.59 (L1.19, 0.00)

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e24642460

3.2. Effect of BDNF Val66Met on cognitive decline and HV in PiBþHAs

Relative to the PiBþVal/Val homozygote group, PiBþMet carriersshowed a greater rate of decline on the episodic memory, executivefunction, and languagecomposite test scoresover36months (Table2,Fig. 1aec), with effect sizes for differences between slopes moderateto large in magnitude. Groups did not differ on the rate of decline onthe attention composite (Fig. 1d). PiBþ Met carriers also showed agreater rate of reduction inHV,with the effect sizes for the differencebetween groups being moderate in magnitude (Table 2, Fig. 2).

Re-analysis with premorbid IQ and depression entered asadditional covariates did not change results, withmagnitudes of thedifference between groups in the rates of decline increasing slightlyfor the episodic memory, executive function, and language com-posites and for rate of reduction in HV (Table 3).

Re-analysis with PiBþ Met/Met homozygotes (n ¼ 2) removeddid not significantly change the results. The magnitude of the ratesof decline between PiBþ Val/Val homozygotes and PiBþ Met car-riers on HV and measures of episodic memory and executivefunction remained significant and moderate in magnitude (i.e.,Cohen’s d ¼ 0.60e0.67).

3.3. Effect of BDNF Val66Met on change in cognition and HV inPiB� HAs

Relative to PiB� Val/Val homozygotes, PiB�Met carriers did notdiffer in their rate of change for any composite measure or HV over

a

c

Fig. 1. Trajectories of change in episodic memory composite (a), executive function compozygotes, Pittsburgh compound B (PiB)� Met carriers, PiBþ Val/Val homozygotes, and PiBþ M95% confidence intervals).

36 months, with the effect sizes for these differences small inmagnitude (Table 2). Re-analysis with premorbid IQ and depressionentered as additional covariates did not change this finding.

3.4. Effect of Ab amyloid on cognitive decline and HV in Met carriers

Relative to PiB� Met carriers, PiBþ Met carriers showed agreater rate of decline on the episodicmemory composite only at 36months (Table 2), with the effect size for this difference large inmagnitude. PiBþMet carriers also showed greater reduction in HV,and the effect size for this difference was large.

3.5. Effect of BDNF Val66Met on the rate of Ab accumulation

Comparison of slopes over 36months indicated no differences inthe rate of Ab accumulation between PiBþ Met carriers and PiBþVal/Val homozygotes, or between PiB� Met carriers and PiB� Val/Val homozygotes (Fig. 2b). However, PiBþ Met carriers showed agreater rate of Ab accumulation than PiB� Met carriers.

4. Discussion

These data support the hypothesis that BDNF Val66Met mod-erates Ab-related memory decline and reduction in HV in HAs. InHAs with high Ab at baseline, Met carriers showed significantlygreater decline in episodic memory, executive function, and lan-guage compared with that in Val/Val homozygotes. For each aspectof cognition that declined, the difference in the rate of decline

b

d

site (b), language composite (c), and attention composite (d), for PiB� Val/Val homo-et carriers, with age, premorbid IQ and depression as covariates (error bars represent

a b

Fig. 2. Trajectories of change in hippocampal volume (a), and Ab amyloid accumulation (b), for Pittsburgh compound B (PiB)� Val/Val homozygotes, PiB� Met carriers, PiBþ Val/Valhomozygotes, and PiBþ Met carriers, with age, premorbid IQ and depression as covariates (error bars represent 95% confidence intervals).

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e2464 2461

between BDNF Val66Met groups was, by convention, moderate tolarge in magnitude. In contrast, HAs with low baseline Ab levelsshowed no decline in cognition over 36 months, irrespective ofBDNF Val66Met polymorphism. Importantly, Val/Val homozygositydid not protect against cognitive decline associated with high levelsof Ab, as HAswith this polymorphismwho also had high Ab showedsubstantial decline in memory when compared with HAs with lowAb (Fig. 1a). The data in the current study also indicate, for the firsttime, that HAs with low Ab, irrespective of BDNF Val66Met poly-morphism, show no decline in cognitive function over 3 years. Theabsence of any decline in cognitive function is consistent with ourobservation of stable cognitive function over 18months in the samegroup of older HAs (Lim et al., 2012).

In HAs with high baseline Ab, the rate of reduction in HV wassignificantly greater in those who carried the Met allele than in Val/Val homozygotes. No effect of BDNF Val66Met on HV was observedfor adults with low Ab. As with episodic memory, Val/Val homo-zygosity did not protect against Ab-related reductions in HV. HVreduced at a greater rate over 36 months in Val/Val homozygoteswho also had high Ab levels than in adults with low Ab, irrespectiveof their BDNF Val66Met polymorphism. Although decline in mem-ory (Lim et al., 2012; Villemagne et al., 2011), and reduction in HV(Chételat et al., 2012; Jack et al., 2009) associated with high levels ofAb have been reported previously in HAs, this study shows for thefirst time that Ab-related changes in cognitive function and hippo-campal structure areworsened by BDNF Val66Met in the preclinicalstage of AD. Importantly, the finding that high Ab was a necessarycondition for BDNF Val66Met-moderated reductions in HV andcognitive decline suggests strongly that BDNF Val66Met is adownstreammoderator of the effect of Ab on hippocampal function.

Given that previous studies of healthy adults suggest a geneedose effect for memory impairment related to the BDNF poly-morphism, it is possible that the data for the 2 Met/Met homozy-gotes in the high Ab groupwould have reflected very large cognitiveimpairment, and that these outliers would have biased estimates ofthe rate of cognitive change in the Met carrier group. However, re-analysis of the data with the 2 Met/Met homozygotes excluded didnot substantially change the results, and the rate of reduction in HVand decline in cognitive function remained moderate and signifi-cantly greater in individuals who carried the Met allele than in Val/Val homozygotes. Furthermore, even though our sample of PiBþMet carriers was relatively small, it is unlikely that any of the effectsobserved herewere driven by individuals with extreme data for anyoutcome measure. First, investigations of data distributions andstatistical checks of the assumptions for parametric analysis sug-gested that there was no unusual variance in outcomemeasures forany group. Second, individuals were assessed repeatedly withmultiple convergent cognitive tests, making it highly unlikely thatdata on any one of the cognitive tests would have influenced con-clusions. Finally, changes in cognition were consistent with declinein HV, an independent measure of AD pathology.

The data also support the hypothesis that BDNF Val66Met wouldnot moderate the rate of Ab accumulation in HAs. Consistent withprevious studies (Chételat et al., 2010; Villain et al., 2012), HAs withhigh baseline levels of Ab (i.e., SUVR > 1.5 at baseline) show higherrates of Ab accumulation than HAs with low levels of baseline Ab,irrespective of their BDNF Val66Met polymorphism. Although therewas a trend for Val homozygotes with high Ab to accumulate Ab at afaster rate, it was not significant. Furthermore, although the effectsize of the difference in the rate of Ab accumulation between high Ab

Table 3Mean (SD) slopes for neuropsychological composites and neuroimaging measures, and magnitude of difference (Cohen’s d) in slopes, with age, APOE genotype, premorbid IQ,and depression as covariates

Composite Mean (SD) slope Cohen’s d (95% CIs)

PiB� Val/Val(n ¼ 72)

PiB� Met carrier(n ¼ 45)

PiBþ Val/Val(n ¼ 35)

PiBþ Met carrier(n ¼ 14)

PiB� Val/Val vs.PiB� Met carrier

PiBþ Val/Val vs.PiBþ Met carrier

PiB� Met carrier vs.PiBþ Met carrier

NeuropsychologicalEpisodic memory �0.057 (0.358) �0.024 (0.374) �0.219 (0.355) �0.468 (0.305) �0.09 (�0.46, 0.28) 0.73 (0.08, 1.35) 1.24 (0.58, 1.86)Executive function �0.095 (0.389) �0.100 (0.408) 0.003 (0.389) �0.297 (0.334) 0.01 (�0.36, 0.38) 0.80 (0.15, 1.43) 0.50 (�0.11, 1.10)Language 0.020 (0.152) 0.059 (0.159) 0.092 (0.150) �0.016 (0.128) �0.25 (�0.62, 0.12) 0.75 (0.10, 1.37) 0.49 (�0.12, 1.09)Attention �0.100 (0.338) �0.024 (0.356) �0.111 (0.339) �0.034 (0.296) �0.22 (�0.59, 0.15) �0.23 (�0.85, 0.39) 0.03 (�0.57, 0.63)

NeuroimagingHippocampal volume �0.034 (0.050) �0.027 (0.053) �0.048 (0.051) �0.084 (0.045) �0.14 (�0.51, 0.24) 0.73 (0.08, 1.35) 1.11 (0.46, 1.73)Ab accumulation 0.018 (0.045) 0.011 (0.045) 0.058 (0.046) 0.037 (0.038) 0.16 (�0.22, 0.53) 0.48 (�0.16, 1.10) L0.60 (L1.20, L0.01)

Values in boldface type indicate significant differences between groups.Key: CI, confidence interval; PiB, Pittsburgh compound B.

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e24642462

Val homozygotes and high AbMet carrierswasmoderate, we believethat the non-significant effect was not due to a lack of power, as thissame sample size yielded statistically significant differences for theother neuropsychological andneuroimagingmeasures.Wehave thusinterpreted the results as showing that the rate of Ab accumulationacross 36 months was not moderated by BDNF Val66Met (Table 3,Fig. 2). The results of this study suggest that although high Ab gaverise to cognitive decline and reduction in HV, the deleterious effectswere less in individuals who carried the BDNF Val66Met poly-morphism that has been associated with greater secretion of BDNF(i.e., Val/Val homozygotes). These data accord with findings fromanimal studies, which have similarly found that the secretion ofmature BDNF plays a crucial role in the neuronal integrity of thehippocampus (Egan et al., 2003; Hariri et al., 2003) and that Ab de-creases BDNF levels by reducing phosphorylated cAMP responseelement binding protein, which in turn regulates BDNF transcriptexpression (Garzon and Fahnestock, 2007). Furthermore, human andanimal neuropathological studies have shown that interactions be-tween BDNF Val66Met and Ab-related synaptic changes occur in thehippocampus, and that these changes are related directly tomemoryfunction (Bueller et al., 2006; Egan et al., 2003; Hariri et al., 2003).Finally, genetic databases do not identify BDNF Val66Met poly-morphismas increasingrisk forAD(Bertrametal., 2007).Collectively,the results suggest that althoughBDNFsecretion isunrelated to levelsof Ab, it doesmoderate the extent towhich Ab affects brain structureand cognitive function, at least in the very early stages of AD.

Although there have been many suggestions that BDNF Val66-Met polymorphism is involved in the clinical presentation of AD(Fukumoto et al., 2010; Matsushita et al., 2005; Ventriglia et al.,2002), there has been no clear evidence for relationships betweenBDNF Val66Met and cognitive impairment or risk for AD fromprevious studies. Although some cross-sectional studies reportedthat the Met allele is associated with memory and executive func-tion impairment, or with neuronal dysfunction (Egan et al., 2003;Hariri et al., 2003; Sambataro et al., 2010), others have found thatVal/Val homozygosity was associated with increased susceptibilityto AD (Voineskos et al., 2011) and greater cognitive decline (Harriset al., 2006; Nagata et al., 2012). Importantly, because none of thesestudies measured Ab levels, differences in the relationships be-tween BDNF Val66Met and brain structure or function may reflectvariability in the number of older adults with high Ab included insamples, especially given that the sample sizes of these studies havebeen small, and neuroimaging studies have shown that up to 30% ofHAs have high Ab levels (Rowe et al., 2010; Villemagne et al., 2011).In this study, Ab levels were determined through PiB-PET neuro-imaging, althoughwe expect that similar results would be observedwith Ab levels measured through cerebrospinal fluid analysis.

Several studies suggest thatdemographicorclinical factors suchaspremorbid intelligence and levels of depression interact significantly

with BDNF Val66Met to influence cognitive function (Bicalho et al.,2012). Re-analysis of our data with premorbid intelligence anddepression controlled statistically did not alter the results. Althoughthis finding may suggest that the effect of BDNF Val66Met on Ab-related cognitive decline, reductions in HV, and progression to mildcognitive impairmentorAD is independentof these factors, this resultmay be because HAs enrolled in the AIBL study are generally of highintelligence and had undergone careful screening for any recent orexisting psychiatric illnesses (Ellis et al., 2009).

An important caveat for interpretation of these results is thatHAs enrolled in AIBL were highly educated and had few existing oruntreated medical or psychiatric illnesses. As such, it would beimportant for these findings to be replicated in individuals withhigh Ab in population-based studies, where it is possible that theeffect of BDNF Val66Met on Ab-related decline may be greater thanthat observed here. A second caveat is that, although we controlledstatistically for APOE ε4 carriage, we investigated only indirect in-teractions between APOE, BDNF Val66Met, Ab and cognitive decline.This was mainly due to the small sample sizes and to the fact thatprevious studies have observed that APOE ε4 does not interact withBDNF levels to affect cognitive function (Ventriglia et al., 2002), andthat APOE ε4 carriage did not moderate Ab-related memory decline(Lim et al., 2012). Nonetheless, further investigation of the rela-tionship between APOE, BDNF Val66Met, and Ab on change incognitive function needs to be conducted in larger prospectivestudies of APOE allele groups. Finally, given the small number ofMet/Met homozygotes who had high Ab (n ¼ 2), we were unable toinvestigate whether there was a Met genee dose effect on Ab-related cognitive decline. When the data were re-analyzed withhigh Ab Met/Met homozygotes excluded, the rates of cognitivedecline and hippocampal atrophy for high Ab Met carriers wereslightly reduced, although the decline remained statisticallydifferent from high Ab Val/Val homozygotes, and of a moderatemagnitude. It will be important for future studies to explore this, asa greater decline in Met/Met homozygotes would further suggestthe protective effects of the Val allele on Ab neurotoxicity in syn-aptic function. Notwithstanding these caveats, results of this studyprovide strong support that Met carriage in combination with highAb may serve as important prognostic markers of acceleratedcognitive decline in individuals in the preclinical stage of AD, andthat methods of increasing BDNF levels may be a potential thera-peutic strategy for mild cognitive impairment and AD.

Disclosure statement

Y.Y.L., K.E., K.H., P.B., O.S., and C.M. report no disclosures. R.P., P.S.,and D.D. are scientific consultants to CogState Ltd. D.D. is a share-holder of stock in CogState Ltd. P.M. is an employee of CogState Ltd.D.A. has served on scientific advisory boards for Novartis, Eli Lilly,

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e2464 2463

Janssen, and Pfizer Inc.; has received funding for travel from Janssenand Pfizer Inc.; has received speaker honoraria from Pfizer Inc. andLundbeck Inc.; and has received research support from Eli Lilly andCompany, GlaxoSmithKline, Forest Laboratories Inc., and Novartis.A.B. is a paid consultant for Synthetic Biologics, and holds equity inPrana Biotechnology Ltd, CogState Ltd., and Eucalyptus BiosciencesPty Ltd. R.M. is a consultant to Alzhyme. C.R. serves on scientificadvisory boards for Bayer Schering Pharma, Elan Corporation, andAstraZeneca; has received speaker honoraria from Bayer ScheringPharma; and receives research support from Bayer ScheringPharma, and Avid Radiopharmaceuticals, Inc. V.V. serves as aconsultant for Bayer Schering Pharma; and receives research sup-port from NEDO, Japan. P.N. is an employee of GlaxoSmithKlinePharmaceuticals, and holds shares in the company.

Acknowledgements

Funding for the study was provided in part by the study partners[Australian Commonwealth Scientific Industrial and research Or-ganization (CSIRO), Edith Cowan University (ECU), Mental HealthResearch institute (MHRI), Alzheimer’s Australia (AA), NationalAgeing Research Institute (NARI), Austin Health, CogState Ltd.,Hollywood Private Hospital, Sir Charles Gardner Hospital]. Thestudy also received support from the National Health and MedicalResearch Council (NHMRC) and the Dementia CollaborativeResearch Centres program (DCRC2), as well as ongoing fundingfrom the Science and Industry Endowment Fund (SIEF). The authorsalso acknowledge the financial support of the Cooperative ResearchCentre (CRC) for Mental Health, from the Australian Government.

Alzheimer’s Australia (Victoria and Western Australia) assistedwith promotion of the study and the screening of telephone callsfrom volunteers. The AIBL team wishes to thank the clinicians whoreferred patients with AD to the study: Associate Professor BrianChambers, Professor Edmond Chiu, Dr Roger Clarnette, Dr MaryDavison, Dr John Drago, Dr Peter Drysdale, Dr Jacqueline Gilbert, DrKwang Lim, Professor Nicola Lautenschlager, Dr Dina LoGiudice, DrPeter McCardle, Dr Steve McFarlane, Dr Alastair Mander, Dr JohnMerory, Professor Daniel O’Connor, Dr Ron Scholes, Dr MathewSamuel, Dr Darshan Trivedi, and Associate Professor MichaelWoodward. We thank all those who participated in the study fortheir commitment and dedication to helping advance research intothe early detection and causation of AD.

References

Albert, M.S., Blacker, D., Moss, M.B., Tanzi, R., McArdle, J.J., 2007. Longitudinal changein cognitive performance among individuals with mild cognitive impairment.Neuropsychology 21, 158e169.

Baker, L.D., Frank, L.L., Foster-Schubert, K., Green, P.S., Wilkinson, C.W.,McTiernan, A., Plymate, S.R., Fishel, M.A., Stennis Watson, G., Cholerton, B.A.,Duncan, G.E., Mehta, P.D., Craft, S., 2010. Effects of aerobic exercise on mildcognitive impairment: a controlled trial. Arch. Neurol. 67, 71e79.

Bateman, R.J., Xiong, C., Benzinger, T.L.S., Fagan, A.M., Goate, A., Fox, N.C.,Marcus, D.S., Cairns, N.J., Xie, X., Blazey, T.M., Holtzman, D.M., Santacruz, A.,Buckles, V., Oliver, A., Moulder, K., Aisen, P.S., Ghetti, B., Klunk, W.E., McDade, E.,Martins, R.N., Masters, C.L., Mayeux, R., Ringman, J.M., Rossor, M.N.,Schofield, P.R., Sperling, R.A., Salloway, S., Morris, J.C., Dominantly InheritedAlzheimer Network, 2012. Clinical and biomarker changes in dominantlyinherited Alzheimer’s disease. N. Engl. J. Med. 367, 795e804.

Bertram, L., McQueen, M.B., Mullin, K., Blacker, D., Tanzi, R.E., 2007. Systematicmeta-analyses of Alzheimer disease genetic association studies: the AlzGenedatabase. Nature Genet. 39, 17e23.

Bicalho, M.A.C., Pimenta, F.A., Bastos-Rodrigues, L., de Oliveira Hansen, E.,Neves, S.C., Melo, M., Rosa, D.V., de Souza, R.P., de Miranda, D.M.,de Moraes, E.N., Romano-Silva, M.A., de Marco, L., 2012. Sociodemographiccharacteristics, clinical factors, and genetic polymorphisms associated withAlzheimer’s disease. Int. J. Geriatr. Psychiatry 28, 640e646.

Bueller, J.A., Aftab, M., Sen, S., Gomez-Hassan, D., Burmeister, M., Zubieta, J.K., 2006.BDNF Val66Met allele is associated with reduced hippocampal volume inhealthy subjects. Biol. Psychiatry 59, 812e815.

Caccamo, A., Maldonado, M.A., Bokov, A.F., Majumder, S., Oddo, S., 2010. CBP genetransfer increases BDNF levels and ameliorates learning and memory deficits ina mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A 107,22687e22692.

Chételat, G., Villemagne, V.L., Pike, K.E., Ellis, K.A., Bourgeat, P., Jones, G., O’Keefe, G.J.,Salvado, O., Szoeke, C., Martins, R.N., Ames, D., Masters, C., Rowe, C.C., AIBLResearch Group, 2010. Independent contribution of temporal b-amyloid depo-sition to memory decline in the pre-dementia phase of Alzheimer’s disease.Brain 134, 798e807.

Chételat, G., Villemagne, V.L., Villain, N., Jones, G., Ellis, K.A., Ames, D., Martins, R.N.,Masters, C.L., Rowe, C., AIBL Research Group, 2012. Accelerated cortical atrophyin cognitively normal elderly with high b-amyloid deposition. Neurology 78,477e484.

Connor, B., Young, D., Yan, Q., Faull, R.L., Synek, B., Dragunow, M., 1997. Brain-derivedneurotrophic factor is reduced in Alzheimer’s disease. Mol. Brain Res. 49, 71e81.

Cotman, C.W., Berchtold, N.C., Christie, L.A., 2007. Exercise builds brain health: keyroles of growth factor cascades and inflammation. Trends Neurosci. 30, 464e472.

Dennis,N.A., Cabeza,R.,Need,A.C.,Waters-Metenier, S., Goldstein,D.B., LaBar, K.S., 2011.Brain-derived neurotrophic factor val66met polymorphism and hippocampalactivationduringepisodicencodingand retrieval tasks.Hippocampus21, 980e989.

Egan, M.F., Kojima, M., Callicott, J.H., Goldberg, T.E., Kolachana, B.S., Bertolino, A.,Zaitsev, E., Gold, B., Goldman, D., Dean, M., Lu, B., Weinberger, D.R., 2003. TheBDNF val66met polymorphism affects activity-dependent secretion of BDNFand human memory and hippocampal function. Cell 112, 257e269.

Ellis, K.A., Bush, A.I., Darby, D., De Fazio, D., Foster, J., Hudson, P., Lautenschlager, N.,Lenzo, N., Martins, R.N., Maruff, P., Masters, C., Milner, A., Pike, K., Rowe, C.,Savage, G., Szoeke, C., Taddei, K., Villemagne, V., Woodward, M., Ames, D., AIBLRsearch Group, 2009. The Australian Imaging, Biomarkers and Lifestyle (AIBL)Study of Aging: methodology and baseline characteristics of 1112 individualsrecruited for a longitudinal study of Alzheimer’s disease. Int. Psychogeriatr. 21,672e687.

Fahnestock, M., 2011. Brain-derived neurotrophic factor: the link between amyloid-b and memory loss. Future Neurol. 6, 627e639.

Fukumoto, N., Fujii, T., Combarros, O., Kamboh, M.I., Tsai, S.J., Matsushita, S.,Nacmias, B., Comings, D.E., Arboleda, H., Ingelsson, M., Hyman, B.T., Akatsu, H.,Grupe, A., Nishimura, A.L., Zatz, M., Mattila, K.M., Rinne, J.O., Goto, Y., Asada, T.,Nakamura, S., Kunugi, H., 2010. Sexually dimorphic effect of the val66metpolymorphism of BDNF on susceptibility to Alzheimer’s disease: new data andmeta-analysis. Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 153B, 235e242.

Garzon, D.J., Fahnestock, M., 2007. Oligomeric amyloid decreases basal levels ofbrain-derived neurotrophic factor (BDNF) mRNA via specific downregulation ofBDNF transcripts IV and V in differentiated human neuroblastoma cells.J. Neurosci. 27, 2628e2635.

Hariri, A.R., Goldberg, T.E., Mattay, V.S., Kolachana, B.S., Callicott, J.H., Egan, M.F.,Weinberger, D.R., 2003. Brain-derived neurotrophic factor val66met poly-morphism affects human memory-related hippocampal activity and predictsmemory performance. J. Neurosci. 23, 6690e6694.

Harris, S.E., Fox, H., Wright, A.F., Hayward, C., Starr, J.M., Whalley, L.J., Deary, I.J.,2006. The brain-derived neurotrophic factor val66met polymorphism is asso-ciated with age-related change in reasoning skills. Mol. Psychiatry 11, 505e513.

Intlekofer, K.A., Cotman, C.W. Exercise counteracts declining hippocampal functionin aging and Alzheimer’s disease. Neurobiol. Dis. epub ahead of print. http://dx.doi.org/10.1016/j.nbd.2012.06.011.

Jack, C.R., Lowe, V.J., Senjem, M.L., Weigand, S.D., Kemp, B.J., Shiung, M.M.,Knopman, D.S., Boeve, B.F., Klunk, W.E., Mathis, C.A., Petersen, R.C., 2008. 11C PiBand structural MRI provide complementary information in imaging of Alz-heimer’s disease and amnestic mild cognitive impairment. Brain 131, 665e680.

Jack, C.R., Lowe, V.J., Weigand, S.D., Wiste, H.J., Senjem, M.L., Knopman, D.S.,Shiung, M.M., Gunter, J.L., Boeve, B.F., Kemp, B.J., Weiner, M.W., Petersen, R.C.,Alzheimer’s Disease Neuroimaging Initiative, 2009. Serial PIB and MRI innormal, mild cognitive impairment and Alzheimer’s disease: Implications forsequence of pathological events in Alzheimer’s disease. Brain 132, 1355e1365.

Lee, S.T., Chu, K., Jung, K.H., Kim, J.H., Huh, J.Y., Yoon, H., Park, D.K., Lim, J.Y., Kim, J.M.,Jeon, D., Ryu, H., Lee, S.K., Kim, M., Roh, J.K., 2012. miR-206 regulates brain-derived neurotrophic factor in Alzheimer disease model. Ann. Neurol. 72,269e277.

Lim, Y.Y., Ellis, K.A., Pietrzak, R.H., Ames, D., Darby, D., Harrington, K., Martins, R.N.,Masters, C.L., Rowe, C., Savage, G., Szoeke, C., Villemagne, V.L., Maruff, P., AIBLResearch Group, 2012. Stronger effect of amyloid load than APOE genotype oncognitive decline in healthy older adults. Neurology 79, 1645e1652.

Masters, C.L., Selkoe, D.J., 2012. Biochemistry of amyloid b-protein and amyloiddeposits in Alzheimer Disease. In: Selkoe, D.J., Mandelkow, E., Holtzman, D.M.(Eds.), The Biology of Alzheimer Disease. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, NY, pp. 181e204.

Matsushita, S., Arai, H., Matsui, T., Yuzuriha, T., Urakami, K., Masaki, T., Higuchi, S.,2005. Brain-derived neurotrophic factor gene polymorphisms and sporadicAlzheimer’s disease. J. Neur. Transm. 112, 703e771.

Nagata, T., Shinagawa, S., Nukariya, K., Yamada, H., Nakayama, K., 2012. Associationbetween BDNF polymorphism (Val66Met) and executive function in patientswith amnestic mild cognitive impairment or mild Alzheimer disease. Dement.Geriatr. Cogn. Disord. 33, 266e272.

Peng, S., Wuu, J., Mufson, E.J., Fahnestock, M., 2005. Precursor form of brain-derivedneurotrophic factor and mature brain-derived neurotrophic factor aredecreased in the preclinical stages of Alzheimer’s disease. J. Neurochem. 93,1412e1421.

Y.Y. Lim et al. / Neurobiology of Aging 34 (2013) 2457e24642464

Rowe, C.C., Ellis, K.A., Rimajova, M., Bourgeat, P., Pike, K.E., Jones, G., Fripp, J., Tochon-Danguy, H., Morandeau, L., O’Keefe, G., Price, R., Raniga, P., Robins, P., Acosta, O.,Lenzo, N., Szoeke, C., Salvado, O., Head, R., Martins, R., Masters, C.L., Ames, D.,Villemagne, V., 2010. Amyloid imaging results from the Australian Imaging,Biomarkers and Lifestyle (AIBL) Study of Aging. Neurobiol. Aging 31, 1275e1283.

Sambataro, F., Murty, V.P., Lemaitre, H.S., Reed, J.D., Das, S., Goldberg, T.E.,Callicott, J.H., Weinberger, D.R., Mattay, V.S., 2010. BDNF modulates normalhuman hippocampal ageing. Mol. Psychiatry 15, 116e118.

Sperling, R.A., Aisen, P.S., Beckett, L.A., Bennett, D.A., Craft, S., Fagan, A.M.,Iwatsubo, T., Jack, C.R., Kaye, J., Montine, T.J., Park, D.C., Reiman, E.M., Rowe, C.C.,Siemers, E., Stern, Y., Yaffe, K., Carrillo, M.C., Thies, B., Morrison-Bogorad, M.,Wagster, M.V., Phelps, C.H., 2011. Toward defining the preclinical stages ofAlzheimer’s disease: recommendations from the National Institute on Agingand the Alzheimer’s Association workgroup. Alzheimers Dement. 7, 280e292.

Ventriglia, M., Bocchio Chiavetto, L., Benussi, L., Binetti, G., Zanetti, O., Riva, M.A.,Gennarelli, M., 2002. Association between the BDNF 196 A/G polymorphism andsporadic Alzheimer’s disease. Mol. Psychiatry 7, 136e137.

Villain, N., Chételat, G., Grassiot, B., Bourgeat, P., Jones, G., Ellis, K.A., Ames, D.,Martins, R.N., Eustache, F., Salvado, O., Masters, C.L., Rowe, C.C., Villemagne, V.L.,AIBL Research Group, 2012. Regional dynamics of amyloid-b deposition inhealthy elderly, mild cognitive impairment and Alzheimer’s disease: a voxel-wise PiBePET longitudinal study [e-pub ahead of print]. Brain 135, 2126e2139.

Villemagne, V.L., Pike, K.E., Chételat, G., Ellis, K.A., Mulligan, R.S., Bourgeat, P.,Ackermann, U., Jones, G., Szoeke, C., Salvado, O., Martins, R., O’Keefe, G.,Mathis, C.A., Klunk, W.E., Ames, D., Masters, C.L., Rowe, C.C., 2011. Longitudinalassessment of Ab and cognition in aging and Alzheimer disease. Ann. Neurol. 69,181e192.

Voineskos, A.N., Lerch, J.P., Felsky, D., Shaikh, S., Rajji, T.K., Miranda, D., Lobaugh, N.J.,Mulsant, B.H., Pollock, B.G., Kennedy, J.L., 2011. The brain-derived neurotrophicfactor val66met polymorphism and prediction of neural risk for Alzheimerdisease. Arch. Gen. Psychiatry 68, 198e206.

Xu, C.L., Wang, Z.H., Fan, M., Liu, B., Song, M., Zhen, X.T., Jiang, T.Z., Alzheimer’sDisease Neuroimaging Initiative, 2010. Effects of BDNF Val66Met polymorphismon brain metabolism in Alzheimer’s disease. Mol. Neurosci. 21, 802e807.