The Neuronal Transporter Gene SLC6A15
Confers Risk to Major Depression
Martin A. Kohli,1,16,* Susanne Lucae,1,16Philipp G. Saemann,1Mathias V. Schmidt,1Ayse Demirkan,2Karin Hek,2
Darina Czamara,1Michael Alexander,3Daria Salyakina,1Stephan Ripke,1David Hoehn,1Michael Specht,1
Andreas Menke,1Johannes Hennings,1Angela Heck,1Christiane Wolf,1Marcus Ising,1Stefan Schreiber,4
Michael Czisch,1Marianne B. Mu ¨ller,1Manfred Uhr,1Thomas Bettecken,1Albert Becker,5Johannes Schramm,6
Marcella Rietschel,7Wolfgang Maier,8Bekh Bradley,9,10Kerry J. Ressler,9,11,12Markus M. No ¨then,13Sven Cichon,3,14
Ian W. Craig,15Gerome Breen,15Cathryn M. Lewis,15Albert Hofman,2Henning Tiemeier,2Cornelia M. van Duijn,2
Florian Holsboer,1Bertram Mu ¨ller-Myhsok,1,17and Elisabeth B. Binder1,9,17,*
1Max Planck Institute of Psychiatry, D-80804 Munich, Germany
2Erasmus University Medical Center, Department of Epidemiology, 300 CA Rotterdam, The Netherlands
3Institute of Human Genetics, Department of Genomics, Life & Brain Center, University of Bonn, D-53127 Bonn, Germany
4Christian-Albrechts University, Department of General Internal Medicine, D-24105 Kiel, Germany
5Institute for Neuropathology, University Clinics Bonn, D-53127 Bonn, Germany
6Department of Neurosurgery, University Clinics Bonn, D-53127 Bonn, Germany
7Department of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, University of Heidelberg, D-68159
8Department of Psychiatry, University of Bonn, D-53105 Bonn, Germany
9Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta, GA 30322, USA
10Howard Hughes Medical Institute, Chevy Chase, MD 20815, USA
11Yerkes National Primate Research Center, Atlanta, GA 30329, USA
12Atlanta VA Medical Center, Atlanta, GA 30033, USA
13Institute of Human Genetics, University of Bonn, D-53115 Bonn, Germany
14Institute of Neurosciences and Medicine (INM-1), Research Center Juelich, D-52425 Juelich, Germany
15Institute of Psychiatry, King’s College London, SE5 8AF London, UK
16These authors contributed equally to this work
17These authors contributed equally to this work
*Correspondence: email@example.com (M.A.K.), firstname.lastname@example.org (E.B.B.)
Major depression (MD) is one of the most prevalent
psychiatric disorders and a leading cause of loss in
work productivity. A combination of genetic and
environmental risk factors probably contributes to
MD. We present data from a genome-wide associa-
tion study revealing a neuron-specific neutral amino
acid transporter (SLC6A15) as a susceptibility gene
for MD. Risk allele carrier status in humans and
chronic stress in mice were associated with a down-
regulation of the expression of this gene in the
hippocampus, a brain region implicated in the patho-
physiology of MD. The same polymorphisms also
volume and neuronal integrity. Thus, decreased
SLC6A15 expression, due to genetic or environ-
mental factors, might alter neuronal circuits related
to the susceptibility for MD. Our convergent data
from human genetics, expression studies, brain
imaging, and animal models suggest a pathophysio-
logical mechanism for MD that may be accessible to
Major depression (MD) is a common psychiatric disorder with
a lifetime prevalence rate of 15%–17% (95% confidence interval
[CI]) (Ebmeier et al., 2006). It is not only a potentially fatal disease
with about 2% of patients committing suicide (Bostwick and
Pankratz, 2000) but also one of the leading causes worldwide
for loss in work productivity (Ebmeier et al., 2006; Ustu ¨n et al.,
2004).Current treatmentsareindispensablebuttheir clinicaleffi-
cacy is still unsatisfactory, as reflected byhigh rates of treatment
resistance and side effects (Fava and Rush, 2006). Identification
of mechanisms causing depression is pertinent for discovery of
better antidepressants. The heritability of this disorder has
been estimated to range from 34%–42% (95% CI) (Ebmeier
et al., 2006) and several attempts to identify susceptibility genes
by linkage and candidate gene approaches have been under-
taken. In candidate gene studies, BDNF, SLC6A4, ACE,
P2RX7, TPH2, PDE9A, PDE11A, DISC1, and GRIK3 have been
reported to be associated with the disease (Levinson, 2006).
Only a few of these initial reports have been confirmed by subse-
quent studies or in meta-analyses. In the last years, the first
genome-wide association (GWA) case-control studies in MD
were published. None reported genome-wide significant results,
and their top hits were difficult to replicate (Lewis et al., 2010;
252 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
et al., 2009; Wray et al., 2010). Phenotypic diversity and genetic
heterogeneity as well as a considerable environmental contribu-
tion inherent to MD have been considered to represent major
obstacles for the identification of causative variants.
Here we present results of a GWA case-control study in
a stringently selected sample of MD inpatients of a tertiary clinic
in Munich, Germany, and matched controls devoid of any life-
Antidepressant Response Signature (MARS) study (Hennings
et al., 2009; Ising et al., 2009). We performed replication of the
results of the GWAS in six additional independent samples of
German, Dutch, United Kingdom (UK), and African American
origin (Binder et al., 2008; Choy et al., 2009; Hofman et al.,
2007; Lewis et al., 2010; Muglia et al., 2010; Rietschel et al.,
2010). The herein reported association results are based on an
overall sample size of 15,089 unrelated individuals.
To further characterize the functional relevance of the identi-
fied locus, we analyzed genotypic influences of associated
SNPs on premortem human hippocampus and lymphoblastoid
cell line expression profiles. We also employed in vivo high-
resolution structural magnetic resonance imaging (MRI) and
proton nuclear magnetic resonance spectroscopy (1H-NMR)
with a focus on the hippocampal formation. We selected this
brain region based on our gene expression results and
because decreased neuronal integrity in this brain region had
previously been identified as a risk factor for major depression
(Frodl et al., 2002). Moreover, we investigated a possible role of
the candidate locus in mediating stress vulnerability by interro-
gating its hippocampal expression in a well-established mouse
model of chronic social stress (Schmidt et al., 2007) as chronic
stress represents an established risk factor for MD (Wang,
SNPs on 12q21.31 Are Associated with MD
We performed a GWA study in a sample of 353 unipolar
depressed German inpatients from the MARS study (Hennings
et al., 2009) and 366 screened controls using Illumina 100k and
300kBeadchips (Manhattan plot, seeFigure S2availableonline).
After applying stringent quality-control criteria (see Experimental
Procedures), 365,676 SNPs entered association analysis.
Neither genomic controls nor Eigenstrat showed evidence for
population stratification in this sample (Figure S1). The common
experiment-wide significance in a recessive mode of inheritance
(AA versus AG+GG) after applying the permutation-based
minimum p method for multiple comparison correction over all
tested SNPs and genetic models (Table 1, Figure 1B, and Fig-
ure S2; n = 353/366, nominal p = 5.53e-08; OR = 2.84 [95% CI
1.92–4.21]). Seven additional common SNPs in linkage disequi-
librium (LD) with rs1545843 located in a region spanning about
450 kb gave nominal p values smaller than 5.0e-04 applying
wise r2values ranged from 0.40 to > 0.99 in controls (Figure 2A
and Figure S2A), suggesting that all eight SNPs might tag the
same underlying causative variant. In fact, rs1545843 and
rs1031681 can be used as tagging SNPs for the associated vari-
ants within this locus in Europeans and fall into two separate
bins, with an interbin r squared of 0.67.
We then genotyped the genome-wide significant SNP
(rs1545843) of the GWA study together with seven to nineteen
SNPs in LD within this locus in five independent samples. These
comprised three German case-control samples, including two
samples for which GWA data have been published (Muglia
et al., 2010; Rietschel et al., 2010). The German samples consist
ofpatients with recurrent MDandmatched controlsscreened for
the absence of lifetime anxiety and mood disorders recruited in
Southern Germany (n = 920/1024) (Muglia et al., 2010), patients
city of Bonn (n = 292/1155), as well as patients and controls
recruited as a follow-up of the discovery sample (n = 300/236).
In addition, the association was tested in a sample from the
Netherlands. In the Erasmus Rucphen Family (ERF) study
subsample (n = 1160) (Choy et al., 2009), symptoms of depres-
sion during the past week were assessed using the Center for
Epidemiologic Studies Depression Scale (CES-D) and the
depression subscale of the Hospital Anxiety and Depression
Scale (HADS-D). To create a proxy for case/control status, we
compared the individuals rating in the upper depression scale
quartile (CES-D R 16.0: cases, indicative of a depressive
disorder [Luijendijk et al., 2008]) with those rating in the lower
quartile (CES-D % 3: controls). Finally, we tested for association
oftheidentified locusinacross-sectionalstudy ofAfrican-Amer-
ican subjects with significant levels of trauma recruited in the
waiting rooms of an urban public hospital in Atlanta (n = 991)
(Binder et al., 2008). Depression was rated by using the quanti-
tative Beck Depression Inventory (BDI). In contrast to popula-
tions of European descent these SNPs displayed much less LD
among each other (Figure 2B). For this study, we also created
a proxy for case-control status. As BDI scores higher than
16 are equated to clinically relevant symptoms of current MD
(Viinama ¨ki et al., 2004), we divided the sample at this cutoff for
a case-control analysis.
Table 1 shows the results of the association in all six samples
for rs1545843 as well as two SNPs in moderate LD with it,
rs1031681 and rs7975057. Testing the recessive model of
rs1545843, we observed nominally significant association in
four of the five replication samples, with the same direction of
the effect in all samples. A meta-analysis conducted across all
samples resulted in a genome-wide significant association with
models) for the recessive model of rs1545843 (see Table 1).
Homozygote carriers of the A-allele of this SNP had a 1.42-
fold-higher risk to suffer from depression and depressive symp-
toms compared to carriers of the two other genotypes.
To replicate the genome-wide significant association of
increased risk for depression in homozygous carriers of the
A-allele of rs1545843, we performed an additional replication
study with the UK cases and controls of the RADIANT study
(Lewis et al., 2010) and added the WTCCC2 control cohorts.
This resulted in a cohort of 1636 cases with recurrent unipolar
depression and 7246 controls. An analysis using logistic regres-
sion showed significant evidence both for an effect of the AA
genotype on risk in the same direction as in the other studies
(OR = 1.344, 95% CI 1.080-1.672, p = 0.008) as well as for an
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 253
Table 1. Association Results of the Discovery GWAS and the First Round of Replication
ERF Study (Dutch
CoAllelicAllelic CoAllelic Allelic Co AllelicAllelic CoAllelicAllelic CoAllelicAllelicCoAllelic AllelicCo AllelicAllelic
CaRecRecCaRecRecCaRecRecCaRecRecCaRecRec Ca RecRecCaRecRec
AllDomDomAllDom DomAllDomDom AllDomDomAllDomDom AllDomDomAllDomDom
366 6.0E-05 1.551022 3.3E-031.19236 1.3E-01 1.151157 4.4E-01 0.99 2906.8E-021.196846.8E-031.29
3755 1.9E-06 1.20
A/G 353 5.5E-08 2.859171.6E-021.27300 4.2E-02 1.472921.4E-01 1.18 2839.3E-031.623072.9E-021.30
2452 2.3E-08 1.42
719 1.8E-01 0.801939 1.1E-020.79 536 3.9E-01 0.951449 1.1E-01 1.195734.3E-010.97991 1.1E-020.59
6207 2.9E-02 0.87
366 1.5E-03 1.42 9981.4E-031.22236 2.9E-02 1.271155 5.0E-01 1.002904.1E-03 1.38675 3.0E-021.20
3720 4.1E-07 1.22
A/G353 3.0E-05 2.438983.2E-021.25300 8.4E-03 1.822911.7E-01 1.172832.2E-031.922993.6E-021.30
2424 2.0E-07 1.43
719 1.7E-01 0.80 1896 1.6E-030.75536 1.8E-01 0.84 1446 2.2E-01 1.115734.9E-020.759741.1E-010.80
6144 1.3E-03 0.83
366 2.0E-03 1.41 10161.6E-031.21236 3.3E-02 1.27 1130 3.8E-01 0.972901.7E-031.42681 3.2E-011.05
3719 1.2E-05 1.18
A/G353 2.5E-04 2.17 9151.2E-021.31300 3.0E-02 1.61289 3.7E-01 1.05283 1.4E-031.98 3035.0E-011.00
2443 4.7E-05 1.32
719 1.0E-01 0.77 19315.7E-030.78536 1.1E-01 0.791419 2.4E-01 1.11573 2.5E-020.71 9842.1E-010.87
6162 1.6E-03 0.83
rs1545843showed genome-wide significant associationwithMDinthediscovery case-control GWAS(MARS) underarecessive model.Thisgenome-wide significant associationwasconfirmed
in a subsequent meta-analysis over a total of six samples from five independent studies. rs1545843 and rs1031681 best tag the region of association with MD on chr12q.21.31 defined by eight
SNPs in moderate to strong linkage disequilibrium with each other in Europeans (Figures 1 and 2). rs7975057 is shown as an example of a third SNP, which was one of the more consistently
associated SNPs across all round 1 replication samples. Abbreviations: Allelic, additive allele dosage model (A versus G); Ca, cases; Co, controls; Dom, dominant model (GG versus AA+AG);
hg18, human genome on UCSC build 18 (NCBI 36. 1); n, number of individuals; OR, odds ratio; p, nominal p value; Rec, recessive model (AA versus AG+GG).
aThe allele shown in bold confers greater odds that thecarrier is a case(risk allele for depression). The direction ofassociation is consistent between samples. There is aminor-major allele switch
between Europeans (minor allele: A, MAF in controls: 0.36–0.45) and African-Americans (minor allele G, MAF in controls: 0.35–0.46).
Association of SLC6A15 with Major Depression
254 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
interaction of sex with this effect (p = 0.0150). The RADIANT/
WTCCC2 study was the only study showing such sex 3 geno-
type interaction on depression. A more detailed description of
thisassociation isgivenintheSupplemental Informationsection.
effect of AA on depression in the RADIANT/WTCCC2 study with
the effects in the previous studies, we arrive at an estimate of an
OR = 1.398 (95% CI 1.254–1.557) with a combined two-sided
p value of 1.41e-09 (Figure 3). Considering only the replication
studies (thus excluding MARS), we have an estimate of 1.315
for the OR (95% CI 1.172–1.477) with a two-sided p value of
Figure 1. Genomic Context of the Associ-
ated Region on 12q21.31
a 3 Mb region comprising the 450 kb region of
association with MD according to the UCSC
Genome Browser: RefSeq annotated genes (blue),
GenBank (black), HapMap Linkage Disequilibrium
(red: high LD, white: low LD), and hotspots of
homologous recombination from SNP genotyping
data provided by HapMap and Perlegen (black).
The associated region did not map to any known
gene (comparewithFigure2B). The flankinggenes
next to the region of association are SLC6A15
(+287 kb), a solute carrier family 6 gene that codes
for a sodium-dependent branched amino acid
the brain, and TMTC2 (?989 kb), the trans-
2 gene of unknown function (see also Table S1).
(B) The negative common logarithm (?log10) of the
best model p values (y axis) of all tested SNPs in
the shown region from genome-wide case-control
association testing in the discovery sample were
plotted against the SNPs’ chromosome positions
(x axis). The horizontal line across the figure indi-
catesthe experiment-widesignificancelevelof the
study. The dot above this line represents the
?log10p value of rs1545843. The corresponding
Manhatten plot over all tested SNPs and chro-
mosomes is shown in Figure S2.
Figure 2. LD Structure of the Eight SNPs Associated with MD on 12q21.31
Presented in (A) German controls of the GWAS in MD (n = 366) and (B) in the African-American control sample (BDI < 14, n = 284). Pairwise r-squared values
multiplied with 100 are shown for each SNP pair. rs1545843 (SNP 2), which reached experiment-wide significance in the GWAS, is in moderate LD with the other
seven associated SNPs in Europeans but in low LD in African-Americans (SNP 1).
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 255
While the association with major depression and depressive
symptoms thus was consistent in samples across different
ethnicities, this did not hold true for incident late-life depression.
The association did not replicate in the Rotterdam study (n =
3512) (Hofman et al., 2007), where subjects older than 55 years
of age and free of dementia were screened with the CES-D
at baseline and two follow-up time points (Luijendijk et al.,
2008). A case-control analysis was performed by comparing
subjects who developed depressive disorders and depressive
syndromes at follow up time points (n = 438) with individuals
without clinically relevant depressive symptoms (n = 3074)
(mean age ± standard deviation [SD] of cases: 72.7 ± 7.4 years
and controls: 73.9 ± 8.3 years). None of the investigated SNPs
reached significant association. The average age of 72 years
at which the index depressive episode in the Rotterdam
sample was diagnosed is substantially older than the average
age in the combined German discovery sample and recurrent
depression sample (50.4 ± 13.9 years), the Dutch ERF sample
(48.7 ± 15.0 years), and the African-American sample (39.3 ±
13.7 years). In fact, in the other samples significant associations
with rs1545843 were only observed when individuals % 55 years
were selected but not in the older age group. A series of studies
indicate that late-life depression is pathophysiologically distinct
from earlier onset MD being more strongly related to vascular
disease and future cognitive impairment (Alexopoulos, 2006).
In summary, as shown in the forest plot in Figure 3, the initial
GWAS revealed a SNP (rs1545843) on chr12q21.31 to be asso-
ciated with MD with experiment-wide significance. This could be
confirmed in a meta-analysis across six additional independent
samples, including one sample of African-American heritage,
with the recessive model of rs1545843 reaching genome-wide
Genomic Context of the Associated Region on 12q21.31
The associated SNP lies within a region of SNPs in moderate LD
that span a gene desert of about 450 kb in size on 12q21.31
mapping neither to any annotated gene nor to predicted human
mRNAs with the exception of some small human expressed
sequence tags (EST, Figure 1A and Table S1). The closest
RefSeq annotated gene is SLC6A15 (NM_182767), which ends
287 kb further distal to the region of association. It belongs
to the solute carrier 6 (SLC6) gene family and codes for a
sodium-dependent branched-chain amino acid transporter
(Bro ¨er, 2006). SLC6A15 gene expression is highest in the human
brain as well as the brain of other vertebrate species (UniGene,
2009; Allen Brain Atlas, 2008). In rodents and humans,
campus (Farmer et al., 2000; Masson et al., 1996). Other genes
distal to SLC6A15 are TSPAN19, LRRIQ1, and ALX1 (Figure 1A).
Their function is largely unknown and their expression levels are
low in the vertebrate brain (UniGene, 2009). The nearest gene on
the proximal side, transmembrane and tetratricopeptide repeat
containing 2 gene (TMTC2, NM_152588), ends 989 kb from the
region of association. It is expressed in a variety of tissues
including the brain, but its function is also unknown. According
Figure 3. Forest Plot of the Combined
rs1545843 remained genome-wide significantly
associated with MD in the meta-analysis after
replication round 1 under the recessive model (AA
versus AG+GG, see Table 1). This association was
further replicated in the RANDIANT/WTCCC2
sample. The combined meta-analysis p value over
Munich antidepressant response signature study,
the German GWAS discovery case-control MD
sample (n = 353/366). Munich recurrent depres-
sion: the Southern German recurrent depression
and control replication sample (n = 917/1022).
ERF: the Dutch Erasmus Rucphen Family study
MDcase-control subsample (n= 283/290). Emory:
the African-American MD case-control subsample
from Emory University in Atlanta (n = 307/684).
Bonn: West German MD case-control replication
sample (n = 292/1157) (Rietschel et al., 2010).
MARS2: additional MD cases and controls from
the MARS study that were recruited after the
GWAS (n = 300/236). RADIANT/WTCCC2: UK
cases and controls of the RADIANT study and
additional controls from the WTCCC2 cohorts
(n = 1636/7246).
Association of SLC6A15 with Major Depression
256 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
to HapMap and Perlegen (Myers et al., 2005) genotyping data,
several hotspots of homologous recombination are predicted
between the associated region and the flanking genes (Fig-
ure 1A), making it unlikely that the underlying functional variant
might directly hit a classical promoter region or the open reading
frame of a known gene. However, long-range regulatory effects
have been described (Kleinjan and van Heyningen, 2005). To
address this issue, we analyzed genome-wide gene expression
data sets of human hippocampus and lymphoblastoid cell lines
(Stranger et al., 2005).
Gene Expression Studies Reveal SLC6A15 as Putative
Candidate Gene within the 12q21.31 Locus
Weanalyzed genome-wide Illumina expression array data onthe
locus associated with MD on 12q21.31 in a premortem human
hippocampus expression study from individuals with temporal
lobe epilepsy of European descent and gene expression from
EBV-transformed lymphoblastoid cell lines of the 210 unrelated
HapMap individuals of different human populations (CEU,
CHB, JPT, YRB) (Stranger et al., 2005). Previous studies
reported that the median distance between SNPs and genes
whose mRNA expression is significantly regulated by them
is approximately 30 kb, ranging up to a maximum of 1 Mb (Myers
et al., 2007). We therefore assessed all five RefSeq annotated
genes within 1.5 Mb proximal to and distal of rs1545843 on
12q21.31 (Figure 1A and Table S1, TMTC2, SLC6A15,
TSPAN19, LRRIQ1, ALX1). Expression levels of all seven avail-
able probes (three for SLC6A15) were related to genotypes of
two of the SNPs associated with MD which best tag the overall
associated SNPs on 12q21.31 for European populations,
rs1545843 and rs1031681 (Table 1). We tested the allelic
and both alternative recessive-dominant genetic models of
rs1545843 and rs1031681 and each probe and applied Bonfer-
roni correction for the number of performed statistical tests.
Both SNPs showed association only with the hippocampal
experiment-wide significance under a recessive model of inher-
itance (AA versus AG+GG: rs1545843: p = 4.3e-04, corrected
p = 1.8e-02, and rs1031681: p = 1.4e-04, corrected p = 6.6e-03,
n = 137). Risk genotype carrier status was associated with less
SLC6A15 transcript (Figures 4A and 4B). These associations
were supported by data from lymphoblastoid cell lines from
the HapMap individuals where expression of the full-length
SLC6A15 transcript was lower in carriers of the depression
risk genotypes (Figure S3) and in an expression data set from
peripheral blood monocytes (Heinzen et al., 2008) but not in
a frontal cortex expression study (Myers et al., 2007), probably
due to the lower expression of this gene in this brain region.
Thus, gene expression experiments, including hippocampus
expression, point toward an effect of the associated locus on
SCL6A15 expression via long-range regulatory mechanisms
(Kleinjan and van Heyningen, 2005).
Presumed Function of SLC6A15
SLC6A15 belongs to the solute carrier 6 (SLC6) gene family,
which also includes the monoamine and gamma-amino butyric
acid (GABA) transporters and codes for a sodium-dependent
branched-chain amino acid transporter (Bro ¨er, 2006). Experi-
mental data from SLC6A15 knockout mice indicate a moderate
contribution of SLC6A15 to total proline and leucine transport
Proline, the amino acid with the highest affinity for SLC6A15,
et al., 2006), and this transporter could thus be involved in the
regulation of glutamate transmission (Tapiero et al., 2002).
Figure 4. SLC6A15 mRNA Expression per rs1545843 Genotype
Measured in premortem human hippocampus from individuals of European
descent with temporal lobe epilepsy.
(A) The MD risk genotype (AA) is associated with reduced full-length (FL, red
boxes in B) SLC6A15 mRNA expression levels compared to the nonrisk
genotypes (AG+GG). None of the other genes flanking the region of associa-
tion with MD showed experiment-wide significant rs1545843 genotype-
specific alterations in expression levels. SLC6A15 S: short mRNA isoform of
in human hippocampus. On the x axis the three genotype groups of rs1545843
are plotted against normalized SLC6A15 mRNA levels on the y axis (group
SLC6A15 transcript. For results of an analogous eQTL analysis in lympho-
blastoid cell lines of HapMap individuals see supplemental Figure S3.
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 257
Effects of Risk Genotypes on Hippocampal Volume
Due to the expression profile of SLC6A15 and its presumed
role in neuronal amino acid transport and glutamate synthesis
(Bro ¨er et al., 2006) and due to reported hippocampal volume
changes in MD (Frodl et al., 2002; Videbech and Ravnkilde,
2004), weinvestigated both
spectroscopy (1H-NMR) markers of hippocampal integrity and
signaling in subsamples of the Southern German discovery
and replication samples (for sample see Supplemental Experi-
We confirmed bilateral hippocampal volume reductions in
recurrent depression (F5,381> 15.128, p < 1.2e-04, n = 204, Table
S2) and found a rs1545843 genotype 3 diagnosis interaction
case-control, genotypes AA versus AG/GG: F5,381= 5.861, p =
the hippocampal formation revealed strongest effects for the
AA versus AG/GG: F5,381= 9.512, p = 0.002, pcorr< 0.05, right:
F5,381= 5.686, p = 0.011, n = 204 cases and 186 controls, Table
S2). For rs1081681, which is highly correlated with rs1545843 in
the MR morphology sample (r = 0.819), diagnosis 3 genotype
interaction effects were even stronger with a similar emphasis
on the left hemisphere and the CA region (Figure 5 and Table
for either polymorphism for the dentate gyrus and the subiculum
of the hippocampus and the control region (precentral gyrus).
Hippocampal morphology is a heritable trait (h2= 0.4) (Sullivan
et al., 2001); nonetheless, it is subject to stronger environmental
nosis Interaction Effects on Hippocampal
(A) Based on cytoarchitectonic probability maps,
automated volumetry of gray matter (GM) of the
total hippocampus (cornu ammonis, subiculum
and dentate gyrus)and respectivesubregionswas
segmentation and coregistration. The resulting
maximum probability maps projected on a stan-
dard brain template in atlas space are shown.
(B) Results of the left total hippocampal GM:
bars show adjusted mean values and one stan-
dard error of the mean for the main effect of
diagnosis and the rs1545843 genotype (AA versus
AG/GG) 3 diagnosis interaction effect. Lowest
mean volumes were seen for patients with the AA
(C) Corresponding depiction for the left cornu
ammonis (*nominal p < 0.05, **Bonferroni cor-
rected p < 0.05.). Results of other subregions and
of right hemisphere are reported in Table S2.
regions (Glahn et al., 2007), and interac-
tions between recurrent depression and
specific genetic predispositions as indi-
cated by our results may thus promote
compared toother brain
hippocampal atrophy, which has been repeatedly reported for
MD (Videbech and Ravnkilde, 2004).
These analyses were complemented by analyzing1H-NMR
markers of hippocampal integrity, including N-acetyl aspartate
(NAA). While NAA serves mainly as a marker of neuronal
viability, it is also regarded as a reservoir for glutamate (Benar-
roch, 2008). To investigate genotype effects of left hippocampal
neurochemistry, we focused on healthy, nonmedicated control
subjects (n = 81) as mood state and medication might influence
hippocampal neurochemistry. Multivariate analysis detected
a significant genotype effect of rs1031681 on hippocampal
metabolites (Wilks’ lambda: 0.683, F2,75= 2.976, p = 0.002)
with univariate comparisons pointing toward NAA (F2,75 =
6.143, p = 0.003, pcorr< 0.05). More specifically, A-risk-allele-
carriers of rs1031681 showed lower levels of hippocampal
neuronal integrity and Glx signaling already in healthy carriers
(NAA: F1,76= 5.575, p = 0.021; Glx: F1,76= 5.752, p = 0.019;
Cr: F1,76= 4.009, p = 0.049, Figure S4B). For NAA, a similar
effect was detected for A-carriers of rs1545843 (F2,75= 5.333,
p = 0.024).
The imaging data thus suggest that risk allele carrier status is
associated with a decrease in hippocampal neuronal integrity
already in healthy controls and that patients with recurrent major
depression and the risk genotype experience an exacerbated
reduction in hippocampal volume.
Evidence for a Role of SLC6A15 in Stress Vulnerability
Epidemiological studies on MD report a 2- to 3-fold risk increase
for individuals exposed to chronic stress (Wang, 2005), and twin
Association of SLC6A15 with Major Depression
258 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
studies clearly point to an increased susceptibility for MD as a
result of a combination of environmental and genetic risk factors
(Kendler et al., 2002). To further validate a role for SLC6A15
in MD, we used microarray gene-expression data from the
hippocampus of mice subjected to chronic stress according to
a recently developed and extensively validated mouse paradigm
of chronic social stress in which susceptible animals show
behavioral, endocrine, and molecular changes reminiscent of
a depression-like phenotype (Schmidt et al., 2007, 2010) (Fig-
ure S5). We selected the six most susceptible and the six most
resilient individuals from a formerly stressed group of 120
mice. Pooled mRNA samples of laser-assisted microdissections
from the CA subregion 1 (CA1) of the hippocampus from both
experimental groups (Supplemental Experimental Procedures)
were analyzed on genome-wide Illumina BeadChips. Expression
data for the probes specific for the genes in the associated
region, TMTC2, SLC6A15, LRRIQ1, and ALX1, were compared
between the two groups. SLC6A15 mRNA levels were reduced
1.9-fold in the CA1 region in stress-susceptible versus stress-
resilient mice. Expression levels of the other genes did not
exceed background noise in the CA1 region and are thus
probably not expressed at higher levels in this brain region
(TableS3).This furthersupports SLC6A15 asthegene ofinterest
within this locus. The reduction of SLC6A15 expression in CA1
could be validated by in situ hybridization in nine stress-suscep-
tible versus nine stress-resilient mice (Figures 6A and 6B).
Figure 6. Reduced Hippocampal SLC6A15
mRNA Expression in Stress-Susceptible
(A) The significant reduction in SLC6A15 mRNA
levels in the CA1 hippocampal region between
stress-resilient (R) and -susceptible (S) mice
detected by microarray analysis could be con-
firmed by in situ hybridization (n = 9/9, ?2.1-fold
(B) Two representative autoradiographs of hippo-
campal slices from one animal per group are
(C and D) SLC6A15 mRNA was also significantly
reduced in the dentate gyrus (DG, ?1.5-fold) and
by trend reduced in the visual cortex (Cx, ?1.7-
fold).+p < 0.06; **p < 0.01; ***p < 0.001. See also
Figure S5 for description of the mouse model and
Table S3 for microarray results. Data are pre-
sented as mean volumes ± standard errors of the
Moreover, a significant reduction in
observed in the dentate gyrus of stress
susceptible animals (Figures 6C and
6D). The demonstrated downregulation
tible mice, most prominent in the CA1
region of the hippocampus, suggests
SLC6A15 to play a role in long-term
effects of chronic stress on neuronal
circuits and is in accordance with the
human MD risk genotype-dependent effects, assayed by
in vivo volumetry, which were also strongest in the CA subregion
of the hippocampus.
We performed a GWA study, with replication of the top hit and
genome-wide significant association in the meta-analysis
across a total of 4,088 patients and 11,001 controls, including
one sample from a different ethnic background. Together with
gene expression data, neuroimaging correlates and evidence
from a mouse model of chronic stress our results point toward
SLC6A15, a neuronal amino acid transporter, as a candidate
gene in the pathophysiology of major depression.
Even though the direction of the association of rs1545843 with
depression and depressive symptoms was the same in all
samples with nongeriatric depression, the effect sizes were
heterogeneous, with a much larger effect in the discovery
sample (OR = 2.8 for the recessive model) as compared to the
other samples with odds ratios ranging from 1.18 to 1.61. The
strong association and low p value in the discovery sample is
probably due to the ‘‘winners’ curse,’’ but this phenomenon
has also been observed for other, now established, disease
loci. For example, the association of a SNP in the FGFR2 gene
with breast cancer was much stronger in the rather small
discovery sample than in any of the subsequent replication
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 259
samples. However, the direction of the association was consis-
tent and reached a p value of 2e-76 in close to 30,000 cases
and controls (Easton et al., 2007). This indicates that heteroge-
neous effect sizes with overestimation of the effect in a small
discovery sample may still be in agreement with a true signal.
In addition to the genome-wide significance (Dudbridge
and Gusnanto, 2008) observed in our study, replication of the
effect in samples of different ethnicities, European and African-
American, might be a further indicator for a true effect.
In addition to replication in independent samples, the func-
tional relevance of the associated locus is supported by results
of gene expression analyses in premortem human hippocampus
and EBV-transformed lymphoblastoid cell lines of the HapMap
individuals (Stranger et al., 2005) and peripheral blood mono-
cytes (Heinzen et al., 2008). While there is a strong indication
of the regulatory relevance of the region associated with MD
for SLC6A15 expression, we cannot exclude that these variants
might also influence the expression of six unspliced brain ESTs
and four spliced ESTs described in nonbrain tissue that have
been mapped to the region of association but were not probed
S1). Additional nonannotated transcripts, as described in the
ENCODE pilot project in regions of the genome previously
thought to be transcriptionally silent (Birney et al., 2007), might
also be functionally relevant for this association.
The imaging genomics results provide evidence that the asso-
ciated SNPs and related functional effects on SLC6A15 expres-
sion might be of relevance for the integrity of brain neurocircuits
shown to be important in MD (Frodl et al., 2002). We found lower
total hippocampal volumes,particularly ofthecornuammonis, in
indicating a higher vulnerability to the well-documented effects
of recurrent depressive episodes on hippocampal volume (Frodl
et al., 2002; Videbech and Ravnkilde, 2004). Further support for
the detrimental effects of the risk allele on neuronal integrity in
this brain region came from1H-NMR spectroscopy. We noted
that healthy risk allele carriers exhibited lower hippocampal
NAA has been reported for different psychiatric disorders and
was also decreased in currently depressed unipolar patients in
this study (Figure S4b). In animal models, hippocampal NAA
can be decreased by chronic stress (Cze ´h et al., 2001; Li et al.,
2008). Thus, a genetic predisposition toward lower hippocampal
NAA, similar to a condition induced by chronic stress experi-
ments, may impair an individual’s resilience to stress which is
a risk factor for MD (Wang, 2005).
While the genetic association data pointed most strongly to
rs1545843, gene expression and imaging data association
were strong with both tag-SNPs of the locus, rs1545843 and
rs1031681. In healthy subjects, genotype effects on hippo-
campal neurochemistry were more prominent for rs1031681
compared with rs1545843, both in terms of effects on NAA
and Glx and in terms of robustness toward multiple test correc-
tion. This is an indication that both SNPs tag the likely underlying
functional variants that still remains to be identified. To this aim,
deep-sequencing analyses are currently underway.
Together with the demonstrated downregulation of SLC6A15
expression in stress-susceptible mice, human gene expression
and imaging data support a role for hippocampal SLC6A15
function in stress sensitivity and the pathophysiology of MD.
This would be in line with a proposed role of the SLC6A15 trans-
porters in neuronal metabolism and the provision of substrates
for neurotransmitters, and specifically glutamate synthesis
(Bro ¨er et al., 2006). Thus, decreased hippocampal NAA and
by extension glutamate neurotransmission (Benarroch, 2008),
related to genetic factors, may limit excitatory signaling capacity
with secondary effects on stress response regulation and hippo-
campal function in general (Herman et al., 2003).
In conclusion, the above presented results from human
genetics, gene expression, volumetric imaging, spectroscopy,
and a mouse model of chronic stress all support the notion
that lower SLC6A15 expression, especially in the hippocampus,
could increase an individual’s stress susceptibility by altering
neuronal integrity and excitatory neurotransmission in this brain
region. Recently, the prokaryotic leucine transporter homolog
(LeuTaa) of SLC6A15 has been crystallized from Aquifex aeolicus
and was shown to bind tricyclic antidepressant drugs that can
directly block leucine transport by closing the molecular gate
for the substrate in a noncompetitive manner (Zhou et al.,
2007). Due to the high degree of phylogenetic conservation of
the antidepressant binding site, these drugs probably also bind
to the human transporter. Because SLC6A15 appears amenable
to drug targeting, our results may lead to the discovery of a novel
class of antidepressant drugs.
MARS (GWAS) Sample
Three hundred and fifty-three unipolar depressive inpatients (155 males,
198 females) were recruited for the Munich Antidepressant Response Signa-
ture (MARS) project (Hennings et al., 2009; Ising et al., 2009) at the Max Planck
Institute of Psychiatry (MPIP) in Munich, Germany. The mean age (±SD) was
49.5 ± 14.3 (males: 48.4 ± 13.4, females: 50.4 ± 15.0) years. See Hennings
et al. (2009) and Ising et al. (2009) for more details on patient recruitment.
Briefly, patients were included in the study within 1–3 days of admission to
the hospital and diagnosis was ascertained according to the Diagnostic and
Statistical Manual of Mental Disorders (DSM) IV criteria. Patients fulfilling
the criteria for at least a moderate depressive episode (HAM-D R 14 on the
21-item Hamilton Depression Rating Scale) entered the analysis. Patients
suffered from a first depressive episode (36.8%) or from recurrent depressive
disorder (63.2%). All included patients were of European descent and 88.7%
were of German origin. Three hundred and sixty-six control subjects were
matched to the patient sample for age, gender, and ethnicity from a randomly
selected Munich-based community sample and underwent a strict screening
procedure for the absence of psychiatric and severe somatic disease (Heck
et al., 2009). The overall inclusion rate of all contacted probands was 50.3%.
These subjects thus represent a group of individuals from the general popula-
tion who have never been mentally ill. Age, gender, and ethnicity did not differ
from the patient sample. This study has been approved by the ethics
committee of the Ludwig-Maximilians-University (LMU) in Munich and written
informed consent was obtained from all subjects.
The Southern German Recurrent Depression Replication Sample
This sample included 920 patients (302 males, 618 females) suffering from
recurrent major depression (Lucae et al., 2006; Muglia et al., 2010) as well
as 1024 controls matched to the patient sample for age, gender, and ethnicity.
The MARS Replication Sample
This sample included an additional 300 unipolar depressed patients and 236
controls, recruited according to the same protocol as the MARS discovery
sample but not genotyped on the initial Illumina platforms.
Association of SLC6A15 with Major Depression
260 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
The Bonn Replication Sample
This sample included patients with a DSM-IV diagnosis of major depression
atry of the University of Bonn, Germany as described in Rietschel et al. (2010).
Of the 604 individuals described in this publication, only the 292 without
a family history of an axis I disorder other than major depression were used
in this analysis. Population-based controls were recruited as described in
Rietschel et al. (2010).
The Erasmus Rucphen Family Study Subsample
This subsample included 1160 participants from the Erasmus Rucphen Family
(ERF) study, part of the Genetic Research in Isolated Population (GRIP)
program (Aulchenko et al., 2004). The Center for Epidemiologic Studies
Depression Rating Scale (CES-D) (Radloff, 1977; Zigmond and Snaith, 1983)
(Spinhoven et al., 1997; Weissman et al., 1977) was used to define depression
using a cutoff of CES-D R 16 as indicative of a depressive disorder (Luijendijk
et al., 2008).
The African-American Replication Sample
This sample included 972 African-Americans (356 males, 616 females) all
screened with the Beck Depression Inventory (BDI) (Beck et al., 1961; Viina-
ma ¨ki et al., 2004). Study design, ascertainment, and rating protocols have
been described elsewhere in more detail (Binder et al., 2008). A BDI score of
16 or greater was considered indicative of current depression.
The Rotterdam Study Subsample
This subsample included 7983 participants from the Rotterdam Study,
a prospective cohort study from 1990 conducted in the Netherlands. All inhab-
itantsaged55and overwereeligible (Hofmanetal.,2007).Depressionwasas-
certained using the CES-D, a semistructured interview with the Present State
Examination (PSE) by a clinician, and GP records and specialist letters.
The UK Replication Sample (RADIANT)
This sample included 1636 patients with a diagnosis of recurrent major
depression (except for 20 with first episode) recruited within the Depression
CaseControl (DeCC) study, theDepressionNetwork (DeNET) affected siblings
linkage study, and the Genome-Based Therapeutics in Depression (GENDEP)
study (Lewis et al., 2010). The matched screened controls described in Lewis
et al. (2010) (n = 1594) and the publicly available controls from the Wellcome
Trust Case Control Consortium 2 (n = 5652) were used for this analysis.
A more detailed description of the study samples can be found in the
Supplemental Experimental Procedures.
Genome-wide SNP genotyping for the MARS discovery sample was per-
formed on Sentrix Human-1 (100k) and HumanHap300 (317k) Genotyping
BeadChips (Illumina, San Diego, USA) according to the manufacturer’s stan-
dard protocols. On the Illumina Human-1 Genotyping BeadChip about
109,000 exon-centric SNPs can be investigated. Nearly 25,000 of the loci
are located in transcripts and more than 73,000 loci are within 10 kb of coding
sequences. The Illumina HumanHap300 Genotyping BeadChip comprises
about 317,000 SNPs. The average call rate achieved was higher than 99%,
with samples below 98% being either retyped or excluded from the study.
Genotyping of the German replication samples (except MARS replication)
and the African-American replication sample was performed on a MALDI-
TOF mass-spectrometer (MassArray system, Sequenom, San Diego, USA)
employing the manufacturer’s AssayDesigner software for primer selection,
multiplexing, and assay design, and the homogeneous mass-extension
(hMe) process for producing primer extension products. MALDI-TOF SNP
genotyping was performed at the Genome Analysis Center (GAC) facility of
able upon request. The individual-wise mean call rate over all plates and these
SNPs was above 98%. Genotypes of all SNPs were in HWE (p < 0.05). To
exclude genotyping errors in the German studies, we regenotyped the two
tagging SNPs (rs1545843 and rs1031681) in more than 95% and 80% of indi-
viduals in the MARS discovery GWAS and the German recurrent depressive
replication sample, respectively, using the MALDI-TOF platform. We obtained
a genotype concordance rate with the genotypes produced by the Illumina
assays of > 99.9%. In the UK studies, all subjects had been genotyped on
the Illumina 610k-Quad Beadchips. In the ERF study 1000 individuals were
genotyped with Illumina 300k, 100 individuals with Illumina 370k arrays, and
200 individulas with the Affymetrix 250k array. The Rotterdam study samples
were genotyped by using the Illumina 550k arrays and the additional MARS
samples using the Illumina 610k array.
Imputation of Genotypes
The imputation of genotypes from ERF and Rotterdam study was performed
using the Maximum Likelihood Method as implemented in the MACH software
v 1.0.16. Release 22 HAPMAP CEU population was used as reference. This
effort yielded a total of 2,500,000 SNPs. Only SNPs with call rates > 98%,
imputations was 0.97 for the 19 SNPs tested within the 450 kb region on
12q21.31. For the MARS replication sample, the genotypes of rs1031681
were imputed using Impute v2.1.0 and the HapMap CEU as a reference pop-
ulation. rs1545843 failed QC in the controls of the UK sample with call
rates < 98% and a p value for differential missingness < 1e-08. Its genotypes
were therefore imputed in both cases and controls using BEAGLE 3.1 (Brown-
ing and Browning, 2009) on HapMap 3.
Power calculations were performed using the Genetic Power Calculator
(Purcell et al., 2003) (http://pngu.mgh.harvard.edu/?purcell/gpc). Given a
prevalence of unipolar depression of 16% (Kessler et al., 2003), a marker in
LD (D’ = 1) with a risk allele R and an alternative protective allele N under an
allelic log-additive, dominant or recessive model and 80% power in our
discovery genome-wide study at a significance level a equal to 1.4 3 10?7
(= 0.05/365,676 SNPs), we would be able to detect an effect with an OR of
2.2 or larger.
Genomic controls (Devlin et al., 2001) for the case-control phenotype were
calculated with R-2.5.0 (http://cran.r-project.org) on a genome-wide level in
the MARS GWAS sample. In addition, population stratification was tested
with EIGENSTRAT implemented in EIGENSOFT (Price et al., 2006) (http://
genepath.med.harvard.edu/?reich/EIGENSTRAT.htm). Neither the genomic
control method (l = 1.023, see Figure S1) nor EIGENSTRAT analysis gave
any indication for population stratification.
The LD pattern and haplotype block delineation were determined by applying
Haploview 4.0 (http://www.broad.mit.edu/mpg/haploview) (Barrett et al.,
2005). Blocks were defined using the confidence interval method described
by Gabriel et al. (Gabriel et al., 2002). Pairwise LD measures (r2and D’) were
calculated in the 366 healthy controls of the GWAS sample and in 284 controls
of the African-American sample for the eight most associated SNPs on
chr12.21.31 (see Figure 2). German controls were also compared to the
HapMap CEU population (CEPH sample consisting of Utah residents with
ancestry from northern and western Europe, n = 60, http://www.hapmap.org)
(Frazer et al., 2007). No deviation in LD could be observed in this comparison
(data not shown).
Genome-wide case-control analyses were conducted by applying the
WG-Permer software (http://www.mpipsykl.mpg.de/wg-permer/). For post-
hoc analyses, applications in R-2.5.0 (http://cran.r-project.org) and SPSS for
Windows (releases 16, SPSS, Chicago, IL, USA) were used. SNPs with geno-
type distributions deviating from HWE at a significance level of 10?5or 0.05
with a call rate below 98% or 95% in the GWAS or German replication sample,
respectively, and SNPs with a MAF below 5% were excluded from statistical
analysis. Autosomal SNPs were tested for association with unipolar depres-
sive disorder in a case-control design using Chi-square test statistics under
allelic and both alternative recessive-dominant modes of inheritance. The level
of significance was set to 5% (family-wise error rate). Nominal p values were
corrected for multiple comparisons by the permutation-based minimum
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 261
et al., 2001) under 104permutations over the three performed genetic models
and all SNPs tested per study. Empirical and nominal p values for all reported
associations did not deviate from each other. Moreover, sample demographic
statistics and post-hoc tests on age, gender, and German origin, life events,
recurrence of MD, age at onset, number of previous depressive episodes,
first-degree family history of MD, and lifetime attempted suicide status were
performed by logistic regression analysis and ANCOVA. P values including
these covariates did not differ from those of the Chi square test statistics for
all reported associations. Thus, none of these additional covariates showed
a significant effect on the reported associations. In the RADIANT study from
the UK, sex was coded as a factorial covariate for the analysis presented in
the main text. The validity of the p values and the distribution of the estimates
were verified using Monte-Carlo (permutation and bootstrap) methods. Below
we give the odds ratios (OR) without sex as a factorial covariate and the ORs in
a gender stratified analysis: OR of all RADIANT cases and RADIANT plus
WTCCC2 controls, sex not included as covariate: 1.082 (95% C.I. 0.951;
1.231), n = 1636 cases and 7261 controls with a p = 0.274. OR of only male
cases and male controls: 1.344 (95% C.I. 1.080; 1.672), n = 485 cases and
0.959 (95% C.I. 0.816; 1.127), n = 1151 cases, 3781 controls with a p = 0.615.
Meta-analyses were conducted using the R library rmeta applying a fixed
effect model. In the first meta-analysis, three genetic models were tested,
the two opposite carrier models and an allelic model resulting in a number of
2.02 effective tests as estimated from 10,000 permutations. In the second
meta-analysis (combining the results of the first meta-analysis with the data
from the RADIANT/WTCCC2 sample), only the recessive model for
rs1545843 was tested. The adjustment for the two tests performed in
RADIANT/WTCCC2 was done by adjusting the standard error of the estimate
Genotype-Specific mRNA Levels (eQTLs)
We used two independent genome-wide SNP/mRNA expression data sets for
SNP-eQTL analyses on 12q21.31. The first data set was from premortem
human hippocampus of 137 individuals involved in the Epilepsy Surgery
Program at Bonn University, Germany. Methods related to the hippocampal
eQTL experiment are detailed in the Supplemental Experimental Procedures.
viduals (http://www.sanger.ac.uk/humgen/genevar/) (Stranger et al., 2005,
Experimental Design and Statistical Analysis
In both data sets, we selected all RefSeq annotated genes (Pruitt et al., 2005)
located within 1.5 megabase on both sides of the genome-wide significant
intersect with the defined genomic region (hybridization probes in brackets,
see also Table S1): TMTC2 (GI_22749210-S), SLC6A15 (GI_33354280-A,
expression variable for each probe was built by regression analysis to correct
for ethnicity. We tested an allelic and both alternative recessive-dominant
genetic models for rs1545843 and rs1031681 for each of the probes (n = 7)
Subsequently, we repeated this analysis by including data of all available
non-RefSeq (EST) gene probes (n = +6: Hs.365699-S, Hs.506230-S,
GI_41149683-S, Hs.208111-S, GI_41149726-S, hmm21473-S, Table S1) for
ESTs from GenBank in the same genomic window. Data of four ESTs were
excluded from the analysis, because their probes did not map completely or
uniquely to any target EST sequence of the current GenBank database
sequences of all probes included in expression analyses mapped uniquely
and completely to the human genome and are all devoid of known common
variations denominated by dbSNP build 129.
Structural MRI Sample Acquisition and Quality Control
Structural MRI with high-resolution T1-weighted images adequate for
and 186 control subjects. MRI was acquired at the MPI of Psychiatry in the
context of the acquisition of the Munich recurrent unipolar depression replica-
tion samples. A detailed description of study participant selection and image
processing is available in the Supplemental Experimental Procedures.
Automated Regional Volumetry
In brief, image preprocessingwas performed as for voxel-based morphometry
to gain gray matter (GM) maps with preserved local volume in stereotactic
space. Histologically validated cytoarchitectonic probability maps (Amunts
et al., 2005) were used to create regional volumetry masks for the left and right
hippocampus and subregions cornu ammonis (CA: CA1–3), subiculum (SUB),
and dentate gyrus (DG). The sum of all modulated GM voxels within the
regional masks was calculated using in-house software programmed in IDL
Analysis of covariance (ANCOVA) was performed for left and right total hippo-
campal GM volume and each three subregions with two-level factors group
(patients, controls), genotype (rs1545843 AA versus AG/GG, equally for
rs1081681), and gender, covarying for age, squared age, total GM volume,
and sequence type. Levene’stests for equality of errorvariances wasexplored
and found nonsignificant for all tests (Figure S4). p values were compared
with a Bonferroni-corrected threshold to adjust for 18 tests (two SNPs, nine
volumetric measurements [including motor cortex as control region]: 0.05/
18 = 0.0028). Both nominal and corrected p values are indicated in Figure 5
and Table S2.
Methods for1H-NMR Spectroscopy
These are discussed in the Supplemental Experimental Procedures.
Male CD1 mice were used for all experiments. Animals were 28 days old at the
day of arrival and were kept on a 12L:12D cycle. Food and water was provided
ad libitum. The experiments were carried out in accordance with European
Communities Council Directive 86/609/EEC. All efforts were made to minimize
animal suffering during the experiments. The protocols were approved by the
committee for the Care and Use of Laboratory Animals of the Government of
Upper Bavaria, Germany.
Chronic Stress Paradigm
The chronic social stress procedure was performed as described previously
Experimental Procedures). In this paradigm, mice are exposed to a highly
unstable social and hierarchical situation during their adolescence and early
adulthood. After the 7 week stress procedure all animals were single housed
for 5 weeks and then sacrificed under basal conditions.
Tissue Dissection and Expression Profiling
Frozen brains were sectioned at the level of the dorsal hippocampus and the
subregions CA1 and dentate gyrus were laser-microdissected using a laser
capture microscope (P.A.L.M. Microlaser Technologies, Bernried, Germany).
Extracted RNA was quality checked on the Agilent 2100 Bioanalyser, sub-
jected to two rounds of linear amplification and hybridized to Illumina Mous-
eRef-8 v1.0 Expression BeadChips according to the manufacturer’s protocol
(see also Supplemental Experimental Procedures). The data discussed in
this publication have been deposited in NCBIs Gene Expression Omnibus
(GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO
Series accession number GSE112211.
Gene Expression Analysis in Stress-Susceptible
versus Stress-Resilient Mice
We chose the same procedure to select genes adjacent to the region of
association for validation in the described mouse experiment as we applied
in the human expression analysis. Expression differences were checked
Association of SLC6A15 with Major Depression
262 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
(XM_137221.4). Differentially expressed genes were validated by in situ
hybridization as described previously (Schmidt et al., 2007). The antisense
cRNA hybridization probe of SLC6A15 was 487 base pairs long (left primer:
TGCCGTGAGCTTTGTTTATG; right primer: CAGTGTTGGGGAACCACTTT
covering exons 11 to 13 of the gene). The slides were exposed to Kodak
Biomax MR films (Eastman Kodak Co., Rochester, NY) and developed.
Autoradiographs were digitized and relative expression was determined by
computer-assisted optical densitometry (Scion Image, Scion Corporation).
The software package SPSS version 16 was used for statistical analysis.
Groupcomparisons wereperformed usingthetwo-tailedpaired t testtodeter-
mine statistical significance (*p < 0.05; **p < 0.01; ***p < 0.001). Data are pre-
sented as mean ± SEM.
The mouse model data discussed in this publication have been deposited in
NCBI’s Gene Expression Omnibus and are accessible through GEO Series
accession number GSE112211.
Supplemental Information includes five figures, three tables, Supplemental
Experimental Procedures, and detailed author contributions and can be found
with this article online at doi:10.1016/j.neuron.2011.04.005.
Thisworkhasbeenfunded bytheExcellence FoundationfortheAdvancement
of the Max Planck Society, the Bavarian Ministry of Commerce, and
the Federal Ministry of Education and Research (BMBF) in the framework of
the National Genome Research Network (NGFN2 and NGFN-Plus, FKZ
01GS0481 and 01GS08145 (Moods)). The Dutch studies are supported by
the Netherlands Organization of Scientific Research (NWO Investments
#175.010.2005.011, 911- 03-012), the Netherlands Genomics Initiative (NGI)/
Netherlands Organization for Scientific Research (NWO project #050-060-
810), the Hersenstichting, and the Centre for Medical Systems Biology
(CMSB). The Atlanta cohort was sponsored by RO1 MH071537-01A1. The
RADIANT study was supported by the UK MRC (G0701420). This study makes
use of data generated by the Wellcome Trust Case-Control Consortium 2 (for
author contributions see http://www.wtccc.org.uk). Funding for the project
was provided by the Wellcome Trust under award 085475. This study was
also supported by NeuroNova (a nonprofit Company for advancement of
Genomics). The authors would like to thank G. Ernst-Jansen, G. Gajewsky,
J. Huber, E. Kappelmann, S. Sauer, S. Damast, M. Koedel, M. Asmus, A.
Sangl, and H. Pfister for their excellent technical support. We further are grate-
ful to R. Hemauer, R. Borschke, and E. Schreiter for excellent MRI data aqui-
sition. We acknowledge the work of Yurii S. Aulchenko, A. Cecile, J.W. Jans-
receives grant support from NIMH, the Behrens-Weise Foundation, and Phar-
maNeuroBoost. F.H. is founder and shareholder of Affectis Pharmaceuticals
and HolsboerMaschmeyer NeuroChemie GmbH. Over the past two years,
B.M.-M. has been a consultant for Affectis. C.M.v.D. discloses her affiliation
to the Centre for Medical Systems Biology (CMSB). Within the last 3 years,
K.J.R. has received research funding support from NIMH, NIDA, Lundbeck,
Burroughs Wellcome Foundation, and NARSAD, and he has an unrelated
agreement with Extinction Pharmaceuticals for NMDA-based therapeutics.
Patent applications: A.M., E.B.B., and F.H. are inventors of means and
methods for diagnosing predisposition for treatment emergent suicidal idea-
tion (TESI), international application number PCT/EP2009/061575. E.B.B.,
F.H., B.M.-M., and M.U. are inventors of (1) FKBP5, a novel target for antide-
pressant therapy, international publication number WO 2005/054500; and (2)
polymorphisms in ABCB1 associated with a lack of clinical response to medi-
caments, international application number PCT/EP2005/005194.
Accepted: March 18, 2011
Published: April 27, 2011
Allen Brain Atlas. (2008). (Seattle, WA: Allen Institute for Brain Science).
Available from http://www.brain-map.org.
Alexopoulos, G.S. (2006). The vascular depression hypothesis: 10 years later.
Biol. Psychiatry 60, 1304–1305.
Amunts, K., Kedo, O., Kindler, M., Pieperhoff, P., Mohlberg, H., Shah, N.J.,
Habel, U., Schneider, F., and Zilles, K. (2005). Cytoarchitectonic mapping of
the human amygdala, hippocampal region and entorhinal cortex: Intersubject
variability and probability maps. Anat. Embryol. (Berl.) 210, 343–352.
Aulchenko, Y.S., Heutink, P., Mackay, I., Bertoli-Avella, A.M., Pullen, J.,
Vaessen, N., Rademaker, T.A., Sandkuijl, L.A., Cardon, L., Oostra, B., and
van Duijn, C.M. (2004). Linkage disequilibrium in young genetically isolated
Dutch population. Eur. J. Hum. Genet. 12, 527–534.
Barrett, J.C., Fry, B., Maller, J., and Daly, M.J. (2005). Haploview: Analysis and
visualization of LD and haplotype maps. Bioinformatics 21, 263–265.
Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., and Erbaugh, J. (1961).
An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561–571.
Benarroch, E.E. (2008). N-acetylaspartate and N-acetylaspartylglutamate:
Neurobiology and clinical significance. Neurology 70, 1353–1357.
Binder, E.B., Bradley, R.G., Liu, W., Epstein, M.P., Deveau, T.C., Mercer, K.B.,
Tang, Y., Gillespie, C.F., Heim, C.M., Nemeroff, C.B., et al. (2008). Association
of FKBP5 polymorphisms and childhood abuse with risk of posttraumatic
stress disorder symptoms in adults. JAMA 299, 1291–1305.
Birney, E., Stamatoyannopoulos, J.A., Dutta, A., Guigo ´, R., Gingeras, T.R.,
Margulies, E.H., Weng, Z.P., Snyder, M., Dermitzakis, E.T., Thurman, R.E.,
et al; ENCODE Project Consortium; NISC Comparative Sequencing
Program; Baylor College of Medicine Human Genome Sequencing Center;
Washington University GenomeSequencing
Children’s Hospital Oakland ResearchInstitute.(2007). Identification and anal-
ysis of functional elements in 1% of the human genome by the ENCODE pilot
project. Nature 447, 799–816.
Bostwick, J.M., and Pankratz, V.S. (2000). Affective disorders and suicide risk:
A reexamination. Am. J. Psychiatry 157, 1925–1932.
Bro ¨er, S. (2006). The SLC6 orphans are forming a family of amino acid trans-
porters. Neurochem. Int. 48, 559–567.
Bro ¨er, A., Tietze, N., Kowalczuk, S., Chubb, S., Munzinger, M., Bak, L.K., and
Bro ¨er, S. (2006). The orphan transporter v7-3 (slc6a15) is a Na+-dependent
neutral amino acid transporter (B0AT2). Biochem. J. 393, 421–430.
Browning, B.L., and Browning, S.R. (2009). A unified approach to genotype
lated individuals. Am. J. Hum. Genet. 84, 210–223.
Choy, W.C., Lo ´pez-Leo ´n, S., Aulchenko, Y.S., Mackenbach, J.P., Oostra,
B.A., van Duijn, C.M., and Janssens, A.C. (2009). Role of shared genetic and
environmental factors in symptoms of depression and body composition.
Psychiatr. Genet. 19, 32–38.
Cze ´h, B., Michaelis, T., Watanabe, T., Frahm, J., de Biurrun, G., van Kampen,
M., Bartolomucci, A., and Fuchs, E. (2001). Stress-induced changes in cere-
bral metabolites, hippocampal volume, and cell proliferation are prevented
by antidepressant treatment with tianeptine. Proc. Natl. Acad. Sci. USA 98,
Devlin, B., Roeder, K., and Wasserman, L. (2001). Genomic control, a new
Drgonova, J., Liu, Q.R., Hall, F.S., Krieger, R.M., and Uhl, G.R. (2007). Deletion
of v7-3 (SLC6A15) transporter allows assessment of its roles in synaptosomal
proline uptake, leucine uptake and behaviors. Brain Res. 1183, 10–20.
Dudbridge, F., and Gusnanto, A. (2008). Estimation of significance thresholds
for genomewide association scans. Genet. Epidemiol. 32, 227–234.
Easton, D.F., Pooley, K.A., Dunning, A.M., Pharoah, P.D., Thompson, D.,
Ballinger, D.G., Struewing, J.P., Morrison, J., Field, H., Luben, R., et al;
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 263
SEARCH collaborators; kConFab; AOCS Management Group. (2007).
Genome-wide association study identifies novel breast cancer susceptibility
loci. Nature 447, 1087–1093.
Ebmeier, K.P., Donaghey, C., and Steele, J.D. (2006). Recent developments
and current controversies in depression. Lancet 367, 153–167.
Farmer, M.K., Robbins, M.J., Medhurst, A.D., Campbell, D.A., Ellington, K.,
Duckworth, M., Brown, A.M., Middlemiss, D.N., Price, G.W., and Pangalos,
M.N. (2000). Cloning and characterization of human NTT5 and v7-3: Two
orphan transporters of the Na+/Cl- -dependent neurotransmitter transporter
gene family. Genomics 70, 241–252.
Fava, M., and Rush, A.J. (2006). Current status of augmentation and combi-
nation treatments for major depressive disorder: A literature review and a
proposal for a novel approach to improve practice. Psychother. Psychosom.
Frazer, K.A., Ballinger, D.G., Cox, D.R., Hinds, D.A., Stuve, L.L., Gibbs, R.A.,
Belmont, J.W., Boudreau, A., Hardenbol, P., Leal, S.M., et al; International
HapMap Consortium. (2007). A second generation human haplotype map of
over 3.1 million SNPs. Nature 449, 851–861.
Frodl, T., Meisenzahl, E.M., Zetzsche, T., Born, C., Groll, C., Ja ¨ger, M.,
Leinsinger, G., Bottlender,R., Hahn,
Hippocampal changes in patients with a first episode of major depression.
Am. J. Psychiatry 159, 1112–1118.
K.,and Mo ¨ller,H.J. (2002).
Gabriel, S.B., Schaffner, S.F., Nguyen, H., Moore, J.M., Roy, J., Blumenstiel,
B., Higgins, J., DeFelice, M., Lochner, A., Faggart, M., et al. (2002). The struc-
ture of haplotype blocks in the human genome. Science 296, 2225–2229.
Glahn, D.C., Thompson, P.M., and Blangero, J. (2007). Neuroimaging endo-
phenotypes: Strategies for finding genes influencing brain structure and
function. Hum. Brain Mapp. 28, 488–501.
Heck, A., Lieb, R., Ellgas, A., Pfister, H., Lucae, S., Erhardt, A., Himmerich, H.,
Horstmann, S., Kloiber, S., Ripke, S., et al. (2009). Polymorphisms in the
angiotensin-converting enzyme gene region predict coping styles in healthy
adults and depressed patients. Am. J. Med. Genet. B. Neuropsychiatr.
Genet. 150B, 104–114.
Heinzen, E.L., Ge, D., Cronin, K.D., Maia, J.M., Shianna, K.V., Gabriel, W.N.,
Welsh-Bohmer, K.A., Hulette, C.M., Denny, T.N., and Goldstein, D.B. (2008).
Tissue-specific genetic control of splicing: Implications for the study of
complex traits. PLoS Biol. 6, e1.
Hennings, J.M., Owashi, T., Binder, E.B., Horstmann, S., Menke, A., Kloiber,
S., Dose, T., Wollweber, B., Spieler, D., Messer, T., et al. (2009). Clinical char-
acteristics and treatment outcome in a representative sample of depressed
inpatients - findings from the Munich Antidepressant Response Signature
(MARS) project. J. Psychiatr. Res. 43, 215–229.
Herman, J.P., Figueiredo, H., Mueller, N.K., Ulrich-Lai, Y., Ostrander, M.M.,
Choi, D.C., and Cullinan, W.E. (2003). Central mechanisms of stress integra-
tion: Hierarchical circuitry controlling hypothalamo-pituitary-adrenocortical
responsiveness. Front. Neuroendocrinol. 24, 151–180.
Hofman, A., Breteler, M.M., van Duijn, C.M., Krestin, G.P., Pols, H.A., Stricker,
B.H., Tiemeier, H., Uitterlinden, A.G., Vingerling, J.R., and Witteman, J.C.
(2007). The Rotterdam Study: Objectives and design update. Eur. J.
Epidemiol. 22, 819–829.
Ising,M., Lucae, S.,Binder, E.B., Bettecken, T., Uhr, M., Ripke, S., Kohli, M.A.,
Hennings, J.M., Horstmann, S., Kloiber, S., et al. (2009). A genomewide asso-
ciation study points to multiple loci that predict antidepressant drug treatment
outcome in depression. Arch. Gen. Psychiatry 66, 966–975.
Kendler, K.S., Gardner, C.O., and Prescott, C.A. (2002). Toward a comprehen-
sive developmental model for major depression in women. Am. J. Psychiatry
Kessler, R.C., Berglund, P., Demler, O., Jin, R., Koretz, D., Merikangas, K.R.,
Rush, A.J., Walters, E.E., and Wang, P.S.; National Comorbidity Survey
Replication. (2003). The epidemiology of major depressive disorder: Results
from the National Comorbidity Survey Replication (NCS-R). JAMA 289,
Kleinjan, D.A., and van Heyningen, V. (2005). Long-range control of gene
expression: Emerging mechanisms and disruption in disease. Am. J. Hum.
Genet. 76, 8–32.
Levinson, D.F. (2006). The genetics of depression: A review. Biol. Psychiatry
Lewis, C.M., Ng, M.Y., Butler, A.W., Cohen-Woods, S., Uher, R., Pirlo, K.,
Weale, M.E., Schosser, A., Paredes, U.M., Rivera, M., et al. (2010). Genome-
wide association study of major recurrent depression in the U.K. population.
Am. J. Psychiatry 167, 949–957.
Li, C.X., Wang, Y., Gao, H., Pan, W.J., Xiang, Y., Huang, M., and Lei, H. (2008).
Cerebral metabolic changes in a depression-like rat model of chronic forced
copy. Neurochem. Res. 33, 2342–2349.
Lucae,S.,Salyakina,D.,Barden,N., Harvey,M.,Gagne ´,B.,Labbe ´,M.,Binder,
E.B., Uhr, M., Paez-Pereda, M., Sillaber, I., et al. (2006). P2RX7, a gene coding
for a purinergic ligand-gated ion channel, is associated with major depressive
disorder. Hum. Mol. Genet. 15, 2438–2445.
F., Hofman, A., Stricker, B.H., and Tiemeier, H. (2008). Incidence and recur-
rence of late-life depression. Arch. Gen. Psychiatry 65, 1394–1401.
Masson, J., Pohl, M., Aı ¨douni, Z., Giros, B., Hamon, M., and el Mestikawy, S.
(1996). The two orphan Na+/Cl(-)-dependent transporters Rxt1 and V-7-3-2
have an overlapping expression pattern in the rat central nervous system.
Receptors Channels 4, 227–242.
Muglia, P., Tozzi, F., Galwey, N.W., Francks, C., Upmanyu, R., Kong, X.Q.,
Antoniades, A., Domenici, E., Perry, J., Rothen, S., et al. (2010). Genome-
wide association study of recurrent major depressive disorder in two
European case-control cohorts. Mol. Psychiatry 15, 589–601.
Myers, S., Bottolo, L., Freeman, C., McVean, G., and Donnelly, P. (2005). A
fine-scale map of recombination rates and hotspots across the human
genome. Science 310, 321–324.
Myers, A.J., Gibbs, J.R., Webster, J.A., Rohrer, K., Zhao, A., Marlowe, L.,
Kaleem, M., Leung, D., Bryden, L., Nath, P., et al. (2007). A survey of genetic
human cortical gene expression. Nat. Genet. 39, 1494–1499.
Price, A.L., Patterson, N.J., Plenge, R.M., Weinblatt, M.E., Shadick, N.A., and
Reich, D. (2006). Principal components analysis corrects for stratification in
genome-wide association studies. Nat. Genet. 38, 904–909.
Pruitt, K.D., Tatusova, T., and Maglott, D.R. (2005). NCBI Reference Sequence
(RefSeq): A curated non-redundant sequence database of genomes, tran-
scripts and proteins. Nucleic Acids Res. 33, 501–504.
Purcell, S., Cherny, S.S., and Sham, P.C. (2003). Genetic Power Calculator:
Design of linkage and association genetic mapping studies of complex traits.
Bioinformatics 19, 149–150.
Radloff, L.S. (1977). The CES-D Scale: A self-report depression scale for
research in the general population. Appl. Psychol. Meas. 3, 385.
Rietschel, M., Mattheisen, M., Frank, J., Treutlein, J., Degenhardt, F., Breuer,
R., Steffens, M., Mier, D., Esslinger, C., Walter, H., et al. (2010). Genome-wide
association-, replication-, and neuroimaging study implicates HOMER1 in the
etiology of major depression. Biol. Psychiatry 68, 578–585.
Schmidt, M.V., Sterlemann, V., Ganea, K., Liebl, C., Alam, S., Harbich, D.,
Greetfeld, M., Uhr, M., Holsboer, F., and Mu ¨ller, M.B. (2007). Persistent neuro-
endocrine and behavioral effects of a novel, etiologically relevant mouse para-
Schmidt, M.V., Scharf, S.H., Sterlemann, V., Ganea, K., Liebl, C., Holsboer,
F., and Mu ¨ller, M.B. (2010). High susceptibility to chronic social stress is
associated with a depression-like phenotype. Psychoneuroendocrinology
Shi, J., Potash, J.B., Knowles, J.A., Weissman, M.M., Coryell, W., Scheftner,
W.A., Lawson, W.B., DePaulo, J.R., Jr., Gejman, P.V., Sanders, A.R., et al.
sive disorder. Mol. Psychiatry 16, 193–201.
Association of SLC6A15 with Major Depression
264 Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc.
Spinhoven, P., Ormel, J., Sloekers, P.P., Kempen, G.I., Speckens, A.E., and Download full-text
Van Hemert, A.M. (1997). A validation study of the Hospital Anxiety and
Depression Scale (HADS) in different groups of Dutch subjects. Psychol.
Med. 27, 363–370.
Sterlemann,V.,Ganea,K.,Liebl,C.,Harbich,D.,Alam,S.,Holsboer, F.,Mu ¨ller,
M.B., and Schmidt, M.V. (2008). Long-term behavioral and neuroendocrine
alterations following chronic social stress in mice: Implications for stress-
related disorders. Horm. Behav. 53, 386–394.
Stranger, B.E., Forrest, M.S., Clark, A.G., Minichiello, M.J., Deutsch, S., Lyle,
R.,Hunt,S., Kahl, B., Antonarakis,S.E., Tavare ´,S., et al. (2005). Genome-wide
associations of gene expression variation in humans. PLoS Genet. 1, e78.
Stranger, B.E., Nica, A.C., Forrest, M.S., Dimas, A., Bird, C.P., Beazley, C.,
Ingle, C.E., Dunning, M., Flicek, P., Koller, D., et al. (2007). Population geno-
mics of human gene expression. Nat. Genet. 39, 1217–1224.
Sullivan, E.V., Pfefferbaum, A., Swan, G.E., and Carmelli, D. (2001). Heritability
of hippocampal size in elderly twin men: Equivalent influence from genes and
environment. Hippocampus 11, 754–762.
Sullivan, P.F., de Geus, E.J.,Willemsen, G., James, M.R., Smit, J.H., Zandbelt,
T., Arolt, V., Baune, B.T., Blackwood, D., Cichon, S., et al. (2009). Genome-
wide association for major depressive disorder: A possible role for the presyn-
aptic protein piccolo. Mol. Psychiatry 14, 359–375.
Tapiero, H., Mathe ´, G., Couvreur, P., and Tew, K.D. (2002). II. Glutamine and
glutamate. Biomed. Pharmacother. 56, 446–457.
UniGene. (2009). http://www.ncbi.nlm.nih.gov/sites/entrez?db=unigene.
Ustu ¨n, T.B., Ayuso-Mateos, J.L., Chatterji, S., Mathers, C., and Murray, C.J.L.
(2004). Global burden of depressive disorders in the year 2000. Br. J.
Psychiatry 184, 386–392.
Videbech, P., and Ravnkilde, B. (2004). Hippocampal volume and depression:
A meta-analysis of MRI studies. Am. J. Psychiatry 161, 1957–1966.
Viinama ¨ki, H., Tanskanen, A., Honkalampi, K., Koivumaa-Honkanen, H.,
Haatainen, K., Kaustio, O., and Hintikka, J. (2004). Is the Beck Depression
Inventory suitable for screening major depression in different phases of the
disease? Nord. J. Psychiatry 58, 49–53.
Wang, J.L. (2005). Work stress as a risk factor for major depressive episode(s).
Psychol. Med. 35, 865–871.
Weissman, M.M., Sholomskas, D., Pottenger, M., Prusoff, B.A., and Locke,
B.Z. (1977). Assessing depressive symptoms in five psychiatric populations:
A validation study. Am. J. Epidemiol. 106, 203–214.
Westfall, P.H., and Young, S.S. (1993). Resampling-Based Multiple Testing
(New York: John Wiley & Sons, Inc.).
Westfall, P.H., Zaykin, D.V., and Young, S.S. (2001). Multiple Tests for
Genetics Effects in Association Studies: Biostatistical Methods (Totawa, NJ:
Wray, N.R., Pergadia, M.L., Blackwood, D.H., Penninx, B.W., Gordon, S.D.,
Nyholt, D.R., Ripke, S., Macintyre, D.J., McGhee, K.A., Maclean, A.W., et al.
(2010). Genome-wide association study of major depressive disorder: new
results, meta-analysis, and lessons learned. Mol. Psychiatry, in press.
Published online November 2, 2010. 10.1038/mp.2010.109.
Zhou, Z., Zhen, J., Karpowich, N.K., Goetz, R.M., Law, C.J., Reith, M.E.A., and
Wang, D.N. (2007). LeuT-desipramine structure reveals how antidepressants
block neurotransmitter reuptake. Science 317, 1390–1393.
Zigmond, A.S., and Snaith, R.P. (1983). The hospital anxiety and depression
scale. Acta Psychiatr. Scand. 67, 361–370.
Association of SLC6A15 with Major Depression
Neuron 70, 252–265, April 28, 2011 ª2011 Elsevier Inc. 265