RESEARCH ARTICLE

Meta-Analysis and Association of Two Common Polymorphisms of

the Human Oxytocin Receptor Gene in Autism Spectrum Disorder
Thorsten M. Kranz, Marnie Kopp, Regina Waltes, Michael Sachse, Eftichia Duketis, Tomasz A. Jarczok,
rgen, Jobst Meyer, Christine M. Freitag, and Andreas G. Chiocchetti
Franziska Degenhardt, Katharina Go
Neuropeptides such as oxytocin (OXT) are known facilitators of social behavior across species. Variants of the OXT
receptor gene (OXTR) have been tested for association with autism spectrum disorder (ASD) across manifold ethnicities, yielding both positive and negative findings. A recent meta-analysis, comprising 16 single nucleotide polymorphisms (SNPs), has corroborated the implication of OXTR in the etiology of ASD. Here, we genotyped and tested two
additional variants (rs237889 and rs237897) for association with ASD in two German predominantly high-functioning
ASD samples. We found nominal over-transmission (OR 5 1.48, CI95 5 1.06-2.08, P 5 0.022) for the minor A allele of
variant rs237889G>A in sample 1 (N 5 135 complete parent-offspring trios, 29 parent child duos), but not in sample 2
(362 trios, 69 duos). Still, in a meta-analysis comprising four different studies including the two unreported German
data sets (N 5 542 families), this finding was confirmed (OR 5 1.12; CI95 5 1.011.24, random effects P 5 0.012). In
addition, carriers of the minor risk allele rs237889-A showed significantly increased severity scores, as assessed through
the autism diagnostic interview revised (ADI-R), with highly significant increases in social interaction deficits. Our
results corroborate the implication of common OXTR variants in the etiology of ASD. There is a need for functional
studies to delineate the neurobiological implications of this and other association findings. (172/250). Autism Res
C 2016 International Society for Autism Research, Wiley Periodicals, Inc.
2016, 00: 000000. V
Keywords: meta-analysis; autism spectrum disorder; oxytocin receptor; genotyping; social interaction; endophenotype; genetics; oxytocin

Introduction
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with an estimated heritability of 80%
[Ronald & Hoekstra, 2011]. ASD is marked by impairment of abilities in social interaction and communication as well as by stereotypical behavior (DSM-5). The
oxytocin receptor gene (OXTR) has been considered as
risk gene for ASD due to its known function in the
facilitation of social bonding and prosocial behavior
[Kumsta & Heinrichs, 2013]. Modahl and colleagues
[1998] first reported that ASD-affected children showed
reduced oxytocin (OXT) levels in plasma in comparison
to healthy controls [Modahl et al., 1998]. The
rs2254298-A allele of OXTR was associated with sociability and increased amygdala volume [Brune, 2012].
In addition, the amygdala was observed to be enlarged
in individuals with ASD [Mosconi et al., 2009]. Interestingly, similar volumetric aberrations were observed in

anxiety disorders, a frequent comorbidity among ASD

cases [Kleinhans et al., 2011]. Furthermore, rs2254298
was found to be a predictor for daughters depression
and anxiety, when mothers had a history of recurrent
major depressive disorder [Thompson et al., 2011]. At
last, also animal studies on the OXT-OXTR system have
corroborated its modulatory qualities on social behavior
[Ferguson et al., 2000; Pobbe et al., 2012].
However, genetic association studies on OXTR and
ASD in human cohorts have yielded ambiguous results:
First, none of the six large genome-wide association
studies identified any variants near or within OXTR as
risk factor for ASD [Poelmans et al., 2013]. Second positive association of or linkage with OXTR variants has
been reported in candidate gene studies in Caucasian
individuals [Campbell et al., 2011; Jacob et al., 2007;
Lerer et al., 2008; Wermter et al., 2010; Yrigollen et al.,
2008]. The largest study investigated a cohort of over
1,200 families with ASD-affected individuals and

From the Department of Neurobehavioral Genetics, University of Trier, Johanniterufer 15, Trier D-54290, Germany (T.M.K., K.G., J.M.,); Department
of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, JW Goethe University Frankfurt, Deutschordenstrae 50, Frankfurt am
Main D-60528, Germany (M.K., R.W., M.S., E.D., T.A.J., C.M.F., A.G.C.,); Institute of Human Genetics, University of Bonn, Sigmund-Freud-Str. 25,
Bonn D-53127, Germany (F.D.)
Received July 24, 2015; accepted for publication December 03, 2015
Address for correspondence and reprints: Andreas G. Chiocchetti, PhD, Department of Child and Adolescent Psychiatry, Psychosomatics and
Psychotherapy, Johann Wolfgang Goethe University, Deutschordenstrae 50, Frankfurt am Main D-60528, Germany. E-mail: andreas.chiocchetti@
kgu.de
Published online 00 Month 2016 in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/aur.1597
C 2016 International Society for Autism Research, Wiley Periodicals, Inc.
V

INSAR

Autism Research 00: 0000, 2016

Table 1.

ASD Categories of Affected Individuals Included in this Study

Early childhood autism (N 5 567)

Atypical autism (N 5 18)
Asperger Syndrome (N 5 72)
High functioning ASD (IQ  70)
Low functioning ASD (IQ <70)
IQ unknown
Male N(%)
Female N (%)

Homburg
N 5 178

Frankfurt
N 5 489

158 (88%)
3 (2%)
17 (10%)

419 (86%)
15 (3%)
55 (11%)

0.521

118 (72.8%)
44 (27.2%)
16

335 (75.5%)
109 (24.5%)
45

0.583

158 (89%)
20 (11)%

389 (80%)
91 (21%)

0.142

Age [years] at diagnosis (Mean SD)

IQ (Mean SD)

10.1 (6.8)
86 (24)

11.0 (5.1)
90 (25)

ADI score social interaction (Mean SD)

ADI score communication (Mean SD)
ADI Score Repetitive Behavior (Mean SD)

17.98 (5.42)
13.86 (4.75)
5.51 (2.66)

20.38 (5.59)
14.72 (4.52)
5.25 (2.58)

0.622
0.112
<0.001
0.045
0.377

N, Numbers; SD, Standard deviation; P, P-value. significant P-values are marked in bold.

reported nominal association of ASD with variants

rs2268493 (P 5 0.043), rs1042778 (P 5 0.037) and
rs7632287 (P 5 0.016) [Campbell et al., 2011]. However,
negative results were also reported [Liu et al., 2010; Wu
et al., 2005].
Third, observed effects differ between ethnicities: In a
Han Chinese and Japanese cohort, respectively, the
major rs2254298-A allele was identified as risk allele
[Liu et al., 2010; Wu et al., 2005], whereas in a Caucasian trio cohort, it was the minor rs2254298-G allele
[Jacob et al., 2007].
Finally, the largest meta-analysis on OXTR and ASD to
date published [LoParo & Waldman, 2015] investigated
16 single nucleotide polymorphisms (SNPs) of OXTR
from 11 independent samples comprising a total of
3,941 individuals and reported an association of four
OXTR SNPs, namely rs7632287, rs237887, rs2268491,
and rs2254298, with ASD [LoParo & Waldman, 2015].
However, this study only included one of the four
SNPs of the haplotype identified by Lerer et al. due to
insufficient number of additional studies performed.
Lerer and colleagues have analyzed 18 tagging SNPs
across the OXTR gene region and identified a fivelocus haplotype block (rs237897-rs13316193-rs237889rs2254298-rs2268494) within the third intron of the
OXTR gene to be highly significantly associated with
ASD and intelligence quotient (IQ) [Lerer et al., 2008].
Four SNPs have not yet been analyzed in any metaanalysis
(rs237897,
rs13316193,
rs237889,
and
rs2268494) but have been previously published in two
or three ASD association studies [Campbell et al., 2011;
Lerer et al., 2008; Tansey et al., 2010; Wermter et al.,
2010]. Here, we aim at testing these SNPs for association with ASD in a meta-analysis by including two

independent German cohorts and all published dataset.

Two of the four SNPs (rs2268494 and rs13316193) are
highly correlated to rs237889 with an r2 > 0.79. Thus,
we selected rs237897 and rs237889 as tagging SNPs for
our analysis. Genotyping two German cohorts and
retrieving effect sizes from published studies allowed us
to perform a random effects meta-analysis across four
independent cohorts for each single SNP. Finally, to
test the hypothesis that variants of the OXTR specifically modulate social interaction we tested both variants
for
association
with
social
interaction,
communication and restrictive, repetitive behavior
severity scores assessed by the Autism Diagnostic Interview Revised [Lord et al., 1994].

Material and Methods

Sample Description, Diagnostic criteria, and Genomic DNA
collection
Parent offspring trios were recruited and diagnosed
(patients only) at the Department of Child and Adolescent Psychiatry, Saarland University Hospital, Homburg,
Germany, and at the Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy,
Johann Wolfgang von Goethe University, Frankfurt/
Main, Germany. All participants or their caregivers gave
written informed consent after oral and written explanation about the aim and scope of the investigation. DNA
was extracted from whole blood following published
protocols [Miller, et al., 1988]. The study was approved
by the respective ethical committees (decision 162/99
(Frankfurt), 73/04 (Homburg), 237/09 (Frankfurt)). All
included children and adolescents fulfilled the

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

INSAR

Table 2.

Primer and PCR Conditions used for OXTR Genotyping

SNPs

PCR Primers (53)

PCR Condition

rs237889

F: TTGGCGTGTGTGTATGTGTG
R: AGCAGAAACTGTGGGTGTCC

3 mM MgCl2
1 pmol primers
TA 5 67.58C

rs237897

F: CCAAGGGAGAGGTGAAGACA
R: CCTCCCCCTCAAACTTGAAT

1 mM MgCl2
7.5 pmol primers
TA 5 598C

Extension Primers
F: [poly-A]7TGATTTGCCGC
TTTCCACAAGTTCCT

rs237889
rs237897

F: [poly-A]25TGCAGAGAGGTGAG
TACTGCAAGGAG

TA, Annealing
nucleotides.

temperature;

[poly-A]n,

number

of

adenosine

diagnostic criteria for either Autism, Asperger Syndrome,

or atypical Autism according to ICD-10 [World Health
Organization, 1993], confirmed by the Autism Diagnostic Observation Schedule (ADOS) [Lord et al., 2000] and/
or the Autism Diagnostic Interview Revised (ADI-R)
[Lord et al., 1994] as published [Waltes et al., 2014]. In
this study, a total of 542 families were included. The
Homburg sample consisted of 135 complete German
parent-offspring trios, 29 duos, and 11 singletons including 175 affected individuals, which were all manually
genotyped.
The Frankfurt cohort consisted of 362 core families,
69 parent child duos, and 15 singletons with a total of
493 affected individuals (47 families had two affected
children). The Frankfurt samples were genotyped on an
Illumina Human Omni Express array. The two cohorts
did not differ with respect to frequency of diagnoses,
mean age at diagnosis, gender distribution and mean
IQ (Table 1).
Genotyping of Variants rs237889 and 237897
Manual genotyping of the Homburg sample was performed by Multiplexing PCR: First, two targeting regions
(PCR 1-2) spanning the two SNPs (PCR1: rs237889 PCR2:
rs237897) were amplified. All PCR reactions were performed in a total volume of 25 lL with 0.2 mM dNTPs,
1 U Taq Polymerase (Life Technologies, USA), 1 X reaction buffer and 100 ng of genomic DNA. Concentrations
of MgCl2 (1-3 mM) and amount of the respective primers (17.5 pmols) were adjusted if necessary. Standard
cycling conditions were applied with annealing temperatures varying between 5967.58C. Amplification was
confirmed by agarose gel separation. For details on primers and PCR conditions see Table 2. Second, 4 lL of
each amplicon were pooled and subjected to degradation of single-stranded DNA (primers) and dephosphorylation of dNTPs in a total volume of 16 lL using

INSAR

0.06 U/lL exonuclease I (ExoI) and 0.03 U/lL shrimp

alkaline phosphatase (SAP), respectively. The enzymatic
reaction was performed at 378C for 1 hr followed by heat
inactivation for 15 min at 808C. The quality of the ExoISAP reaction was checked by agarose gel separation.
Third, genotypes were determined via single base
extension using custom-designed interrogation primers
ending precisely one base prior to each SNP and the
ddNTP containing SNPStart kit (Beckman Coulter Inc.,
Krefeld, Germany) according to the manufacturers protocol. The cycling conditions were as follows: 27 cycles
at 948C for 10 sec and 608C for 20 sec. Prior to the SNP
detection, the remaining ddNTPs were dephosphorylated with 0.75U SAP at 378C for 45 min followed by
inactivation at 658C for 15 min. The incorporated fluorescent dye-labeled terminators (ddNTPs) were determined on a CEQ 8000 sequencer (Beckman Coulter,
Krefeld, Germany). The two SNPs were distinguished by
size based on the variable length of poly-A extensions
of the interrogation primers (Table 2). In total 474 samples were successfully genotyped. No Mendelian error
was identified using Haploview.
The Frankfurt sample was genotyped on an Illumina
HumanOmniExpress_12v1-H bead array at Life and
Brain (Bonn). SNP calling was performed using
GenomeStudio software and GenTrain-Algorithmus 2.0
with a GenCall-Threshold of 0.2. Only individuals with
call rates above 98% were included. Samples were
excluded if they were discordant between genetically
inferred gender and annotated gender, as well as doublets (identity by state IBS 2) and samples within families with IBS estimates >1.6. Population outliers were
not excluded here as we performed family based association tests that are not biased by population stratification effects. We quality checked SNP calling by
comparing calls to 15 manually genotyped variants
including previously published data [Waltes et al.,
2014] and achieved a concordance of 99.98%. Finally,
SNPs rs237889 and rs237897 were extracted from the
whole dataset and tested for Mendelian errors. A total
of 1,239 individuals out of 1,388 samples passed quality
control (QC) and were included in the Frankfurt cohort
here. For additional details on statistical analysis see
section below.
Statistical Analysis
Power analyses were performed using Quanto v 1.2.4
May, 2009 [Gauderman & Morrison, 2006]. The Homburg Sample had a power of 1-b 0.8 (P 5 0.05; twosided, log-additive model) to detect effect sizes of
OR 2.5 under the assumption of a MAF 5 0.05 and of
OR 2.1 with MAF 5 0.1. The Frankfurt sample was sufficiently powered (1-b 0.8; P 5 0.05; two-sided, logadditive model) to detect ORs 1.8 (MAF 5 0.05) and

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

Table 3.
Allele

TDT Analysis of OXTR Alleles

Trans

Untrans

% Geno

HWE P

OR minor

95%Lo

95%Hi

P-value

Homburg
rs237889
G
A

166
100

195
71

100

0.564

1.48

1.06

2.08

0.022

rs237897
C
T

165
99

187
77

99.40

0.684

1.43

1.01

2.02

0.042

Frankfurt
rs237889
G
A

411
267

459
251

98.80

0.212

1.13

0.92

1.38

0.245

rs237897
C
T

339
279

441
269

99.00

0.584

1.1

0.89

1.36

0.364

Combined
rs237889
G
A

579
369

658
322

99.10

0.201

1.22

1.03

1.46

0.0245

rs237897
C
T

567
379

632
346

99.10

0.389

1.18

0.99

1.41

0.0656

Trans, Transmitted; Untrans, Untransmitted; Geno, genotyped; HWE P, Hardy-Weinberg Equilibrium P-value; OR, Odds-ratio; 95Lo and 95Hi, lower
and higher boundary of 95 confidence interval.

ORs 1.6 (MAF 5 0.1). Primary genetic analyses on

genotyping efficiency, estimation of minor allele frequencies, linkage disequilibrium, Hardy-Weinberg equilibrium (HWE) and identification of Mendelian errors
were performed using Haploview v4.2 [Barrett, Fry, Maller, & Daly, 2005]. Families with inconclusive genotype
segregation patterns were excluded from analysis. Both
markers were within HWE in both samples (all P > 0.1).
Single marker and haplotype association analyses were
performed using transmission disequilibrium tests
(TDT) in UNPHASED [Dudbridge, 2008]. Descriptive statistics and meta-analysis on published odds ratios (OR)
were calculated in the R-program version 3.0.2. For
meta-analysis we used the rmeta package and applied
the meta.summary() function with a random
DerSimonian-Laird (DSL) model, as we observed a significant heterogeneity among studies. Only studies with
available odds ratios (OR) and 95% confidence intervals
(CI95) were included, and standard errors were calculated. Studies were tested for homogeneity of ORs using
Woolfs test.
Studies Included in Meta-Analysis
In total 10 studies [Campbell et al., 2011; de Krom et al.,
2009; Jacob et al., 2007; Kelemenova et al., 2010; Lerer
et al., 2008; Liu et al., 2010; K. E. Tansey et al., 2010;

Wermter et al., 2010; Wu et al., 2005; Yrigollen et al.,

2008] have investigated SNPs in the OXTR gene for association with ASD. SNP rs237889 was investigated in three
independent studies [Campbell et al., 2011; Lerer et al.,
2008; K. E. Tansey et al., 2010]. SNP rs237897 was studied
by three groups [Campbell et al., 2011; Lerer et al., 2008;
Wermter et al., 2010]. If effect sizes were not available or
imputable from the original publication (i.e., OR, CI95, b,
standard errors or transmission rates, respectively), the
corresponding authors were contacted and provided original data [Campbell et al., 2011]. This data was processed
using Haploview and UNPHASED as described above.
Meta-analyses were exclusively performed on the newly
analyzed data. Wermter et al. provided OR and 95CI (personal communication). No effect sizes could be retrieved
from Lerer and colleagues.

Results
We tested both SNPs rs2378889 and rs237897 for
transmission disequilibrium in the Homburg and the
Frankfurt sample. The two samples did not differ with
respect to diagnoses, Gender, IQ, and Age at Diagnosis. However, differences between ADI-R severity
scores were observed for social interaction and
communication.

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

INSAR

Table 4.

Haplotype Association Analysis

rs237889-rs237897

Trans

Untrans

OR

95%Lo

95%Hi

P-value

Homburg
Main P 5 0.0157
G-C
A-T
G-A
A-C

113
52
24
27

132
23
34
27

ref
2.01
0.8
1.06

ref
1.21
0.44
0.6

ref
3.33
1.45
1.87

0.107
0.002
0.155
0.948

Frankfurt
Main P 5 0.152
G-C
A-T
G-A
A-C

246
134
88
74

301
129
79
57

ref
1.16
1.35
1.41

ref
0.88
0.95
0.98

ref
1.52
1.91
2.04

0.041
0.636
0.263
0.155

Combined
Main P 5 0.051
G-C
A-T
G-A
A-C

361
187
112
102

437
152
113
84

ref
1.34
1.2
1.31

ref
1.06
0.89
0.97

ref
1.7
1.61
1.78

0.007
0.039
0.827
0.224

Trans, Transmittted; Untrans, Untransmitted; OR, Odds-ratio; 95%Lo and 95%Hi, Lower and higher boundary of 95% confidence interval.

Single marker analysis revealed a nominal significant

over transmission for both variants in the Homburg
sample for the minor allele rs237889-A (OR 5 1.48,
CI95 5 1.062.08; P 5 0.022) and the minor allele of
rs237897-T (OR 5 1.43; CI95 5 1.012.02, P50.042),
respectively (Table 3). However, these findings could
not be replicated in the Frankfurt sample (all variants
nominal P > 0.1).
In the combined sample the association of rs237889
again
reached
nominal
statistical
significance
(OR 5 1.22, CI95 5 1.031.46; P 5 0.025), which could
be potentially driven by Homburg sample alone.

Haplotype Analysis
A haplotype analysis was conducted to have a more
robust risk explanation for the association of both variants with ASD than just considering the single
markers, especially in case of a haplotype including
both risk alleles (Table 4). The haplotype comprising
both risk alleles was significantly over-transmitted
(OR 5 2.01; CI95 5 1.213.33, P 5 0.002; main effect
P 5 0.015) in the Homburg sample compared with the
major G-C haplotype. This finding was not replicated
in the Frankfurt sample (main effect P > 0.1) and

rs237889

rs237897

random effects meta-analysis

random effects meta-analysis

Homburg
Frankfurt
Tansey et al 2010
Campbell et al 2011

Homburg
Frankfurt
Wermter et al 2010
Campbell et al 2011

Summary

Summary

0.89 1.00 1.12

1.41

1.78

OR
summary OR(95CI): 1.12 ( 1.01 1.24 ); p.val: 0.0365
het pval: 0.364

0.63

0.79

1.00

1.26

1.58

2.00

OR
summary OR(95CI): 1.05 ( 0.88 1.25 ); p.val: 0.594
het pval: 0.100

Figure 1. Meta-analysis of two OXTR variants: SNP rs237889 was significantly associated with ASD using four studies. The dataset
from Tansey et al. [2010] are the effect sizes calculated using the combination of the three samples included in their study. Here
we used effect sizes of the combined samples only. For Campbell et al. [2011] only newly analyzed data was used. Significant heterogeneity was observed for both SNPs, thus meta-analysis was performed applying random effects models.

INSAR

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

Figure 2. Association of genotypes with phenotypic traits. (A) SNP rs237889 and (B) rs237897. Noncarriers are homozygous for
the major allele, whereas carriers have at least one minor allele. Dots correspond to individual data-points of affected cases; thick
bars depict sample median, whereas thin lines depict the interquartile range. Nominal p-values are shown. Wilcoxon test was performed comparing data points of ADI-R Algorithm severity scores, age at Diagnosis (Diag) and IQ. Chi-square test was performed
comparing gender distribution. Asterisk (*) marks p-values that survived correction for multiple testing.

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

INSAR

remained significant in the combined sample (A-T haplotype OR 5 1.34; CI95 5 1.061.70; nominal P 5 0.039).
However, the 95% confidence intervals overlapped with
the OR observed for the individual markers and thus
the A-T haplotype does not explain a higher risk propensity than an individual marker.
Meta-Analysis
A meta-analysis combining the results of our and previous association studies in ASD cohorts of SNPs rs237889
and rs237897 was conducted (Fig. 1). Given the significant heterogeneity between the samples (P < 0.05,
Woolfs test), a random effects DerSimonian-Laird (DSL)
meta-analysis was performed. SNP rs237889 but not
rs237897 is significantly associated with ASD with
OR 5 1.12 (CI95 5 1.011.24, DSL P 5 0.037).
Given the results of the meta-analysis and to evaluate
the negative finding in the Frankfurt cohort, we performed a post-hoc power analysis under the premises of
a minor allele frequency of MAF 5 0.32, a log additive
model and a two-sided a 5 0.05. The Frankfurt cohort
had a power of 1-b 5 0.37 and the combined German
cohort had a power of 1-b 5 0.44 to detect an effect size
of OR 5 1.2, respectively, and were thus underpowered.
Phenotypic Association Study
As OXT-OXTR interaction is related to social behavior
[Chen et al., 2011; Ebstein et al., 2009], the role of
the ASD-risk allele rs237889-A and the nonassociated
allele 237897-T on ASD severity regarding the three
dimensions social interaction, communication and
repetitive stereotypical behavior (Fig. 2) were additionally explored.
Distributions of gender, age, and IQ were similar
between carriers of the minor alleles of both variants,
respectively. Carriers of the risk rs273889-A allele (genotypes AA/AG) showed a nominally significantly (Wilcoxon P < 0.05) higher score on all three ADI-R
algorithm scores (Figure 2A). The largest mean difference was observed for the social interaction algorithm
score (Mean [SD] GG: 18.9 [5.6] AG/AA: 20.3 [5.6], Wilcoxon P 5 0.0048). This association was still significant
after correction for multiple testing (Bonferroni
P 5 0.0288). We did not observe any association of ADIR severity scores with rs237897 genotype (Figure 2B).

Discussion
In this study we found an association of rs237889-A
with ASD by meta-analysis of four different association
studies, including two newly genotyped German samples and two unrelated published cohorts [Campbell
et al., 2011; K. E. Tansey et al., 2010]. Interestingly, this
SNP is in the same linkage block as the variant

INSAR

rs2268491 reported by LoParo and colleagues with a

high D value (D >0.99) but with a low r2 (r2 5 0.08)
due to differing MAF [International HapMap et al.,
2010]. Our observation provides a rationale to include
rs237889 in further genetic analysis of the OXTR: The
low r2 between these two SNPs shows that one variant
cannot be tagged or substituted by the other although
being located in the same linkage block. This would
mean that the association reported here is independent
of the SNPs identified in the meta-analysis.
This study also has some limitations. First, the small
sample size of the two German cohorts provided limited statistical power to identify small odds ratios, and
did not allow us to investigate population specific
effects. However, the family based TDT itself is not
affected by population stratification, as the transmitted
alleles are independent from ethnicity. Second, the significant difference in the ADI-R severity score for social
interaction between the Frankfurt and the Homburg
sample may have biased the replication of the genetic
association with ASD. Third, the number of studies
included in the meta-analysis was still limited.
However, odds ratios for the rs237889-A allele were
above 1 in all studies, indicating the same direction of risk
effects across all four studies. In addition, rs237889 has
previously been reported to be part of a haplotype block
of five SNPs, significantly associated with ASD (P 5 0.009)
[Lerer et al., 2008]. Lerer et al. did not identify any phenotypic association with Vineland Adaptive Behavior Scale
and IQ. However, they did not test ADI-R derived scores.
In this study, we found a nominal significant association for rs237889-A with ADI-R algorithm scores. The
association with social interaction even was significant
after correction for six statistical tests. Therefore, we
suggest that rs237889 or linked OXTR variants may
have a general modulatory role on the ASD phenotypes
with a more pronounced effect on features needed for
social interaction.
The impact of OXTR and related pathway genes on
social functions is highly conserved across animal evolution [Feldman et al., 2015]. Further support is provided by studies on OXTR distribution in the CNS.
Brain regions like the limbic system, which is involved
in emotional processing and spatial orientation, reveal
high OXTR expression levels in animal models [Baribeau & Anagnostou, 2015]. Thus, polymorphisms of
OXTR related to reduced sociability in ASD could affect
protein expression levels of OXTR or binding affinity of
oxytocin and thereby lead to decreased activation of
specific neuronal networks within the limbic brain system. In addition, OXTR activates MAPK signaling pathway [Feldman et al., 2015] which is responsible for
neuronal survival and differentiation, and thus variants
of OXTR may impair these regulatory processes. One
study has reported increased methylation of the OXTR

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

promoter region in ASD cases resulting in decreased

OXTR expression [Gregory et al., 2009]. Thus, a reduced
OXTR pathway activation may underlie the etiology of
ASD and/or of aberrant social communication. The
finding that administrating intranasal oxytocin has
improved emotional recognition and social cognition
in ASD cases [Guastella et al., 2010; Hollander et al.,
2007] supports this hypothesis. Hence, oxytocin delivery to increase OXTR activation in the CNS seems a
promising therapeutic strategy to improve core deficits
of ASD.
Effects on social interaction have been previously
reported for other OXTR variants: In a study about
decision-making, Israel et al. [2008] investigated the
influence of 15 tagging SNPs of the OXTR gene on prosocial behavior. In 203 healthy individuals they showed
significant association of three variants (rs1042778
P 5 0.001, rs2268490 P 5 0.011, and rs237887 P 5 0.005)
and of 12 haplotypes (including two haplotypes spanning the here identified variant rs237889) with prosocial behavior [Israel et al., 2008].
Several functional studies support the modulatory
role of the OXT-OXTR system on social traits [Alves
et al., 2015]. In one study in virgin female rats, the
optogenetic activation of OXTR in neurons within
hypothalamic structures projecting to the central amygdala led to the release of oxytocin, which had a suppressing effect on contextual freezing in the previously
fear-conditioned rats [Knobloch et al., 2012]. Oxytocinpositive axons were identified to form projections from
hypothalamus and paraventricular nucleus to various
forebrain structures, especially the prefrontal cortex. As
it has been shown in humans that these brain regions
are involved in active decision making and formation
of personality, the findings of Knobloch et al. support
the importance of the OXT-OXTR system on social
behavior modulation [Yates, 2012]. In a further study,
the authors demonstrated that overexpression of OXTR
in the lateral septum results in increased anxiety in
mice that previously underwent social defeat. The
authors concluded that the OXTR has a modulatory
role on social cognitive tuning, depending on the location where the OXT-OXTR system gets activated in the
human brain [Guzman et al., 2013]. Furthermore, a
study on OXTR conditional knockout mice identified
its importance in the modification of social reward in
the mesocortico-limbic pathway and in the observed
concerted action of OXTR and the serotonin receptor
1B (5-HT1B) in the nucleus accumbens through oxytocinergic projections emerging from the dorsal raphe
nucleus [Dolen et al., 2013]. Electrophysiological experiments or immunohistochemical analysis of the OXTROXT system within these brain regions partially elucidated the pathophysiological mechanisms. Interestingly, mapping of the OXTR distribution in the brain of

rhesus macaques revealed that nonhuman primates

(NPH) and rodents differ from each other. Whereas the
highest OXTR expression levels in rodents are found in
brain regions processing social cues like the olfactory
bulb and the medial amygdala, rhesus macaques show
the highest OXTR levels in cholinergic brain regions
such as the nucleus basalis of Meynert (NBM), which
sends massive projections to the limbic brain system
and the neocortex [Freeman et al., 2014]. However, an
electrophysiological study in rat brain demonstrated a
contrasting mode of action of OXT: higher responses to
OXT were observed in the medial amygdala in comparison to the central amygdala, although the OXTR density
was significantly lower in the medial amygdala as proven by autoradiograms [Terenzi & Ingram, 2005]. Thus,
OXT-OXTR signaling is not exclusively dependent on
expression levels, but yet on the interaction with other
still undiscovered signaling pathways. These secondary
pathways may also be involved in the modulation of
social interaction. These secondary pathways such as
serotonin signaling [Dolen et al., 2013] may be involved
in the modulation of social interaction and ASD.
It seems thus likely that rs237889-A increases ASD risk
and severity mainly through a modulatory role in the
downstream activation of the OXTR pathway. The exact
mechanism and the functional role of this or other
SNPs in high linkage disequilibrium needs to be clarified
by additional studies. It is noteworthy that several
intronic variants of the OXTR gene have been associated
with ASD in the meta-analysis [LoParo & Waldman,
2015]. Thus, intronic OXTR regions need to be investigated for their regulatory role on OXTR function.
Taken together, we show that an intronic variant of
the OXTR gene has low but significant effects on ASD
risk and phenotypes. Thus, we provide further evidence
for OXTR as candidate risk gene in ASD. However, to
identify these or similar genetic variants in genomewide screens much larger samples will be needed. Additionally, using functional in vitro and in vivo assays
such as OXTR/OXT specific immunohistochemical or
electrophysiological experiments in associated brain
regions will help to gain a better understanding of
the complexity of the OXTOXTR neurotransmitter
system.

Funding
German Research Foundation (DFG) (ME 1923/5-1, ME
1632/5-3, and GRK 1389/1 to J.M.); Saarland University
(T6 03 10 00-45 to C.M.F); The European Commission
r Bildung und
and the German Bundesministerium fu
Forschung BMBF (ERA-NET NEURON project: EUHFAUTISM; EUHFAUTISM-01EW1105 to C.M.F).

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

INSAR

Acknowledgments
The authors thank the participating individuals for taking part in the study. All authors declare no conflict of
interest.

References
Alves, E., Fielder, A., Ghabriel, N., Sawyer, M., & BuismanPijlman, F.T. (2015). Early social environment affects the
endogenous oxytocin system: A review and future directions. Front Endocrinol (Lausanne), 6, 32. doi:10.3389/
fendo.2015.00032
Baribeau, D.A., & Anagnostou, E. (2015). Oxytocin and vasopressin: Linking pituitary neuropeptides and their receptors
to social neurocircuits. Frontiers in Neuroscience, 9, 335.
doi:10.3389/fnins.2015.00335
Barrett, J.C., Fry, B., Maller, J., & Daly, M.J. (2005). Haploview:
Analysis and visualization of LD and haplotype maps.
Bioinformatics, 21(2), 263265. doi:10.1093/bioinformatics/
bth457
Brune, M. (2012). Does the oxytocin receptor polymorphism
(rs2254298) confer vulnerability for psychopathology or
differential susceptibility? insights from evolution. BMC
Medicine, 10(1), 38. doi:10.1186/1741-7015-10-38
Campbell, D.B., Datta, D., Jones, S.T., Batey Lee, E., Sutcliffe,
J.S., Hammock, E.A., & Levitt, P. (2011). Association of oxytocin receptor (OXTR) gene variants with multiple phenotype domains of autism spectrum disorder. Journal of
Neurodevelopment Disorders, 3(2), 101112. doi:10.1007/
s11689-010-9071-2
Chen, F.S., Kumsta, R., Dawans, B.v., Monakhov, M., Ebstein,
R.P., & Heinrichs, M. (2011). Common oxytocin receptor
gene (OXTR) polymorphism and social support interact to
reduce stress in humans. Proceedings of the National Academy of Sciences of the United States of America, 108(50),
1993719942. doi:10.1073/pnas.1113079108
de Krom, M., Staal, W.G., Ophoff, R.A., Hendriks, J., Buitelaar,
J., Franke, B., . . . van Ree, J.M. (2009). A common variant in
DRD3 receptor is associated with autism spectrum disorder.
Biological Psychiatry, 65(7), 625630. doi:10.1016/
j.biopsych.2008.09.035
Dolen, G., Darvishzadeh, A., Huang, K.W., & Malenka, R.C.
(2013). Social reward requires coordinated activity of
nucleus accumbens oxytocin and serotonin. Nature,
501(7466), 179184. doi:10.1038/nature12518
Dudbridge, F. (2008). Likelihood-based association analysis for
nuclear families and unrelated subjects with missing genotype data. Human Heredity, 66(2), 8798. doi:10.1159/
000119108
Ebstein, R.P., Israel, S., Lerer, E., Uzefovsky, F., Shalev, I.,
Gritsenko, I., . . . Yirmiya, N. (2009). Arginine vasopressin
and oxytocin modulate human social behavior. Annals of
the New York Academy of Sciences, 1167, 87102.
Feldman, R., Monakhov, M., Pratt, M., & Ebstein, R.P. (2015).
Oxytocin pathway genes: Evolutionary ancient system
impacting on human affiliation, sociality, and psychopathology. Biological Psychiatry. 79(3), 174184 doi:10.1016/
j.biopsych.2015.08.008

INSAR

Ferguson, J.N., Young, L.J., Hearn, E.F., Matzuk, M.M., Insel,

T.R., & Winslow, J.T. (2000). Social amnesia in mice lacking
the oxytocin gene. Nature Genetics, 25(3), 284288. doi:
10.1038/77040
Freeman, S.M., Inoue, K., Smith, A.L., Goodman, M.M., &
Young, L.J. (2014). The neuroanatomical distribution of
oxytocin receptor binding and mRNA in the male rhesus
macaque (Macaca mulatta). Psychoneuroendocrinology, 45,
128141. doi:10.1016/j.psyneuen.2014.03.023
Gauderman, W., & Morrison, J. (2006). QUANTO 1.1: A computer program for power and sample size calculations for
genetic-epidemiology studies. Retrieved from http://biostats.usc.edu/software, last accessed Jan 4th 2016.
Gregory, S.G., Connelly, J.J., Towers, A., Johnson, J., Biscocho,
D., Markunas, C., . . . Pericak-Vance, M.A. (2009). Genomic
and epigenetic evidence for OXTR deficiency in autism.
BMC Medicine, 7, 62.
Guastella, A.J., Einfeld, S.L., Gray, K.M., Rinehart, N.J., Tonge,
B.J., Lambert, T.J., & Hickie, I.B. (2010). Intranasal oxytocin
improves emotion recognition for youth with autism spectrum disorders. Biological Psychiatry, 67(7), 692694. doi:
10.1016/j.biopsych.2009.09.020
Guzman, Y.F., Tronson, N.C., Jovasevic, V., Sato, K., Guedea,
A.L., Mizukami, H., . . . Radulovic, J. (2013). Fear-enhancing
effects of septal oxytocin receptors. Nature Neuroscience,
16(9), 11851187. doi:10.1038/nn.3465
Hollander, E., Bartz, J., Chaplin, W., Phillips, A., Sumner, J.,
Soorya, L., . . . Wasserman, S. (2007). Oxytocin increases
retention of social cognition in autism. Biological Psychiatry, 61(4), 498503.
International HapMap, C., Altshuler, D.M., Gibbs, R.A.,
Peltonen, L., Altshuler, D.M., Gibbs, R.A., . . . McEwen, J.E.
(2010). Integrating common and rare genetic variation in
diverse human populations. Nature, 467(7311), 5258. doi:
10.1038/nature09298
Israel, S., Lerer, E., Shalev, I., Uzefovsky, F., Reibold, M.,
Bachner-Melman, R., . . . Ebstein, R.P. (2008). Molecular
genetic studies of the arginine vasopressin 1a receptor
(AVPR1a) and the oxytocin receptor (OXTR) in human
behaviour: From autism to altruism with some notes in
between. Advances in Vasopressin and Oxytocin: From
Genes to Behaviour to Disease, 170, 435449. doi:Doi
10.1016/S0079-6123(08)00434-2
Jacob, S., Brune, C.W., Carter, C.S., Leventhal, B.L., Lord, C., &
Cook, E.H. Jr. (2007). Association of the oxytocin receptor
gene (OXTR) in Caucasian children and adolescents with
autism. Neuroscience Letters, 417(1), 69. doi:10.1016/
j.neulet.2007.02.001
Kelemenova, S., Schmidtova, E., Ficek, A., Celec, P., Kubranska,
A., & Ostatnikova, D. (2010). Polymorphisms of candidate
genes in Slovak autistic patients. Psychiatric Genetics,
20(4), 137139. doi:10.1097/YPG.0b013e32833a1eb3
Kleinhans, N.M., Richards, T., Johnson, L.C., Weaver, K.E.,
Greenson, J., Dawson, G., & Aylward, E. (2011). fMRI evidence of neural abnormalities in the subcortical face processing system in ASD. Neuroimage, 54(1), 697704. doi:
10.1016/j.neuroimage.2010.07.037
Knobloch, H.S., Charlet, A., Hoffmann L.C., Eliava, M.,
Khrulev, S., Cetin, A.H., . . . Grinevich, V. (2012). Evoked
axonal oxytocin release in the central amygdala attenuates

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

fear response. Neuron, 73(3), 553566. doi:http://dx.doi.

org/10.1016/j.neuron.2011.11.030
Kumsta, R., & Heinrichs, M. (2013). Oxytocin, stress and social
behavior: Neurogenetics of the human oxytocin system.
Current Opinion in Neurobiology, 23(1), 1116. doi:
10.1016/j.conb.2012.09.004
Lerer, E., Levi, S., Salomon, S., Darvasi, A., Yirmiya, N., &
Ebstein, R.P. (2008). Association between the oxytocin
receptor (OXTR) gene and autism: Relationship to Vineland
Adaptive Behavior Scales and cognition. Molecular Psychiatry, 13(10), 980988. doi:DOI 10.1038/sj.mp.4002087
Liu, X., Kawamura, Y., Shimada, T., Otowa, T., Koishi, S.,
Sugiyama, T., . . . Sasaki, T. (2010). Association of the oxytocin receptor (OXTR) gene polymorphisms with autism spectrum disorder (ASD) in the Japanese population. Journal of
Human Genetics, 55(3), 137141. doi:10.1038/jhg.2009.140
LoParo, D., & Waldman, I.D. (2014). The oxytocin receptor gene
(OXTR) is associated with autism spectrum disorder: A metaanalysis. Molecular Psychiatry. doi:10.1038/mp.2014.77
LoParo, D., & Waldman, I.D. (2015). The oxytocin receptor
gene (OXTR) is associated with autism spectrum disorder: A
meta-analysis. Molecular Psychiatry, 20(5), 640646.
Lord, C., Risi, S., Lambrecht, L., Cook, E.H., Leventhal, B.L.,
DiLavore, P.C., . . . Rutter, M. (2000). The autism diagnostic
observation schedule-generic: A standard measure of social
and communication deficits associated with the spectrum
of autism. Journal of Autism and Developmental Disorders,
30(3), 205223.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnostic interview-revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive
developmental disorders. Journal of Autism Developmental
Disorders, 24(5), 659685.
Miller, S.A., Dykes, D.D., & Polesky, H.F. (1988). A simple salting out procedure for extracting DNA from human
nucleated cells. Nucleic Acids Research, 16(3), 1215.
Modahl, C., Green, L., Fein, D., Morris, M., Waterhouse, L.,
Feinstein, C., & Levin, H. (1998). Plasma oxytocin levels in
autistic children. Biol Psychiatry, 43(4), 270277.
Mosconi, M.W., Cody-Hazlett, H., Poe, M.D., Gerig, G.,
Gimpel-Smith, R., & Piven, J. (2009). Longitudinal study of
amygdala volume and joint attention in 2- to 4-year-old
children with autism. Archives of General Psychiatry, 66(5),
509516. doi:10.1001/archgenpsychiatry.2009.19
Pobbe, R.L.H., Pearson, B.L., Defensor, E.B., Bolivar, V.J.,
Young, W.S., Lee, H.J., . . . Blanchard, R.J. (2012). Oxytocin
receptor knockout mice display deficits in the expression of
autism-related behaviors. Hormones and Behavior, 61(3),
436444. doi:10.1016/j.yhbeh.2011.10.010
Poelmans, G., Franke, B., Pauls, D.L., Glennon, J.C., &
Buitelaar, J.K. (2013). AKAPs integrate genetic findings for

10

autism spectrum disorders. Translational Psychiatry, 3,

e270. doi:10.1038/tp.2013.48
Ronald, A., & Hoekstra, R.A. (2011). Autism spectrum disorders
and autistic traits: A decade of new twin studies. American
journal of medical genetics. Part B, 156B(3), 255274 Neuropsychiatric genetics : The official publication of the International Society of Psychiatric Genetics. doi:10.1002/
ajmg.b.31159
Tansey, K.E., Brookes, K.J., Hill, M.J., Cochrane, L.E., Gill, M.,
Skuse, D., . . . Anney, R.J. (2010). Oxytocin receptor (OXTR)
does not play a major role in the aetiology of autism:
Genetic and molecular studies. Neuroscience Letters,
474(3), 163167. doi:10.1016/j.neulet.2010.03.035
Terenzi, M.G., & Ingram, C.D. (2005). Oxytocin-induced
excitation of neurones in the rat central and medial amygdaloid nuclei. Neuroscience, 134(1), 345354. doi:10.1016/
j.neuroscience.2005.04.004
Thompson, R.J., Parker, K.J., Hallmayer, J.F., Waugh, C.E., &
Gotlib, I.H. (2011). Oxytocin receptor gene polymorphism
(rs2254298) interacts with familial risk for psychopathology
to predict symptoms of depression and anxiety in adolescent girls. Psychoneuroendocrinology, 36(1), 144147. doi:
10.1016/j.psyneuen.2010.07.003
Waltes, R., Gfesser, J., Haslinger, D., Schneider-Momm, K.,
Biscaldi, M., Voran, A., . . . Chiocchetti, A. (2014). Common
EIF4E variants modulate risk for autism spectrum disorders
in the high-functioning range. Journal of Neural Transmission, 121(9), 11071116. doi:10.1007/s00702-014-1230-2
Wermter, A.K., Kamp-Becker, I., Hesse, P., Schulte-Korne, G.,
Strauch, K., & Remschmidt, H. (2010). Evidence for the
involvement of genetic variation in the oxytocin receptor
gene (OXTR) in the etiology of autistic disorders on highfunctioning level. American Journal of Medical Genetics B
Neuropsychiatric Genetics, 153B(2), 629639. doi:10.1002/
ajmg.b.31032
World Health Organization. (1993). The ICD-10 classification
of mental and behavioural disorders: Diagnostic criteria for
research. World Health Organization, Geneva, Switzerland.
Wu, S., Jia, M., Ruan, Y., Liu, J., Guo, Y., Shuang, M., . . .
Zhang, D. (2005). Positive association of the oxytocin
receptor gene (OXTR) with autism in the Chinese Han population. Biological Psychiatry, 58(1), 7477. doi:10.1016/
j.biopsych.2005.03.013
Yates, D. (2012). Behavioural neuroscience: Taking axonal
delivery of oxytocin. Nature Review Neuroscience, 13(4),
220221. doi:10.1038/nrn3213
Yrigollen, C.M., Han, S.S., Kochetkova, A., Babitz, T., Chang,
J.T., Volkmar, F.R., . . . Grigorenko, E.L. (2008). Genes controlling affiliative behavior as candidate genes for autism.
Biological Psychiatry, 63(10), 911916.

Kranz et al./Genetic variability of oxytocin receptor in autism spectrum disorder

INSAR