Background
Autism spectrum disorders (ASD), characterized by defects in social reciprocity, impairment in communication and restricted and repetitive stereotyped behavioral patterns, are the most prevalent childhood neurodevelopmental disorders. They affect all racial, ethnic and socioeconomic groups equally, with a worldwide prevalence of approximately 0.6% [
1,
2]. The genetic influences in the etiology of ASD have been demonstrated in family and twin studies [
3,
4], along with discoveries of common and rare genetic variants and pronounced chromosomal abnormalities [
5]. Recently,
de novo rare variants with a large effect size were found to increase ASD susceptibility [
6,
7]. However, generation of the ASD phenotype requires interaction between environmental factors, and inherited and
de novo genetic variants [
8]. Furthermore, the pivotal role of epigenetic regulatory mechanisms involved in the pathogenesis of Rett syndrome, fragile X syndrome and the identification of ASD-associated genetic defects in imprinted regions lends strength to the hypothesis that epigenetic factors are causative in ASD etiology [
9].
Epigenetic mechanisms involving post translational modification of histone lysine methylation influence numerous biological processes, including transcription, replication and chromosome maintenance, all of which are tightly regulated by methyltransferases and demethylases [
10]. Among them, methylation of lysine 9 in histone H3 (H3K9), marks a conserved epigenetic signal; by heterochromatin formation through trimethylation (H3K9me3) and transcriptional silencing through dimethylation (H3K9me2) [
11]. The formation of H3K9me1 and H3K9me2 are mediated by a Suv39h subgroup of histone methyl transferases, namely G9a/KMT1C and GLP/KMT1D, both having Su(var)3-9-Enhancer of zeste-Trithorax (SET) domain, through which they form homomeric and heteromeric complexes [
12]. The G9a-GLP heteromeric complex is known to interact with Wiz, a multi-zinc finger-containing molecule, resulting in a stable and dominant intracellular heteromeric methyltransferase complex [
13].
Regulation of H3K9 methylation has a powerful impact on neurological function and disease, as exemplified in Kleefstra syndrome. This disease is characterized by severe intellectual disability, developmental delay and psychiatric disorders, and is the result of a 9q34 subtelomeric deletion and loss-of-function mutations in
EHMT1[
14,
15]. In
Ehmt1 heterozygous knockout mice, the typical autistic-like features including reduced exploration, increased anxiety, altered social behavior, deficits in fear extinction, and learning and object recognition (novel and spatial) are observed [
16,
17]. Furthermore, the lack of postnatal and neuron-specific GLP/G9a expression in mouse models dysregulates neuronal transcriptional, resulting in behavioral abnormalities, such as impaired learning, motivation and environmental adaptation [
18].
Therefore, the autistic-like features and behavioral abnormalities precipitated by defects in histone methyltransferases provide a powerful case for examining their involvement in ASD pathogenesis. We put forward that rare functional variants in these genes may be associated with ASD. Since G9a-GLP-Wiz forms a stable and dominant heteromeric methyltransferase complex in H3K9 methylation, we set out to resequence the EHMT1, EHMT2 and WIZ genes coding for GLP, G9a and WIZ, respectively, in Japanese ASD case and control samples.
Discussion
Disruption of histone lysine methylation plays an important role in the pathogenesis of neurological disorders and cancer, as evidenced by the reports of genomic aberrations in histone methyltransferases in these diseases [
10]. Since defective G9a and GLP histone lysine methyltransferases, give rise to autistic phenotypes [
21], we searched for loss of function variants in the genes involved in H3K9 methylation, concentrating on rare mutations that show enrichment in ASD subjects. We focused on the variants located in the functional domains that are important in the formation of multimeric enzyme complex, and we identified the EHMT1 ankyrin repeat domain variant (Lys968Arg) and EHMT2 SET domain variant (Thr961Ile), which were present only in ASD cases and not in 1,140 control subjects. Although these two mutations were found exclusively in cases, case–control comparisons found no statistically significant association. Thus, our results did not support a role for these rare variants in ASD. This is in keeping with
in silico analyses which predicted that the effects for both the
EHMT1 (Lys968Arg) and
EHMT2 (Thr961Ile) mutations would to be ‘neutral’ and ‘tolerated’ by Provean and SIFT, respectively, although PolyPhen2 predicted a ‘possibly damaging’ phenotype.
Since a large number of ‘loss of function’ variants are present in healthy human genomes [
22], we speculate that the variants we identified may be private, owing to their lack of ‘predicted functional defects’, consistent through the three algorithms. On the other hand, balanced chromosomal abnormalities seen in ASD and related neurodevelopmental disorders are reported to disrupt the
EHMT1 gene [
23]. In addition, a
de novo deletion and rare inherited loss of function mutation in
EHMT1 were observed in a sporadic ASD trio sample [
24] and in ASD families [
25], respectively. It is clear that to understand the exact role of our identified variants, it will be necessary to examine them using much larger sample sets and more sophisticated functional assessments.
Interestingly, we observed an overexpression of the
EHMT2 gene in peripheral blood cells from ASD patients pointing towards a role of restricted chromatin state in ASD pathogenesis. A recent study showed increased expression of the
EHMT2 gene in lymphocytes and the
EHMT1 gene in both postmortem parietal cortex and lymphocyte samples, from patients with schizophrenia [
26]. The study also found that a diagnosis of schizophrenia was a significant predictor for increased expression of histone methyltransferases. Therefore, the present results are interesting, given the genetic overlap between schizophrenia and ASD [
27]. However, no significant changes in the expression levels of
EHMT1, EHMT2 or
WIZ were observed in the postmortem brain samples from BA09, BA21, BA40 and DoRN region, between ASD subjects and controls. Additionally, we detected no differential expression of
EHMT1 and
EHMT2 isoforms in the prefrontal cortex (BA09) between the two subject groups. The results suggest an absence of common variants in the regulatory genomic elements of these genes associated with ASD.
Mutations in the chromatin remodeling enzymes have been reported in psychiatric diseases, which disrupt the chromatin regulation leading to altered neuronal function and behavioral abnormalities [
28]. But in our study, such a loss of function mutation was not observed. Moreover, the identified mutations did not have a cogent effect in ASD pathogenesis, either through functional deficits or changes in expression levels. Therefore, it can be concluded that the loss of function mutations in histone methyltransferases may constitute a rare event in ASD pathogenesis, which is supported by the fact that H3K9 modifying enzymes have fewer reported mutations, when compared to other chromatin regulators [
29].
Since
EHMT2 overexpression correlates with the increased H3K9me2 levels [
30], it could result in the repressed transcription of the genes/genetic network relevant to ASD pathogenesis. However, the results from expression analysis of peripheral blood cells should be interpreted cautiously because peripheral blood chromatin may not essentially provide information specific to a brain region or neuronal phenotype [
31]. Future studies are warranted to profile the global H3K9 (mono and di) methylation status in ASD brain to delineate the genetic networks, which are dysregulated in ASD.
Although the present study did not show statistically significant enrichment of variants in ASD, their possible contribution to disease cannot be ruled out, due to the relatively small sample size restricting the statistical power of this study and also the absence of identified patient-specific mutations in global databases for the control population. From the available three-dimensional structures, it would appear that both mutations are located on the surface of the proteins (Additional file
3: Figure S1 (B and C), implying a potential role for the variants in complex formation. Recent whole genome and exome sequencing studies have clearly shown a heterogeneous genetic basis for ASD and have identified a large number of candidate genes, converging on functional pathways of neuronal signaling and development, synapse function and chromatin regulation [
32]. It is also known that SETDB1 and Suv39h1 co-exist in the H3K9 methylation multimeric complex, with interdependent functionality [
33]. Therefore, the polygenic burden of ASD may mask the effects of single rare variants, obscuring their individual contribution to disease pathogenesis [
34].
Acknowledgments
This study was supported in part by Grant-in-Aid for Scientific Research on Innovative Areas (TY) from the Japan Society for the Promotion of Science (JSPS), Japan, and by CREST (Core Research for Evolutionary Science and Technology) (YS and TY) from the Japan Science and Technology Agency (JST), Japan. In addition, this study was supported by RIKEN Brain Science Institute Funds (TY). Sections of this study was conducted as part of the ‘Development of biomarker candidates for social behavior’ (TY) and ‘Integrated research on neuropsychiatric disorders’ (NM) projects, carried out under the Strategic Research Program for Brain Sciences by the Ministry of Education, Culture, Sports, Science and Technology of Japan. A part of this work was also supported by a grant ‘Platform for Drug Discovery, Informatics, and Structural Life Science’ (MO and SF) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SB participated in the study design, performed the experiments, data analysis, interpreted the data and drafted the manuscript. YoI performed the experiments and data analysis. MM recruited participants, undertook the clinical evaluation and collected DNA samples. TT recruited participants, undertook the clinical evaluation and collected DNA samples. MTo recruited participants and collected DNA samples. CS recruited participants and collected DNA samples. KY recruited participants, undertook the clinical evaluation and collected DNA samples. YaI recruited participants, undertook the clinical evaluation and collected DNA samples. KS recruited participants, undertook the clinical evaluation and collected DNA samples. MTs recruited participants, undertook the clinical evaluation and collected DNA samples. MO performed in silico protein structure analysis. SF performed in silico protein structure analysis. TO analyzed and interpreted the data. KE analyzed and interpreted the data. MI interpreted the data. NM participated in the study design. YS conceived the study and participated in the study design. TY conceived the study and participated in the study design, interpreted the data and prepared the manuscript. All authors read and approved the manuscript.