Investigation of Rare Single-Nucleotide PCDH15 Variants in Schizophrenia and Autism Spectrum Disorders

Both schizophrenia (SCZ) and autism spectrum disorders (ASD) are neuropsychiatric disorders with overlapping genetic etiology. Protocadherin 15 (PCDH15), which encodes a member of the cadherin super family that contributes to neural development and function, has been cited as a risk gene for neuropsychiatric disorders. Recently, rare variants of large effect have been paid attention to understand the etiopathology of these complex disorders. Thus, we evaluated the impacts of rare, single-nucleotide variants (SNVs) in PCDH15 on SCZ or ASD. First, we conducted coding exon-targeted resequencing of PCDH15 with next-generation sequencing technology in 562 Japanese patients (370 SCZ and 192 ASD) and detected 16 heterozygous SNVs. We then performed association analyses on 2,096 cases (1,714 SCZ and 382 ASD) and 1,917 controls with six novel variants of these 16 SNVs. Of these six variants, four (p.R219K, p.T281A, p.D642N, c.3010-1G>C) were ultra-rare variants (minor allele frequency < 0.0005) that may increase disease susceptibility. Finally, no statistically significant association between any of these rare, heterozygous PCDH15 point variants and SCZ or ASD was found. Our results suggest that a larger sample size of resequencing subjects is necessary to detect associations between rare PCDH15 variants and neuropsychiatric disorders.


Introduction
Schizophrenia (SCZ) and autism spectrum disorders (ASD) are neurodevelopmental in origin.While SCZ and ASD are regarded as separate clinical entities, etiological, clinical, and genetic overlap between them have been discovered [1,2].Genetic factors make substantial contributions to the etiology of both conditions; heritability is estimated to be a minimum of 80% for each [3].Thousands of trait-and disease-associated common genetic variants confer increased risk of developing either condition [4,5,6], however, they may explain less than half of the total variation in risk of SCZ [7,8] and ASD [9].Recent studies suggest that rare copy-number variants (CNVs) and rare single-nucleotide variants (SNVs) may explain additional disease risk or trait variability [10,11,12,13,14,15].A significant excess of rare, disruptive SNVs has been detected in those genes that have previously been implicated as candidate risk genes for SCZ and/or ASD [16,17,18].Thus, deep sequencing of candidate genes might be a good way for elucidating the pathogenesis of these neuropsychiatric disorders [19].
PCDH15 is a member of the largest group in the cadherin superfamily that involved in generating neural diversity for neuronal differentiation and synapse formation [20].PCDH15 is primarily recognized as a gene that forms tip-link filaments in sensory hair cells and associated with Usher syndrome type 1F (OMIM 602083) [21].Notably, more than 20% of these patients exhibit neuropsychiatric symptoms [22].A GWAS identified PCDH15 as relevant to neurocognitive processes [23].In mice, PCDH15 is expressed throughout the brain and central nervous system (CNS) during embryogenesis [24], and influences serotonin transporter expression in the adult CNS [25].Rare, exonic CNVs in PCDH15 were recently identified in ASD [26] and bipolar disorder (BD) patients [27,28].These findings strongly suggest that PCDH15 is a promising candidate risk gene for neuropsychiatric disorders because several neuropsychiatric disorders including SCZ, ASD, and BD share genetic risk factors [3,4,5,6,13,27,29,30].To our knowledge, however, no published study has focused on rare PCDH15 variants in cases of neuropsychiatric disorders.
Our hypothesis was that rare PCDH15 variants might confer susceptibility to neuropsychiatric pathogenesis.To increase statistical power and detect shared risk, we combined SCZ and ASD samples in a study cohort [6,31,32].First, we performed targeted-region sequencing of PCDH15 coding exons in 562 Japanese patients; we then conducted single-variant association analysis in an independent case-control set comprising 4,013 samples to identify putative variants with large effect.

Study samples
Two independent Japanese sample groups were used in this study.For the targeted-resequencing discovery cohort, 370 SCZ (mean age ± SD, 49.7 ± 14.8 years; 53.0% male) and 192 ASD (mean age ± SD = 16.3 ± 8.4 years; 77.6% male) individuals participated.For genetic association analysis, the case control sample set comprised 1,714 SCZ (46.3 ± 15.1 years; 51.2% male), 382 ASD (19.6 ± 10.7 years; 77.8% male), and 1,917 control subjects (44.7 ± 14.7 years; 55.3% male).All subjects were unrelated, living on the mainland of Japan, and self-identified as Japanese.All patients fulfilled the criteria listed in Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) for SCZ or ASD.Healthy control subjects were selected from the general population and had no history of mental disorders based on questionnaire responses from the subjects themselves during the sample inclusion step.The study was explained to all participants and/or their parents both verbally and in writing.Written informed consent was obtained from the participants and from the parents of the patients under 20 years old.All procedures performed in this study involving human participants were approved by the Ethics Committee of the Nagoya University Graduate School of Medicine and conducted in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Sample preparation
DNA was extracted from peripheral blood or saliva from each SCZ, ASD, and control participant.For DNA extraction, we used the QIAamp DNA Blood Kit or Tissue Kit (Qiagen Ltd.Hilden, Germany).The quantity of extracted DNA was estimated using the Qubit 1 dsDNA BR Assay Kit (Life Technologies, Carlsbad, CA, USA) on a Qubit 1 2.0 Fluorometer (Life Technologies, Carlsbad, CA, USA) following the manufacturer's recommended protocol.

Library preparation and resequencing
The next-generation sequencing technology on the Ion Torrent PGM™ was used to resequence the PCDH15 coding regions (Ensembl Transcript ID: ENST00000320301; 1995 amino acids) via the protocols described in the Ion AmpliSeq™ Library Preparation User Guide (Thermo Fisher Scientific, Rev.5; MAN0006735), Ion PGM™ Template OT2 200 Kit (Thermo Fisher Scientific, Rev. 5; MAN0007220), and Ion PGM™ Sequencing 200 Kit (Thermo Fisher Scientific, Rev. 3; MAN0007273).After target-specific PCR amplification, amplicons were purified and pooled.Libraries were then prepared to obtain 200-bp PCR fragments flanked by adaptor and barcode sequences; these sequences allowed sequencing and sample identification respectively.The concentration of each library was determined with the Ion Library TaqMan Quantitation Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA).Amplified libraries were subjected to emulsion PCR and subsequent enrichment for template-positive Ion Sphere™ particles (ISPs) with the Ion OneTouch™ system (Life Technologies, Carlsbad, CA, USA).ISPs were enriched and sequenced in a 200-bp configuration run using 318 chips (Life Technologies, Carlsbad, CA, USA).
Sanger sequencing with the ABI 3130xl Genetic Analyzer (Life Technologies, Carlsbad, CA, USA), and standard methods were used to confirm each candidate variant.Sequence analysis software version 6.0 (Applied Biosystems, Foster City, CA, USA) was used to analyze all sequence data.Primer sequences for validating each variant are available in S1 Table.

Genetic association analysis
An ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA) and TaqMan assays with custom probes were used to genotype putative deleterious variants.Custom probe sequences are available in S2 Table .Each 384-microtiter plate contained two non-template controls and two samples with the variant.The reactions and data analysis were performed using Genotyping Master Mix and Sequence Detection Systems, respectively, according to the standard protocols (Applied Biosystems, Foster City, CA, USA).Differences in genotype distribution between cases and controls were tested with one-sided, Fisher's exact tests.
We computed the effective sample size and statistical power using a web browser program, Genetic Power Calculator developed by Purcell et al. (http://pngu.mgh.harvard.edu/~purcell/gpc/) [38].

Variation screening of all PCDH15 coding exons
Nucleotide sequence data reported have been deposited in the DNA Data Bank of Japan (DDBJ) databases (http://www.ddbj.nig.ac.jp) under the accession number DRA004490.
We sequenced PCDH15 exon and exon-intron boundary in genomic DNA isolated from Japanese patient sample (n = 562).Of 17 SNVs and three indels detected by the Ion Torrent PGM™, one SNV and three indels were not validated by Sanger sequencing.In total, we evaluated one splice-site variant and 15 missense variants (Table 1).Analyzing the frequency of rare SNVs in ASD and SCZ individuals, 8.9% of ASD (17/192) and 5.6% of SCZ (17/370) were identified as carriers, pointing to a higher frequency of rare SNVs in ASD (p = 0.037).Nonsense and frameshift variants were not found.Each variant detected was heterozygous.Each of the 15 missense SNVs was located in the coding region of the extracellular domain (Fig 1).
Of the 15 missense variants, 12 were predicted to be damaging with in silico prediction tools (Table 1).No variants were located in conserved sequences of the cadherin-specific motifs (XEX, DXD, DYE, XDX, and DXNDN) required for calcium binding and rigidification of the extracellular cadherin (EC) domains [39].Based on the in silico predictions, the 15 missense variants were not expected to affect splicing, but one splice-site variant (c.3010-1G>C) was (result not shown).
We were able to determine inheritance status for 16 cases.Among these 16 cases, 10 involved mother-to-son variant transmission, three involved father-to-son transmission, and three others involved father-to-daughter transmission (Table 1).An affected brother shared the p.T281A variant (S1 Fig) .An unaffected brother shared the p.G1151R variant, but an unaffected sister of the same patient did not.No de novo variants were found in these 16 cases.
We regarded four missense variants (p.R219K, p.T281A, p.V469A, p.D642N) and the splice-site variant (c.3010-1G>C) as novel ones because they were predicted to be damaging or to affect splicing based on in silico predictions and because each was not registered in any of the public databases described in the Materials and Methods.p.M60I was included in the association analysis because it was previously detected in a Japanese boy with developmental delay and hearing loss [40], although it was neither classified as damaging in the in silico analysis nor absent from the Human Genetic Variation Database.Each of these six SNVs was located in a genomic region that is highly conserved among eight vertebrate species (Table 2).Brief information and results of segregation analysis are in S1 Fig.

Genetic association analysis
For our sample set of cases (n = 2,096) and controls (n = 1,917), we computed a statistical power of >80% using the following parameters: disease prevalence of 0.01, observed rare-allele frequency of 0.0021, odds ratio for dominant effect of 3.59, and type I error rate of 0.0083 (using a Bonferroni correction by a factor of 6, based on the 6 SNVs investigated).An overview and each phenotype of genetic association analysis of the six novel variants are presented in Table 3 and S3   remained as singleton observations after genotyping of all cases and controls.We found no statistically significant association for any of the six rare heterozygous point variants in PCDH15 with case-control analysis.Post hoc calculation of statistical power based on a minor allele frequency of 0.00048 (p.M60I, Table 3) revealed that good power accrues for odds ratio 8.96 or with an increase of the sample size to nearly 20,000 individuals (cases + controls).

Discussion
To our knowledge, this is the first study to investigate the contribution of rare PCDH15 variants to neuropsychiatric disorders and susceptibility to these disorders.We conducted targeted resequencing of coding exons in PCDH15 for 562 Japanese patients and detected 16 heterozygous SNVs as condition-related candidate genes.More rare SNVs were detected from samples of ASD than those of SCZ.Of these 16 SNVs, five SNVs (p.R219K, p.T281A, p.V469A, p. D642N, c.3010-1G>C) were selected because they were both predicted to be protein-damaging by in silico analysis and not registered in public databases or found with a very low frequency in ExAC.p.M60I was selected because it previously implicated in developmental delay [40].An independent association analysis was then performed with a cohort comprising 2,096 cases and 1,917 controls.Our a priori calculation indicated our sample size was appropriately powered to determine statistical significance of SNVs.To assume the odds ratio for dominant effect of rare SNVs would be more than 3.59 seems reasonable according to previous studies that reported odds ratios from 1.88 [41] to 7.1 [32].Of these six SNVs, four variants (p.R219K, p. T281A, p.D642N, c.3010-1G>C) were not detected in our case control samples, in public databases or found with a very low frequency in ExAC.Although a number of similar studies have identified statistical associations between rare SNVs and SCZ [32,41,42,43], we found no statistically significant association between any of these rare heterozygous PCDH15 SNVs and either neuropsychiatric disorder.We find it interesting that all protein-coding SNVs observed in the resequencing cohort were located within the PCDH15 extracellular domain (Fig 1 ), which may interact with other proteins.Of the 15 protein-damaging SNVs, 12 predicted by in silico analysis might change the biological functions of PCDH15.PCDH15 plays an essential role in maintenance of normal retinal and cochlear function [39].Atypical processing of peripheral sensory inputs plays a crucial role in both SCZ and ASD pathology [44,45,46].Taken together, deleterious protein changes will induce sensory processing differences contribute to SCZ and ASD symptoms.Notably, splicing misregulation has been implicated in neuropsychiatric disorders [36,47]; c.3010-1G>C also might be a promising candidate for a causal variant in these disease etiopathologies.
In this study, all inheritance statuses were either from apparently unaffected parents or of unknown origin, suggesting variable penetrance (Table 1; S1 Fig) .Each candidate variant of maternal origin was transmitted to an affected son; this finding is similar to previous findings [48].While de novo variants have been the focus of research on SCZ and ASD pathogenesis, inherited variants also contribute substantially to these complex diseases [49].In addition, evolutionary theory predicts that deleterious alleles are likely to be especially rare because of purifying selection [50,51].Recent large-scale genetic studies report that ultra-rare, private, and inherited-truncating variants in conserved genes are highly enriched in patient populations, especially in genes that closely involved in neurodevelopment [8,17,48,52,53].The inherited ultra-rare variants (p.R219K, p.T281A, p.D642N, c.3010-1G>C) within highly conserved regions (Table 2) could increase susceptibility to development of a neuropsychiatric disorder.
There are several explanations for our inability to find statistical evidence for a causative role in SCZ and/or ASD for any of these rare PCDH15 SNVs.First, due to extremely low minor-allele frequencies (< 0.0005) as revealed by the association analysis or to odds ratios lower than expected, we could neither confirm nor dismiss the significance of rare PCDH15 variants in either neuropsychiatric disorder.Post hoc calculations revealed that a larger, higherpowered sample should be sought to reveal relationships between neuropsychiatric disorders and PCDH15 variants.Secondly, we focused on the shared genetic risk to increase the statistical power in this study.Considering that the burden of rare PCDH15 was statistically greater in ASD cases than in SCZ, further research may be needed to provide similarities and differences between SCZ and ASD.Thirdly, we focused on the ENST00000320301 transcript, but PCDH15, like many other neuronal proteins, is structurally diversified through the differential inclusion and exclusion of exons.We did not cover the promoter, untranslated regions, or intronic regions of PCDH15, which contain potentially disease-relevant regions.Fourthly, the lack of DNA from a sufficient number of patient family members prevented us from monitoring variant segregation.Finally, although 81% of the SNVs (13/16) identified in this study were predicted to be protein-disrupting or splicing-altering based on in silico analysis, the exact molecular mechanisms and networks affected by PCDH15 variants in SCZ and ASD remain unclear.Useful model systems that can address these questions will be needed to assess the impact of the SNVs discovered here.

Conclusions
We explored the role of rare PCDH15 SNVs in Japanese SCZ and ASD patients.We found four ultra-rare variants (p.R219K, p.T281A, p.D642N, c.3010-1G>C) that may increase disease susceptibility.No statistically significant association between any rare, heterozygous point PCDH15 variant and neuropsychiatric disorders was detected.A much larger sample size is needed to elucidate the relevance of rare PCDH15 variants to neuropsychiatric disorders.

Fig 1 .
Fig 1. Location for each variant of interest.PCDH15 protein structure is based on NCBI Reference Sequence NP_149045.Each variant was located in the extracellular domain.EC: extracellular cadherin repeat, TM: Transmembrane.doi:10.1371/journal.pone.0153224.g001

Table 2 .
Multiple alignments of amino acid sequences for PCDH15 eight vertebrate homologs.

Table 3 .
Association analysis of novel rare SNVs.
b homozygous for a minor allele / heterozygote / homozygous for a major allele; C P values were calculated by one-tailed Fisher's exact test doi:10.1371/journal.pone.0153224.t003