Genetics of obsessive-compulsive disorder

Abstract

Obsessive-compulsive disorder (OCD) has been seen to run in families and genetics help to understand its heritability. In this review, we summarize older studies which focused on establishing the familial nature of OCD, including its various dimensions of symptoms, and we focus on recent findings from studies using both the candidate gene approach and genome-wide association study (GWAS) approach. The family studies and twin studies establish the heritability of OCD. Candidate gene approaches have implicated genes in the serotonergic, glutamatergic, and dopaminergic pathways. GWAS has not produced significant results possibly due to the small sample size. Newer techniques such as gene expression studies in brain tissue, stem cell technology, and epigenetic studies may shed more light on the complex genetic basis of OCD.

INTRODUCTION

Obsessive-compulsive disorder (OCD) is characterized by recurrent and persistent thoughts, urges, or images that are intrusive, unwanted, and resisted (obsessions), and by repetitive behaviors or mental acts that the affected individual feels driven to perform, usually to reduce the distress associated with obsessions (compulsions). These symptoms are time-consuming, distressing, and impairing (Diagnostic and Statistical Manual of Mental Disorders-V). Once thought to be a rare condition, OCD is now estimated to occur in 0.8%–3.3% of the population of India,[1] with similar prevalence in other countries throughout the world.[2] There is strong evidence for a biological basis of OCD, including occurrence in other neurological disorders, response to specific neurotropic medications, and neuroimaging studies implicating cortico–striatal–thalamic neural circuits. Moreover, considerable evidence for a genetic contribution to the disorder has been provided by family and twin studies, genetic linkage studies, candidate gene association studies, genome-wide association studies (GWAS), and genome sequencing studies. In this review, we summarize salient findings from each of these approaches, which have provided insights, and paved the way for further research, into the pathophysiology of OCD.

FAMILY STUDIES

Clinicians have long observed that many of their patients with OCD have relatives who also are affected with the disorder. Based on reports of family informants, early family history studies found that 4%–8% of relatives of cases had OCD, and 20%–40% of first-degree relatives had “obsessional traits.”[3,4] Several later family studies also found that the prevalence of OCD in directly interviewed first-degree relatives of OCD cases was considerably greater than the prevalence of OCD in the community.[5,6,7]

These results have been supported by findings from several more methodologically rigorous family studies, which used explicit diagnostic criteria, direct assessment of relatives, structured or semi-structured assessment instruments, inclusion of families of non-OCD affected individuals as a comparison group, and assessment of relatives by examiners blind as to the OCD status of the proband (i.e., the index case). For example, Pauls et al. found that the prevalence of OCD was significantly greater in first-degree relatives of adult OCD cases compared to relatives of controls (10% vs. 2%),[8] as did Nestadt et al. (12% vs. 3%)[6] and Grabe et al. (10% vs. 1%).[5] Of note, the first modern family study of OCD did not find evidence for familiarity of OCD; however, when a best estimate diagnostic procedure was used, the prevalence of definite or probable OCD was found to be greater in relatives of case versus control probands (11% vs. 4%).[7,9] In a study using Swedish national registry data, the odds of OCD in case versus control relatives increased with the degree of genetic relatedness from 4.6 to 5.0 in first-degree relatives; 1.5–2.3 in second-degree relatives; and 1.4 in third-degree relatives. Interestingly, nonbiological relatives (spouses or partners having a child together) also had about a 2.5-fold increased risk of OCD, suggesting possible assortative mating in individuals with OCD.[10]

Studies with child and adolescent probands have found even stronger evidence for familial aggregation of OCD.[11,12] Moreover, studies in the US and India have found that the prevalence of OCD is substantially greater in relatives of probands with early onset (i.e., ≤19 years old) than later onset.[6,8,13]

Several studies have investigated the familiality of OCD symptom dimensions. Brakoulias et al. found greater familiarity for hoarding and contamination/cleaning symptoms than for other dimensions.[14] A study from our center showed symmetry-related symptoms to be more prevalent in familial OCD than sporadic OCD.[15]

Family studies also have been useful for investigating the spectrum of disorders that may be related to OCD. The Johns Hopkins OCD Family Study, in particular, evaluated a variety of disorders for their potential familial relationship to OCD.[6] For example, Nestadt et al.[16] found that generalized anxiety disorder, agoraphobia, panic disorder, separation anxiety disorder, and major depression were significantly more prevalent in relatives of case than control probands. Bienvenu et al.[17] found that body dysmorphic disorder and grooming disorders (nail biting, skin picking, or trichotillomania) were significantly more prevalent in case than control relatives. Grados et al.[18] found that tic disorders, including chronic motor or vocal tic disorders, were more prevalent in OCD case than control relatives. Among the case relatives, tic disorders were more prevalent in those with a male than with a female proband, and in those whose proband had onset of OCD before 18 years of age. Samuels[19] investigated personality disorders and general personality traits in these families. They found that obsessive-compulsive personality disorder was significantly more prevalent in case than control relatives. They also found that mean scores on neuroticism were significantly higher in case than control relatives.

TWIN STUDIES

Monozygotic twins are genetically (nearly identical), whereas dizygotic twins share about 50% of their genes, on average. Twin studies, which compare concordance of symptoms in the two types of twins, provide further support for a genetic contribution to OCD. For example, using the Maudsley Twin Registry, Carey and Gottesman (1981) found concordance for “obsessive symptoms or features” was 87% in 15 identical twin pairs, as compared to 47% in fraternal twin pairs.

More recently, investigators have evaluated the heritability of obsessive-compulsive symptoms assessed dimensionally, typically by questionnaire. For example, in a study of 4564 4-year-old twin pairs, Eley et al. estimated the heritability of obsessive-compulsive behaviors at 65%, with the remaining 35% of variance in these behaviors attributed to unshared environmental influences.[20] Similarly, in 4246 child twin pairs, Hudziak et al. estimated the heritability of 45%–61% for obsessive-compulsive symptoms, with unique environmental influences of 42%–55%.[21] In a review of OCD twin studies reported from 1929 to 2005, van Grootheest et al. concluded that genetic factors account for 27%–47% in adult-onset OCD cases, 45%–65% of the heritability in child-onset cases of OCD, and 27%–47% in adult-onset OCD cases.[22] Subsequent twin studies have supported this conclusion, and also found shared genetic overlap between OCD and other disorders, including tic disorders, anxiety disorders, and attention-deficit hyperactivity disorder.[23,24,25,26] Iervolino et al. found that concordance rates were higher for identical than for dizygotic twins across all five OCD symptom dimensions, although the genetic contribution was greatest for the hoarding dimension. Data from GWAS have also been used to estimate the heritability of OCD and have produced similar results of 37%.[27]

CANDIDATE GENE STUDIES

Candidate gene studies are guided by information about the physiological and functional aspects of the illness. Particular genes of relevance are chosen based on the understanding of the illness. The variations in the gene are then analyzed for any association with the observed phenotype. This approach has its limitations in that it is poorly replicable and multiple studies exist with conflicting results. Furthermore, the choice of genes to be studied depends on the prior knowledge of the illness which itself may not be complete or sufficient. It is also possible that in a multifactorial disorder, the gene may only have a small contribution to the overall genetic risk. Despite these drawbacks, multiple genes have been examined for their association with OCD. In OCD, the most widely studied genes are those related to serotonin neurotransmission. This is in due to robust evidence for the treatment of OCD with serotonin reuptake inhibitors (SRI).[28]

Pauls has reviewed the gene association studies in OCD through 2010 and notes that more than 80 genes have been examined in over 10 years since this approach was first used for OCD.[29] In this review, we summarize the genes that have been studied in various pathways in the last decade.

Genes related to the serotonergic system

The gene which has received the most attention in OCD has been the SLC6A4 gene coding for the serotonin transporter protein. The variant most often studied is the variation in the promoter region of the gene known as 5HTTLPR. The region exists either as short (S) or long (L) allele based on the number of repeat elements. It was discovered that there was a single-nucleotide polymorphism (SNP) of A to G within the L allele. This makes the gene triallelic.[30] L(A) increases the expression of the gene, whereas both L(G) and S allele decrease the expression of the gene.[31] In recent meta-analyses, when the tri-allelic nature of the gene is taken into consideration, it has been shown that the L(A) allele is associated with OCD.[32,33]

The serotonin receptor 2A (HTR2A) is a protein on the postsynaptic membrane. Two SNP's rs6311 (G/A) and rs6313 (T/C) have been studied from the promoter and the coding region of the gene, respectively. Allele A of rs6311 and allele T of rs6313 have been associated with OCD in a meta-analysis.[34] A most recent study has found no association between HTR2A gene polymorphisms and OCD but found an association with a variable number tandem repeat (VNTR) polymorphism in the intron 2 of the SLC6A4 gene.[35]

Genes related to the glutamatergic system

Glutamate signaling has been implicated in the pathogenesis of OCD[36,37] and a glutamate-mediated reversible dysfunction in the neurocircuitry has been shown to be associated with OCD.[38]

The postsynaptic glutamate transporter protein is coded by the gene SLC1A1 and has aroused much attention in OCD as it has been associated with OCD in multiple studies;[39,40,41,42] however, it is to be noted that different studies have studied different variants within the gene. GWAS, however, has failed to support the role of these genes.[43] Some rare variants have also been found in the gene containing a missense mutation (Wu et al., 2013). An SNP rs10491734 was found to be associated with early-onset OCD.[44] Studies which have looked into the functioning of the isoforms of the gene have noted that they are differentially expressed in participants with OCD and are responsible for reduced functioning of the transporter.[45] This suggests that reduced glutamate transport into the postsynaptic neuron may be responsible for the pathology in OCD. A recent study has shown that SLC1A1 plays a role in basal ganglia-dependent repetitive behaviors in mice models of OCD.[46]

Another gene, DLGAP3 which encodes a scaffolding protein is known to affect glutamate functioning.[47] The murine homolog of the DLGAP3 known as SAPAP3 has been studied in mice model of OCD. Gene knockout mice for SAPAP3 display compulsive behavior- and anxiety-like behaviors which respond to treatment with SRI, optogenetic studies have proposed a role for the protein in the lateral orbito–fronto–striatal pathway.[47,48,49] Another related gene DLGAP1 has also been studied following preliminary evidence from GWAS studies but has not shown further association with OCD.[43,50]

Genes related to the dopaminergic system

Although a majority of genetic studies have studied serotonergic and glutamatergic systems, some have suggested a role for dopamine in the pathophysiology of OCD.[51] The most positive associations for genes in this system have been for the DRD4 gene.[29] A VNTR sequence has been noted in this gene whose 7R allele was found to be associated with OCD, and also the 2R allele was noted to be associated with the dimension of symmetry of the OC symptom dimensions.[52] Other genes SLC6A3 and DRD3 have also been explored and have shown association with white matter changes that occur in young-onset OCD.[53] Most studies for dopamine transporter gene in OCD have produced negative results.[54]

LINKAGE STUDIES

Genetic linkage studies of OCD seek to identify chromosomal regions that contain genes for the disorder, by statistically testing if alleles of specific genetic markers cosegregate (i.e., “travel together”) with OCD in families. Subsequently, “fine mapping” is often conducted by including additional markers to further narrow the linkage region, and family-based association analyses may be conducted to evaluate associations between specific SNPs within these regions and OCD. The first reported linkage scan in OCD, which studied 56 individuals in seven families with a pediatric OCD proband, identified a linkage peak in the 9p24 region on chromosome 9. Interestingly, this region harbors the glutamate transporter gene SLC1A1, which is involved in neurotransmission and is a potential candidate gene for OCD.[55] This finding was replicated by Willour et al. in a linkage analysis in 50 OCD pedigrees, using the same 13 genetic markers in the region used by the Hanna group, and pedigree-based association analyses identified two markers associated with OCD in this region.[56] These linkage findings stimulated further studies of SLC1A1 as a candidate gene for OCD.

Another genome-wide linkage study was conducted in 219 families with multiple relatives affected with OCD, mostly affected sibling pairs, as part of the OCD collaborative genetics study. The strongest suggestive linkage signal was found on chromosome 3, in the 3q27-28 region; other suggestive linkage signals were found in regions on chromosomes 1, 6, 7, and 15.[57] Further analyses provided evidence for different linkage patterns in these families, depending on specific phenotypic characteristics. For example, families were stratified into those with two or more individuals with compulsive hoarding behavior, and those with one or fewer hoarding relatives. In the hoarding-loaded families, there was a significant linkage on the chromosome 14q23-32 region.[58]

WHOLE-GENOME ASSOCIATION STUDIES

Whole genome or GWAS, involves scanning hundreds of thousands, or even millions, of SNPs across the genome to detect relatively common genetic variants associated with the disorder. The studies have compared the variants in individuals with and without the disorder, or have focused the analysis on “trios,” affected individuals, and their parents, whether affected or unaffected. This approach has been successful in the study of “complex” disorders, which are hypothesized to be due to multiple genes, each of which makes a small contribution to the overall risk of the disorder.[59,60]

Three GWAS studies of OCD have been reported to date. The International OCD Foundation Genetics Collaborative analyzed almost 1500 OCD cases, 5600 controls, and 400 trios with nearly 500,000 SNPs. In the case–control analysis, the two most significantly associated SNPs were located within the DLGAP1 gene, although no association had “genome-wide significance” after correcting for the large number of comparisons. In the analysis of trios, but not in the combined case–control-trio analysis, an SNP near the BTBD3 gene was found to significantly associated with OCD.[43]

The OCD Collaborative Genetic Association Study analyzed over 1000 families and 500,000 SNPs. Although no association was found with genome-wide significance, the most significant finding was for SNP on chromosome 9, near the protein tyrosine phosphate receptor D (PTPRD) gene. Several additional candidate genes emerged from these two studies, including Fas apoptotic inhibitory molecule 2 (FAIM2), glutamate ionotropic receptor N-methyl-D-aspartate type subunit 2B, and the cadherin genes CDH9 and CDH10. Several of these genes appear to be involved in glutamatergic neurotransmission.[61]

A meta-analysis of the above-mentioned studies was conducted, including those with European ancestry (2688 patients and 7037 controls), the findings were similar with no SNP reaching genome-wide significance. The genes from top signal regions were ASB13, RSPO4, DLGAP1, PTPRD, GRIK2, FAIM2, and CDH20; these included some of the genes which had shown trends toward significance in the original GWAS studies.[62]

The third GWAS analyzed almost 7000 individuals and 31,000,000 SNPs as part of the Netherland National Twin Registry, using a quantitative measure of obsessive-compulsive symptoms. A genome-wide significant finding was found for SNP in the myocyte enhancer factor 2B neighbor (MEF2BNB) gene in region 9p13 on chromosome 19. Additional gene-based testing found four significantly associated genes in the same region, MEF2BNB, MEF2B, and MEF2BNB-MEF2B and RFXANK.[63]

RARE VARIANT STUDIES

OCD, similar to other genetically complex disorders, is likely to have a heterogeneous etiology. Therefore, less common variants with strong effect may be found in a subset of cases. Three types of these variants have been reported in OCD as follows: chromosomal rearrangements, copy number variants (CNV), and rare variants identified by “deep sequencing” of the genome.

McGrath et al. investigated for the presence of large (>500 kb), rare (1%) CNVs in OCD. They found that the proportion of individuals with deletions in known pathogenic neurodevelopmental loci was four times greater in the patients with OCD than in controls. Certain of these have been implicated in other neuropsychiatric conditions.[64] Further studies are now underway to search for de novo variants in OCD.

Cappi et al. found a higher than expected rate of de novo “nonsynonymous variants” (i.e., those that alter the amino acid sequence of a protein) in OCD cases in a study sequencing 20 OCD cases and relatives. The analysis of the protein–protein interaction network suggested an enrichment of genes involved in immunological and central nervous system functioning and development.[65]

CONCLUSIONS AND FUTURE DIRECTIONS

Evidence from epidemiological and molecular genetic studies strongly suggests a complex genetic etiology for OCD. To date, except for a few candidate genetic variants such as SLC1A1, DLGAP1, and PTPRD, no established variant has been identified for OCD. Larger samples are required for many of the empirical approaches described above.

Beyond the identification of associated genetic variants, it will be critical to understand the functional basis for the relationship between the protein and the phenotype additional strategies are in the pipeline or should be such as genetic expression studies in brain tissues and application of induced pluripotent stem cell technology. Epigenetic studies may contribute by elucidating how modification of gene expression contributes to the risk of OCD. The encyclopedia of DNA elements program of the US National Institutes of Health is an example of a program that will contribute greatly in the future.[66]

The phenotype of OCD and related conditions is likely to be modified because of these studies. We expect that additional clinical subtypes will emerge and an extended boundary of traits involving additional psychiatric disorders will occur. Most optimistically, we hope that these research endeavors will lead to the development of rational treatments and preventive measures for OCD.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.