Family-based association studies support a sexually dimorphic effect of COMT and MAOA on genetic susceptibility to obsessive-compulsive disorder

Abstract

Background: Obsessive-compulsive disorder (OCD) is a common and severe psychiatric illness that affects 1–3% of the population and presents a well-established co-morbidity with major depressive disorder (MDD). Twin and family studies have suggested a genetic component in the etiology of OCD, although the mode of inheritance is unknown. Pharmacotherapy of the disease implicates both serotonergic and dopaminergic pathways. Previously, guided by the 22q11 microdeletion-related psychiatric phenotype, we provided evidence for a sexually dimorphic association between OCD and the gene for catechol-O-methyltransferase (COMT). In this report, we use 110 nuclear OCD families to analyze the inheritance of variants of COMT and monoamine oxidase-A (MAOA), another gene modulating monoamine metabolism.

Methods: A sample of 110 nuclear OCD families was collected, and lifetime diagnoses were ascertained using the Diagnostic Interview for Genetic Studies (DIGS). DNA was genotyped for functional variants of the COMT and MAO genes, and allele inheritance was examined using the Transmission Disequilibrium Test (TDT) and Haplotype-based Haplotype Relative Risk (HHRR) test.

Results: We provide evidence supporting the previously reported sexually dimorphic association between low COMT enzymatic activity and OCD. We also provide evidence for a similar sexually dimorphic association between OCD and an allele of the MAOA gene, previously linked to high MAO-A enzymatic activity. In agreement with the well-established action of MAO-A inhibitors as antidepressants, this association is particularly marked among male OCD probands with co-morbid MDD, who represent more than 50% of our male OCD sample.

Conclusions: Our analysis indicates that variants of two genes modulating monoamine metabolism contribute significantly to OCD susceptibility. Most importantly, an unexpected sexually dimorphic pattern of genetic susceptibility to OCD is revealed and suggests the possibility that profound gender differences in genetic predisposition may exist not only for other OCD susceptibility genes, but for an array of other psychiatric disorders as well.

Introduction

Obsessive-compulsive disorder (OCD) is a common and severe psychiatric illness that affects 1–3% of the population Karno et al 1988, Weissman et al 1994. Patients afflicted with OCD experience intrusive, disturbing, repetitive thoughts (obsessions) and an uncontrollable urge to repeatedly enact stereotypic rituals (compulsions) (American Psychiatric Association 1994). When severe, cycles of obsessions and compulsions can require hours each day to complete. Twin and family studies have suggested a genetic component in the etiology of OCD Lenane et al 1990, Pauls et al 1995, although the mode of inheritance is unknown.

Despite its prevalence and severity, very little is known about the disorder’s pathogenesis. The partial efficacy of selective serotonin reuptake inhibiting agents (SSRIs) has led to the hypothesis that OCD may be associated with dysregulation of serotonergic neurotransmission (Murphy et al 1995); however, a substantial proportion (30–40%) of OCD patients do not respond to SSRIs Greist et al 1995, McDougle et al 1994. Treatment augmentation with dopamine receptor blockers has proven useful among these patients, thus implicating involvement of dopaminergic pathways as well.

The rates of co-morbid diagnoses in OCD patients have varied according to the population ascertained and the methodology employed. More often than not, however, studies have suggested that depressive symptoms are OCD’s most common complication Goodwin et al 1969, Rapoport 1989, suggesting a shared neuropathophysiology. For example, lifetime co-morbid major depressive disorder (MDD) was reported by 67% of OCD patients interviewed in a semistructured format and by 78% of patients assessed with a structured diagnostic instrument (Rasmussen and Eisen 1994). In fact, medications with specific serotonergic properties (e.g., clomipramine, fluoxetine, fluvoxamine) are most efficacious in treating both OCD and MDD. While the specific biologic relationships between MDD and OCD have not been specified (Zohar and Pato 1991), clinical reports have described the two illnesses functioning in synchrony, with both improving during pharmacologic treatment (Pato and Zohar 1991).

Although no linkage studies have been reported for OCD, it is likely that only a portion of the loci harboring OCD susceptibility genes would be identified. The limitations to detecting many of the remaining genes could be overcome by association studies that are more suitable for detecting genes of weak effect (Risch and Merikangas 1996). Association studies are based on the assumption that complex psychiatric disorders are likely to be associated with low penetrance but relatively common, functional variations (alleles) in a number of susceptibility genes. Two designs are used more frequently to test this assumption. In one design, the distribution of a specific allele of a polymorphic locus is tested in affected individuals compared with control subjects who have been matched by ethnicity, sex, and age and preferably have been psychiatrically evaluated (case/control studies). Problems with this design include stratification effects, where a section of the population contains a disease and a marker allele more commonly than expected without a causal relationship. Another design tests the transmission of a particular allele from a parent to the affected individual using the other untransmitted allele from the parent as the “control”; this test mitigates the need for ethnically well-matched control subjects. In both methodologies, differences in the distribution of alleles may suggest that the marker itself may have some direct influence on susceptibility to the disease or, alternatively, that the marker allele may be closely linked to the disease allele. Association studies have several practical advantages over linkage studies, namely that families with multiple affected individuals, which are relatively scarce, are not required and that no assumptions are made about the mode of inheritance of the disease. In addition, association studies have considerable statistical power to detect genes of weak effect, unlike linkage studies in families. A limitation of association studies is that functional polymorphisms within candidate genes, or at least polymorphisms in strong disequilibrium, must be identified before the test can be performed. The reason is that in heterogeneous populations, departure from the expected distribution is only detectable very close to or at the “associated” polymorphism. Finally, when a sufficient number of patients are available, the transmission of alleles and genotypes among them can also be examined in relation to several variables, such as gender, family history, presence of co-morbid conditions, response to treatment, etc. It is likely that this kind of analysis will eventually provide objective criteria to define etiological subtypes of psychiatric disorders.

The q11 band of human chromosome 22 has been previously reported to be hemizygously deleted in a subgroup of patients with schizophrenia (Karayiorgou et al 1995). In addition, two independent studies raised the possibility that the 22q11 genomic region may harbor genes predisposing an individual to OCD. These studies reported increased rates of co-morbid OCD or obsessive-compulsive symptoms (OCS) among schizophrenic patients with the 22q11 microdeletion (Pulver et al 1994) and, similarly, increased rates of anxiety, OCD, and OCS in children and adults with the 22q11 microdeletion in the absence of schizophrenia (Papolos et al 1996). Previously, in an effort to isolate OCD-predisposing genes from this region, we used association studies to examine the role of the gene for catechol-O-methyltransferase (COMT). In humans, a common functional polymorphism is associated with a three- to fourfold variation in COMT enzyme activity. This variation in activity is due to a G → A transition at codon 158 of the COMT gene that results in a valine (high-activity allele, COMT∗H) to methionine (low-activity allele, COMT∗L) substitution Grossman et al 1992, Karayiorgou et al 1998, Lachman et al 1996, Lotta et al 1995. We showed that the COMT∗L allele was significantly associated in a recessive manner with susceptibility to OCD, particularly in male patients (Karayiorgou et al 1997). The COMT∗L / COMT∗L genotype appeared to be a risk factor for OCD, with an approximate relative risk of 5.91 (95% confidence intervals: 2.40, 14.53) estimated for genotype COMT∗L / COMT∗L versus nonCOMT∗L / COMT∗L. COMT is a Mg²⁺-dependent enzyme that catalyzes the transfer of methyl groups from S-adenosyl methionine to a hydroxyl group of a catecholic substrate: Dopamine (DA) is converted into 3-methoxytyramine, and norepinephrine (NE) is converted into normetanephrine. COMT is widely distributed in the mammalian brain. The enzyme is absent from the dopaminergic terminals, and it is thought to participate in the catabolism of extraneuronal DA in glial cells and/or postsynaptic neurons (Napolitano et al 1995).

Monoamine oxidases (MAOs) are flavin-containing enzymes that degrade a variety of biogenic amines, including the neurotransmitters NE, DA, and serotonin (5-HT) (Weyler et al 1990). Two forms of the enzyme, MAO-A and MAO-B, differ in molecular weight, substrate affinities, inhibitor sensitivities, and immunological properties. These two proteins are encoded by two adjacent genes (Hsu et al 1989) located at the p11.23–11.4 region of the X chromosome (Levy et al 1989).

The MAOA locus is a logical candidate to investigate for involvement in susceptibility to OCD because of its participation in COMT-controlled pathways, the association between its activity variations and behavioral effects, and the pharmacologic evidence from OCD patients. More specifically, MAO-A is localized both in serotonergic and catecholaminergic neurons in the brain (Thorpe et al 1987) and, along with COMT, comprises the major mammalian enzymes involved in the metabolic degradation of DA, NE, and epinephrine. Results from pharmacologic studies, however, suggest that the relative importance of COMT (methylation) versus MAO-A (deamination) in the metabolic degradation of catecholamines varies in different brain regions; for example, methylation accounts for about 15% of released DA in striatum and nucleus accumbens and for more than 60% in frontal cortex (Karoum et al 1994). Unlike COMT, MAO-A is also involved in the metabolic degradation of 5-HT, and, in rodents, pharmacologic inhibition of activity or targeted deletion of the gene for MAO-A leads to an accumulation of 5-HT and NE, but not DA, in the brain Campbell et al 1979, Cases et al 1995. Variations in MAO activity appear to affect behavioral traits in humans and animals. Submicroscopic deletions of the Xp11 region resulting in the absence of both MAO genes have been reported in atypical, male Norrie disease patients (de la Chapelle et al 1985), who present marked changes in monoamine metabolism (Sims et al 1989), accompanied by severe neurologic and behavioral dysfunctions and profound mental retardation (Brunner et al 1993a). Most interestingly, Brunner et al (1993b) described a large Dutch kindred with an X-linked borderline mental retardation and pronounced behavioral disturbance in male members, where a point mutation in the eighth exon of the MAOA gene changing a glutamine to a stop codon was detected in all affected men (Brunner et al 1993b). With regard to OCD, MAO-A inhibitors seem to have beneficial effects for a subset of OCD patients Dominguez and Mestre 1994, Jenike et al 1997, Liebowitz et al 1990, McDougle et al 1993.

MAO-A activity is under strong genetic control, with levels ranging more than 50-fold among control subjects, as measured in cultured skin fibroblasts, where MAO-A accounts for 80–100% of MAO activity (Breakefield et al 1980). A T → G substitution at the third base of codon 297/exon 8 of MAO-A (position 941 of MAOA cDNA, GenBank Accession No.: M69226, [Hsu et al 1988]) results in a Fnu4H1 restriction length polymorphism (RFLP) site (Hotamisligil and Breakefield 1991). Unlike the COMT codon 158 variation, this substitution does not result in a change in amino acid at this position; however, a marginally significant association between Fnu4H1 alleles and MAO-A activity levels has been previously described (Hotamisligil and Breakefield 1991). More specifically, this MAOA RFLP has been associated with high MAO-A activity in vitro when the Fnu4H1 site is present (MAOA∗297CGG, [Hotamisligil and Breakefield 1991]) and is probably in linkage disequilibrium with another (yet unidentified) functional variant in the MAOA locus.

In this report, we analyze variants of these two genes (COMT and MAOA) in relation to OCD. The analysis of the transmission of the COMT alleles was designed as a replication of our previous findings. On the other hand, no positive associations have been reported between OCD and MAOA, and the direction of a possible association could not be predicted a priori (especially since MAO-A participates in both DA and 5-HT metabolism).