Bipolar disorder is a severe, chronic, and disabling mental illness characterized by recurrent episodes of hypomania or mania and depression. It is a clinically defined nosological entity with multi-factorial but poorly understood etiologic mechanisms. The evidence from twin, family, and adoption studies provide compelling evidence for a strong genetic predisposition to BPD with heritability estimated to be as high as ≥80%. Given BPD is a heterogeneous disease with substantial phenotypic and genetic complexities, the identification for BPD risk loci has proven to be difficult. Some researchers have proposed that dissecting BPD into clinical subgroups with distinct sub-phenotypes may result in genetically homogeneous cohorts to facilitate the mapping of BPD susceptibility genes. Among the sub-phenotypes, early-onset BPD is of particular interest as several independent cohort studies have demonstrated their existence. Comparing to the non-early-onset BPD, the early-onset sub-type is associated with a more severe form of clinical manifestations characterized by frequent psychotic features, more mixed episodes, greater psychiatric comorbidity such as drug and alcohol abuse and anxiety disorders, higher risk of suicide attempt, worse cognitive performance, and poorer response to prophylactic lithium treatment. In addition, the pattern of disease inheritance seems to differ between early‐ and late‐onset BPD families, with the former involving greater heritability. These observations indicate that early-onset BPD may be a genetically homogenous subset and thus could be used for genetic study to identify its susceptibility genes. A number of BPD genes identified by genome-wide association study have been widely replicated and intensively studied, but these studies did not include early-onset BPD. Over the past two decades, a host of studies have investigated genetic loci responsible for early-onset BPD through linkage-analyses, candidate–gene association, analyses of copy number variants , and GWAS, but findings are inconclusive. Candidate–gene association studies have identified a number of genes potentially associated with early-onset BPD,the rack of clones including glycogen synthase kinase 3-β gene, circadian clock gene Per 3, serotonin transporter gene, brain-derived neurotrophic factor gene, and gene coding synaptosomal-associated protein SNAP25.
However, very few positive findings of these studies have been replicated independently. Findings from linkage studies suggested chromosomal regions harboring the susceptibility genes at 3p14 and 21q22, plus the loci at 18p11, 6q25, 9q34 and 20q11 with nominal significance. Studies of CNVs in early-onset BPD were based on relatively small effect sizes and were irreproducible, suggesting that CNVs are unlikely the major source of liability. Finally, GWAS failed to find any risk variant with genome-wide statistical significance in Caucasian populations, despite some variants showed suggestive significance. In previous genetic studies, the definition of early-onset in BPD typically ranged from 15 to 25 years of age. These association studies were largely conducted with small sample size and were under powered. Most of them compared early-onset BPD vs. healthy control. Such a case–control design is more likely to identify susceptibility gene for BPD per se, but not for the early-onset sub-type. The optimal strategy to identify gene for the early-onset BPD is to include the non-early-onset BPD group for comparison. Different definitions for early onset of BPI have been proposed in previous work. In this paper, we reported findings from a GWAS with high-density SNP chips on early-onset, defined as ≤20 years of age, BPI patients of Han Taiwanese descent.The clinical phenotype assessment of manic and depressive episodes was performed by well-trained psychiatric nurses and psychiatrists using a cross culturally validated and reliable Chinese version of the Schedules for Clinical Assessment in Neuropsychiatry, supplemented by available medical records. All of them were diagnosed according to the DSM-IV criteria for BPI disorder with recurrent episodes of mania with or without depressive episode. The assessment of onset age was based on a life chart prepared with graphic depiction of lifetime clinical course for each of the BPI patient recruited. This life chart included largely all the mood episodes with date of onset , duration, and severity . The construction of this life chart was based on integrated information gathered from direct interview with patients and their family members, interviews with in-charge psychiatrists, and a thorough medical chart review.
Different definitions for early onset of BPI have been proposed in previous work. Our selection of 20 as the age threshold was based on a systematic review for pediatric BPD. The age at onset was defined by the first mood episode . Of all patients, 1306 with genotyping data available were included in the discovery group for GWAS and the rest 473 without genotyping data were included in the replication group.In this paper, we have reported one of the largest GWASs to investigate genetic susceptibility to early-onset BPI with the first mood episode occurring ≤20 years of age. We employed standardized psychiatric interview and constructed a life chart with detailed clinical history to ensure the accuracy and homogeneity of phenotype for genotyping. Our GWAS with high-density SNP chips identified the SNP rs11127876 in CADM2 gene to be associated with early-onset BPI in both discovery and replication groups, and meta-analysis for the association was close to genome-wide significance . The gene CADM2 on chromosome 3 encodes a synaptic cell adhesion molecule that is prominently expressed in neurons, and plays key roles in the development, differentiation, and maintaining synaptic circuitry of the central nervous system. In previous GWASs, CADM2 has been found to be associated with a number of mental health related traits, including alcohol consumption, cannabis use, reduced anxiety, neuroticism and conscientiousness, and increased risk-taking behavior. CADM2 was also reported to be associated with executive functioning and processing speed, general cognitive function, and educational attainment. Though there have been no investigations examining the risk-taking phenotype in early-onset relative to non-early-onset BPD, Homes et al. showed that BPD patients with a past history of alcohol abuse or dependence had a higher risk-taking propensity, suggesting a relationship between early-onset BPD and risk-taking propensity. Of note, Morris et al. suggested that CADM2 variants may not only link with psychological traits, but also influence metabolic-related traits, such as body mass index, blood pressure, and C-reactive protein. In addition, they found that CADM2 variants had genotype specific effects on CADM2 expression levels in adult brain and adipose tissues.
The finding highlights the potential pleiotropy of CADM2 gene, i.e., the genetic variants may influence multiple traits, and shared biological mechanisms across brain and adipose tissues through the regulation of CADM2 expression. Given that the metabolic comorbidities are prevalent in patients with early-onset BPD, it is conceivable that CADM2 variants may influence both psychological and physical traits, further contributing to a more severe clinical sub-type of BPD with accompanying risk of metabolic adversities. In addition, an association between risk-taking and obesity has also been implicated in previous research, which suggests that risk takers tend to overlook health-related outcomes and are prone to aberrant reward circuitry predisposing them to poor dietary choices and excessive intake. Collectively, in line with the characteristics found to be associated with CADM2 variants, it is likely that CADM2 may exert an effect on the constellation of clinical features related to early-onset BPD with greater symptom severity . Therefore,grow shelves with lights our findings suggest that CADM2 genetic variants may have significant effects on a sub-type of BPD with early-onset. Two previous GWASs comparing early-onset BPD patients with healthy controls did not find any genetic variants reaching genome-wide significance. Our study included a larger sample of early-onset BPI patients to conduct GWAS using high-density genotyping . The statistical power was calculated using Post-hoc Power Calculator , combining the allelic frequencies of both discovery and replication groups. In this study of two independent samples of BPI with dichotomous endpoint, the power reached 99.4% and 18.2% under type I error = 0.05 and = 5 × 10−8 , respectively. Results of our study are also likely to be under powered under the stringency setting of type I error. However, the frequency of risk allele T was higher in patients with onset ≤20 than in patients with onset >20 in both discovery and replication groups. We believe all these have provided strong evidence to confirm the association of this SNP with early onset BPD. In Table 2, the minor allele frequencies differ quite a bit between the discovery and replication cohorts. In the NCBI SNP database, minor allele frequency of rs11127876 is 0.08 in Han Chinese in Beijing, close to our results and suggest that the different allele frequencies observed in Table 2 may mainly result from our sampling. The over-representative minor allele frequency in replication group may have come from random sampling or effects of hidden characters of our patients recruited. Genetic variant of CADM2 has been reported to be associated with behavioral and metabolic traity, which were not assessed in this study. Though the minor allele frequencies of rs11127876 were different in discovery and replication groups, the same direction of ORs of rs11127876 minor allele supports the reliability of our findings. The SNP rs75928006 located in the upstream of MIR522 reached genome-wide significance in discovery group but failed to show statistical significance in replication group.
MIR522 promotes glioblastoma cell proliferation, but there was no evidence to suggest its association with any psychiatric disorders. One major limitation of this study is the possibility of recall bias about the exact onset age of the first mood episode of BPI, particular when there was a long history of the illness. The preparation of life chart containing detailed clinical course and treatment based on a semi-structured clinical interview plus a thorough medical chart review for individual patients should have overcome this potential limitation satisfactorily. In summary, we have identified a genetic locus rs11127876 in CADM2 gene to be associated with early onset BPI. The finding has reflected the co-sharing genetic features of psychiatric disorders and behavioral traits. Further investigations of the CADM2 biological function in BPI are warranted.The fatty acid amide, palmitylethanolamide, was previously found to inhibit formalin-evoked pain behavior in Ž . rodents Calignano et al., 1998; Jaggar et al., 1998 . In the present study, we have further characterized the antinociceptive activity of this endogenous lipid molecule in several models of phasic and tonic pain. Our initial structure–activity relationship studies suggest that the ability of palmitylethanolamide to reduce formalin-evoked nociception may have distinct structural requirements, insofar as small variations in chemical structure were found to produce substantial losses of biological Ž activity. For example, myristylethanolamide shorthand . fatty acid designation, 14:0 and palmitoleylethanolamide Ž16:1D9 cis. displayed no significant antinociceptive activity, despite their close structural resemblance with Ž . palmitylethanolamide 16:0 . These findings, which need to be extended by testing a wider range of structural analogs, are consistent with the possibility that the effects of palmitylethanolamide are mediated by activation of a selective receptor site. This hypothesis is further supported by the ability of the cannabinoid CB receptor antagonist, 2 SR144528, to prevent palmitylethanolamide-evoked Ž . antinociception Calignano et al., 1998; present study . The relationship of the putative receptor activated by palmitylethanolamide with the cloned cannabinoid CB re- 2 ceptor subtype, if any, currently remains undetermined. There is general agreement that palmitylethanolamide does not productively interact with the cloned cannabinoid CB2 Ž . receptor Devane et al., 1992; Griffin et al., 2000 , a Ž negative finding that we have reproduced in our lab S. . Kathuria and D. Piomelli, unpublished observations . However, activation of cannabinoid CB receptors by the 2 selective agonist, HU308, was recently shown to inhibit Ž . pain in the formalin model Hanus et al., 1999 . Oleylethanolamide was found to exert a weak antinociceptive effect in the formalin test, which was reduced by systemic administration of either cannabinoid CB or CB 1 2 receptor antagonists. These results suggest that oleylethanolamide may reduce nociception through a dual mechanism. Oleylethanolamide may weakly interact with a receptor site sensitive to the cannabinoid CB receptor 2 antagonist SR144528. Moreover, by inhibiting anandamide Ž inactivation Desarnaud et al., 1995; Di Tomaso et al., ´ . 1996; Piomelli et al., 1999 , oleylethanolamide may cause anandamide to accumulate in the injected paw and activate local cannabinoid CB receptors. The possibility that 1 oleylethanolamide directly binds to and activates cannabinoid CB receptors is unlikely, because oleylethanolamide 1 Ž displays no affinity for these receptors in vitro S. Kathuria . and D. Piomelli, unpublished observations . In previous experiments, using an identical protocol, we failed to observe the antinociceptive effects of oleylethanolamide Ž .Ž 50 mg per animal, intraplantar in mice Calignano et al., . 1998 .