IS THE ERA OF CANDIDATE GENES X CANNABIS USE REALLY DEAD?

Background: Historically, gene X environment examinations in psychotic disorders have employed candidate gene methods and environmental determinants impacting on similar biological mechanisms. However, genome wide association studies (GWAS) show that many variants associated with schizophrenia have a modest effect size on risk. In this respect, it is unclear whether the effect of cannabis on psychosis phenotypes is modified by a few genes, e.g. those involved in dopamine signalling, or by the overall genetic susceptibility to schizophrenia. Indeed, candidate gene approaches might be complementary to GWAS to test gene X cannabis interaction. We aimed to investigate the interactive effects of cannabis use on variants in DRD2 gene on: 1) the risk of developing first episode psychosis (FEP); and 2) the frequency of positive symptoms at FEP. We undertook a replication study of previous Gene X Cannabis interactions for DRD2 (rs1076560); further, we tested interactions between cannabis use and any SNP associated with schizophrenia within DRD2 gene in the last Psychiatric Genomics Consortium (PGC) GWAS. Methods: We genotyped ∼830 FEP patients and ∼1200 controls recruited across six countries as part of the large EUGEI study. OPerational CRITeria system and Cannabis Experience Questionnaire were used for evaluating psychopathology and patterns of cannabis use. Dimensions of psychopathology, which included a specific dimension of positive symptoms, were estimated using multidimensional item response modelling in Mplus. We tested for an interaction between risk allele count and daily cannabis use on: 1) the risk of psychotic disorder in the case-control study; 2) the positive symptom dimension in the case-only sample. Only one SNP in DRD2 was significantly associated with schizophrenia in the PGC study, so we tested for interaction with rs2514218, in addition to rs1076560. These regression models, conducted in STATA 14, were adjusted for age, gender, ethnicity, 10 principal components (PC) for population stratification and SNP-environment and SNP-PC interaction terms. Results were corrected for multiple testing of four SNPs (p<0.0125 as significance threshold). Results: We found a significant interaction between cannabis use and the rs2514218 in DRD2: daily cannabis users with one risk allele showed a 2.5-fold increased probability to suffer a psychotic disorder (OR = 2.43; 95% confidence interval [CI]: 1.52–3.89). We did not replicate previous findings of an interaction between cannabis use and rs1076560 in DRD2 for increasing psychosis risk (p=0.092). Higher levels of positive symptoms were seen in FEP who were daily cannabis users with either rs2514218 (B=0.34; 95% CI: 0.04–0.64; p=0.024) or rs1076560 risk alleles (B=0.54; 95% CI: 0.15–0.93; p=0.006). Discussion: Our results suggest that dopamine signaling is implicated in cannabis associated psychosis, and, more specifically, that variation within the DRD2 gene may modulate the effect of cannabis use on psychosis phenotypes. Such findings require a replication for rs2514218.