A pilot genome-wide association study for exploring clinical response in schizophrenia patients treated with clozapine

106

Introduction Schizophrenia is a chronic psychiatric disease with complex neurodevelopmental and genetic origins. Clozapine has better clinical response on reduction of psychotic symptoms than other type of antipsychotics in treatment-resistant patients with a response rate ranged from 40% to 70%. Although many studies focused on genetic predictors of clozapine response in schizophrenia, the studies were based on candidate gene approach rather than genome-wide association study (GWAS); the result may be either non-reproducible or minimal associated. Thus, the aim of this study was (1) to identify novel susceptibility loci for a fair clinical response to clozapine, i.e., symptomatic remission using GWAS; (2) to develop a predictive model using single nucleotide polymorphisms (SNPs) singled out from GWAS, and to forecast a fair clinical response to clozapine in schizophrenia patients. Material and Methods All of 462 schizophrenia patients were enrolled. We clustered the subjects into two groups according the clinical symptoms, i.e., symptomatic remission and non-remission, which are assessed with the Chinese Mandarin version of the Positive and Negative Syndrome Scale (CMV-PANSS). Subjects with symptomatic remission include those with full remission and partial remission on clozapine for longer than 3 months. The platform of Axiom Genome-Wide TWB Array which specially designed for Han Chinese residing in Taiwan was used. Three stages of analysis were carried out: Discovery phase: all of 279 subjects were for discovery of related SNPs. For the whole genome association scan, we used additive model logistic regression with PLINK software to control for confounders such as age, sex, daily dose of clozapine and the presence of use of mood stabilizer. SNPs related to response to clozapine were singled out with a p-value less than 10-4. Replication phase: all of 183 subjects were for replication which focused on SNPs singled out from discovery phase. Joint analysis: 462 subjects were tested for the association with the SNPs singled out from discovery phase. To get a model predicting response to clozapine, we used a propensity score model, in which age, sex, daily dosage of clozapine, presence of use of mood stabilizers, and SNPs singled out from GWAS were included and formed a propensity score. The value of specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV), and accuracy to predict the clinical response to clozapine were obtained Results The mean age was 52.9 years. The majority of participants were male (77.3%). The percentage of patients with a fair response to clozapine was 31.2% (144/462). All of 42 SNPs were singled out from discovery phase. There were no SNPs significantly associated with a fair response to clozapine in replication phase. In joint analysis, 5 SNPs were singled out. The putative loci which had nominally significant association with the response to clozapine were as follows: rs60930171 within the gene of vav guanine nucleotide exchange factor 1 (VAV1) on chromosome 19 (p = 3.48E-05), rs17286495 within the gene of chromosome 20 open reading frame 196 (C20orf196) on chromosome 20 (p = 4.20E-05), rs6581362 on chromosome 12 (p = 5.79E-05), rs1103659 on chromosome 16 (p = 8.37E-05), and rs184952 on chromosome 19 (p = 9.21E-05). Concerning the predictive model, for a given individual, if the propensity score > 0.291, then the model was predictive of a fair response to clozapine. The value of specificity was 0.7, sensitivity 0.72, PPV 0.71, NPV 0.72, and accuracy 0.71. Discussion and conclusion Our findings revealed that rs60930171 within VAV1, rs17286495 within C20orf196, and other 3 SNPs were nominally associated with a fair clinical response in schizophrenia patients treated with clozapine, and the VAV1 as well as C20orf196 might play a key role in the treatment of refractory schizophrenia. Out study has paved the way to use GWAS to predict clinical response to clozapine in schizophrenia patients, as well as providing clues for further studies exploring novel treatment targets and elucidating the molecular mechanisms of antipsychotics in the treatment of schizophrenia. For obtaining accurate genetic variants related to clinical response to clozapine in patients with schizophrenia, more studies such as replication incorporating the database in current research, fine mapping and functional analysis of putative loci found in current study, are warranted in the future.