T38 - Preliminary Model For The Genetic Prediction Of Clozapine Response

Abstract

Background Despite its clinical utility, the antipsychotic clozapine (CLZ) has not yet had a promising pharmacogenetic test developed for prediction of response. In the past 20 years, multiple genetic variants (e.g. dopamine, serotonin) have been suggested to be associated with CLZ response. Arranz et al (2000) proposed a model to predict response, but it was not replicated in independent samples. Since then, no other studies have sought to create a genetic panel predicting response to CLZ. Thus, we reinitiated the effort to develop a genetic model for CLZ response incorporating the most promising findings from our group’s repository of CLZ response studies. Methods Our sample consisted of 151 Caucasian subjects with schizophrenia (SCZ) (DSM-III) treated with CLZ for six months. Response was assessed using the Brief Psychiatric Rating Scale (BPRS), and evaluated using absolute score change and binary response (Kane et al. 1988 criteria), with baseline score as a covariate. A total of 99 polymorphisms were tested from a range of candidate genes. Variants showing at least a nominal statistical trend (p Results Four markers from genes encoding for dopamine D2 receptor (DRD2), serotonin-6 receptor (5-HT6), brain-derived neurotrophic factor (BDNF), and neurexin-1 (NRXN1) were included in the model. We observed a statistically significant association between genetic risk score with BPRS score change (p=0.000039, Adjusted R^2=0.565) and binary response (p=0.004, Nagelkerke R^2=0.097) assuming a linear increase in response for each additional risk allele. The model had an accuracy of 62%, a sensitivity of 70%, and a specificity of 47%. The model was not significantly associated with response in the independent CATIE European Caucasian subsample treated with other antipsychotics (p=0.10). Discussion We have developed a preliminary genetic model for CLZ response that includes genes with strong rationale for involvement. NRXN1 is a synaptic membrane cell-adhesion protein that has been suggested to modulate NMDA receptor activity, which is indirectly regulated by CLZ. The 5-HT6 receptor is involved in neurite growth and has lower expression in the hippocampus of SCZ patients compared to healthy controls. It mediates cognitive function, anxiety, and positive symptom improvement in animal models of SCZ. The model does not appear to generalize to other antipsychotics. For CLZ response prediction per se, replication in independent studies of CLZ response in SCZ is required to confirm the validity of these findings.