Genetic variants in COMT and ESR1 genes shape treatment response to raloxifene in schizophrenia-spectrum disorders

Background/objective: Raloxifene, a selective estrogen receptor modulator (SERM), may improve symptoms and cognition in schizophrenia spectrum disorders (SSD). Studies have shown inconsistent efficacy, especially in men with SSD. We assessed whether single nucleotide polymorphisms (SNPs) on genes involved in the pharmacodynamics (ESR1 and COMT) and pharmacokinetics (UGT1A8) of raloxifene can explain the heterogeneous treatment response to raloxifene augmentation in patients with SSD. Methods: We used a subsample of the participants of a previously published randomized controlled trial (RCT) on the effects of 12-week raloxifene augmentation on symptom severity in SSD. The subsample consisted of 83 participants (28 % female), of which 40 were randomized to receive raloxifene 120 mg/day and 43 to placebo. Saliva samples for DNA-analysis were collected at baseline, symptom severity was measured with the Positive and Negative Syndrome Scale (PANSS). Participants were genotyped for two SNPs on ESR1, one on UGT1A8, and four on COMT using the Agena MassArray system. Linear mixed-effect models were used to assess the effect of treatment-by-genotype as the primary analysis and treatment-by-genotype-by-sex as a secondary analysis. Results: We found interactions of treatment-by-genotype for ESR1 rs2234693 (chi(2) = 6.32, p < 0.05), and COMT rs4818 (chi(2) = 4.08, p < 0.05), indicating that for these polymorphisms, the effect of raloxifene differed per genotype. Pairwise comparisons revealed a beneficial effect of raloxifene on general symptom severity in participants with ESR1 rs2234693 TT genotype but not CT and CC genotypes (LSM -3.19 [95 % CI -6.38-0.00]; p = 0.050). Furthermore, mean change in positive symptom severity was greater with raloxifene in participants with COMT rs4818 CG genotype but not CC genotype compared to placebo (LSM -2.18 [-3.93 to -0.43]; p = 0.016). Secondary sex-specific analysis indicated an interaction effect of treatment-by-genotype-by-sex for COMT rs737865 on total (chi(2) = 10.90, p < 0.05) and negative symptom severity (chi(2) = 11.99, p < 0.05). In men, genotype CT but not TT was associated with beneficial effects of raloxifene on total symptoms (LSM -5.46 [-10.43 to -0.48]; p = 0.032), whereas in women, genotype TT but not CT was associated with a beneficial effect of raloxifene on negative symptoms (LSM -7.80 [-12.70 to -2.89]; p = 0.005). Conclusion: Our results suggest that treatment response to raloxifene may depend on ESR1 and COMT gene variants, while UGT1A8 SNP variation did not affect treatment response. These findings provide evidence that genetic variants may explain the heterogeneous response to raloxifene augmentation in SSD, suggesting that raloxifene may have beneficial effects in genetic subgroups of SSD patients. Our findings warrant further research on the pharmacogenetic effects of raloxifene in SSD.