Pharmacogenomics and Prediction of Precise Treatment Outcomes in Major Depressive Disorder

Abstract

Selective serotonin reuptake inhibitors (SSRIs) are the most commonly used medication class for major depressive disorder (MDD); however, response to SSRI treatment varies considerably between patients. A number of genome-wide association studies (GWAS) of SSRI treatment outcomes have been performed with the goal of identifying genetic variants associated with SSRI response. The hope is that knowledge of the genetic variants contributing to treatment outcomes will facilitate more precise estimation of each patient’s probability of having a favorable treatment outcome, thus allowing the physician to select an appropriate treatment plan – This concept of precise estimation of treatment outcomes is at the heart of precision medicine. However, the prior GWAS of antidepressant response have had limited success, with no replicated genome-wide significant findings. Moreover, despite all the attempts to identify genetic variants to enable more accurate prediction of treatment response, prior studies have largely used broad (and thus imprecise) definitions of treatment outcomes, typically based on total scores on depression rating scales. We postulate that different genetic factors may play a role in how specific symptoms of depression change in response to a particular medication. For example whether a person’s mood improves after treatment may be influenced by different genes than whether a person’s sleep quality improves after treatment. Specific components of depression and antidepressant response can be derived from individual items assessed by depression rating scales such as the Hamilton Rating Scale for Depression (HRSD). By viewing depression as a multifaceted illness, and response to antidepressants as a complex, multi-component treatment outcome, we aim to identify genomic contributors to more precisely-defined aspects of treatment response, including items such as depressed mood, anxiety, interests/activity, and insomnia. We illustrate this concept using data from several existing studies of SSRI response to identify novel pharmacogenomic effects on specific aspects of treatment outcomes, demonstrating that decomposition of the total scores into specific components can increase the power of pharmacogenomics studies of antidepressants and provide more clinically useful information.