Metabolomic Signatures Predictive of Outcomes to Treatments for Major Depression



 DESCRIPTION (provided by applicant): Major Depressive Disorder (MDD) is a significant public health problem worldwide. However, despite the availability of medication and psychotherapeutic treatments for MDD, fewer than 40% of patients achieve remission after initial treatment. It would represent a major advance for Precision Medicine if we were able to individualize the therapy of MDD. Recent advances in analytical chemistry have led to the emergence of Metabolomics, a discipline that allows the simultaneous measurement of 100's to 1000's of metabolites to map perturbations in metabolic pathways and networks, thus potentially enabling a systems approach to the study of MDD and its treatment. Our work over the past decade has pioneered the application of metabolomics to study selective serotonin reuptake inhibitors (SSRIs). We have mapped metabolic pathways implicated in SSRI response and discovered novel mechanisms associated with that response. In this proposal, we set out to apply metabolomics to greatly expand our knowledge of MDD treatment response by the use of previously collected samples and comprehensive clinical data from two large studies, the Emory PReDICT and the Mayo Pharmacogenomics Research Network (PGRN) trials. Both of these independent studies used the SSRI escitalopram and the serotonin-norepinephrine reuptake inhibitor duloxetine as treatments. The PReDICT study also included a non-pharmacologic treatment arm, cognitive behavior therapy (CBT). Our goal is to leverage these large investments made by the NIH by applying an integrated metabolomics- genomics- neuroimaging approach to characterize and functionally validate the biological systems predictive of MDD treatment outcomes. In Specific Aim 1, we will define metabolomic signatures of exposure to the 3 therapies in the treatment-naïve MDD patients from the PReDICT study. In Aim 2, we will evaluate and model the metabolomic signatures associated with improvement during treatment with escitalopram, duloxetine, and CBT, and then replicate our findings in the Mayo study. In Aim 3 we will use "pharmacometabolomics-informed pharmacogenomics" both to compare the biomarkers discovered in the Mayo study with those identified in PReDICT, and to identify the metabolite-associated genes and single nucleotide polymorphisms involved in mechanisms associated with variation in treatment response using cell-line based systems. Finally, our Exploratory Aim will examine linkages between central nervous system function and peripheral metabolomic signatures. This proposal is innovative because it applies the novel tool of metabolomics to the question of treatment selection in MDD, and because it integrates metabolomics with genomics and neuroimaging data to enable a deeper understanding of therapeutic mechanisms of action. It is also highly significant because it will add to evidence-based methods for the selection of optimal treatments for individual MDD patients and will expand our understanding of biological mechanisms underlying response to the therapy for this major disease.