Precision medicine for depression: identifying multi-Omic, immunological markers for treatment-resistant depression

Background

Major Depression is predicted to be the leading cause of disability by 2030. While antidepressants are frontline treatments, approximately 30% of MDD patients are treatment resistant [1]. This creates significant stress and uncertainty for individual well-being, healthcare costs and the global labour market. Researchers have limited understanding of why some individuals fail to respond to antidepressants. Novel molecular findings are urgently needed to develop targeted treatment for these vulnerable patients.

The immune-psychiatry hypothesis suggests that chronic immune system overactivation in some people with depression could predict a poorer response to treatment, as well as a higher risk of suicide and longer hospital stays [2]. By using a multi-omics approach to study these immune markers, we can potentially identify patients with treatment-resistant depression earlier, paving the way for targeted interventions and the development of new drugs. To achieve this, several challenges remain to be addressed.

First, depression is heritable and substantial progress has been made recently in using polygenic scores to predict depression and symptom severity. However, it remains unknown whether polygenic scores can be used to predict treatment resistance, and specifically, whether scores based on variants within immune pathways may predict treatment response.

Second, we have previously identified inflammatory protein markers associated with antidepressant exposure [3]. These findings indicate that plasma protein abundance is highly sensitive to antidepressant use. Yet, it's unclear whether inflammatory proteomic markers can help us understand treatment resistance, or further, help identify individuals who may be treatment resistant.

Third, our studies have shown that genomic functions are determined by multiple factors including DNA methylation, protein abundance and metabolomic levels [4]. Integrating multi-omic datasets can help capture the complexity of antidepressant response from upstream genetic underpinning to downstream metabolism. Recent advances in machine learning for genomic data offer new ways to predict antidepressant response using multi-omic, inflammatory markers. Previously, this research was constrained by limited datasets. Now, the availability of population-scale, multi-omic data in, for example, Generation Scotland and UK Biobank, creates promising new research opportunities.