Transcriptomic signaling pathways involved in a naturalistic model of inflammation-related depression and its remission

Abstract

This study aimed at identifying molecular biomarkers of inflammation-related depression in order to improve diagnosis and treatment. For this, we performed whole-genome expression profiling from peripheral blood in a naturalistic model of inflammation-associated major depressive disorder (MDD) represented by comorbid depression in obese patients. We took advantage of the marked reduction of depressive symptoms and inflammation following bariatric surgery to test the robustness of the identified biomarkers. Depression was assessed during a clinical interview using Mini-International Neuropsychiatric Interview and the 10-item, clinician-administered, Montgomery-Asberg Depression Rating Scale. From a cohort of 100 massively obese patients, we selected 33 of them for transcriptomic analysis. Twenty-four of them were again analyzed 4-12 months after bariatric surgery. We conducted differential gene expression analyses before and after surgery in unmedicated MDD and non-depressed obese subjects. We found that TP53 (Tumor Protein 53), GR (Glucocorticoid Receptor), and NFκB (Nuclear Factor kappa B) pathways were the most discriminating pathways associated with inflammation-related MDD. These signaling pathways were processed in composite z-scores of gene expression that were used as biomarkers in regression analyses. Results showed that these transcriptomic biomarkers highly predicted depressive symptom intensity at baseline and their remission after bariatric surgery. While inflammation was present in all patients, GR signaling over-activation was found only in depressed ones where it may further increase inflammatory and apoptosis pathways. In conclusion, using an original model of inflammation-related depression and its remission without antidepressants, we provide molecular predictors of inflammation-related MDD and new insights in the molecular pathways involved.

Fig. 1. Gene expression analyses at baseline and after bariatric surgery in MDD and non-depressed patients.

A Heatmap of the results of the sparse least square analysis (sPLS) on the 33 patients at baseline showing the strength of correlation between the 2000 gene probes the most correlated to depression (color code at the top left); each line corresponds to a gene probe and in each column a phenotypic value of depression. B Venn diagrams of the three differential expression analyses showing significantly differentially expressed genes.

Fig. 2. Upstream regulator analyses using Ingenuity Pathway Analysis.

A Mechanistic network connecting upstream regulators found in the differential expression analysis between MDD patients after versus before surgery. B Comparison analysis of upstream regulators in the three differential expression analyses. Upstream regulators are sorted by activation z-score with color code at the top of the figure.

Fig. 3. TP53, NR3C1, and RELA pathways differentiate MDD (orange symbols) from non-depressed patients (blue symbols) at baseline and are normalized after surgery.

Two-way ANOVA followed by Bonferroni post hoc tests, ***p < 0.001.

Fig. 4

TP53, NR3C1, and RELA pathways predict the intensity of inflammation-related depressive symptoms at baseline (A) and their decrease after bariatric surgery (B). Multivariate linear regression analyses. MDD patients (orange symbols), non-MDD patients (blue symbols).