Correlating plasma protein profiles with symptomatology and treatment response in acute phase and early remission of major depressive disorder

Abstract

Objectives: This study aimed to explore the relationship between plasma proteome and the clinical features of Major Depressive Disorder (MDD) during treatment of acute episode.

Methods: In this longitudinal observational study, 26 patients hospitalized for moderate to severe MDD were analyzed. The study utilized Liquid Chromatography with Tandem Mass Spectrometry (LC-MS/MS) alongside clinical metrics, including symptomatology derived from the Montgomery-Åsberg Depression Rating Scale (MADRS). Plasma protein analysis was conducted at the onset of acute depression and 6 weeks into treatment. Analytical methods comprised of Linear Models for Microarray Data (LIMMA), Weighted Correlation Network Analysis (WGCNA), Generalized Linear Models, Random Forests, and The Database for Annotation, Visualization and Integrated Discovery (DAVID).

Results: Five distinct plasma protein modules were identified, correlating with specific biological processes, and uniquely associated with symptom presentation, the disorder's trajectory, and treatment response. A module rich in proteins related to adaptive immunity was correlated with the manifestation of somatic syndrome, treatment response, and inversely associated with achieving remission. A module associated with cell adhesion was linked to affective symptoms and avolition, and played a role in the initial episodes and treatment response. Another module, characterized by proteins involved in blood coagulation and lipid transport, exhibited negative correlations with a variety of MDD symptoms and was predominantly associated with the manifestation of psychotic symptoms.

Conclusion: This research points to a complex interplay between the plasma proteome and MDD's clinical presentation, suggesting that somatic, affective, and psychotic symptoms may represent distinct endophenotypic manifestations of MDD. These insights hold potential for advancing targeted therapeutic strategies and diagnostic tools.

Limitations: The study's limited sample size and its naturalistic design, encompassing diverse treatment modalities, present methodological constraints. Furthermore, the analysis focused on peripheral blood proteins, with potential implications for interpretability.

Keywords: LC-MS/MS; biomarker in depression; immune response; major depressive disorder; plasma proteomics; symptom presentation; treatment response.

Figure 1

Volcano plot displaying changes in protein groups between week 0 and week 6. Protein groups that are significantly different (p<0.05) and have a log2(fold change) (Week 0/Week 6) of less than -0.5 are highlighted in red, and those with a log2(fold change) greater than 0.5 are highlighted in yellow. Each point represents an individual protein group. Points marked with the corresponding coding gene symbols represent the most significantly up- or down-regulated differentially expressed protein groups.

Figure 2

Heat map representation of module-trait relationships at week 0. Each column represents a module eigenvectors, and each row represents demographic/clinical traits. Each cell contains Pearson’s correlation and p-values. The color is coded by correlation with the legend on the right. BMI: body mass index, MADRS: total Montgomery–Åsberg Depression Rating Scale (MADRS) score, MADRS1-MADRS10: MADRS item 1-10 score, MADRS cluster A (Exaggerated reactivity to negative information): MADRS1+MADRS3+MADRS9+MADRS10 score, MADRS cluster B (Altered reward processing): MADRS8 score, MADRS cluster C (Deficits in cognitive control): MADRS6 score, MADRS cluster D (Somatic syndrome): MADRS4+MADRS5+MADRS7 score, MADRS dysphoria: MADRS1+MADRS9+MADRS10 score, MADRS retardation: MADRS2+MADRS6+MADRS7+MADRS8 score, MADRS vegetative: MADRS3+MADRS4+MADRS5 score.

Figure 3

Heat map representation of module-trait relationships at week 6. Each column represents a module eigenvectors, and each row represents demographic/clinical traits. Each cell contains Pearson’s correlation and p-values. The color is coded by correlation with the legend on the right. BMI: body mass index, MADRS: total Montgomery–Åsberg Depression Rating Scale (MADRS) score, MADRS1-MADRS10: MADRS item 1-10 score, MADRS cluster A (Exaggerated reactivity to negative information): MADRS1+MADRS3+MADRS9+MADRS10 score, MADRS cluster B (Altered reward processing): MADRS8 score, MADRS cluster C (Deficits in cognitive control): MADRS6 score, MADRS cluster D (Somatic syndrome): MADRS4+MADRS5+MADRS7 score, MADRS dysphoria: MADRS1+MADRS9+MADRS10 score, MADRS retardation: MADRS2+MADRS6+MADRS7+MADRS8 score, MADRS vegetative: MADRS3+MADRS4+MADRS5 score.

Figure 4

Heat map representation of module-trait relationships across week 0 and week 6. Each column represents a module eigenvectors, and each row represents demographic/clinical traits. Each cell contains Pearson’s correlation and p-values. The color is coded by correlation with the legend on the right. BMI: body mass index, MADRS: total Montgomery–Åsberg Depression Rating Scale (MADRS) score, MADRS1-MADRS10: MADRS item 1-10 score, MADRS cluster A (Exaggerated reactivity to negative information): MADRS1+MADRS3+MADRS9+MADRS10 score, MADRS cluster B (Altered reward processing): MADRS8 score, MADRS cluster C (Deficits in cognitive control): MADRS6 score, MADRS cluster D (Somatic syndrome): MADRS4+MADRS5+MADRS7 score, MADRS dysphoria: MADRS1+MADRS9+MADRS10 score, MADRS retardation: MADRS2+MADRS6+MADRS7+MADRS8 score, MADRS vegetative: MADRS3+MADRS4+MADRS5 score.