Adjunctive minocycline for major depressive disorder: A sub-study exploring peripheral immune-inflammatory markers and associated treatment response

Abstract

Background: Adjunctive minocycline shows promise in treating affective and psychotic disorders; however, the therapeutic mechanism remains unclear. Identifying relevant biomarkers may enhance the efficacy of novel adjunctive treatment candidates. We thus investigated the peripheral immune-inflammatory profile in a randomized controlled trial (RCT) of minocycline in major depressive disorder (MDD).

Methods: This sub-study investigated serum samples from a RCT evaluating minocycline (200 mg/day, 12 weeks) in addition to treatment as usual for MDD (ACTRN12612000283875). Of the original sample (N = 71), serum assays were conducted in 47 participants (placebo n = 24; minocycline n = 23) targeting an array of 46 immune-inflammatory analytes including cytokines, chemokines, and acute-phase reactants. General estimating equations (GEE) were used to assess whether analyte concentration at baseline (effect modification) and change in analytes (change association) influenced change in Montgomery-Åsberg Depression Rating Scale (MADRS) score over time. The Benjamini-Hochberg approach was applied when adjusting for false discovery rates (FDR).

Results: GEE models revealed several interaction effects. After adjusting for FDR several change association-models survived correction. However, no such models remained significant for effect modification. Three-way group × time × marker interactions were significant for complement C3 (B = -10.46, 95%CI [-16.832, -4.095], q = 0.019) and IL-1Ra (B = -9.008, 95%CI [-15.26, -2.751], q = 0.036). Two-way group × biomarker interactions were significant for ICAM-1/CD54 (B = -0.387, 95%CI [-0.513, -0.26], q < 0.001) and IL-8/CXCL8 (B = -4.586, 95%CI [-7.698, -1.475], q = 0.036) indicating that increases in the serum concentration of these analytes were associated with an improvement in MADRS scores in the minocycline group (compared with placebo).

Conclusions: Change in complement C3, IL-1Ra, IL-8/CXCL8, and ICAM-1 may be associated with greater change in depressive scores following adjunctive minocycline treatment in MDD. Further investigations are needed to assess the utility of these biomarkers.

Keywords: Biomarkers; Depression; Inflammation; Minocycline; Neuroscience; Psychiatry.

Fig. 1

Pearson's correlations heatmap. Negative correlation coefficients represented by red shading; positive coefficients represented by blue shading. (A)baseline analyte vs. baseline MADRS score - A positive coefficient suggests higher marker concentration at baseline correlated with higher initial depression score; a negative coefficient suggests lower marker concentration at baseline correlated with higher depression score. (B)baseline analyte vs. change in MADRS score- A positive coefficient suggests a lower concentration of analyte at baseline correlated with improvement in MADRS between baseline and week 12; a negative coefficient suggests higher baseline marker correlated with improvement in MADRS. (C)change in marker vs. endpoint MADRS score - A positive coefficient suggests an increase in marker concentration (change between baseline and week 12) correlated with worse depression score at end; a negative correlation suggests that an increase in marker concentration between baseline and week 12 correlated with better depression score at endpoint. (D)Change in marker vs. change in MADRS score - A positive correlation coefficient indicates that a reduction in marker concentration correlated with improvement in depression score between baseline and week 12; a negative coefficient indicates that an increase in marker concentration (change) correlated with improvement in depression score. *p < 0.05; **p < 0.01, unadjusted. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

(A) Mean MADRS total scores ± standard deviation over time for this cohort subset. (B–J). Box and whiskers plots (min to max) for notable markers following analyses using GEE models. In all graphs the placebo group represented by blue shading, and minocycline group represented in red shading. Full descriptive statistics provided in Table S4. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)