Comments on “A Cross-sectional Trait Versus State Biomarker Analysis of Inflammatory Cytokines and miRNAs in Patients with Major Depressive Disorder”

Dear Editor,

Depression, as we know, is a complex condition often linked to immune system dysregulation, and its etiopathogenesis is potentially influenced by inflammation and deranged immune markers. Immune markers influence symptom severity and underlying vulnerability and serve as both state and trait biomarkers, respectively. ¹ Rockson et al. ² evaluated state versus trait biomarker status for interleukin (IL) levels- IL-6, IL-1β, and IL-10, tumor necrosis factor alpha, high sensitivity C-reactive protein, gene expressions of plasma micro-ribonucleic acids (miRNAs)–miR-16, miR-132, and miR-1202, total oxidant-antioxidant status in patients with major depressive disorder (MDD) compared to their first-degree relatives (FDRs) and healthy controls (HCs). This work marks a significant effort in expanding the study of immune biomarkers beyond IL to include miRNAs and oxidant levels. Using multiple linear regression, they found statistical significance for increased IL-6 levels and reduced miRNA-1202 expression in the model that compared MDD and HCs (model-1). However, the model comparing FDRs and HCs (model-2) did not show statistical significance. Based on this statistical modeling, they conclude that these immune markers are state markers but not trait markers.

While we endorse the strengths of the study that the authors state, we raise specific critiques on the observations made in this study. First, the results of the regression analysis (models 1 and 2) contrast with those of the univariate analysis (Table 2), in which IL-6 levels are comparable between MDD and HCs and lower in the FDRs. When viewed more closely, the contrast between univariate and regression results seems to have been driven by the gender differences (lower female-male ratio in FDRs) across the three groups corrected for in the regression models. Literature supports that females exhibit increased levels of IL-6 compared with males, attributed to greater monocyte expression. ³ This aspect needs more discussion.

Second, the regression model plan, especially its sequence of it, also needs a close relook from the perspective of the consecutive hurdle method for identification of endophenotype or trait markers, proposed by Irving Gottesman and colleagues.^4,5According to them, ⁴ to be defined as an endophenotype or trait-marker, that particular marker, which is found in the affected population (or in other words, it is significantly different in the affected population compared with healthy individuals), has to be “found in nonaffected family members at a higher rate than in the general population,” before proceeding for further characterization. Therefore, when model-2 (FDRs vs. HCs) was found to be statistically not significant, which was the case for all variables assessed, interpreting data using model-3 (MDD vs. FDRs) may not have been necessary. As per the ‘‘Gottesman criteria,’’ statistical significance in model-1 (MDD vs. HCs) is sufficient for the ‘‘state marker’’ interpretation, apart from being used for ‘‘assay sensitivity’’ as deemed by the authors.

Interestingly, however, the ‘‘Gottesman criteria’’ were met for the total oxidant status (TOS). Table 2 shows that both model-1 (β = −7.07, p = .003) and model-2 (β = −13.93, p = .01) were statistically significant. Further, TOS also showed statistical significance in model-3, implying the need for further discussion. The authors did not highlight this result. More importantly, the MDD group was found to have the lowest TOS, suggesting reduced inflammation. This unexpected finding contradicts the study’s hypothesis that excessive oxidative stress and diminished antioxidant defenses are typically associated with increased inflammation and pathogenesis of depression. This point is indeed ‘‘unusual’’ as endorsed by the authors but requires more discussion, which is lacking in the article.

We also raise three minor comments: one relates to the selection of specific ILs in the study. A meta-analysis by Osimo et al. ⁶ highlighted the involvement of nearly 14 ILs in the etiology of depression. Justifying the selection of the ILs would have further enhanced the study’s credentials. This also holds true for the selection of miRNAs. Perhaps the ILs selected for this study are marked as “commonly examined inflammatory markers” in a recent meta-analysis. ⁷ Second, the fact that only 31 of the expected 50 FDRs were included for analysis lowers the power of the study results, particularly about the trait markers. A supplementary pair-wise analysis of MDD-FDRs for the 31 pairs for which the data is available might provide further insights. Third, the article would benefit from a more detailed description of sample collection, storage, and analysis techniques, thereby providing opportunities for accurate reproduction of the method in future studies and averting speculations regarding its accuracy.

Overall, Rockson et al. offer important insights into immune biomarkers, which can help improve diagnostic accuracy, predict severity, and monitor treatment responses.