Response to immune-based augmentation treatment for depression: a potential role of immunosenescence

Immunological processes are increasingly in focus as factors contributing to major depressive disorder (MDD). Especially subtypes such as immunometabolic depression phenotypes have been linked to low-grade inflammation. Augmentation with immune-based therapies, e.g., celecoxib, is being tested for its efficacy in treating depression. Many physiological processes during life are also linked to immunological changes, particularly aging. As results from current trials with celecoxib augmentation remain inconclusive, we tested the hypothesis that age affects treatment efficacy. In total, 113 individuals with a diagnosis of major depressive disorder (M _age = 44, 56% women, M _MADRS = 27.7) had biomarkers available and were included in our analyses. Patients were recruited as part of a randomized controlled trial (RCT) and stratified by hsCRP (>3 mg/L or < 3 mg/L). All patients were treated with vortioxetine and randomized to receive either celecoxib (N = 55) or placebo (N = 58) by a randomized design. Patients were treated for 6 weeks (M _{MADRS 6W} = 20.2). We tested one main hypothesis: that age affects the treatment outcome of celecoxib as an immunomodulatory agent. To further explore the meaning of our results, we used epigenome-wide DNA methylation data (Illumina Infinium MethylationEPIC 850k BeadChip) available to estimate cell type compositions of neutrophils, monocytes, B-cells, CD4⁺ and CD8⁺ Lymphocytes, and natural killer cells (NK), using the Houseman method. The analyses were performed using linear regression and ANCOVA models, corrected for sex, hsCRP, years of education, BMI, and depression severity at baseline. Our analysis showed a statistically significant interaction between age and treatment condition on depression outcome (beta = -0.24; p = 0.045), with significant main effects for both variables in the model (intervention: beta = 10.88; p = 0.048, age: beta = 0.23; p = 0.01). Sex, hsCRP, and years of education were no statistically significant contributors. BMI was a marginally significant contributor in the model (beta = -0.25; p = 0.059). The intersection was identified at 45.5 years. Younger individuals treated with celecoxib showed a more pronounced reduction in MADRS than older individuals treated with celecoxib (F(1,52) = 5.74, p = 0.020). Further exploration showed that for individuals younger than 45 years, neutrophils at baseline might be associated with better treatment outcome (beta = 25.87; p = 0.051). For individuals older than 45 years, this was the case for B-cells, NK cells, and, suggestively, CD8⁺ cells (beta = -146.29; p = 0.014, beta = 129.39; p < 0.001, and beta = -20.23; p = 0.055, respectively). Our results indicate that response to celecoxib as an augmentation treatment for depression can be age-dependent, with younger patients responding better to treatment. In addition, immunological cell type profiles at baseline differ between both age groups, supporting the hypothesis that changes in treatment efficacy might follow from immunological changes that happen during aging. Our result fits with the growing body of literature focusing on immunosenescence and that aging of the immune system is relevant for treatment response and personalized treatment choices. Replication in an independent sample is needed to confirm the role of age in immune-focused treatment strategies for depression.