Cytokine correlation analysis based on drug perturbation

Abstract

Cytokines and chemokines play a crucial role in regulating the immune system. Understanding how these molecules are co-regulated is important to understand general immunology, and particularly their role in clinical applications such as development and evaluation of novel drug therapies. Cytokines are today widely used as therapeutic targets and as biomarkers to monitor effects of drug therapies and for prognosis and diagnosis of diseases. Therapies that target a specific cytokine are also likely to affect the production of other cytokines due to their cross-regulatory functions and because the cytokines are produced by common cell types. In this study, we have perturbated the production of 17 different cytokines in a preclinical rat model of autoimmune arthritis, using 55 commercially available immunomodulatory drugs and clinical candidates. The majority of the studied drugs was selected for their anti-inflammatory role and was confirmed to inhibit the production of IL-2 and IFN-γ in this model but was also found to increase the production of other cytokines compared to the untreated control. Correlation analysis identified 58 significant pairwise correlations between the cytokines. The strongest correlations found in this study were between IL-2 and IFN-γ (r=0.87) and between IL-18 and EPO (r=0.84). Cluster analysis identified two robust clusters: (1) IL-7, IL-18 and EPO, and (2) IL-2, IL-17 and IFN-γ. The results show that cytokines are highly co-regulated, which provide valuable information for how a therapeutic drug might affect clusters of cytokines. In addition, a cytokine that is used as a therapeutic biomarker could be combined with its related cytokines into a biomarker panel to improve diagnostic accuracy.

Keywords: Clustering; Compound; PCA; Profile; T cell.