A Methodology and Application for Statistical Polarization in the Immune Response

Abstract

Abstract Polarization in the CD4+ mediated T-helper cell immune response is a very well characterized phenomenon in immunology; towards either Th1, Th2, or Th17 pathways. However, to-date, this concept has neither been developed nor demonstrated in a robust statistical framework. In this study, we provide a novel statistical methodology to facilitate the idea of polarization within a multivariate setting, and apply it to luminal fluid cytokine data of 90 women from the Peruvian Amazon. The panel of cytokines and chemokines for this study were assayed using separate BioRad Multiplex Array kits. In our application, we show that soil transmitted helminth (STH) infection induces statistically significant Th2 polarization in the female genital tract. To provide statistical evidence of Th2 polarization in STH-infected women at the cervix, we compare the clustering of cytokines among STH-infected versus STH-uninfected women while examining the shift in the strength of correlation for Th2 cytokines with IL-4, which is seen as a central Th2 cytokine. The proposed statistical test is one that tests if, significantly, k>1 for f(L1)=f(L2) where L1 and L2 are the loadings in the principal component analysis (PCA) framework for the STH-infected and STH-infected groups respectively. The proposed method is a confirmatory one and could be supported by visual representation of the Th polarization through heat maps. It has the potential to become the reference or gold standard when studying the polarization phenomenon in immunology as it tailors the clinical characteristics and mechanisms associated with the entire panel of cytokines and chemokines in the context of a system.