Multidimensional scaling in Euclidean space of cytokine responses to document hyperinflammation in cystic fibrosis cells

Introduction: Studies of hyperinflammation in Cystic Fibrosis (CF) frequently yield complex high-dimensional data due to comparisons of CF cells with wildtype or treated controls, and multiple cytokine analyses. This warrants utilization of computational approaches beyond traditional statistics. We previously (Pediatr Pulmonol 2020;55/S2:56) utilized unsupervised clustering to assess a single-cytokine response in CF cells over two culture conditions. The Aim: of this study was to evaluate a composite cytokine response in CF cells under two culture conditions, in comparison with a known hyperinflammatory control. Methods: CFF16HBEge (delF508 CFTR) and 16HBE14o- (wildtype CFTR) were the cell models, tested either under submerged conditions or after differentiation at the air-liquid interface. The BEAS2B cells (wildtype CFTR) provided hyperinflammatory control data points, as per our previous work (Am J Respir Crit Care Med 2020;201:A2918). Basal and stimulated secretions of nine cytokines were quantified by multiplex ELISA (Millipore). The combined dataset, including the hyperinflammatory control, comprised 720 data points. All data points were log2 transformed and scaled to yield an Euclidean distance matrix for the multidimensional scaling (MDS) analysis. Results: The MDS analysis demonstrated proximity (i.e., similarity) of the composite inflammatory response of differentiated CFF16HBEge to that of the hyperinflammatory control. Conclusion: The developed workflow and MDS analysis support the presence of hyperinflammation in differentiated CF airway epithelial cells. Funding: Internal funds, DZL, Krieble Found., Gosselin Found., RSR-QC. ND, WA, ACH: co-first authors