Deep immunophenotyping reveals that auto-immune and auto-inflammatory disorders are spread along two immunological axes capturing disease inflammation levels and types
TCHITCHEK N. 1,2, ROUX A. 1,2, BINVIGNAT M. 1, PITOISET F. 1,2, DUBOIS J. 1,2, MARGUERIT G. 1,2, SAADOUN D. 1,2,3, CACOUB P. 1,2,3, SELLAM J. 1,2,4, BERENBAUM F. 1,2,4, HARTEMANN A. 1,2,5, AMOUYAL C. 1,2,5, LORENZON R. 1,2, MARIOTTI-FERRANDIZ E. 1,2,6, ROSENZWAJG M. 1,2, KLATZMANN D. 1,2
1 Sorbonne Université, INSERM, Immunology-Immunopathology-Immunotherapy (i3), Paris, France; 2 Assistance Publique Hôpitaux de Paris, Hôpital Pitié-Salpêtrière – Charles Foix Hospital, Clinical Investigation Center for Biotherapies (CIC-BTi) and Immunology-Inflammation-Infectiology and Dermatology Department (3iD), Paris, France; 3 Assistance Publique Hôpitaux de Paris, Pitié-Salpêtrière – Charles Foix Hospital, Department of Internal Medicine and Clinical Immunology and Immunology-Inflammation-Infectiology and Dermatology Department (3iD), Paris, France; 4 Rheumatology Department, Sorbonne University, Inserm U938, Saint-Antoine University Hospital, AP–HP center, Paris, France; 5 Diabetology-Metabolism Dpt, Sorbonne Université, APHP, Institut Hospitalo-Universitaire de Cardiometabolisme et Nutrition (ICAN), Pitié-Salpêtrière-Charles Foix Hospital, Paris, France; 6 Institut Universitaire de France (IUF), Paris, France
Objectives: Auto-immune and auto-inflammatory diseases (AID) are complex disorders in which the body's immune system mistakenly attacks healthy tissues. A study from McGonagle and McDermott using genetic association suggested that these conditions form a continuum ranging from pure auto-immune to pure auto-inflammatory diseases (McGonagle et al., 2006). We propose here to evaluate such a hypothesis using deep immunophenotyping.
Methods: We collected blood from 443 patients having 15 different AIDs (such as arthritis, bowel, metabolism, muscle, or vasculitis -associated diseases) and from 72 healthy volunteers. Deep phenotyping of blood immune cells was performed using 13 flow cytometry panels of 10 cell markers, each covering more than 800 innate and adaptive cell parameters (Pitoiset et al., 2018). Principal component analysis (PCA) and classification decision tree analyses were conducted to identify clusters of diseases along with their specificities and commonalities.
Results: We found that AIDs were categorized into five clusters, which gathered diseases by inflammation levels and type, and their spread followed two immunological axes captured by the PCA. The first axis was determined by the ratio of LAG3+ and ICOS+ in Tregs, and segregated diseases based on inflammation levels. The second axis was determined by the increase of activated Treg subsets and the decrease of ILC3s, and segregated diseases based on their type or localization of immune dysregulations. Finally, we identified a signature consisting of 30 cell populations that accurately characterized the identified disease clusters.
Conclusions: Rather than a mono-dimensional continuum of diseases characterized by innate and adaptive immune responses, we identified two immunological axes associated with disease inflammation levels and disease localizations. These results call for further investigations on the role of the LAG3+/ICOS+ expression balance by Tregs and the contribution of ILC3s in AIDs.