AUTHOR=Garantziotis Panagiotis , Nikolakis Dimitrios , Doumas Stavros , Frangou Eleni , Sentis George , Filia Anastasia , Fanouriakis Antonis , Bertsias George , Boumpas Dimitrios T. TITLE=Molecular Taxonomy of Systemic Lupus Erythematosus Through Data-Driven Patient Stratification: Molecular Endotypes and Cluster-Tailored Drugs JOURNAL=Frontiers in Immunology VOLUME=13 YEAR=2022 URL=https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2022.860726 DOI=10.3389/fimmu.2022.860726 ISSN=1664-3224 ABSTRACT=Objectives

Treatment of Systemic Lupus Erythematosus (SLE) is characterized by a largely empirical approach and relative paucity of novel compound development. We sought to stratify SLE patients based on their molecular phenotype and identify putative therapeutic compounds for each molecular fingerprint.

Methods

By the use of whole blood RNA-seq data from 120 SLE patients, and in a data-driven, clinically unbiased manner, we established modules of commonly regulated genes (molecular endotypes) and re-stratified patients through hierarchical clustering. Disease activity and severity were assessed using SLEDAI-2K and Lupus Severity Index, respectively. Through an in silico drug prediction pipeline, we investigated drugs currently in use, tested in lupus clinical trials, and listed in the iLINCS prediction databases, for their ability to reverse the gene expression signatures in each molecular endotype. Drug repurposing analysis was also performed to identify perturbagens that counteract group-specific SLE signatures.

Results

Molecular taxonomy identified five lupus endotypes, each characterized by a unique gene module enrichment pattern. Neutrophilic signature group consisted primarily of patients with active lupus nephritis, while the B-cell expression group included patients with constitutional features. Patients with moderate severity and serologic activity exhibited a signature enriched for metabolic processes. Mild disease was distributed in two groups, exhibiting enhanced basic cellular functions, myelopoiesis, and autophagy. Bortezomib was predicted to reverse disturbances in the “neutrophilic” cluster, azathioprine and ixazomib in the “B-cell” cluster, and fostamatinib in the “metabolic” patient subgroup.

Conclusion

The clinical spectrum of SLE encompasses distinct molecular endotypes, each defined by unique pathophysiologic aberrancies potentially reversible by distinct compounds.