1601 Scoring personalized molecular portraits of systemic lupus erythematosus patients to predict treatment responses, flares, and prognosis

Abstract

Objectives

Systemic Lupus Erythematosus is a complex autoimmune disease that leads to important worsening of the quality of life and significant suffering to those affected. Currently, therapies used are partially inefficient, mainly due to the molecular heterogeneity of the disease, being personalized medicine the big promise for the future of autoimmunity. With this work we intend to take a step further in that direction by developing MyPROSLE, a system capable of measuring the molecular portrait of individual patients.

Methods

We defined co-expressed and functionally annotated gene-modules conserved across two longitudinal datasets with 158 and 301 patients. The dysregulation magnitude of each gene-module was calculated at the patient level using averaged z-scores. We analyzed the association between gene-modules, clinical manifestations and the evolution of the disease by ANOVA, Student’s t-test and Cox proportional-hazard models. Drug responses to hydroxychloroquine and mycophenolate was analyzed by comparing each individual´s molecular portraits. A third dataset of 1760 patients was used to compare the response to Tabalumab.

Results

The system allows to quantify the dysregulation of 30 gene-modules individually with respect to healthy distributions. We show that dysregulation of certain gene-modules is strongly associated with different clinical manifestations and with predicting the time when remissions and relapses of the disease are to occur in the short term. We also demonstrate how the analyzed drugs act specifically on patients with strong dysregulation of gene-modules related with plasma cells.

Conclusion

MyPROSLE allows to extract key information for medical practice and may be a support for more precise therapeutic decisions in the future.