Background/Purpose: : Systemic Lupus Erythematosus (SLE) follows a characteristic relapsing-remitting course with unpredictable flares interspersed with periods of relative quiescence. Mycophenolate mofetil (MMF) has emerged as a cornerstone immunosuppressive therapy in SLE management, the challenges of balancing effective disease control against medication-associated toxicities remain significant. Our recent clinical trial (Chakravarty, et al. Lancet Rheum 2024) suggests that MMF discontinuation in clinically quiescent patients may allow remission rates similar to continuous therapy in select patients. However, an integrated set of clinically and biologically relevant biomarkers capable of identifying patients at the highest risk of disease flare following MMF withdrawal is needed to optimize clinical outcomes and guide clinical care.

Methods: Safety of Estrogens in Lupus Erythematosus National Assessment – Systemic Lupus Erythematosus Disease Activity Index (SELENA-SLEDAI) flares were tracked for 60 weeks post-MMF withdrawal. Serum (baseline) samples from SLE patients (n=50) with clinically quiescent disease (physician defined and SLEDAI< 4) on long-term MMF from the “Randomized MMF Withdrawal in Systemic Lupus Erythematosus (SLE) (ALE06)” trial (Chakravarty, et al. Lancet Rheum. 2024) with available samples and appropriate follow-up were collected prior to and after MMF taper. Soluble serum proteins were profiled using Olink Explore HT and an iterative machine learning approach (ML) used to elucidate soluble protein signatures that predict disease flare post-MMF withdrawal.

Results: In total, 31 SLE patients experienced a SELENA-SLEDAI flare within 60 weeks of MMF withdrawal, compared to 19 who did not. Using 283 known secreted protein levels predictive ML modeling was able to identify disease flares post-MMF withdrawal with 70.3% accuracy and an AUC of 88.4 ± 9 using 34 proteins (Fig 1A). Among the top predictors, TGF-β2, a less recognized immune modulatory isoform of TGF-B that can directly regulate T cell functions, was lower in patients who flared (Fig 1B). Of note, T cells are one of the cell lineages directly affected by MMF usage. Protein set enrichment analysis of ML model proteins show elevations of ERK1/2 signaling proteins, B cell proteins such as FcγR2b a moderator of B Cell activity with known deficiencies in lupus, and phagocytosis-associated proteins in patients who flared post MMF withdrawal (Fig 1C). Together, these findings suggest that immunologic pathway dysregulation is present in clinically quiescent patients at the highest risk of disease flares after MMF withdrawal.

Conclusion: Our findings demonstrate that a secreted protein-based proteomic signature can robustly predict imminent SLE flares following MMF withdrawal, with TGF-β2 and FcγR2b emerging as key biomarkers associated with flare risk. These results also highlight the value of proteomic profiling for risk stratification and provide new insights into the mechanisms underlying disease reactivation after immunosuppressant withdrawal.

Fig. 1: Proteomic biomarkers are associated with subsequent disease flare following MMF withdrawal in SLE patients. (A) Random forest model using baseline serum levels of secreted proteins (identified via the Human Protein Atlas) predicts SELENA-SLEDAI flare within 60 weeks of MMF discontinuation. (B) Ranked importance of the top secreted protein predictors for disease flare following MMF withdrawal, as determined by the random forest model. (C) Pathways with significant enrichment in patients with disease flare following MMF withdrawal. (D) Volcano plot of phagocytosis-related proteins showing differential abundance between patients who experienced a disease flare compared to those who did not following MMF withdrawal.

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/distinct-proteomic-signature-predicts-post-mmf-withdrawal-flares-in-systemic-lupus-erythematosus-patients/