PI-123 - A MECHANISTIC QUANTITATIVE SYSTEMS PHARMACOLOGY FRAMEWORK FOR MODEL-INFORMED TRANSLATIONAL ASSESSMENT OF MESENCHYMAL STEM CELL THERAPY IN LUPUS NEPHRITIS

C. Zhao¹, S. Zhang²; ¹Nanjing Medical University, Nanjing, Jiangsu, China, ²Nanjing Medical University, Nanjing, China.

Background: Lupus nephritis (LN) is a complex autoimmune disease. Recently, mesenchymal stem cells (MSCs), given their immune-modulatory potential, have emerged as a promising route for the treatment of LN, but their clinical translation still faces many challenges. This study aimed to build the first QSP model of LN that can facilitate the clinical development of new MSC-based therapies, with application scenarios such as translational efficacy evaluation, dosing regimen design, and assessment of patient stratification biomarkers.

Methods: The QSP model was mechanistically formulated based on mutiscale disease mechanisms of LN and considered MSCs as well as standard of care (SoC) treatments including MMF, prednisone and CTX. For MSCs, given its unique in vivo biodistribution and mechanisms, the model included a PBPK module that has been validated against MSC datasets. The overall QSP model was integratively calibrated and utilized to simulate the clinical efficacy biomarkers in LN patients receiving different SoC regimens and SoC+MSC, and the results were compared with published data as well as early phase LN trial data of a new MSC product. The modeling framework was then used to predict the efficacy of different MSC dosing regimens and the potential of various biomarkers for patient stratification.

Results: The QSP model was rigorously calibrated, and the model-predicted clinical biomarker and renal response data were quantitively consistent with an extensive set of clinical data (involving 10+ trials). Model simulations predicted that LN patients with MSCs infused intravenously at 3-month intervals can achieve the optimal clinical response given a 6-month trial duration, and this regimen was adopted in the phase 2 design of our MSC product of interest. In addition, our model identified high CXCL10 plus low UPCR as a set of biomarkers that can potentially enhance patient response to MSC+SoC, which requires further validation.

Conclusion: We have developed the first QSP model for LN that integrated multiscale disease progression mechanisms, multi-modal drug-disease regulation as well as extensive preclinical-clinical data. The model was validated and utilized to predict translational efficacy of novel MSC-based therapies for LN and provided critical insights on the optimal dosing regimen as well as new patient stratification biomarkers.