Standardizing MSC Therapies: A Framework for Addressing Donor Variability and Benchmarking Functional Heterogeneity

Abstract

Background & Aim

Inter- and intra-donor variability poses a significant challenge to ensuring consistent quality and performance in cell therapies. This challenge is emphasized in MSC therapies, which rely on the cells' inherent, multifaceted therapeutic capacities. Compared to cells with defined therapeutic targets, MSC efficacy is more sensitive to donor and batch variations, complicating standardization. MSC priming has emerged as a strategy to enhance therapeutic potency and consistency. Previously, we introduced a primed MSC therapy (MesenCure™) that reduced mortality by 68% and the need for invasive ventilation by 57% in respiratory distress patients. Here, we characterize the distinct phenotype of primed MSCs and introduce a framework for assessing functional heterogeneity, enabling the evaluation of cell bank performance against established benchmarks.

Results & Conclusion

Results: RNAseq analysis of primed MSCs from 3 donors revealed 6,909 differentially expressed genes (Fig. 1A, q<0.01) compared to unprimed MSCs. PCA of genes accounting for 99.9% of variability demonstrated distinct clustering of donors by priming status (Fig. 1B). Similar patterns were observed in PCA of cellular and secretome proteomic data from 3 batches derived from a single donor (Figs. 1C, 1D), reaffirming the unique phenotype of primed MSCs.

Subsequently, donor variability was assessed in 15 donor-derived primed cell batches, examining parameters related to immunomodulatory capacity (IC50 of CD4+ T-cell activation), regenerative potential (FGF7 secretion), and cellular durability (autophagy levels) (Fig. 2A). PCA of these parameters identified a principal component (PC1) linked to enhanced performance, clustering batches into low-, medium-, and high-performing groups (Fig. 2B).

To evaluate our clinical cell bank performance, we simulated 1,000 normally distributed samples (Fig. 2C) based on observed averages for IC50 (N = 67), FGF7 (N = 16), and autophagy (N = 4), overlaid in Fig. 2A. PC1 values for the simulated samples, calculated using the same transformation that identified the three clusters, were overlaid onto these clusters in Fig. 2B. The distribution of simulated PC1 values demonstrated alignment of clinical cell bank batches with the high-performing cluster (Figs. 2B, 2D).

Conclusion

In conclusion, these findings present a robust blueprint for assessing functional heterogeneity among MSC batches, offering a practical framework to benchmark cell bank performance and advance the standardization of MSC-based therapies