Back

Responders vs. non-responders to mesenchymal stromal cells in knee osteoarthritis patients: mechanistic correlates of donor cell attributes and patient features

Robb, K. P.; Rabani, R.; Vohra, S.; Fazio, A.; Singh, S.; Chahal, J.; Audet, J.; Perruccio, A. V.; Gandhi, R.; Viswanathan, S.

2025-07-20 cell biology
10.1101/2025.07.17.665028 bioRxiv
Show abstract

Mesenchymal stromal cells (MSCs) have resulted in heterogenous clinical effectiveness in knee osteoarthritis (KOA). Herein, a framework that dually correlates KOA patient responsiveness to baseline bone marrow-derived MSC(M) donor batch attributes and baseline clinical and biomarker features is provided. Using previous clinical trial data, we demonstrated that MSCs with increased immunomodulatory potency are more efficacious. Multivariable MSC(M) genes correlated strongly with responder status and to 12- and 24-month improvements in Knee Osteoarthritis and Outcome Scores. Responder MSC(M) donor batches had differential microRNA expression and ability to polarize CD14+ monocytes in vitro. KOA Responders had more severe OA and lower physical activity. Baseline biomarkers showed trending correlations to patient responsiveness. 62.5% of KOA patients remained Responders at 8-10 years, speaking to longer-term effects of single MSC(M) injections. Thus, our analytical methodology defines critical quality attributes of potent MSC donor batches and identifies putative KOA patient theratypes to MSC treatments. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=160 HEIGHT=200 SRC="FIGDIR/small/665028v3_ufig1.gif" ALT="Figure 1"> View larger version (49K): org.highwire.dtl.DTLVardef@f47f11org.highwire.dtl.DTLVardef@1c25354org.highwire.dtl.DTLVardef@ecb4cdorg.highwire.dtl.DTLVardef@5219e0_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.