Back

Molecular and spatial profiling identifies immune endotypes for the stratification of OA patients

Gaigeard, N.; Cardon, A.; Guiho, R.; Cubuk, C.; Ouattara, A.; Lamothe, C.; Brouard, J.; De Lima, J.; Ahmed, M.; Le Mercier, M.; Defois, A.; Danet, L.; Perrot, J.; Benzerdjeb, N.; Vinatier, C.; Danger, R.; Delbos, L.; Brouard, S.; Degauque, N.; Pagliuca, S.; Lewis, M. J.; Fossati-Jimack, L.; Nerviani, A.; Waast, D.; Le Goff, B.; Blanchard, F.; Moulin, D.; Pitzalis, C.; Guicheux, J.; Marie-Astrid, B.

2026-01-12 immunology
10.64898/2026.01.10.698810 bioRxiv
Show abstract

Osteoarthritis (OA) is a prevalent and heterogeneous joint disease in which synovial inflammation drives structural progression and pain. Despite the recognized heterogeneity of OA, the cellular and molecular organization of synovial tissue remains poorly characterized and defining distinct histological and immune endotypes could guide precision medicine and therapeutic targeting. We show that histologically defined synovial pathotypes are conserved across independent cohorts and correspond to distinct molecular immune endotypes. Integration of bulk and spatial transcriptomics with proteomics revealed niche-specific gene and protein signatures, reflecting the anatomical and functional diversity of OA synovium. The lympho-myeloid pathotype was characterized by mature ectopic lymphoid structures containing CD21+CD23+ follicular dendritic cells, spatially organized T and B cell zones, and clonally expanded T and B cells with shared immune cell receptor motifs, consistent with local adaptive immune activity correlating with radiological joint damage. These findings highlight how immune organization and cellular composition shape OA pathogenesis and provide a framework for endotype-guided stratification and therapeutic targeting.

Matching journals

The top 5 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.