Back

Microgeography of staphyloccoci in human tissue explains antibiotic failure

Tripathi, V.; Morin, B.; Antelo Varela, M.; Loriol, A. A.; Kurg, K.; Jakonia, S.; Ehsan, A.; Herbert, S.; Ferrand, A.; Sauteur, L.; Marro, F. C.; Henselmann, N.; Künnecke, J.; Estermann, F.; Strenger, K.; Iizuka, A.; Woischnig, A.-K.; Claudi, B.; Söderholm, S.; Keller, P. M.; Baumhoer, D.; Bock, T.; Jantarung, K.; Rivera Fuentes, P.; Meylan, S.; Morgenstern, M.; Clauss, M.; Khanna, N.; Kuehl, R.; Bumann, D.

2025-09-12 microbiology
10.1101/2025.09.12.675782 bioRxiv
Show abstract

Summary paragraphBacterial infections remain a major health threat, yet pathogen biology in human tissues is poorly understood. Using AI-guided imaging, we mapped [~]15,500 Staphylococcus aureus cells in biopsies from 33 patients undergoing surgery for musculoskeletal infections. Despite substantial interindividual variability, consistent patterns emerged. Most bacteria resided within non-classical monocytes/macrophages, challenging models of primarily extracellular pathogenesis. Both intra- and extracellular bacteria were predominantly isolated single cells or doublets with low rRNA content, suggesting limited replication. Complementary proteomics implicated inflammation-associated hypoxia and host glucose-to-lactate metabolism as growth constraints. Preoperative antibiotic therapy failed to clear bacteria across microenvironments and cluster sizes, challenging assumptions that antibiotic tolerance is confined to intracellular niches or biofilms and underscoring the clinical need for debridement. In vitro models replicating diverse tissues conditions impaired antibiotic activity, indicating multifactorial resilience. Together, these findings redefine S. aureus infection biology in musculoskeletal infections and establish a framework for mechanism-based prevention and therapy.

Matching journals

The top 3 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.