The study, published in the journal Materials Today Bio, centers on Lactobacillus rhamnosus GG, one of the most extensively characterized probiotic strains and one that holds Generally Recognized as Safe status from the US Food and Drug Administration. Like many bacteria, LGG continuously releases extracellular vesicles, or BEVs, nanoscale sacs of membrane, typically 30 to 150 nanometers across, that package proteins, lipids, and microRNAs and ferry them between cells. Bacterial vesicles have attracted growing interest as therapeutics because, unlike the vesicles shed by mammalian stem cells, they can be produced at scale by fermentation, quickly, cheaply, and with far less batch-to-batch variability. The researchers grew LGG in broth, spun out the bacteria, and pelleted the vesicles by ultracentrifugation, yielding particles with a peak diameter of about 153 nanometers and the classic spherical bilayer structure seen under the electron microscope.

When injected into the knee joints of mice whose anterior cruciate ligament had been surgically transected to induce osteoarthritis, the natural vesicles already showed real promise. Micro-computed tomography revealed fewer bone spurs, or osteophytes, and healthier subchondral bone architecture in treated animals, while histological staining of the joint surfaces showed better-preserved cartilage, lower disease scores on the standard OARSI grading system, and reduced synovial inflammation. Molecular markers told a consistent story: levels of type II collagen, the key structural protein of healthy cartilage, recovered, while MMP13, an enzyme that chews up cartilage matrix, declined. Chondrocytes, the resident cartilage cells, readily took up the vesicles, a process the team showed is driven mainly by caveolae-mediated endocytosis.

But the natural vesicles had a fundamental weakness. Within roughly three days of injection, most had been cleared from the joint, swept into the bloodstream or lymphatics by synovial fluid, and the dense meshwork of the cartilage matrix kept the survivors away from the deep chondrocyte layers where they were needed most. The team’s answer was to graft a targeting peptide onto the vesicle surface. They chose WYRGRL, a six-amino-acid peptide that binds specifically to the alpha 1 chain of type II collagen, the scaffold protein of articular cartilage, and which previous work has shown can boost joint targeting by as much as 72-fold. Using a chemical conjugation strategy with methacrylic anhydride to create reactive alkene groups on the vesicle membranes, the researchers covalently anchored the peptide, achieving a modification efficiency of 73.5 percent as measured by nano-flow cytometry.

The engineering left the vesicles’ morphology intact but changed their surface chemistry noticeably. Untreated vesicles carried a zeta potential of about minus 16 millivolts; the modified particles shifted to roughly plus 8 millivolts, and their peak diameter crept up to about 160 nanometers. Safety testing showed no cytotoxicity in chondrocyte cultures by either CCK-8 assay or live-dead staining, and no pathological changes in the heart, liver, spleen, lung, or kidney of injected mice. Crucially, the peptide did not hinder uptake: fluorescently labeled engineered vesicles were internalized by chondrocytes even more efficiently than the natural particles, while uptake by fibroblasts stayed unchanged, confirming that the modification conferred genuine cartilage specificity rather than a blanket increase in cell binding.

The in vivo results were striking. Whereas unmodified vesicles vanished from the joint within three days, the engineered particles persisted for seven to ten days and were only fully cleared by day fourteen. Imaging of dissected organs three days after injection showed strong fluorescence accumulating in the livers of mice that received natural vesicles, but essentially none in mice given the engineered version, indicating that the peptide kept the particles anchored in the joint instead of drifting into the reticuloendothelial system. Joint sections revealed that the modified vesicles not only clung to the cartilage surface but penetrated into its deeper layers, reaching chondrocytes that unmodified particles never touched.

Functionally, the engineered vesicles outperformed their natural counterparts at every level tested. In chondrocytes stimulated with the inflammatory cytokine IL-1 beta to mimic osteoarthritic conditions, the engineered particles more effectively reversed the disease signature, restoring expression of the matrix-building proteins type II collagen and aggrecan while suppressing the matrix-degrading enzymes MMP13 and ADAMTS5. Even after the treatment medium was replaced to simulate the washout dynamics of synovial fluid, cells that had internalized the engineered vesicles continued to benefit, staining more strongly for proteoglycans with Safranin O and Toluidine Blue, migrating faster in wound-healing assays, and proliferating at higher rates. In the mouse model, eight weeks of weekly injections produced joints that looked markedly closer to healthy controls: smoother cartilage surfaces, uniform matrix thickness, fewer and smaller osteophytes, and significantly lower expression of destructive enzymes and collagen X, a marker of pathological chondrocyte hypertrophy.

To understand how the vesicles worked, the team sequenced their microRNA cargo. Among the twenty most abundant microRNAs were several with established links to cartilage biology, including miR-21, miR-27, miR-122, and miR-199, all previously reported to influence osteoarthritis. Enrichment analyses pointed to target genes concentrated in the Wnt, TGF-beta, and HIF-1 signaling pathways, which govern inflammation, immune regulation, and cellular stress responses in the joint. Network analysis highlighted miR-27b-3p, miR-26b-5p, and miR-16-5p as key mediators, acting through the TGF-beta pathway, TLR3 signaling, and the transcription factor SMAD3 to shape synovial inflammation and chondrocyte metabolism. The vesicles also carry immunomodulatory proteins and lipoteichoic acid, which engages the TLR2 pathway implicated in osteoarthritis inflammation, suggesting the particles act through a combination of RNA-based and protein-based mechanisms.

The work also carries conceptual weight beyond the drug itself. A growing body of research on the gut-joint axis has established that intestinal microbes communicate with joint tissue and influence arthritis progression, and vesicles released by probiotics have been proposed as one of the molecular couriers mediating this crosstalk. By isolating the vesicles from a single standardized probiotic strain and converting them into an injectable, cartilage-targeted formulation, the researchers effectively compressed the gut-joint axis into a nanoscale therapeutic, sidestepping the variability inherent in fecal microbiota preparations. The authors note that this is the first report of modifying probiotic-derived vesicles for osteoarthritis treatment, and that the same platform could in principle be adapted to carry other drugs or to target other degenerative tissues.

Significant hurdles remain before patients could benefit. The study relied on a single surgically induced mouse model with five animals per group, and the authors acknowledge that whether the vesicles can reach and modulate the subchondral bone microenvironment, another important driver of osteoarthritis, is still unknown. Larger-animal studies will be needed to confirm efficacy and safety, and the team plans to screen vesicles from additional gut bacteria associated with the disease to find potentially more potent candidates. Still, the combination of scalable bacterial fermentation, a simple chemical targeting strategy, and demonstrated retention in one of the body’s most drug-resistant compartments makes this a notable proof of concept. If the approach survives the translational gauntlet, the humble vesicles of a yogurt bacterium could become a precision delivery vehicle for joints that medicine has long struggled to reach.

Subject of Research: Cartilage-targeting engineered probiotic bacterial extracellular vesicles for osteoarthritis treatment

Article Title: Cartilage-targeting engineered probiotic bacterial extracellular vesicles treat osteoarthritis through modulation of matrix homeostasis

Article References: Niu, L., Chen, W., Liu, T., Liang, X., Song, P., Liu, H., Su, T., Chen, X., Liu, Z., Jing, Y., Cui, J., & Su, J. (2026). Cartilage-targeting engineered probiotic bacterial extracellular vesicles treat osteoarthritis through modulation of matrix homeostasis. Materials Today Bio, 41, Article 103668. https://doi.org/10.1016/j.mtbio.2026.103668

Image Credits: AI Generated

DOI: 10.1016/j.mtbio.2026.103668

Keywords: osteoarthritis, bacterial extracellular vesicles, probiotics, Lactobacillus rhamnosus GG, cartilage targeting, WYRGRL peptide, microRNA, chondrocytes, extracellular matrix, gut-joint axis, drug delivery, nanomedicine