Gene therapy for cartilage repair has long promised to coax stem cells into rebuilding the tissue that osteoarthritis destroys, but the molecular choreography behind that transformation has remained largely hidden. Now, researchers reporting in Arthritis Research & Therapy have mapped, at the level of thousands of individual proteins, what happens inside human bone marrow-derived adherent stromal cells (BM-MSCs) from osteoarthritis patients when they are equipped with an insulin-like growth factor I (IGF-I) gene and pushed toward a cartilage-forming fate. The study, published on 10 October 2026, is described by its authors as the first proteomic characterization of recombinant adeno-associated virus (rAAV)-mediated IGF-I gene therapy in this setting.
The experimental system was deliberately rigorous. Bone marrow aspirates were collected from four donors undergoing total knee arthroplasty at Saarland University Medical Center in Germany, with ethical approval and informed consent. Three donors supplied cells for the proteomic analysis: a 63-year-old woman, a 66-year-old woman and a 71-year-old man. For each donor, cells were grown in three-dimensional micromass pellets and transduced with rAAV vectors carrying either the human IGF-I gene or the lacZ gene encoding beta-galactosidase as an active vector control. Because both groups underwent identical transduction and differentiation procedures, differences between them could be attributed specifically to IGF-I overexpression rather than to the vector itself. One pellet was lost during medium changes, leaving ten samples for proteomic analysis, five per condition.
The vectors themselves were single-stranded rAAV2 particles produced in HEK293 cells and purified by dialysis, with the IGF-I construct carrying a 536-base-pair cDNA under the control of the cytomegalovirus immediate-early promoter. Cells received a multiplicity of infection of four transgene copies per cell, and chondrogenic differentiation ran for 21 days in defined medium containing TGF-beta3, dexamethasone and ascorbate. After differentiation, proteins were extracted, digested with trypsin and analysed by shotgun label-free mass spectrometry on a timsTOF Pro instrument using data-independent acquisition with the Parallel Accumulation-Serial Fragmentation (PASEF) method, a state-of-the-art platform that enabled deep proteome coverage.
The scale of the measurement was striking. In total, 4,284 proteins were identified in the micromass proteome. Principal component analysis separated the IGF-I group cleanly from the lacZ controls, with the first principal component accounting for 51.68 percent of the variance, and hierarchical clustering likewise split the samples by condition. After quality filtering, an exploratory chi-squared screen of protein detection patterns flagged 54 proteins with nominal differences in detection frequency, including IGF-I itself, beta-catenin and the NF-kappa-B p105 subunit. Separately, a quantitative differential expression analysis using linear models with empirical Bayes statistics, donor identity as a blocking factor and a threshold of adjusted p at or below 0.05 with a fold change of at least 1.5, identified 58 significantly regulated proteins. Together, 112 proteins showed notable changes in detection patterns or abundance associated with IGF-I.
The direction of those changes told a coherent metabolic story. Thirty-eight proteins were significantly upregulated, with fold changes from 6.85 to 1.57, including apolipoprotein E, fatty aldehyde dehydrogenase and the COPII coat GTPase SAR1A. Twenty proteins were significantly downregulated, with fold changes from minus 5.18 to minus 1.50, and these were mostly associated with mitochondrial translation, including several mitochondrial ribosomal proteins. Pathway and protein-protein interaction enrichment analyses grouped the 112 regulated proteins into 15 clusters spanning six overarching pathways: protein processing and organization, intracellular transport and localization, apoptosis, proliferation and survival, metabolism, and migration. Three densely connected network components emerged, linked respectively to Golgi vesicle transport and ER-to-Golgi trafficking, mitochondrial ribosomal function, and carbohydrate metabolism including the pentose phosphate pathway.
The authors interpret this pattern as consistent with a shift toward an anabolic, biosynthetically active cellular state, the kind of metabolic adaptation chondrogenic differentiation demands to sustain extracellular matrix production. They emphasize, however, that the enrichment of glycolysis and carbohydrate metabolism pathways is hypothesis-generating: no dedicated metabolic assays were performed, and functional confirmation awaits future work. Proteins tied to glycosylation, including O-linked beta-N-acetylglucosamine transferase, along with extracellular glycoproteins such as biglycan and glypicans, also appeared among the regulated proteins, suggesting coordinated regulation of protein synthesis and post-translational processing with potential relevance to cartilage matrix organization. Components of the ESCRT endosomal sorting machinery, including PDCD6IP, were likewise identified, possibly reflecting increased vesicular trafficking.
Two proteomic findings were validated independently. Beta-catenin, a central effector of canonical Wnt signalling, was detected in all IGF-I-transduced samples but only two controls in the exploratory screen. Immunohistochemistry on micromasses from a fourth donor, a 58-year-old woman, showed a significantly higher beta-catenin-positive area in IGF-I-transduced cultures compared with controls, 33.7 versus 9.6 percent, and an antibody specific for the active, dephosphorylated form of beta-catenin confirmed an increase, 25.0 versus 4.67 percent, with nuclear staining visible by immunofluorescence. Exportin-1 (XPO1), the principal nuclear export receptor, was significantly upregulated in the quantitative proteomic analysis with a fold change of 2.24, and immunostaining showed a marked increase in XPO1-positive area, 54.3 versus 12.4 percent. Sirtuin-1, a known XPO1 cargo, showed increased staining with a predominantly extranuclear distribution, supporting increased activity of the XPO1-SIRT1 nucleocytoplasmic transport axis.
The chondrogenic phenotype of the model itself was confirmed histologically. Toluidine Blue staining revealed glycosaminoglycan-rich matrix in both conditions, and type II collagen immunostaining was markedly stronger in IGF-I-transduced micromasses, 87.6 versus 5.8 percent positive area, a statistically significant difference supporting an IGF-I-associated enhancement of cartilage-specific matrix deposition. GAG deposition showed a numerical increase in the IGF-I group, 51.7 versus 16.7 percent, but this did not reach statistical significance. IGF-I expression itself was confirmed by immunostaining, with 89.8 percent positive area in transduced micromasses versus 14.9 percent in controls.
The authors are candid about the limitations. Only three donor-derived cell populations fed the proteomic analysis, constraining statistical power and generalizability, and the validation analyses came from a single additional donor, making them supportive evidence rather than proof of reproducibility across donors. Standard immunophenotypic characterization of the isolated BM-MSCs was not performed because of limited cell yield, no non-transduced control group was included, and transduction efficiency was not quantified per donor. The cells came from osteoarthritis patients and may differ biologically from cells of healthy donors. Most importantly, as an entirely in vitro study, the findings cannot be extrapolated to conclusions about cartilage regeneration or therapeutic efficacy in patients.
What the study delivers instead is a molecular map. It shows that IGF-I gene transfer in osteoarthritis-derived BM-MSCs is accompanied by broad proteomic rearrangements touching metabolism, protein processing, ER homeostasis, intracellular transport and Wnt signalling, alongside enhanced cartilage matrix marker deposition, and it validates two of those signals, beta-catenin and XPO1, with independent methods. The authors frame these as pathways of potential relevance to cartilage repair whose functional significance must now be tested in larger cohorts, more advanced three-dimensional models and in vivo systems before any therapeutic implication can be drawn.
Subject of Research: Proteomic changes induced by rAAV-mediated IGF-I gene transfer in osteoarthritis-derived human bone marrow stromal cells undergoing chondrogenesis
Article Title: Proteomic landscape of IGF-I gene transfer in osteoarthritis-derived human bone marrow stromal cells undergoing chondrogenesis
Article References: Quaranta, P., Venkatesan, J. K., Calamia, V., Fernández-Puente, P., Riva-Mendoza, S., Blanco, F. J., Cucchiarini, M., & Ruiz-Romero, C. (2026). Proteomic landscape of IGF-I gene transfer in osteoarthritis-derived human bone marrow stromal cells undergoing chondrogenesis. Arthritis Research & Therapy, 28(1), Article 180. https://doi.org/10.1186/s13075-026-03919-2
Image Credits: AI Generated
DOI: 10.1186/s13075-026-03919-2
Keywords: osteoarthritis, gene therapy, IGF-I, proteomics, chondrogenesis, mesenchymal stromal cells, rAAV vectors, cartilage repair, beta-catenin, XPO1, mass spectrometry, DIA-PASEF
News Source: Juliet Wilcox. (October 30, 2026). IGF-I Gene Therapy Reshapes the Proteome of Cartilage-Forming Stem Cells from Osteoarthritis Patients. Scienmag.
