Rheumatoid arthritis (RA) is marked by a breakdown of immunological tolerance to synovial self-proteins, but the initiating events that create novel neo-antigens remain incompletely defined. The authors investigated whether human cathepsin K (hCatK), a cysteine protease implicated in joint matrix degradation in RA, can generate covalently spliced peptides via cis- and trans-peptide ligation. The hypothesis tested was that protease-mediated transpeptidation could produce peptide sequences compatible with MHC class II presentation and therefore serve as a source of neo-antigens in RA.
The study used high-resolution liquid chromatography–tandem mass spectrometry (LC-MS/MS) combined with database-assisted de novo sequencing to detect and identify spliced peptides produced by hCatK. Substrates included major RA-associated self-proteins and a foreign antigen: type II collagen, fibrinogen, and the SARS-CoV-2 Spike protein. Using these methods, the authors identified over 90 unique spliced peptides formed by hCatK-catalyzed reactions.
Biochemical characterization revealed that splicing efficiency differed markedly from classic hydrolytic activity. Specifically, splicing activity peaked at near-neutral pH values (6.5–7.5). This profile contrasts with the typical hydrolytic behavior of proteases and indicates that hCatK transpeptidation is favored under conditions closer to neutrality rather than strictly acidic environments.
Detailed biochemical profiling showed pronounced subsite selectivity in hCatK-mediated splicing. There was a striking enrichment for small, aliphatic, and hydroxyl-containing residues at the P1 position of spliced junctions, with Gly, Thr, and Ser highlighted as common P1 residues. This residue preference suggests mechanistic constraints on the transpeptidation reaction and informs which sequence contexts are most likely to yield spliced products.
Splicing was not uniformly distributed across substrates but preferentially targeted flexible, intrinsically disordered regions. For fibrinogen, splicing events were highly clustered: 81% of observed fibrinogen splicing occurred within the αC domain. The study also reports splicing events derived from type II collagen and the SARS-CoV-2 Spike protein, indicating that hCatK can act on both self and foreign antigens to produce spliced peptides.
To assess potential immunological relevance, the authors performed in silico binding predictions for the RA-susceptibility allele HLA-DRB1*04:01, which carries the shared epitope linked to RA risk. These predictions indicated that numerous hCatK-generated spliced peptides exhibit predicted affinities that exceed those of established immunogenic and canonical genomic sequences. The results therefore suggest that some spliced peptides formed by hCatK could be presented by HLA-DR4 molecules with high affinity, supporting their plausibility as MHC class II autoantigens.
Collectively, the data identify protease-mediated transpeptidation by hCatK as a novel post-translational modification capable of generating peptide sequences with strong predicted MHC class II-binding properties. Because hCatK is active in joint tissues and the spliced peptides derive from RA-relevant proteins and a viral antigen, these findings raise the possibility that hCatK-catalyzed splicing contributes to the pool of neo-antigens that can breach tolerance and drive pathogenic CD4+ T cell responses in RA. The pH dependence, subsite specificity, and regional clustering provide mechanistic constraints that narrow which spliced peptides are most likely to be produced in vivo.
The authors report associated code and data resources. The manuscript cites a GitHub repository and a PRIDE archive accession for project data and mass spectrometry files. Funding for the work was declared from the Canadian Institutes of Health Research (CIHR; PJT-155979). The authors declared no competing interests.
This study demonstrates that human cathepsin K can catalyze both cis- and trans-splicing to form a diverse set of peptides from RA-associated self-proteins and a foreign antigen. Over 90 unique spliced peptides were identified by LC-MS/MS, splicing was most efficient at near-neutral pH, and there was pronounced preference for Gly, Thr, and Ser at P1. Splicing concentrated in flexible protein regions such as the fibrinogen αC domain, and many spliced products are predicted to bind HLA-DRB1*04:01 with high affinity. These observations support protease-mediated transpeptidation as a putative mechanism generating MHC class II–compatible neo-antigens relevant to RA. Further experimental validation of immunogenicity and presentation in vivo would be required to confirm pathogenic roles, details of which were not reported in the source article.