OBJECTIVE: A deficiency in C5 protects against arthritis development. However, there is currently no approach successfully translating these findings into arthritis therapy, as by targeting the key component, C5a. The aim of this study was to develop a vaccination strategy targeting C5a as therapy for patients with rheumatoid arthritis. METHODS: An anti-C5a vaccine was generated by incorporating the unnatural amino acid p-nitrophenylalanine (4NPA) into selected sites in the murine C5a molecule. C5a-4NPA variants were screened for their immunogenicity in mice on different arthritis-susceptible class II major histocompatibility complex (MHC) backgrounds. A candidate vaccine was tested for its impact on disease in a murine model of collagen-induced arthritis (CIA). Immunity toward endogenous C5a as well as type II collagen was monitored and characterized. RESULTS: Replacing a single tyrosine residue in position 35 (Y(35) ) with 4NPA allowed the generation of an anti-C5a vaccine, which partly protected mice against the development of CIA while strongly ameliorating the severity of clinical disease. Although differing in just 3 atoms from wild-type C5a (wtC5a), C5aY(35) 4NPA induced loss of T cell and B cell tolerance toward the endogenous protein in mice expressing class II MHC H-2(q) molecules. Despite differential B cell epitope recognition, antibodies induced by both wtC5a and C5aY(35) 4NPA neutralized C5a. Thus, anti-wtC5a IgG titers during arthritis priming were potentially of critical importance for disease protection, because high titers of C5a-neutralizing antibodies after disease onset were unable to reverse the course of arthritis. CONCLUSION: The results of this study suggest that the most effective anti-C5a treatment in arthritis can be accomplished using a preventive vaccination strategy, and that treatment using conventional biologic or small molecule strategies targeting the C5a/C5aR axis may miss the optimal window for therapeutic intervention during the subclinical priming phase of the disease.