Abatacept modulates CD28-mediated T cell costimulation and is efficacious in the treatment of rheumatoid arthritis (RA). Its mechanism of action has not been fully elucidated but will likely reveal critical pathologic pathways in RA. We show that abatacept substantially modulated Ag-specific T and B cell responses in vivo. Ag-specific T cell proliferation was reduced, and the acquisition of an activated phenotype, characterized by upregulation of CD69, OX40, ICOS, and programmed death-1 and downregulation of CD62L, was suppressed. Furthermore, abatacept suppressed the production of inflammatory cytokines, such as IFN-γ and IL-17. These effects were associated with a failure of Ag-specific T cells to acquire the CXCR5+ ICOS+ T follicular helper cell phenotype. This, in turn, led to a failure of these cells to enter B cell follicles, resulting in reduced specific Ab responses, despite normal B cell clonal expansion. To test the pathologic significance of this, we used a novel model of RA associated with breach of self-tolerance to self-Ag and demonstrated that abatacept prevented the emergence of self-reactivity. Thus, CD28-dependent signaling is required for optimal T follicular helper cell maturation and expansion, and its inhibition prevents loss of self-tolerance in a model of articular pathology. Thus, we provide a novel mode of action for abatacept with profound implications for its potential usefulness in early inflammatory arthropathies associated with autoantibody expression.