The understanding of immunological tolerance has been greatly aided by the development of transgenic animal models in which expression of a specific T cell receptor (or B cell receptor) and its cognate self antigen is experimentally controlled. In most cases, expression of the self antigen was constitutive and did not allow for variation of its time- and dose-dependent expression pattern, parameters which are known to influence the balance of tolerance versus immunity. We describe a transgenic model in which expression of human C-reactive protein (hCRP), an acute-phase protein, is tightly controlled at basal levels (female mice express around 10−9 M and male mice 5 × 10−7 M circulating hCRP) and is highly inducible (induction factor of 25–500). T cells from C57BL/6 mice recognize two epitopes of hCRP termed A (residues 79–95) and B (residues 87–102). Different efficacies of presentation in vitro and in vivo identify epitope A as subdominant and epitope B as dominant. T cells of non-induced hCRP transgenic mice are tolerant to the dominant epitope, but reactive to the subdominant epitope. A hCRP-specific IgG antibody response is detectable in transgenic mice, but is weaker than in littermates. Upon induction of hCRP, both T cell epitopes are presented by thymic and splenic antigen-presenting cells (APC) in vivo. Kinetics of presentation by splenic APC closely match serum kinetics of hCRP, whereas presentation in the thymus is considerably prolonged. This model enables epitope-specific T cell tolerance to be studied as a function of time- and dose-dependent expression of the self antigen.