High numbers of Toxoplasma gondii oocysts in the environment are a risk factor to humans. The environmental contamination might be reduced by vaccinating the definitive host, cats. An experimental challenge model is necessary to quantitatively assess the efficacy of a vaccine or drug treatment. Previous studies have indicated that bradyzoites are highly infectious for cats. To infect cats, tissue cysts were isolated from the brains of mice infected with oocysts of T. gondii M4 strain, and bradyzoites were released by pepsin digestion. Free bradyzoites were counted and graded doses (1000, 100, 50, 10), and 250 intact tissue cysts were inoculated orally into three cats each. Oocysts shed by these five groups of cats were collected from faeces by flotation techniques, counted microscopically and estimated by real time PCR. Additionally, the number of T. gondii in heart, tongue and brains were estimated, and serology for anti T. gondii antibodies was performed. A Beta-Poisson dose-response model was used to estimate the infectivity of single bradyzoites and linear regression was used to determine the relation between inoculated dose and numbers of oocyst shed. We found that real time PCR was more sensitive than microscopic detection of oocysts, and oocysts were detected by PCR in faeces of cats fed 10 bradyzoites but by microscopic examination. Real time PCR may only detect fragments of T. gondii DNA without the presence of oocysts in low doses. Prevalence of tissue cysts of T. gondii in tongue, heart and brains, and anti T. gondii antibody concentrations were all found to depend on the inoculated bradyzoite dose. The combination of the experimental challenge model and the dose response analysis provides a suitable reference for quantifying the potential reduction in human health risk due to a treatment of domestic cats by vaccination or by therapeutic drug application.