Oncolytic herpes simplex virus vectors are a promising strategy for cancer therapy, as direct cytotoxic agents, inducers of antitumor immune responses, and as expressers of anticancer genes. Progress is dependent upon representative preclinical models to evaluate therapy. In this study, two families of oncolytic herpes simplex virus vectors (G207 and NV1020 series) that have been in clinical trials were examined for the treatment of breast cancer, using the C3(1)/T-Ag transgenic mouse model. Female mice spontaneously develop mammary carcinomas, and the C3(1)/T-Ag-derived tumor cell line M6c forms implantable tumors. Both in vitro and in vivo, G47D, derived from G207 by deletion of ICP47 and the US11 promoter, was more efficacious than G207. Whereas NV1023, derived from NV1020 by deletion of ICP47 and insertion of LacZ, was as cytotoxic to M6c cells in vitro as G47D, it did not inhibit the growth of s.c. M6c tumors but did extend the survival of intracerebral tumor bearing mice. In contrast, NV1042, NV1023 expressing interleukin 12, inhibited s.c. M6c tumor growth to a similar extent as G47D, but was less effective than NV1023 in intracerebral tumors. In the spontaneously arising mammary tumor model, when only the first arising tumor per mouse was treated, G47D inhibited the growth of a subset of tumors, and when all tumors were treated, G47D significantly delayed tumor progression. When the first mammary tumor was treated and the remaining mammary glands removed, NV1042 was more efficacious than G47D at inhibiting the growth and progression of injected tumors. (Cancer Res 2005; 65(4): 1532-40)