Oncolytic adenoviruses have been safe in clinical trials but the efficacy has been mostly limited. All published trials have been performed with serotype 5 based viruses. The expression level of the Ad5 receptor CAR may be variable in advanced tumors. In contrast, the Ad3 receptor remains unclear, but is known to be abundantly expressed in most tumors. Therefore, we hypothesized that a fully serotype 3 oncolytic adenovirus might be useful for treating cancer. Patients exposed to adenoviruses develop high titers of serotype-specific neutralizing antibodies, which might compromise re-administration. Thus, having different serotype oncolytic viruses available might facilitate repeated dosing in humans. Ad3-hTERT-E1A is a fully serotype 3 oncolytic adenovirus controlled by the promoter of the catalytic domain of human telomerase. It was effective in vitro on cell lines representing seven major cancer types, although low toxicity was seen in non-malignant cells. In vivo, the virus had anti-tumor efficacy in three different animal models. Although in vitro oncolysis mediated by Ad3-hTERT-E1A and wild-type Ad3 occurred more slowly than with Ad5 or Ad5/3 (Ad3 fiber knob in Ad5) based viruses, in vivo the virus was at least as potent as controls. Anti-tumor efficacy was retained in presence of neutralizing anti-Ad5 antibodies whereas Ad5 based controls were blocked. In summary, we report generation of a non-Ad5 based oncolytic adenovirus, which might be useful for testing in cancer patients, especially in the context of high anti-Ad5 neutralizing antibodies. © 2011 Nature America, Inc. All rights reserved.