Top of pageAbstract Conditionally replicating adenoviruses (CRADs) are a novel approach for treatment of cancer. The oncolytic potency of CRADs is determined by their capability to infect tumor cells. Most adenoviruses used for gene therapy have been based on serotype 5 (Ad5). Unfortunately, expression of the primary receptor for Ad5 is highly variable on ovarian cancer cells. Genetic fiber pseudotyping is an approach to circumvent this problem by using the alternative Ad serotype 3 receptor for entry into, and killing of ovarian cancer cells. Furthermore, restriction of the CRAD replication to tumor cells minimizes nontumor tissue injury. One way to gain this selectivity is to control replication regulation genes with promoters active only in specific tumor cells. In this study, we constructed a fiber-modified CRAD containing the secretory leukoprotease inhibitor (SLPI) promoter to control viral replication via the E1A gene (Ad5/3SLPI). SLPI has been shown to be over-expressed in various tumors including ovarian cancer but minimally active in normal tissues. Viral replication and oncolysis were assessed in various ovarian cancer cell lines. Replication was also analyzed with quantitative PCR on three-dimensional primary tumor cell sheroids purified from fresh patient samples. Moreover, in a therapeutic orthotopic mouse model of peritoneal carcinomatosis, dramatically enhanced survival was noted with Ad5/3SLPI. To assess the selectivity of Ad5/3SLPI in comparison with Ad5wt, Ad5/3wt and a non replicative control virus Ad5/3luc, we evaluated the ability of the viruses to replicate in non-target liver cells in vivo. Ad5/3Δ24, a type 1 CRAD replicates in cancer cells inactive in the Rb/p16 pathway, was included as a control. Serum levels of liver enzymes, liver pathology, viral E1A DNA copy numbers and E1A RNA levels were evaluated 48 hours after i.v. injection of viruses into immunocompetent mice. Little to no hepatotoxicity was seen with the non replicative E1-deleted vector whereas pronounced effects were seen with the replication competent Ad5/3wt vector. Toxicities included submassive hepatic necrosis, high E1A DNA copy number and E1A RNA levels. Ad5/3SLPI demonstrated significantly lower liver pathology and E1A DNA and RNA copy numbers in comparison to Ad5/3wt and Ad5/3Δ24. In summary, Ad5/3SLPI is a promising candidate for treating metastatic ovarian cancer and showed robust virus replication, oncolysis, and in vivo therapeutic efficacy. Ad5/3SLPI showed comparatively low liver toxicity and therefore holds great potential for clinical testing.