Background Inhibition of tumor growth by endostatin has been shown to be an effective strategy in cancer therapy in mice However, its widespread application has been hampered by difficulties in a large scale production of the recombinant endostatin protein, rapid loss bioactivity of the protein, and the cumbersome daily administration These limitations could be resolved by in vivo delivery and expression of the endostatin gene In this study, we observed the effect and advantage of endostatin gene therapy mediated by a recombinant adenoviral vector (Ad/hEndo) on the growth of hepatocellular carcinoma BEL 7402 xenografted tumors, comparison with recombinant endostatin protein Methods Hepatocellular carcinoma BEL 7402 cells were inoculated subcutaneously in the flank of Balb/c nude mice Nine days after tumor cell inoculation, animals were given a cycle of four courses of intra tumoral injections of Ad/hEndo of 5×10 8 pfu (low dose group) and 1×10 9 pfu (high dose group) at intervals of six days, respectively Recombinant human endostatin protein (rhEndo) was administrated daily subcutaneously at a dose of 10 mg·kg 1 ·d 1 at a site nearby the tumor for ten days The expression of endostatin mRNA in tumor tissue was analyzed by reverse transcription polymerase chain reaction (RT PCR) after Ad/hEndo injection Dynamic changes of concentration of endostatin protein in tumor tissue were quantitated by enzyme linked immunosorbent assay (ELISA) Results After 4 courses of treatment, the tumor growth rates of high dose treated group with 1×10 9 pfu of Ad/hEndo were inhibited by 42 26% compared with the Ad/ LacZ control group ( P =0 001) and by 46 26% compared with the NIH buffer control group ( P =0 003), respectively However, in this study, Ad/hEndo at low dose of 5×10 8 pfu failed to demonstrate significant inhibition of tumor growth, compared with control groups After daily administration of recombinant human endostatin protein (rhEndo) for 9 days, the ratio of T/C (rhEndo group versus PBS group) was less than 47% However, two days after rhEndo treatment ceased, the ratio of T/C was more than 50% The peak of expression of endostatin mRNA in tumor tissue was at 2 or 3 days after administration intratumorally with Ad/hEndo of 1×10 9 pfu and gradually dropped undetectable by day 7 Dynamic analysis of endostatin concentration in tumor tissue showed that the highest level of mRNA is up at the third day after injection, and dropped to basal level three weeks later Conclusions Endostatin gene therapy mediated by a recombinant adenoviral vector had significantly inhibited the growth of hepatocellular carcinoma BEL 7402 xenografted tumors at a high dose of 1×10 9 pfu compared with other groups The analysis of dynamic expression of endostatin in vivo indicated that Ad/hEndo had acquired a high level, relatively long term expression in vivo and bioactivity capability Background Inhibition of tumor growth by endostatin has been shown to be effective strategies in cancer therapy in mice However, its widespread application has been hampered by difficulties in a large scale production of the recombinant endostatin protein, rapid loss bioactivity of the protein, and the cumbersome daily administration These measures could be resolved by in vivo delivery and expression of the endostatin gene In this study, we observed the effect and advantage of endostatin gene therapy mediated by a recombinant adenoviral vector (Ad / hEndo) on the growth of hepatocellular carcinoma BEL 7402 xenografted tumors, comparison with recombinant endostatin protein Methods Hepatocellular carcinoma BEL 7402 cells were inoculated subcutaneously in the flank of Balb / c nude mice Nine days after tumor cell inoculation, animals were given a cycle of four courses of intra-tumoral injections of Ad / hEndo of 5 × 10 8 pfu (low dose group) and 1 × 10 9 pfu (high dose group) at intervals of six days, respectively, Recombinant human endostatin protein (rhEndo) was administrated daily subcutaneously at a dose of 10 mg · kg 1 · d 1 at a site nearby the tumor for ten days The expression of endostatin mRNA in tumor tissue was analyzed by reverse transcription polymerase chain reaction (RT PCR) after Ad / hEndo injection Dynamic changes of concentration of endostatin protein in tumor tissue were quantitated by enzyme linked immunosorbent assay (ELISA) Results After 4 courses of treatment, the tumor growth rates of high dose treated group with 1 × 10 9 pfu of Ad / LacEndo were inhibited by 42 26% compared with the Ad / LacZ control group (P = 001 001) and by 46 26% compared with the NIH buffer control group (P = 0 003), respectively However, in this study, Ad / hEndo at low dose of 5 × 10 8 pfu failed to demonstrate significant inhibition of tumor growth, compared with control groups After daily administration of recombinant human endostatin protein (rhEndo) for 9 days, the ratio of T / C ( rhEnd o group versus PBS group was less than 47% However, two days after rhEndo treatment ceased, the ratio of T / C was more than 50% The peak of expression of endostatin mRNA in tumor tissue was at 2 or 3 days after administration intratumorally with Ad / hEndo of 1 × 10 9 pfu and abded undetectable by day 7 Dynamic analysis of endostatin concentration in tumor tissue showed that the highest level of mRNA is up at the third day after injection, and dropped to basal level three weeks later Conclusions Endostatin gene therapy mediated by a recombinant adenoviral vector had significantly inhibited the growth of hepatocellular carcinoma BEL 7402 xenografted tumors at a high dose of 1 × 10 9 pfu compared with other groups The analysis of dynamic expression of endostatin in vivo indicated that Ad / hEndo had acquired a high level, relatively long term expression in vivo and bioactivity capability