The new dioxolane, (-)-2'-deoxy-3'-oxacytidine (BCH-4556, troxacitabine), has activity against pancreatic human tumor xenografts.

There is a great need for new therapeutic agents for patients with advanced pancreatic cancer. The new dioxolane analogue troxacitabine was evaluated in two human pancreatic cancer xenograft models. The models used included the Panc-01 and MiaPaCa pancreatic cancer cell lines. Whereas there is certainly no absolute evidence that either of the in vivo models is predictive for clinical activity, there is at least some evidence that they may be helpful in selecting agents for clinical trials in patients with pancreatic cancer. Troxacitabine was administered i.v. to the animals at doses of 10 and 25 mg/kg on a daily x 5 regimen. Gemcitabine was used as a positive control. The end points for the study included tumor growth inhibition (TGI), final weight, and the number of partial and complete tumor responses in the animals. Troxacitabine was highly active against the Panc-01 model (n = 8), with TGI levels of 88.5% and 84.3% at the 10 and 25 mg/kg doses, respectively. The mean final tumor weights for animals given troxacitabine were also significantly smaller (P < 0.001) compared with vehicle controls. At the 10 mg/kg dose, there were three partial tumor shrinkages and one complete tumor shrinkage, whereas at the 25 mg/kg dose, there were three partial tumor shrinkages. Troxacitabine had less activity against the MiaPaCa model (n = 10) and, by traditional response criteria, would be considered inactive, with TGIs of 4% and 22.7% at the 10 and 25 mg/kg dose level, respectively. Of note is that in comparison with gemcitabine, troxacitabine was more efficacious against Panc-01 and was equally active against MiaPaCa. These in vivo results are encouraging and support the prospect of performing Phase II and perhaps Phase III trials with troxacitabine in patients with advanced pancreatic cancer.