Background: Cancer of the exocrine pancreas is the fourth leading cause of cancer-related deaths in the United States. We have previously shown that combination therapy with parthenolide, a nuclear factor-kappaB (NF-κB) inhibitor, and sulindac, a cyclooxygenase (COX) inhibitor, enhanced the growth inhibition of pancreatic cancer cells in vitro. A bioavailable analog of parthenolide, LC-1, has recently been synthesized. The aim of the present study was to evaluate the efficacy of LC-1 and the COX-2 inhibitor celecoxib in vivo. Methods: To determine the dose of celecoxib for in vivo studies, the effect of celecoxib on prostaglandin E2 (PGE2) levels was determined ex vivo in homografted hamster PC-1 pancreatic tumors following a single administration of celecoxib. For in vivo studies, a carcinogen-induced developmental model of pancreatic cancer was employed. Syrian golden hamsters (n=140) were injected with 10mg/kg N-nitroso-bis[2-oxopropylamine] (BOP), once weekly for six weeks. Following the first BOP injection, hamsters were randomized into seven groups: placebo, low/high dose LC-1 (20 or 40 mg/kg/day), low/high dose celecoxib (10 or 50mg/kg/day), or combination LC-1/celecoxib (low/low or high/high). Hamsters were sacrificed after 32 weeks, and the presence of tumors in pancreata and other organs was noted. Pancreatic intraepithelial neoplastic lesions (PanIN1, 2, or 3) were identified microscopically following hematoxylin/eosin staining of fixed tissues. COX-2 expression and mucin production were localized by immunohistochemistry. Tissue PGE2 and plasma PGEM levels were determined by ELISA. Results: In ex vivo studies, 20 hours after treatment with 12.5 or 50mg/kg celecoxib, PGE2 levels decreased ~30% or 70% respectively relative to placebo. The in vivo trial showed that 40mg/kg LC-1 or 50mg/kg celecoxib alone decreased the size of gross pancreatic cancers relative to placebo, reaching significance for the LC-1 group (3.8 +/- 0.72 vs. 8.7 +/- 2.1mm, P<0.05). The number of gross pancreatic tumors was not significantly affected by any of the treatments. Similarly, the incidence of PanIN lesions did not differ except for an increase in PanIN3 incidence in the low dose combination group relative to placebo. Metastasis to other organs was only noted in the placebo and low dose LC-1 groups. COX-2 and mucin were localized to pancreatic lesions. Tissue PGE2 levels were significantly decreased in the high dose LC-1/celecoxib group relative to placebo (0.52 +/- 0.05 vs. 0.82 +/- 0.09 pg/μg protein, P<0.05). Plasma PGEM levels were significantly decreased by 50mg/kg celecoxib relative to placebo (78.9 +/- 15.9 vs. 141.6 +/- 22.5 pg/ml, P<0.05), confirming suppression of COX-2 activity. Conclusions: LC-1 or celecoxib alone decreased the size of pancreatic tumors and inhibited their metastatic progression. Therefore, these agents may be effective as chemotherapy for pancreatic cancer but were insufficient to prevent pancreatic cancer.

98th AACR Annual Meeting-- Apr 14-18, 2007; Los Angeles, CA