Background: 5-fluorouracil(5FU) has been the most widely used chemotherapeutic agent in the treatment of GI malignancies. Understanding 5FU's precise mechanism of action would help guide clinicians and may lead to drug optimization. 5FU inhibits thymidylate synthase(TS) and thus causes a ‘thymidine-less’ death in proliferating cancer cells. 5FU sensitivity has been linked to several tumor markers with a major focus on TS. In general, clinical studies show that high TS expression is correlated with 5FU resistance. TS genetic alterations in cancer cells have been reported as well. Others have claimed p53 status, which is mutated in over 75% of cancers, is a determinant of 5FU sensitivity.
Methods: To study the importance of different levels of TS expression and mutations and p53 genetic status on 5FU sensitivity, we utilized outbred GI cancer cell lines and cancer knock out models. Previously, we showed that 5FU inhibits TS in yeast similarly as in cancer cells. Hence, TS knockout yeast cells were used as a model to test 5FU toxicity in a TS-null system. In search for TS-independent targets of 5FU, we employed drug sensitivity studies with the addition of exogenous uridine and thymidine. We also treated cancer cells with C14-labeled 5FU to screen for novel targets covalently modified by 5FU metabolites.
Results: Previously, we reported only extremely low TS expression levels made cancer cells hypersensitive to 5FU (a model of LOH with only 1 TS allele). Surprisingly, yeast cells null for TS were more 5FU sensitive compared to wild-type yeast (TS intact). All previously described TS mutations and polymorphisms did not influence 5FU sensitivity. In a knockout model and pancreatic and colon outbred cell lines, we found no connection between p53 genetic status and 5FU sensitivity. Uridine, and not thymidine, rescued cancer and yeast cells from 5FU toxicity, underscoring the significance of RNA-related, 5FU targets. Studies using C14-labeled 5FU revealed two candidate targets apart from TS.
Conclusion: In GI cancers where one TS allele exists, cancer cells are hypersensitive to 5FU, and may be a predictive marker for a subset of cancers. However, our work shows that most forms of TS and p53 in cancer cells are not predictive of 5FU sensitivity. Taken together, our work suggests that RNA-related targets of 5FU is more informative than p53 and TS for predicting 5FU sensitivity in GI cancers.