We have reported the elevation of uridine phosphorylase (UPase) in many solid tumors and the presence of a variant phosphorolytic activity in breast cancer tissues (M. Liu et al., Cancer Res., 58: 5418-5424, 1998). To better understand the biological and pharmacological significance of these findings, we have developed an UPase gene knockout embryonic stem (ES) cell model by specific gene targeting techniques. In this cellular model, we establish the critical role of UPase as an important anabolic enzyme in 5-fluorouracil (5-FU) activation and pyrimidine salvage pathway regulation. It has long been known that UPase regulates the plasma concentration of uridine; however, little is known of the role of UPase in the activation and metabolism of 5-FU and its derivatives, mainly because of the lack of an appropriate model system. The experimental data indicate that the disruption of UPase activity in murine ES cells leads to a 10-fold increase in 5-FU IC(50) and a 2-3-fold reduction in its incorporation into nucleic acids, whereas no differences in toxicity is seen with other pyrimidine nucleoside analogues such as 5-fluorouridine, 2'-deoxy-5-fluorouridine, and 1-beta-D-arabinofuranosylcytosine compared with WT (wild-type) ES cells. Benzylacyclouridine can specifically prevent the WT ES cells from the sensitivity of 5-FU. Our data also shows the effect of UPase on the cytotoxicity of 5'-deoxy-5-fluorouridine (5'DFUR), a 5-FU prodrug. The IC(50) is increased almost 16-fold in the knockout cells compared with the wild type cells, demonstrating the role of UPase in catalyzing the conversion of 5'DFUR to 5-FU. These findings additionally elucidate the tumor-specific selectivity of capecitabine, the oral fluoropyrimidine prodrug approved for the treatment of metastatic breast and colorectal cancers. Not only do the knockout cells present a decreased incorporation of 5-FU into nucleic acids but also an increased reliance on the pyrimidine salvage pathway. The reduced dependence of UPase knockout cells on the pyrimidine de novo synthesis is reflected in the apparent resistance to phosphonacetyl-L-aspartic acid, a specific inhibitor of pyrimidine pathway, with a 5-fold elevation in its IC(50) in UPase-nullified cells compared with WT. In summary, we have successfully generated an UPase gene knockout cell model that presents reduced sensitivity to 5-FU, 5'DFUR, and phosphonacetyl-L-aspartic acid, although it does not affect the basic cellular physiology under normal tissue culture conditions. Considering the role of UPase in 5-FU metabolism and the elevated expression of this protein in cancer cells compared with paired normal tissues, additional investigation should be warranted to firmly establish the clinical role of UPase in the tumor selective activation of 5-FU and capecitabine.