Purpose: In colorectal, breast, and head and neck cancers, response to 5-fluorouracil is associated with low expression of thymidylate synthase. In contrast, tumors with high expression of thymidylate synthase may be more sensitive to prodrugs such as 1-(2-deoxy-2-fluoro-beta-D-arabinofuranosyl) uracil (FAU) that are activated by thymidylate synthase. These studies were designed to evaluate FAU as a potential therapeutic and diagnostic probe.
Experimental design: [18F]-FAU and [3H]-FAU were synthesized with >97% radiochemical purity. [3H]-FAU or [18F]-FAU was administered intravenously to severe combined immunodeficient mice bearing either HT29 (low thymidylate synthase) or LS174T (high thymidylate synthase) human colon cancer xenografts. Four hours after [3H]-FAU dosing, tissue distribution of total radioactivity and incorporation of 1-(2-deoxy-2-fluoro-beta-D-arabinofuranosyl) 5-methyluracil (FMAU), derived from thymidylate synthase activation of FAU, into tumor DNA was measured. Positron emission tomography (PET) images were obtained for 90 minutes after injection of [18F]-FAU. Thymidylate synthase activity was determined in vitro in tumors from untreated mice by [3H] release from [3H]dUMP. Each cell line was incubated in vitro with [3H]-FAU or [3H]-FMAU in the absence or presence of 5-fluoro-2'-deoxyuridine (FdUrd) and then was analyzed for incorporation of radiolabel into DNA.
Results: Thymidylate synthase enzymatic activity in LS174T xenografts was approximately 3.5-fold higher than in HT29 xenografts, and incorporation of radioactivity derived from [3H]-FAU into LS174T DNA was approximately 2-fold higher than into HT29 DNA. At 240 minutes, radioactivity derived from [3H]-FAU was approximately 2-fold higher in tumors than in skeletal muscle. At times up to 90 minutes, PET imaging detected only small differences in uptake of [18F]-FAU between the tumor types. Fluorine-18 in skeletal muscle was higher than in tumor for the first 90 minutes and plateaued earlier, whereas [18F] in tumor continued to increase during the 90-minute imaging period. For both cell lines in vitro, FdUrd decreased the rate of incorporation of [3H]-FAU into DNA, whereas the incorporation of [3H]-FMAU was increased.
Conclusions: These results for FAU incorporation into DNA in vitro and in vivo further support clinical evaluation of FAU as a therapeutic agent in tumors with high concentrations of thymidylate synthase that are less likely to respond to 5-fluorouracil treatment. The high circulating concentrations of thymidine reported in mice may limit their utility in evaluating FAU as a PET probe.