Reply

We agree with Cobben et al. (1) that competition between endogenous thymidine and FLT for nucleoside carrier proteins is likely to decrease the absolute amount of FLT transported into tissues. The absolute amount of FLT transported into tissues can be enhanced by degrading blood thymidine with i.v. TP as suggested by Cobben et al. (1) or by feeding rodents a thymidine-deficient diet. Such strategies may, however, produce complications in interpretation. For instance, by catalyzing the reversible phosphorolytic cleavage of thymidine, TP can enhance the sensitivity of tumors to TS inhibitors (such as 5-fluorouracil used in our studies; Ref. 2). Because endogenous levels of thymidine in the plasma/tissues will compete for tissue uptake of FLT, this needs to be accounted for. Ideally, one would wish to express FLT uptake via the salvage pathway as the flux of thymidine from plasma into DNA precursors (the product of thymidine concentration in plasma and FLT flux constant; μmol/min/g). This is not too dissimilar a position as [¹⁸F]fluorodeoxyglucose relative to glucose. For this, a lumped constant that defines the differential volumes of distribution and TK₁-dependent phosphorylation of FLT versus thymidine is also required. Given the encouraging result with respect to therapy response monitoring, investment is now needed to determine the tissue extraction ratio between FLT and thymidine. This will be required for normal and neoplastic tissue over a range of endogenous thymidine plasma levels. The resulting data will allow studies to be interpreted irrespective of species.

Thymidine levels are not the only known differences between rodents and humans that affect FLT kinetics. FLT, like its congener 3′-azido-3′deoxythymidine, does not undergo substantial glucuronidation in the mouse within 60 min (2). Glucuronidation increases in the order mouse < dog < humans (2, 3, 4). Studies of 3′-azido-3′deoxythymidine metabolism with liver microsomes and hepatocyte cultures suggests that this phenomenon may be attributable to interspecies differences in isozyme expression of uridine diphosphoglucuronosyltransferase (5). In this regard, FLT studies in mice systematically underestimate the degree of glucuronidation in humans. Interspecies differences would be dealt with systematically with the above strategy together with metabolite profiling.

A major advantage for using FLT in measuring cellular proliferation is its lack of degradation. The degradation of pyrimidine nucleosides is initiated within the sugar ring by TP. The fact that ring-substituted nucleosides such as FLT and 5-[methyl-¹¹C]-2′-fluoro-5-1-β-d-arabinofuranosyl uracil are stable to degradation is evidence that these compounds are not substrates for TP. Thus, the discrepancy between TK₁ levels and FLT uptake seen in our studies (2), i.e., after a single dose of 5-fluorouracil TK₁ levels decreased at 24 h and increased beyond control levels at 48 h, cannot be explained by differences in TP levels.

As with thymidine, FLT is a substrate for TK₁. Evidence for this has been provided in vitro (6, 7) and in vivo (8, 9, 10). TK₁ catalyzes the phosphorylation of FLT to FLT-phosphate. Because it lacks a 3′-hydroxyl group, very little FLT is incorporated into DNA. Thus, FLT measures an early event in DNA synthesis rather than DNA incorporation. TK₁ activity is up-regulated in cells entering S phase; the protein is nearly undetectable in growth arrested cells (11). In the study by Dittmann et al. (12), FLT uptake decreased at 24 and 72 h after 4 h of cisplatin treatment but increased after treatment with the TS inhibitors methotrexate and 5-fluorouracil (there is an error in the reporting of this fact by Cobben et al.). The cisplatin effect was caused by cell cycle arrest. The TS inhibitors effect was caused by enhanced utilization of the salvage pathway. We have demonstrated this phenomenon in vitro (13) and in vivo in man (14) using radiolabeled thymidine. FLT is a promising PET radiotracer for imaging antiproliferative activity. However, more work is required to develop its full imaging potential.