Sulfonyl-Containing Nucleoside Phosphotriesters and Phosphoramidates as Novel Anticancer Prodrugs of 5-Fluoro-2´-Deoxyuridine-5´- Monophosphate (FdUMP) Yuan-Wan Sun, Kun-Ming Chen, and Chul-Hoon Kwon†,* †Department of Pharmaceutical Sciences, College of Pharmacy and Allied Health Professions, St. John’s University, Jamaica, New York 11439 * To whom correspondence should be addressed. Department of Pharmaceutical Sciences, College of Pharmacy and Allied Health Professions, St. John’s University, 8000 Utopia parkway, Jamaica, NY 11439. Tel: (718)-990-5214, fax: (718)-990-6551, e-mail: [email protected]. Abstract A series of sulfonyl-containing 5-fluoro-2´-deoxyuridine (FdU) phosphotriester and phosphoramidate analogues were designed and synthesized as anticancer prodrugs of FdUMP. Stability studies have demonstrated that these compounds underwent pH dependent β-elimination to liberate the corresponding nucleotide species with half-lives in the range of 0.33 to 12.23 h under model physiological conditions in 0.1M phosphate buffer at pH 7.4 and 37 °C. Acceleration of the elimination was observed in the presence of human plasma. Compounds with FdUMP moiety (4-9) were considerably more potent than those without (1-3) as well as 5-fluorouracil (5-FU) against Chinese hamster lung fibroblasts (V-79 cells) in vitro. Addition of thymidine (10 µM) reversed the growth inhibition activities of only 5-FU and the compounds with FdUMP moiety, but had no effect on those without. These results suggested a mechanism of action of the prodrugs involving the intracellular release of FdUMP. Introduction 5-Fluoro-2´-deoxyuridine-5´-monophosphate (FdUMP) is the major metabolite responsible for the anticancer activity of 5-FU (Chart 1). FdUMP inhibits enzyme thymidylate synthase (TS, EC2.1.1.45) which catalyzes the reductive methylation of 2´-deoxyuridine-5´-monophosphate (dUMP) to 2´-deoxythymidine-5´-monophosphate (dTMP) in the de novo synthesis of 2´-deoxythymidine triphosphate (dTTP).1 Since TS catalyzed one of the rate-limiting steps in DNA synthesis, modulation of TS has often been targeted in the design of anticancer agents.2 Direct administration of FdUMP could potentially bypass the disadvantages of 5-FU including the toxic side effects, poor response rates and the development of drug resistance.3 FdUMP, however, does not readily penetrate cell membranes and is susceptible to degradation by nonspecific phosphohydrolase. Several nucleotide prodrug (pronucleotide) approaches have been attempted to circumvent the delivery problems of FdUMP.4-9 The major challenge of developing FdUMP prodrug is to select a suitable phosphate masking group so that the prodrugs could resist the degradation by phosphohydrolase and penetrate cell membranes to release sufficient FdUMP inside the tumor cells. We have previously exploited substituted sulfonylethyl groups as pro-moieties in the design of novel lipophilic prodrugs of phosphoramide mustard (PM, pKa =4.5), the ultimate alkylating species of the anticancer agent cyclophosphamide (CP) (Scheme 1).10 The sulfonyl-containing prodrugs were chemically stable, more cytotoxic than PM in vitro, and were substantially superior to CP against CP-resistant P388 cell line in vivo. Based on these results, the substituted sulfonylethyl groups may also serve as useful transport pro-moieties that could be employed in the delivery of anticancer agent FdUMP intracellularly. In this study, we have designed sulfonyl pronucleotides in the forms of phosphotriester, phosphoramidate monoester and bifunctional bis(2-chloroethyl) phosphoramidate monoesters. Analogous to sulfonyl PM prodrugs except that the leaving group is FdUMP instead of PM, the designed sulfonyl pronucleotides were expected to penetrate cell and then undergo β-elimination to produce mono anionic intermediates, which in turn would liberate FdUMP after further decomposition by either additional β-elimination or P-N bond hydrolysis (Scheme 2). Cleavage of P-N bond was presumably through the action of enzyme phosphoramidase. However, the bis(2-chloroethyl) phosphoramidate anion intermediate was expected to form spontaneously an aziridinium species which would undergo either P-N bond hydrolysis or alkylation with nucleophiles followed by cleavage of P-N bond to liberate FdUMP as reported. 11 Result and Discussion Selected substituted sulfonyl-containing nucleoside phosphotriesters and phosphoramidates were prepared for this study (chart 2). The normal nucleoside 2´-deoxyuridine (dU) was used to synthesize model compounds 1, 2 and 3. It was assumed that 5´-substituent on the pyrimidine ring of nucleoside moiety would have little effect on the β-elimination rate. In this study, we present the synthesis and growth inhibitory activities of these compounds and provide mechanistic evidence to suggest that these sulfonyl pronucleotides deliver FdUMP intracellularly and this account for their potent cytotoxicity. Chemistry The 3´-protected nucleosides 10 and 11 were prepared by treating 2´-deoxynucleosides with dimethoxytrityl chloride and tert-butyldimethylsilyl chloride followed by detritylation (Scheme 3). The 2-thiothanol starting materials were prepared by reaction of thiocresol or p-nitrothiophenol with 2-chloroethanol to produce 12 and 13 respectively. Reaction of thiocresol with propylene oxide yielded 14 (Scheme 4). The synthetic procedures used for sulfonyl phosphotriester and phosphoramidate compounds were based on P(III) chemistry. The phosphoramidite intermediates 15-19 were prepared via a two-step sequence involving phosphorylating of thioethanol followed by an in situ displacement of diisopropylamine with an appropriate thioethanol (12, 13 or 14) or 3´-protected nucleoside (10 or 11) (scheme 5). Subsequent conversion of phosphoramidite intermediates (15, 16 and 17) to phosphotriesters was accomplished by treatment with the appropriate nucleoside and 1H-tetrazole followed by in situ oxidation with t-butylhydroperoxide to obtain 1b, 4b, 7b and 9b. Reaction of phosphoramidites 18 and 19 with 1H-tetrazole in acetonitrile/H2O produced H-phosphonate intermediates 20 and 21, respectively. The H-phosphonate intermediates were then converted to phosphoramidate intermediates 2b, 3b, and 5b by oxidative coupling with ammonium or bis(2-chloroethyl)amine hydrochloride. However, the coupling reaction of 20 with bis(2-chloroethyl)amine hydrochloride resulted a very low yield of 3b (<10%). To improve the yield, an alternative one-pot strategy was applied (scheme 6). Phosphorus trichloride was reacted with the corresponding thioethanol in the presence of diisopropylethylamine followed by reaction with bis(2-chloroethyl)amine hydrochloride to generate a highly reactive monochloro intermediate which was then reacted with the FdU in situ and oxidized by tert-butylhydroperoxide to yield 6a and 8a. Intermediates 1a-5a, 7a and 9a were obtained by treatment of 1b-5b, 7b and 9b with tetra-n-butylammonium fluoride, respectively. The target sulfonyl compounds were prepared by oxidation of intermediates 1a-9a with H2O2 in the presence of catalyst ammonium molybdate. However, compound 7 was not able to be obtained because it decomposed during the process of oxidation. Phosphoramidate analogues were obtained as a mixture of diasteromers without further purification because they produced the identical nucleotides or phosphoramidate enantiomers after β-elimination activation. Certain compounds turned out to be very hydroscopic. Stability Studies Model sulfonyl compound 1 underwent base-catalyzed β-elimination of dU with half life (t1/2) of 58.8 min at pH 7.4, 5.28 h at pH 6.4 and 21.15 h in 0.1 M phosphate buffer at 37 °C (Table 1). Additionally, a comparison of pH-dependent profiles of 1 and 4 indicated that the influence of 5´-substitution pyrimidine to the rate of β-elimination was insignificant. Having established that compound 1 undergoes β-elimination at physiological pH, several sulfonyl pronucleotides (chart 2) were synthesized to examine their ability to liberate nucleotide intracellularly. The rate of β-elimination of these sulfonyl pronucleotides was determined by measuring the rate of disappearance of pronucleotides upon incubation in 0.1 M phosphate buffer, pH 7.4 and 37 °C with the use of quantative TLC technique. Results are presented in Table 2. The effects of R substitute and leaving group to the rate of β-elimination have been discussed in our previously studies. Compound 5 and 6 exhibited greater stability than 4, apparently due to the change of the nature of the leaving group (e.g. phosphoramidic acid vs. phosphoric acid) or the number of sulfonylethyl group accessible for β-elimination. Moreover, compound 5 containing P-NH2 phosphoramidate linkage showed a longer half-life (~2×) than the corresponding bis(2-chloroethyl) phosphoramidate analogue 6. Consistent with our previous findings, the electron donating p-tolylsulfonyl substituent in analogue 6 reduces acidity of the α-proton, thereby decreasing the β-elimination rate as compared to 8 (0.33 h vs. 5.73 h). The β-methyl substitution has dramatically prolonged the half-life of compound 9 over 4 (7.05 h vs. 0.95 h). The electron-inductive effects and steric hindrance of β-methyl may explain the slower β-elimination in 9. Our previous studies have shown that the catalytic effect of human serum albumin (HSA) on the β-elimination rates of Aldo and the sulfonylethyl analogues of PM are essentially equivalent10. Consistent with this finding on sulfonylethyl PM analogues, the sulfonylethyl pronucleotides displayed a similar