Search

Our laboratory recently reported the development of novel prodrug chemistry for the intracellular delivery of phosphotyrosine mimetics. This chemistry has now been adapted for the synthesis of a prodrug that delivers the nonhydrolyzable difluoromethylphosphonate moiety intracellularly. Activation of the prodrug generates a difluoromethylphosphonamidate anion that undergoes subsequent cyclization and hydrolysis with a t1/2 = 44 min. A highly potent and selective inhibitor of protein tyrosine phosphatase 1B (PTP1B) with a nanomolar Ki has been reported, but this bis(difluoromethylphosphonate) lacks potential utility due to its exceedingly low membrane permeability at physiological pH. A prodrug of this inhibitor has been synthesized and evaluated in a cell-based assay. The prodrug exhibits nanomolar PTP1B inhibitory activity in this assay, confirming the efficacy of intracellular phosphonate delivery using this prodrug approach.

Both enantiomers of protected and unprotected 2-hydroxymethylaziridines are efficiently and enantiospecifically synthesized by using a combination of enzymatic and synthetic methods. PPL was used for lipase-catalyzed desymmetrization of N-protected serinol. [reaction: see text].

A novel approach to the intracellular delivery of aryl phosphates has been developed that utilizes a phosphoramidate-based prodrug approach. The prodrugs contain an ester group that undergoes reductive activation intracellularly with concomitant expulsion of a phosphoramidate anion. This anion undergoes intramolecular cyclization and hydrolysis to generate aryl phosphate exclusively with a t(1/2) = approximately 20 min. Phosphoramidate prodrugs (8-10) of phosphate-containing peptidomimetics that target the SH2 domain were synthesized. Evaluation of these peptidomimetic prodrugs in a growth inhibition assay and in a cell-based transcriptional assay demonstrated that the prodrugs had IC50 values in the low micromolar range. Synthesis of phosphorodiamidate analogues containing a P-NH-Ar linker (16-18) was also carried out in the hope that the phosphoramidates released might be phosphatase-resistant. Comparable activation rates and cell-based activities were observed for these prodrugs, but the intermediate phosphoramidate dianion underwent spontaneous hydrolysis with a t(1/2) = approximately 30 min.

Previously a series of 2- and 3-substituted indolequinone phosphoramidate prodrugs was synthesized, and the compounds were shown to be nanomolar inhibitors of cell proliferation. The activation of these compounds following both one- and two-electron reduction has been investigated. (31)P NMR experiments demonstrated that both series of compounds undergo rapid activation following two-electron reduction. Additionally, the 3-series of compounds undergo rapid activation following one-electron reduction, while activation of the 2-series of compounds via this mechanism is very slow. The activation of these prodrugs by direct displacement using sulfur nucleophiles such as glutathione has been examined. Activation via this route is rapid for the 3-regioisomers, but is considerably slower for the 2-substituted analogues under similar conditions. Together these findings suggest that drug delivery via two-electron reduction from the 2-position is the more selective prodrug strategy.

A series of 2- and 3-substituted indolequinone phosphoramidate prodrugs targeted to DT-diaphorase (DTD) have been synthesized and evaluated. These compounds are designed to undergo activation via quinone reduction by DTD followed by expulsion of the phosphoramide mustard substituent from the hydroquinone. Chemical reduction of the phosphoramidate prodrugs led to rapid expulsion of the corresponding phosphoramidate anions in both series of compounds. Compounds substituted at the 2-position are excellent substrates for human DTD (k(cat)/K(M) = (2-5) x 10(6) M(-1) s(-1)); however, compounds substituted at the 3-position are potent inhibitors of the target enzyme. Both series of compounds are toxic in HT-29 and BE human colon cancer cell lines in a clonogenic assay. There was a correlation found between cytotoxicity and DTD activity for the 2-series of phosphoramidates; however, there was no correlation between cytotoxicity and DTD activity in the 3-series of compounds. This finding suggests the presence of an alternative mechanism for the activation of these compounds.

[reaction--see text] Sugar nucleoside diphosphates have been prepared using an efficient phosphate coupling reaction that employs a highly reactive zwitterionic phosphoramidate intermediate as the phosphorylating species.

A series of naphthoquinone and benzimidazolequinone phosphorodiamidates has been synthesized and studied as potential cytotoxic prodrugs activated by DT-diaphorase. Reduction of the quinone moiety in the target compounds was expected to provide a pathway for expulsion of the phosphoramide mustard alkylating agent. All of the compounds synthesized were excellent substrates for purified human DT-diaphorase (k(cat)/K(m) = 3 x 10(7) - 3 x 10(8) M(-1) s(-1)). The naphthoquinones were toxic to both HT-29 and BE human colon cancer cell lines in a clonogenic assay; however, cytotoxicity did not correlate with DT-diaphorase activity in these cell lines. The benzimidazolequinone analogues were 1-2 orders of magnitude less cytotoxic than the naphthoquinone analogues. Chemical reduction of the naphthoquinone led to rapid expulsion of the phosphorodiamidate anion; in contrast, the benzimidazole reduction product was stable. Michael addition of glutathione and other sulfur nucleophiles provides an alternate mechanism for activation of the naphthoquinone phosphorodiamidates, and this mechanism may contribute to the cytotoxicity of these compounds.

A series of novel nitroheterocyclic phosphoramidates has been prepared, and the cytotoxicity of these compounds has been evaluated in clonogenic assays against B16, wild-type and cyclophosphamide-resistant MCF-7, and HT-29 cells under aerobic conditions and HT-29 cells under hypoxic conditions. All compounds were comparable in toxicity to wild-type and resistant MCF-7 cells and were also selectively toxic to HT-29 cells under hypoxic conditions (selectivity ratios 1.7 to >20). Analogues lacking the nitro group were not cytotoxic. Electron-withdrawing substituents increased cytotoxicity under aerobic conditions and thereby decreased hypoxic selectivity. In contrast, an electron-donating substituent markedly decreased both aerobic and hypoxic cytotoxicity but enhanced hypoxic selectivity. Chemical reduction of the nitro group resulted in rapid expulsion of the cytotoxic phosphoramide mustard. The most potent of these compounds show significant cytotoxicity under both aerobic and hypoxic conditions.

A series of alkylating phosphoramidate analogs of 5-fluoro-2'-deoxyuridine has been prepared and their growth inhibitory activity evaluated against murine L1210 leukemia and B16 melanoma cells in vitro. These compounds were designed to undergo intracellular release of the phosphoramidate anions, which it was hoped would function as irreversible inhibitors of thymidylate synthase. The expectation was that binding of the nucleoside moiety would be followed by alkylation of the enzyme via the phosphoramidate. The chloride, bromide, iodide, and tosylate analogs were highly potent inhibitors of L1210 cell proliferation, with increased inhibition observed at both higher drug concentrations and longer exposure times. Addition of thymidine completely reversed the inhibition for all compounds, suggesting that these compounds are acting via inhibition of thymidylate synthase. Although the nonalkylating morpholine analog 1f was ca. 50-fold less potent than the methyl(chloroethyl)amino compound, the piperidine analog 1g was only 2-fold less potent, confirming that nitrogen basicity may be as important as the presence of an alkylating group. Addition of thymidine reversed the growth inhibition of the morpholine and piperidine analogs, suggesting that these compounds may also undergo intracellular conversion to 5-fluoro-2'-deoxyuridine 5'monophosphate. The thymidine and deoxyuridine derivatives 2 and 3 showed minimal growth inhibition in the L1210 assay. The alkylating analogs showed modest cytotoxicity against B16 melanoma cells, and the potency of the analogs was more dependent upon the alkylating moiety than on the 5-substituent.

A series of novel nitrobenzyltetrakis(chloroethyl)phosphorodiamidates has been prepared, and its cytotoxicity has been evaluated against HT-29 cells under aerobic and hypoxic conditions and against murine bone marrow progenitor cells under aerobic conditions. All compounds were selectively toxic to HT-29 cells under hypoxic conditions, and the selectivity ratios varied from 1.6 to > 90. Analogs lacking either the nitro group or the tetrakis(chloroethyl) moiety were not cytotoxic, confirming that the presence of both nitro and incipient alkylating groups are essential for activity. Surprisingly, some analogs were far more toxic to bone marrow progenitors than to HT-29 cells under aerobic conditions, suggesting that other activation mechanisms must exist in these hematopoietic cells. Cytotoxicity increased with increasing depth in the HT-29 spheroid model, consistent with the preferential hypoxic toxicity of these compounds. Alkaline elution experiments showed a greater number of DNA interstrand cross-links under hypoxic compared to aerobic conditions. The extent of cross-linking in hypoxic cells was essentially identical to that produced by an equitoxic dose of melphalan, suggesting that the cytotoxicity of these compounds results from phosphorodiamidate release and alkylation of DNA.

31P NMR spectroscopy was used to study the solvolysis kinetics of a novel series of alkylating monoester phosphoramidates (4a-d) under model physiologic conditions. Halide ion kinetics were used to determine the rate of aziridinium ion formation. The solvolysis rates showed the expected dependence upon substitution at the reactive nitrogen; comparison of 4a with phosphoramide mustard (1a) indicated that replacement of the amino group by alkoxy decreased the solvolysis rate by approximately 10-fold. The rate of conversion of starting compound (4a-d) to solvolysis product was essentially equal to the rate of halide ion release, suggesting that the aziridinium ion is a short-lived intermediate. 1H NMR and 31P NMR kinetics experiments performed in the absence and presence of trapping agent (dimethyldithiocarbamate) confirmed that the aziridinium ion was too short-lived to be observed via NMR. These compounds were also tested for cytotoxicity against L1210 leukemia and B16 melanoma cells in vitro; the monoalkylators 4c and 4d showed no activity, 4a was weakly cytotoxic, and 4b was comparable in activity to phosphoramide mustard.

4-Hydroxycyclophosphamide (2/3) of unknown stereochemistry is the initial metabolite formed after administration of cyclophosphamide (1). Ultimate conversion to the cytotoxic metabolite phosphoramide mustard (6) is initiated by ring opening of 4-hydroxycyclophosphamide to produce aldophosphamide (4). The ring-opening reaction and subsequent equilibration of 2-4 are subject to general-acid catalysis, and the equilibrium composition is independent of buffer structure and pH. In contrast, formation of 6 from 4 proceeds by general-base-catalyzed beta-elimination. trans-4-Hydroxycyclophosphamide undergoes ring opening ca. 4 times faster than the cis isomer, and cyclization of 4 favors the trans isomer by a factor of ca. 3 over the cis isomer. The rapid equilibration of 2-5 and the absence of elimination to give 6 at pH approximately 5 provides a convenient method to prepare a stable equilibrium mixture of activated cyclophosphamide metabolites suitable for in vitro use.

cis-4-Hydroxycyclophosphamide (2) and aldophosphamide (4) were generated in aqueous phosphate or cacodylate buffer by dimethyl sulfide reduction of cis-4-hydroperoxycyclophosphamide (8) and by sodium periodate cleavage of 3,4-dihydroxybutyl N,N-bis(2-chloroethyl)phosphorodiamidate (9), respectively; the reactions of 2 and 4 were examined by 1H and 31P NMR. Within 30-60 min (pH or pD 7.0, 25 degrees C) the same pseudoequilibrium mixture was established in both reactions, with cis- and trans-4-hydroxycyclophosphamide (2 and 3), aldophosphamide (4), and its hydrate (5) present in the approximate ratio of 4:2:0.3:1. Structures of the intermediates were assigned unambiguously based upon analysis of the chemical shifts and coupling constants in the proton spectra determined in D2O buffers, and the 31P assignments followed by correlation of component ratios at equilibrium. Free energy differences of 0.4, 0.4, and 0.7 kcal/mol at 25 degrees C were estimated between 2, 3, 5, and 4, respectively, with 2 being the most stable. The aldehyde 4 reacted most rapidly with water to give hydrate 5; cyclization of 4 to 3 occurred faster than to 2, and the rate of cyclization to 2 was comparable to that for elimination to 6. Compound 5 is formed much faster than 3 from the diol cleavage, but 5 and 3 are produced at comparable rates from 2, suggesting that conversion of 2 to 3 can proceed by a mechanism other than ring opening. The rate of equilibration appears to be independent of buffer structure, indicating that bifunctional catalysis is not important in the ring-opening reaction.(ABSTRACT TRUNCATED AT 250 WORDS)