Department of Pharmacology and Therapeutics
Roswell Park Cancer Institute
Elm and Carlton Streets
Buffalo NY USA 14263
Tel: 716 - 845 - 8249
Fax: 716 - 845 - 8857
Email: john.mcguire@roswellpark.org

Program Antifolates
Antimetabolites
Enzymology
Drug Resistance

This laboratory explores biochemical and molecular aspects of: (1) folic acid (folate) metabolism in tumor and normal tissues; (2) the action of antifolates currently used in cancer chemotherapy and of experimental antifolates; and (3) acquired and natural antifolate resistance. Novel antifolate drugs or therapeutic modalities involving folate metabolism are developed and characterized based on concepts derived from these studies.

Fundamental biochemical and molecular studies are designed to elucidate aspects of folate metabolism relevant to cancer therapeutics. An area of ongoing interest is the synthesis, regulation, and function of poly(g-glutamyl) metabolites of folates and antifolates in proliferation and in resistance. Folylpolyglutamates are essential for normal folate metabolism and hence for cell viability. Thus, agents interfering with folylpolyglutamate synthesis and function may be therapeutically useful and could also be used as probes of their function. Folylpolyglutamate synthetase (FPGS) is being characterized from a recombinant human source expressed in Sf9 insect cells, from human tumor cell lines, and from normal murine intestinal epithelium to provide a rational basis for design of selective FPGS inhibitors. Agents designed based on these studies are synthesized by collaborating chemists (primarily Dr. J. K. Coward, University of Michigan, Ann Arbor and Dr. A. Gangjee, Duquesne University). Three classes of agents have been designed and synthesized thus far: (1) direct FPGS inhibitors; (2) folate analogs that are nonsubstrates for FPGS; and (3) agents with enhanced ability to form short chain length polyglutamates. Biochemical and pharmacological properties of new agents are examined in detail and the results are used to refine the structures. The subcellular compartmentalization of FPGS expression and factors regulating FPGS expression are also of interest. These studies should provide important information about the role of subcellular compartmentalization of folate metabolism.

Biochemical and molecular studies of drug sensitivity and resistance may lead to increased efficacy in clinical antifolate use. The mechanism of action of methotrexate (MTX) and other “classical” antifolates and their metabolism to poly(g-glutamates) are being explored in antifolate-sensitive and antifolate-resistant human tumor cell lines. These studies may identify factors relevant to natural and acquired MTX resistance as well as provide data allowing more judicious clinical use of these agents.

Progress

Mutational deletion of FPGS activity is lethal, thus FPGS is a potential target for cancer chemotherapy. We pioneered a two-track strategy to design FPGS inhibitors. One track identifies changes in the acceptor Glu substructure of folate analogs that: (1) directly impart FPGS inhibition; or (2) inform about the acceptor Glu subsite of the FPGS active site. The second track is to define other structural requirements, generally in the extended heterocycle/pteroate moiety, for increasing potency and/or specificity of lead inhibitors identified in the first track; this involves extensive studies of human FPGS substrate specificity. Progress on both tracks has been expedited since obtaining from Dr. B. Shane (UC Berkeley) a baculovirus shuttle vector containing the human cytosolic FPGS (cFPGS; below) cDNA. High-level expression in Sf9 insect cells has been obtained using plaque-purified recombinant baculovirus. cFPGS is purified using techniques developed in this laboratory.

Detailed knowledge of the enzyme mechanism of human FPGS will be useful in inhibitor design. In addition to natural folates, FPGS catalyzes the ATP-dependent ligation of glutamic acid to the anticancer drug 5,10-dideaza-5,6,7,8-tetrahydrofolate (Lometrexol™, (6R)-DDATHPteGlu1), which is known to be readily converted to polyglutamates in vivo. Kinetic characteristics of the mono- and polyglutamate forms were determined so that the effects of the presence of polyglutamate products on FPGS activity might be understood. Synthesis of (6R)-DDATHPte[14C]Glu, and unlabeled mono- and various polyglutamates, DDATHPteGlun (6R, n=1-6; 6S, n=1-2) was effected from (6R)- or (6S)-5,10-dideazatetrahydropteroyl-azide and [14C]glutamic acid, glutamic acid, or H-Glu-g-Glun-g-Glu-OH (n=0-4), respectively. These compounds were evaluated as FPGS substrates to determine steady-state kinetic constants for subsequent mechanistic studies. Saturation kinetics were observed for (6R)-DDATHPteGlu1, the isomer corresponding to natural tetrahydrofolate (H4PteGlu), whereas marked substrate inhibition was observed for (6S)-DDATHPteGlu1 and all DDATHPteGlun polyglutamate substrates, except for DDATHPteGlu6. Multiple ligation of glutamate renders a quantitative analysis of these data difficult. However approximate values of Km = 0.65 –1.6 µM and KI = 144–417 µM for DDATHPteGlun were obtained using a simple kinetic model. Preliminary mechanistic studies, including time course, substrate trapping, and pulse-chase experiments, have provided strong evidence in favor of a processive, rather than distributive, mechanism of multiple glutamate ligations. The degree of processivity is dependent on the concentration of the folate substrate, thus suggesting a mechanism for regulation of folate polyglutamate synthesis in cells.

With regard to track 1 of inhibitor design, we have previously examined the FPGS inhibitory potency of two classes of phosphorous-containing pseudopeptide folate analogs (Tsukamoto et al. Arch. Biochem. Biophys. 355: 109-118, 1998; McGuire et al. Biochem. Pharmacol., 65: 315-318, 2003) synthesized by Dr. Coward that were designed to mimic the tetrahedral intermediate formed in the ATP-dependent reaction catalyzed by folylpolyglutamate synthetase (FPGS). To allow direct comparison, both classes were synthesized with the same heterocycle as MTX. Methotrexate-phosphinate (MTX-phosphinate; 4-amino-10-methyl-pteroyl-L-Glu-g-[Y{P(O)(OH)-CH2}]glutarate) is a more highly potent (Kis, 3.1 nM), competitive inhibitor of recombinant human cFPGS than is MTX-phosphonate (4-amino-10-methyl-pteroyl-L-Glu-g-[Y{P(O)(OH)-O}]glutarate; Kis, 46 nM). For both inhibitors, FPGS inhibition is not time-dependent and preincubation of FPGS, inhibitor, and ATP does not potentiate inhibition, within experimental limits. These results suggest that slow phosphorylation to produce a more potent inhibitor form is not involved. Neither MTX-phosphinate nor MTX-phosphonate is growth inhibitory to human CCRF-CEM leukemia cells at 1 µM (70-fold above the concentration of MTX giving 50% growth inhibition), probably because of poor transport. Because of its exceedingly high potency as an FPGS inhibitor, MTX-phosphinate represents a lead structure from which cell-permeable analogs may be developed to test the hypothesis that FPGS inhibition is therapeutically efficacious. Further studies are aimed at determining: (a) the potency of the two isomers of the chiral phosphinate moiety; (b) the effect of the position of the phosphinate group in analogs of longer polyglutamates; and (c) the optimal heterocycle to ligate to the phosphinate to achieve potent and specific FPGS inhibition.

With regard to track 2 of inhibitor design, a number of novel heterocycles and substituted variations have been studied for their FPGS substrate efficiency as leads for increasing potency and/or specificity of phosphinate inhibitors. Substrate efficiency, as we have shown, is a surrogate for inhibitory potency when the heterocycle is then appended to an inhibitory Glu analog (above). Analogs evaluated were synthesized by Dr. A. Gangjee (Duquesne U). We previously identified the 6,5-fused ring furo[2,3-d]pyrimidine and pyrrolo[2,3-d]pyrimidine hetero-cycles as enhancing FPGS substrate activity. Further substitutions in these two heterocycles were evaluated in an attempt to optimize affinity for FPGS. In particular it was shown that 5-substitution of 2,4-diamino-furo[2,3-d]pyrimidines with a C4 chain in the bridge to the p-aminobenzoylglutamate (pABGlu) most common in polyglutamylatable antifolates gave substrates with Km values ≤ 2 µM with little effect on Vmax. Other substitutions of furo[2,3-d]pyrimidine and pyrrolo[2,3-d]pyrimidine heterocycles were shown to be less advantageous. In addition, tricyclic structures, in which the C9-N10 bridge was incorporated into the third ring, were FPGS substrates (although with diminished affinity) suggesting that conformational restriction might be exploited in design of FPGS inhibitors. Several of the analogs had unusual properties that might be of therapeutic significance. For example, pyrrolo[2,3-d]pyrimidine-containing heterocycles were examined that appeared to inhibit both thymidylate synthase (TMPS) and dihydrofolate reductase (DHFR), similar to the activity of the newly FDA-approved agent pemetrexed. Other modifications allow binding of inhibitor to TMPS while TS binds to its own message and thus do not cause the translational induction of TMPS that occurs with the TMPS inhibitors now used clinically, Tomudex and pemetrexed. Other modifications changed the locus of action to purine synthesis. The furo[2,3-d]pyrimidine- and pyrrolo[2,3-d]pyrimidine-based antifolates were thus recognized as a diverse group with the potential to inhibit most, if not all, of the recognized folate-dependent pathways recognized as critical in cancer chemotherapy.

We previously identified human mitochondrial (mFPGS) and cytosolic (cFPGS) FPGS isoforms and others have shown that they are encoded by one gene. We also showed that the two forms have different electrophoretic mobilities in SDS-PAGE. Because of the potential role(s) of cFPGS and mFPGS isoforms in antifolate sensitivity and resistance, we have continued our study of these isoforms. In order to determine the submitochondrial location of mFPGS, we adapted literature methods to purify mitochondria from CCRF-CEM human leukemia cells in order to be able to isolate mFPGS free from cFPGS. Since there are no well-described, published methods for isolating pure human mitochondria, considerable effort was expended developing these methods. We are now able to obtain a good yield of intact human mitochondria from CCRF-CEM cells. In addition, methods for submitochondrial fractionation with digitonin, which are best described for rat liver where large quantities of mitochondria can be obtained, had to be adapted for human cell culture sources. Human mitochondria were fractionated with increasing concentrations of digitonin to successively extract the four submitochondrial compartments. Western analyses of the fractions using protein markers specific for each compartment suggest that mFPGS is distributed in the matrix and/or inner membrane compartments. Interaction of any FPGS with membranes has not previously been reported, although there is literature precedent for matrix proteins that are partially associated with the inner membrane. Further support for an interaction of mFPGS with the inner mitochondrial membrane is provided by localization of about half of the mFPGS in the mitochondrial membrane fraction obtained by freeze-thaw of intact mitochondria; the remaining mFPGS is located in the soluble fraction. Resistance of about half of the mFPGS in whole mitochondria to alkaline carbonate extraction suggests that its interaction with the inner membrane is more similar to an integral, than a peripheral, membrane protein. The data suggest that human mFPGS is at least in part strongly associated with the inner mitochondrial membrane. Because of the success of our method, we were able to collaborate with Dr. B.J. Dolnick of this center in his efforts to characterize the subcellular distribution of the rTSb, a protein involved in growth regulation that is translated from a transcript that is a partial antisense of the TMPS gene. Bioinformatics analysis indicated that some transcripts encoded N-terminal mitochondrial signaling sequences. Using our methods, we were able to show for the first time that rTSbis expressed both in the cytosol and in mitochondria. Further studies are planned to investigate the submitochondrial location of rTSb and its functional significance.

Selected Publications

McGuire JJ. Anticancer Antifolates: Current Status and Future Directions. Curr. Pharm. Design. 9: 2593-2613, 2003.

McGuire JJ and Coward JK. Folylpolyglutamate synthetase as a target for therapeutic intervention. Drugs of the Future 28: 967-974, 2003.

Gangjee A, Zeng Y, McGuire JJ, Mehraein F, and Kisliuk, RL. Synthesis of classical, three carbon bridged 5-substituted furo[2,3-d]pyrimidine and 6-substituted pyrrolo[2,3-d]pyrimidine analogues as antifolates. J. Med. Chem. 47: 6893-6901, 2004.

Gangjee A, Jain HD, McGuire JJ, and Kisliuk, RL. Benzoyl ring halogenated classical 2-amino-6-methyl–3,4-dihydro-4-oxo-5-substituted thiobenzoyl-7H-pyrrolo[2,3-d]-pyrimidine antifolates as inhibitors of thymidylate synthase and as antitumor agents. J. Med. Chem. 47: 6730-6739, 2004.

Jayakumar Nair R and McGuire JJ. Submitochondrial localization of human mitochondrial folylpolyglutamate synthetase. Manuscript in preparation.