Discovery of Tumor-Targeted 6-Methyl Substituted Pemetrexed and Related Antifolates with Selective Loss of RFC Transport

Pemetrexed and related 5-substituted pyrrolo[2,3-d]pyrimidine antifolates are substrates for the ubiquitously expressed reduced folate carrier (RFC), and the proton-coupled folate transporter (PCFT) and folate receptors (FRs) which are more tumor-selective. A long-standing goal has been to discover tumor-targeted therapeutics that draw from one-carbon metabolic vulnerabilities of cancer cells and are selective for transport by FRs and PCFT over RFC. We discovered that a methyl group at the 6-position of the pyrrole ring in the bicyclic scaffold of 5-substituted 2-amino-4-oxo-pyrrolo[2,3-d]pyrimidine antifolates 1–4 (including pemetrexed) abolished transport by RFC with modest impacts on FRs or PCFT. From molecular modeling, loss of RFC transport involves steric repulsion in the scaffold binding site due to the 6-methyl moiety. 6-Methyl substitution preserved antiproliferative activities toward human tumor cells (KB, IGROV3) with selectivity over IOSE 7576 normal ovary cells and inhibition of de novo purine biosynthesis. Thus, adding a 6-methyl moiety to 5-substituted pyrrolo[2,3-d]pyrimidine antifolates affords tumor transport selectivity while preserving antitumor efficacy.

O ne-carbon (C1) metabolism offers an important therapeutic target for many cancers, reflecting its role in the biosynthesis of purines, thymidylate, serine, and methionine, and in supporting biological methylation reactions from S-adenosylmethione. 1 Chemotherapy with antifolates has been a cornerstone of cancer therapy for over 60 years. 2 Clinically used antifolates include methotrexate (MTX), pemetrexed (PMX), pralatrexate (PTX), and raltitrexed (RTX) (Figure 1).While antifolates serve important roles in the therapeutic armamentarium for cancer, they manifest clinical challenges, most notably dose-limiting toxicities and emergence of drug resistance. 2 Membrane transport of antifolates is integral to their therapeutic efficacy in treating a variety of malignancies and nonmalignant conditions. 2,3−9 Both RFC and PCFT are facilitative transporters; RFC is a folateorganic anion antiporter, 5 and PCFT is a folate-proton symporter. 7FRs are glycosyl phosphatidylinositol-modified proteins that mediate cellular uptake of (anti)folates by receptor-mediated endocytosis. 4,10C is ubiquitously expressed and is the major tissue transporter of folates. 5Dose-limiting toxicities of classic antifolates are partially due to their membrane transport by RFC in both normal tissues and tumors. 5In addition, loss of transport in tumors due to decreased expression or mutation of RFC results in insufficient levels of intracellular antifolates to inhibit metabolic targets and sustain polyglutamate synthesis. 5,11CFT is the major mechanism for intestinal folate uptake. 12lthough PCFT is expressed in a select number of other tissues including the choroid plexus, kidney, liver, and spleen, 13,14 PCFT is not a major tissue transporter of folates.Further, transport by PCFT is optimal at acidic pHs not commonly associated with most tissues, 13,15 and the modest level of PCFT transport that occurs at neutral pH is suppressed by bicarbonate. 16Notably, PCFT is expressed in solid tumors including malignant pleural mesothelioma, 17 pancreatic adenocarcinoma, 18 nonsmall cell lung cancer, 19 and epithelial ovarian cancer 20 and is highly active in the acidic tumor microenvironment. 6,13Rα is expressed in normal epithelial tissues such as kidney, lung, choroid plexus, and placenta and in several tumors including ovarian cancer, nonsmall cell lung cancer, triple negative breast cancer, and kidney, endometrial, and colorectal cancers. 4,9,21However, only FRα in tumors exhibits the basolateral exposure required for access to the circulation. 4Rβ is expressed in placenta, mature neutrophils, and activated monocytes and macrophages, along with acute myeloid leukemia (AML) blasts. 4,9MX (1) is a 5-substituted pyrrolo [2,3-d]pyrimidine benzoyl antifolate with a 2-carbon bridge (Figure 1).PMX primarily inhibits thymidylate synthase (TS) with secondary inhibition at glycinamide ribonucleotide formyltransferase (GARFTase) and 5-aminoimidazole-4-carboxamide (AICA) ribonucleotide formyl transferase (AICARFTase) in de novo purine biosynthesis. 22The 5-substituted pyrrolo [2,3-d]pyrimidine benozyl compound 2 (3-carbon bridge) and the analogous pyrrolo [2,3-d]pyrimidine thienoyl compound 3 (2carbon bridge) (Figure 2) were reported to inhibit GARFTase and AICARFTase. 23,24While these compounds would presumably circumvent resistance due to alterations in TS, the 5-substituted pyrrolo[2,3-d]pyrimidine compounds includ-ing PMX all exhibit promiscuous transport in that they are substrates for RFC as well as for PCFT and FRs. 23,24−31 We envisage that cytotoxic C1 inhibitors that are selectively transported by PCFT or FRs with limited transport by RFC would be far less toxic to normal tissues and would also circumvent resistance due to loss of RFC.
We reasoned that if a simple structural modification could be introduced into targeted antifolates to abolish RFC transport, while preserving uptake by FRs and/or PCFT along with intracellular target engagement, a more effective agent would be realized.One such structural modification involves adding a 6-methyl group to the 5-subsituted pyrrolo [2,3-d]pyrimidine scaffold.We hypothesized that a strategically placed methyl moiety could sterically hinder drug attachment and net transport by RFC.As long as the net impact was greater toward RFC than on either PCFT or FRs, this would result in a more tumor-selective analogue than the corresponding nonmethylated compound.In addition to a potential impact on substrate binding to RFC, 6-methyl substitution also effects conformational restriction on the flexibility of the 5-substituted pyrrolo[2,3-d]pyrimidine compounds; this could further   the fluorinated analogue of PMX (1) and compound 5, the 6methyl analogue of PMX (1), were published previously; 32,33 however, there is no report in the literature regarding their transport properties.
In Figure 3, compounds 1 (green) and 5 (purple) and the residues within 5 to 6 Å of the C6 position are shown as spacefilled models.The folate binding sites of PCFT (Figure 3a,b), FRα (Figure 3c,d), and FRβ (Figure 3e,f) can all accommodate a 6-methyl group, without any steric hindrance.However, in the binding site for RFC (Figure 3g,h), the 6-methyl group of 5 (panel h) causes a serious steric clash with Tyr126, while breaking the edge-to-face π interaction of the pyrrole ring with Tyr126.Thus, the 6-methyl moiety of compound 5 seems likely to interfere with cellular uptake by RFC, although the 6methyl group does not appear to compromise the binding and transport of 5 by PCFT or FRs α or β.Similar simulations with compounds 6−8 provided identical conclusions (see Figure S1 in the Supporting Information).
For the synthesis, nucleophilic aromatic substitution of the commercially available 2-amino-6-chloropyrimidin-4(3H)-one 9 (Scheme 1) with hydrazine hydrate afforded the intermediate hydrazine 10. 32 Palladium-catalyzed Heck coupling of commercially available starting materials 11a and 11c (Scheme 2) and alkenol 12 gave the coupled, unsaturated secondary alcohols that rearranged to the vinyl alcohols and tautomerized to afford the ketones 17a and 17c.A similar Heck coupling of 11a with alkenone 13 provided an unsaturated ketone 14 which underwent catalytic hydrogenation to give 17b.The ketone intermediate 17d was synthesized using a modified Heck coupling reaction between 11d and alkenol 12, employing an electron-rich phosphine ligand.
Table 1 also shows the inhibition of proliferation of KB human nasopharyngeal carcinoma cells and IGROV1 epithelial ovarian cancer cells, both of which express RFC, PCFT, and FRα. 27,42Results are compared to those with IOSE 7576 normal ovary cells 43 which express RFC comparable to IGROV1 cells accompanying low levels of PCFT (∼30% of IGROV1) and undetectable FRα.For the tumor cells, except the des-methyl compound 4 with IGROV1 cells, all compounds were inhibitory.The 6-methyl substituted compounds (5, 6, and 8) inhibited KB cell proliferation with nanomolar IC 50 values only slightly less than those for the nonmethylated 5-substituted (1, 2, and 4) pyrrolo [2,3d]pyrimidine compounds.For both KB and IGROV1 cells, compound 7 was the most potent inhibitor with 242-fold and 127-fold, respectively, greater potencies than the corresponding 6-desmethyl analogue 3.
In conclusion, we established that 6-methyl substitutions on the pyrrole ring of 5-substituted pyrrolo [2,3-d]pyrimidine analogues preserved or increased inhibitor potencies and tumor selectivity, while positively impacting drug uptake by FRs and PCFT over RFC.Improved target selectivity, potency, and modified mechanisms-of-action (including multienzyme inhibition) provide compelling evidence of the profound impact of minor structural alterations in antitumor drug polypharmacology for tumor-targeted pyrrolo[2,3-d]pyrimidine antifolates.Clearly, the 6-methyl group provides an elegantly simple structural solution that would afford significant advantages over current clinically used antifolates.
pyrimidines 19a−d by thermolysis in diphenyl ether.1 H NMR showed no evidence of the presence of the other possible 6-substituted regioisomer (a key determinant is the presence of a C6-CH 3 chemical shift).The aromatic esters of 19a−d were subjected to base-catalyzed hydrolysis to afford pteroic acids 20a−d.Compounds 20a−d were subsequently coupled with L-glutamate diesters to afford 21a−d.Final saponification of diesters 21a−d with 1 N NaOH, followed by neutralization and acidification to pH 4 in the cold, provided target compounds 5−8.For desmethyl compound 4, a different synthetic strategy was used.A Sonogashira coupling reaction between bromofluorobenzoate 11d and unsaturated alkynol 22 afforded 23 (Scheme 4), followed by hydrogenation and oxidation to obtain aldehyde 25.Bromination afforded 26, which was immediately condensed with 2,6-diamino-4-oxo-pyrimidine at 45 °C to afford the 5-substituted pyrrolo[2,3-d]pyrimidine 27.

a
Scheme 1 a

a
Scheme 2 a

a
Scheme 4 a