Home | Molecular Cancer Therapeutics - Novel Pyrrolo[3,2-d ......Small Molecule Therapeutics Novel...

Small Molecule Therapeutics

Novel Pyrrolo[3,2-d]pyrimidine CompoundsTarget Mitochondrial and Cytosolic One-carbonMetabolism with Broad-spectrum AntitumorEfficacyAamod S. Dekhne1, Khushbu Shah2, Gregory S. Ducker3, Jade M. Katinas4,JenniferWong-Roushar4, Md. JunayedNayeen2, Arpit Doshi2,ChangwenNing5, XunBao1,Josephine Fr€uhauf1, Jenney Liu6, Adrianne Wallace-Povirk1, Carrie O'Connor1,Sijana H. Dzinic1, Kathryn White1, Juiwanna Kushner1, Seongho Kim1, Maik H€uttemann6,Lisa Polin1, Joshua D. Rabinowitz3, Jing Li1, Zhanjun Hou1, Charles E. Dann III4,Aleem Gangjee2, and Larry H. Matherly1

Abstract

Folate-dependent one-carbon (C1) metabolism is com-partmentalized into the mitochondria and cytosol andsupports cell growth through nucleotide and amino acidbiosynthesis. Mitochondrial C1 metabolism, including ser-ine hydroxymethyltransferase (SHMT) 2, provides glycine,NAD(P)H, ATP, and C1 units for cytosolic biosyntheticreactions, and is implicated in the oncogenic phenotypeacross a wide range of cancers. Whereas multitargeted inhi-bitors of cytosolic C1 metabolism, such as pemetrexed, areused clinically, there are currently no anticancer drugs thatspecifically target mitochondrial C1 metabolism. We usedmolecular modeling to design novel small-moleculepyrrolo[3,2-d]pyrimidine inhibitors targeting mitochondri-al C1 metabolism at SHMT2. In vitro antitumor efficacy wasestablished with the lead compounds (AGF291, AGF320,AGF347) toward lung, colon, and pancreatic cancer cells.

Intracellular targets were identified by metabolic rescue withglycine and nucleosides, and by targeted metabolomicsusing a stable isotope tracer, with confirmation by in vitroassays with purified enzymes. In addition to targetingSHMT2, inhibition of the cytosolic purine biosyntheticenzymes, b-glycinamide ribonucleotide formyltransferaseand/or 5-aminoimidazole-4-carboxamide ribonucleotideformyltransferase, and SHMT1 was also established.AGF347 generated significant in vivo antitumor efficacy withpotential for complete responses against both early-stageand upstage MIA PaCa-2 pancreatic tumor xenografts,providing compelling proof-of-concept for therapeutic tar-geting of SHMT2 and cytosolic C1 enzymes by this series.Our results establish structure–activity relationships andidentify exciting new drug prototypes for further develop-ment as multitargeted antitumor agents.

IntroductionMetabolic reprogramming to support tumor progression

has emerged as a hallmark of cancer (1). Of the many alteredmetabolic pathways associated with the malignant phenotype,one-carbon (C1) metabolism is particularly notable (2–4). C1metabolism depends on an adequate supply of tetrahydrofolate(THF)metabolites and generates critical purine, thymidylate, andglycine metabolites essential for cell proliferation and tumorprogression (3–5). Purine nucleotides are synthesized de novo inthe cytosol, whereas thymidylate is synthesized in both thecytosol and nucleus (3–6). C1 enzymes such as thymidylatesynthase (TS) and the purine biosynthetic enzymesb-glycinamideribonucleotide (GAR) formyltransferase (GARFTase) and5-aminoimidazole-4-carboxamide (AICA) ribonucleotide (AICAR)formyltransferase (AiCARFTase) (the third and ninth steps inde novo purine biosynthesis, respectively) are important therapeu-tic targets for cancer (7–9). Serine biosynthesis from glycine in thecytosol involves serine hydroxymethyltransferase (SHMT) 1 anduses C1 units from 5,10-methylene-THF (5,10-me-THF) (2–4).SHMT1, like TS, is localized to both the cytosol andnucleus (5, 6);however, nuclear SHMT1 does not appear to be an importantsource of C1 units for TS (6, 10).

1Department of Oncology, Wayne State University School of Medicine, and theBarbaraAnnKarmanos Cancer Institute, Detroit, Michigan. 2Division ofMedicinalChemistry, Graduate School of Pharmaceutical Sciences, Duquesne University,Pittsburgh, Pennsylvania. 3Department of Chemistry/Lewis-Sigler Institute forIntegrative Genomics, Princeton University, Princeton, New Jersey. 4Depart-ment of Chemistry, Indiana University, Bloomington, Indiana. 5Biochemistry andMolecular Biology, Jilin University, Changchun, Jilin Province, China. 6Center forMolecular Medicine and Genetics, Wayne State University School of Medicine,Detroit, Michigan.

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

Corresponding Authors: Larry H. Matherly, Molecular Therapeutics Program,Barbara Ann Karmanos Cancer Institute, 4100 John R Street, Detroit, MI 48201.Phone: 313-578-4280; Fax: 313-578-4287; E-mail: [email protected];Aleem Gangjee, Duquesne University, 600 Forbes Avenue, Pittsburgh, PA15282. Phone: 412-396-6070; [email protected]; and Charles E. Dann III, IndianaUniversity, 800 E. Kirkwood Ave, Bloomington IN 47405. Phone: 812-856-1704;[email protected]

Mol Cancer Ther 2019;18:1787–99

doi: 10.1158/1535-7163.MCT-19-0037

�2019 American Association for Cancer Research.

MolecularCancerTherapeutics

www.aacrjournals.org 1787

on July 4, 2021. © 2019 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst July 9, 2019; DOI: 10.1158/1535-7163.MCT-19-0037

http://crossmark.crossref.org/dialog/?doi=10.1158/1535-7163.MCT-19-0037&domain=pdf&date_stamp=2019-9-19http://crossmark.crossref.org/dialog/?doi=10.1158/1535-7163.MCT-19-0037&domain=pdf&date_stamp=2019-9-19http://mct.aacrjournals.org/

Cytosolic and mitochondrial C1metabolic pathways are inter-connected by exchange of serine, glycine, and formate (Fig. 1).Extracellular folates are transported into cells by the reducedfolate carrier (RFC), proton-coupled folate transporter (PCFT),and folate receptors (FR; refs. 11, 12). Whereas cytosolic folatesare transported into mitochondria via the mitochondrial folatetransporter (SLC25A32; refs. 13, 14), mitochondrial folates donot exchange with those in the cytosol (15). In cancer cells, the3-carbon of serine is the major source of C1 units, and inmitochondria, serine catabolic enzymes including SHMT2,5,10-me-THF dehydrogenase 2 (MTHFD2), and 10-formyl-THFsynthetase (reverse; MTHFD1L) generate glycine and C1 units(i.e., formate) to sustain C1-dependent nucleotide and aminoacid biosynthesis in the cytosol (Fig. 1). 10-Formyl-THF is syn-thesized from formate in the cytosol by the trifunctional enzymeMTHFD1. 10-Formyl-THF is utilized for purine nucleotidebiosynthesis and can be further converted by MTHFD1 to 5,10-me-THF for TS and SHMT1.

Several studies have implicated mitochondrial C1 metabolismas critical to the malignant phenotype (16–19). A study (18) ofmessenger RNA profiles for 1,451 metabolic enzymes spanning1,981 tumors across nineteen different cancer types and 951matched normal tissues identified SHMT2 and MTHFD2 amongthe top five most differentially expressed genes, highlighting anoncogenic role of mitochondrial C1 metabolism. Metabolomicsanalyses of 219 extracellular metabolites from the NCI-60 cancercell lines showed that glycine consumption and the glycinebiosynthetic pathway correlated with cell proliferation (16).These findings, combined with evidence of functional shortagesof amino acids (e.g., glycine) in tumors (20), suggested a ther-apeutic opportunity for SHMT2 targeting in cancer.

SHMT2 is induced by hypoxic stress in Myc-transformedcells (21) and is critical to tumor cell survival in the hypoxic,nutrient-poor tumor microenvironment (17, 21). SHMT2 (or

MTHFD2) knockout (KO) cells are viable and tumorigenic (albeitwith decreased growth rates) in nutrient-rich conditions,as reversal of cytosolic SHMT1 (serine-to-glycine) providessufficient C1 units to sustain some level of de novo nucleotidebiosynthesis (22). However, SHMT1 only restores a smallfraction of the C1 pools in wild-type (WT) cells (22). Further-more, SHMT1 does not generate sufficient glycine for protein,nucleotide, and glutathione biosynthesis, rendering bothSHMT2 and MTHFD2 KO cells as glycine auxotrophs (4). Thiswas the impetus for studies of pyrazolopyran compounds (e.g.,SHIN1) with dual SHMT1/SHMT2 inhibition (23). Whilestructurally unrelated to folates, these compounds bound tothe folate-binding site in SHMT2 and showed in vitro antitumorefficacy (particularly with B-cell cancers).

In this report, we describe a novel series of 5-substitutedpyrrolo[3,2-d]pyrimidine compounds that inhibit SHMT2, andalso SHMT1, GARFTase, and/or AICARFTase. Our lead com-pounds (AGF291, AGF320, and AGF347) show broad-spectrum in vitro antitumor efficacies against H460 non–smallcell lung cancer (NSCLC), HCT116 colon cancer, andMIA PaCa-2pancreatic cancer cells. For AGF347, in vitro findings were extend-ed in vivo to MIA PaCa-2 tumor xenografts, providing compellingproof-of-concept of the therapeutic potential of ourmultitargetedSHMT2 therapeutics for cancer.

Materials and MethodsChemicals

[2,3,3-2H]L-Serine (98%) was purchased from CambridgeIsotope Laboratories, Inc. Leucovorin [(6R,S)5-formyl-THF]was provided by the National Cancer Institute (Bethesda, MD).Pemetrexed (Alimta; PMX) was purchased from LC Laboratories.Gemcitabine (Gemzar) was purchased from Fresenius Kabi USA,LLC. Serine-, glycine- and folate-free RPMI1640 was custom-

Figure 1.

C1 metabolism is compartmentalized inthe cytosol and mitochondria. Folatesenter the cell through the plasmamembrane–facilitated folate transporters,PCFT and RFC, and enter themitochondria via SLC25A32. Serinecatabolism in the mitochondria beginningwith SHMT2 generates glycine andformate, the latter of which is required fordownstream cytosolic de novo purinebiosynthesis (by GARFTase andAICARFTase) as 10-formyl-THF, andthymidylate biosynthesis, followingconversion to 5,10-me-THF by MTHFD1.SHMT1 catalyzes conversion of glycine toserine in the cytosol. AICA is metabolizedto AICAR (ZMP), the AICARFTasesubstrate which circumvents theGARFTase step. The arrows denote thenet flux of C1 metabolism in proliferatingcells, althoughmost reactions arereversible. Both SHMT1 and TS are alsolocalized to the nucleus; however, nuclearSHMT1 is not an important source of C1units for TS (6, 10).

Dekhne et al.

Mol Cancer Ther; 18(10) October 2019 Molecular Cancer Therapeutics1788



http://mct.aacrjournals.org/

ordered from Thermo Fisher Scientific and supplemented withtissue-culture grade glycine (Thermo Fisher Scientific) or serine(Sigma-Aldrich), as needed. Other chemicals were obtained fromcommercial sources in the highest available purities. Chemicalsynthesis of the pyrrolopyrimidine compounds is described inthe Supplementary Methods.

Molecular modeling and computational studiesMolecular modeling was performed for all analogues with the

human SHMT2 crystal structure (PDB: 5V7I; ref. 23) using theinduced fit docking protocol of Maestro (refs. 2, 3; see Supple-mentary Methods). The compounds were also docked into rabbitSHMT1 (PDB: 1LS3; ref. 24) binding sites. The docking scores ofthe analogues are reported in Supplementary Table S1.

Cell culture and proliferation/protection assaysThe HCT116 cell lines including the SHMT1, SHMT2, and

MTHFD2 KO cells were described previously (22, 23). Thewild-type HCT116 and H460 human tumor cell lines wereobtained from the ATCC, whereas MIA PaCa-2 cells were provid-ed byDr. Yubin Ge (Karmanos Cancer Institute, Detroit, MI). Celllines were frozen in aliquots in liquid nitrogen upon receiptwith aliquots from the original freezes thawed and used forexperiments. As warranted, cell lines were reverified by STRanalysis by Genetica DNA Laboratories (H460, November2015; MIA PaCa-2, April 2019).

MTXRIIOuaR2-4 (i.e., R2) Chinese hamster ovary (CHO) cellswere provided by Dr. Wayne Flintoff (University of WesternOntario, Ontario, Canada; ref. 25). From this parental R2 cellline, human RFC and PCFT were individually transfected togenerate the isogenic CHO cell lines designated PC43-10 (RFC)and R2/PCFT4 (PCFT; refs. 26–28). Other than to confirm phe-notypic expression of individual RFC or PCFT expression levelsand/or inhibitor sensitivities, CHO cells were not furtherauthenticated.

Cell lines were tested for Mycoplasma by PCR using a Myclo-plasma Testing Kit (Venor GeM Mycoplasma Detection Kit,Sigma; February 2018). As warranted, cells were treated (Plasmo-cin, Invivogen) and retested for Mycoplasma. To avoid cell line"drift," fresh cultures were reconstituted no less than everymonthfrom the original freezer stocks (above).

Human tumor cell lines were cultured in folate-free RPMI1640supplemented with 10% dialyzed FBS (Sigma-Aldrich), 1% pen-icillin/streptomycin solution, 2 mmol/L L-glutamine, and25 nmol/L leucovorin in a humidified atmosphere at 37�C inthe presence of 5% CO2 and 95% air. The CHO cell lines werecultured in a-MEM supplemented with 10% bovine calf serum,1% penicillin/streptomycin solution, and 2mmol/L L-glutamine.The transfectedCHOcell lines (i.e., R2/PCFT4 andPC43-10)werecultured in the presence of 1 mg/mL of G418.

Proliferation assays in CHO cells were performed as describedpreviously (29), except glycine-free folate-free RPMI1640 wasused. For the HCT116, H460, and MIA PaCa-2 tumor cell lines,an identical procedure was followed except that themaximal drugconcentration was increased to 10 mmol/L from 1 mmol/L. For theHCT116 KO cells, glycine (130 mmol/L) was included.

Glycine/nucleoside rescue experiments in CHO and tumor celllines were performed in folate- and glycine-free RPMI1640/10%dialyzed FBS, supplemented with 25 nmol/L leucovorin withoutadditions, or in the presence of adenosine (60 mmol/L),thymidine (10 mmol/L), glycine (130 mmol/L), and/or AICA

(320mmol/L). Growthofmetabolite-treated cells was normalizedto controls treated with metabolites and vehicle (i.e., DMSO),both singly and in combination.

Generation of H460 SHMT2 knockdown cell lineH460 SHMT2 knockdown (KD) cells were generated by trans-

duction with MISSION lentivirus particles (Sigma-Aldrich) con-taining shRNA targeting SHMT2 (TRCN0000034805) or non-targeted control (NTC) lentivirus as described previously (29).Single clones were isolated and cultures expanded. The extent ofSHMT2 KD was confirmed by Western blotting (below).

Real-time PCRCells were harvested from 60-mm dishes or T25 flasks at

approximately 80% confluence and RNAs extracted using TRI-zol reagent (Invitrogen). cDNAs were synthesized with randomhexamers, MuLV reverse transcriptase, and RNase inhibitor(Applied Biosystems), and purified with a QIAquick PCRPurification Kit (Qiagen). Quantitative real-time RT-PCR wasperformed using a Roche LightCycler 480 (Roche Diagnostics)with gene-specific primers to the major folate transportersand Universal Probe Library probes (SHMT2-#83, RFC-#32,PCFT-#89, FRa-#65; Roche Diagnostics). Transcript levels werenormalized to b-actin and GAPDH. Primer sequences areavailable upon request.

Gel electrophoresis and Western blotsH460 WT, H460 NTC, H460 SHMT2 KD, HCT116 WT, and

HCT116 SHMT2 KO cell lines were plated (1� 106 cells/dish) in60-mm dishes and harvested when the cells were approximately80% confluent. Cells were disrupted by sonication and cell debrisremoved by centrifugation (1,800 rpm, 5 minutes, 4�C). Thesoluble cell fraction was assayed for protein (30). Equal amountsof protein (37 mg) from each sample were electrophoresed on10% polyacrylamide gels with SDS (31) and transferred to poly-vinylidene difluoride membranes (Thermo Fisher Scientific;ref. 32). To detect SHMT2, membranes were incubated for72hourswith rabbit anti-SHMT2primary antibody [#12762(22);Cell Signaling Technology]. The blots were developed by incu-bating in IRDye800CW-conjugated goat anti-rabbit IgG second-ary antibody (LICOR Biosciences) for 90 minutes and scanningwith an Odyssey infrared imaging system (LICOR Biosciences).Protein loading was normalized to b-actin using anti-b-actinmouse antibody (Sigma-Aldrich).

In vitro–targeted metabolomics by liquid chromatography-mass spectrometry

Cells were seeded in triplicate 60-mm dishes in folate-freeRPMI1640 (contains glycine and unlabeled serine) supplemen-ted with 10%dialyzed FBS, 1%penicillin/streptomycin, 2mmol/L L-glutamine, and 25 nmol/L leucovorin. The cells were allowedto adhere for 24 hours. The media were aspirated and replacedwith culture media (contains 25 nmol/L leucovorin, glycine, andunlabeled serine) and 10 mmol/L AGF291, AGF320 or AGF347,or vehicle (DMSO; with or without 1 mmol/L formate). After16 hours, the cells were washed with PBS (3�), and the mediawere replaced with folate- and serine-free media (containingglycine) supplemented with 10% dialyzed FBS, 25 nmol/L leu-covorin, and [2,3,3-2H]serine (250mmol/L), including 10mmol/Ldrug or vehicle. The cells were incubated for 24hours.Metaboliteswere extracted and analyzed (22, 23) with normalization to total

Targeting Mitochondrial and Cytosolic C1 Metabolism

www.aacrjournals.org Mol Cancer Ther; 18(10) October 2019 1789




protein from the post-extraction pellet (solubilized with 0.5 NNaOH) by the Folin–phenolmethod (30). Samples were run on aThermoElectron Corporation Exactive mass spectrometer oper-ating in negative ion mode (33). Separation used reverse-phaseion-pairing chromatography on a 100 mm/2.5 mm SynergiHydro-RP C18 column (Phenomenex). Individual metaboliteswere identified from their exact masses and comparison of reten-tion times with standard compounds using the MAVEN softwaresuite (34). Values below the limit of detection were assigned avalue of 100 for normalization.

Enzyme expression and purificationN-terminal His-tagged GARFTase (formyltransferase domain;

residues 100-302) was expressed and purified (35). Full-lengthhuman AICARFTase/IMP cyclohydrolase (ATIC), SHMT1 (resi-dues 7-478, Uniprot ID P34896), SHMT2 (residues 42-504,Uniprot ID 34897), and MTHFD2 (residues 36-333, Uniprot IDP13995) with N-terminal, cleavable hexahistidine tags wereexpressed and purified by essentially the same protocol. Sampleswere further purified on €AKTA FPLC (GE Healthcare) on aSuperdex 200 16/60 (GE Healthcare) column. Further detailsare in the Supplementary Methods.

In vitro enzymatic assays and Ki determinationsAICARFTase and GARFTase activities were measured by

monitoring formation of THF spectrophotometrically from10-formyl-THF in the presence of various concentrations ofinhibitor (36, 37). SHMT1/2 activities were assayed by a coupledreaction with His-MTHFD2 in 200-fold molar excess with NADHproduction monitored by fluorescence. Kis were calculated fromthe IC50 values [Ki ¼ IC50/([S]/KMþ1)], using the Km and sub-strate concentrations for 10-formyl-THF. The calculatedKmvaluesfor 10-formyl-THF with His-ATIC and His-GARFTase were100 mmol/L and 84.8 mmol/L, respectively. The calculatedKm values for THF with His-SHMT1 and His-SHMT2 were62.8 mmol/L and 108 mmol/L, respectively. To confirm thatMTHFD2 was not inhibited by our inhibitors, MTHFD2 activitywas evaluated with an NAD(P)H-Glo Detection System Kit(Promega, Ref G9061). Further details are in the SupplementaryMethods.

In vivo efficacy trial with MIA PaCa-2 pancreatic cancerxenografts

Methods for in vivo maintenance of MIA PaCa-2 tumor xeno-grafts and drug efficacy evaluations are analogous to thosedescribed previously (36, 38–41). The mouse study in this reportwas approved by theWayne State University Institutional AnimalCare and Use Committee (IACUC). Female NCr severe compro-mised immunodeficient (SCID) mice were purchased fromCharles River Labs. For the early-stage trial, chemotherapy wasinitiated one day posttumor implantation with AGF347; for theupstage trial, chemotherapy was initiated 7 days posttumorimplantation (when tumors had grown to 100–150 mg) withAGF347. For both designs, dosing for AGF347 was 15 mg/kg/injection every 2 days � 8 (total dose of 120 mg/kg); for gemci-tabine, dosing was 120 mg/kg/injection every 4 days � 4 (totaldose of 480mg/kg). Both drugs were administered intravenously(0.2 mL/injection). The tumor masses from both flanks of eachmouse were added together, and the total mass per mouse wasused for calculations of antitumor activity. Quantitative end-points include: (i) tumor growthdelay [T-C,where T is themedian

time in days required for the treatment group tumors to reach apredetermined size (e.g., 1000 mg), and C is the median time indays for the control group tumors to reach the same size; tumor-free survivors are excluded from these calculations]; and (ii) grosslog10 cell kill (LCK), determined by the formula LCK ¼ (T-C;tumor growth delay in days)/3.32� Td (tumor doubling time indays determined by growth plot). Qualitative analysis includeddetermination of T/C values (in percent) on all days of tumormeasurement using themedian total tumor burden for treatment(T) and control (C) groups. The endpoint %T/C value for thisstudy corresponds to the first measurement taken post last treat-ment (day 16 for early stage or day 21 for upstage), when controltumors were still in exponential growth phase (i.e., 500–1,250 mg). Further details are described in the SupplementaryMethods.

From the upstage in vivo study, a separate AGF347 treatmentarm (3 mice) was designated for metabolomics analysis in com-parison with the untreated control (3 mice). Six hours after thesixth dose of AGF347, mice were sacrificed, and tumors removedand immediately frozen in liquid nitrogen. Frozen isolatedtumorswereweighed; approximately 50mgpieceswere disruptedwith a cryomill (Retsch) in 1-mL ice-cold acetonitrile:methanol:water (40:40:20). Solids were collected by centrifugation, thenreextracted with 1 mL lysis solution (above). Combined super-natants were dried down and then resuspended in water to aconcentration of 50 mg/mL (original tissue mass) for LC-MSanalysis, as described above.

Statistical analysisAll data shown reflect at least three biological replicates unless

noted otherwise. Statistical analyses were performed by theKarmanos Cancer Institute Biostatistics Core. The expressionlevels were assessed for the normality assumption. The log2transformation was used when all values were positive and thesquare root transformation was used when values included zero.Statistical tests were carried out using an unpaired t test. P valueswere not adjusted for multiple comparisons.

ResultsRationale for design of 5-substituted pyrrolo[3,2-d]pyrimidinecompounds targeting mitochondrial C1metabolism at SHMT2

Given the association of mitochondrial C1 metabolism withmalignancy (16–19), including reports of SHMT2 as a potential"onco-driver" (17, 42), it was of interest to develop inhibitors ofSHMT2 as antitumor agents. We initially looked for potentialSHMT2 inhibitors among prototypes of different classes of GARF-Tase and AICARFTase inhibitors from our published stud-ies (26, 27, 36, 39, 41, 43, 44). We tested these compounds(Supplementary Fig. S1) for inhibition of proliferation of CHOcells engineered from the transporter-null R2 CHO subline (25)to express human PCFT (R2/PCFT4; ref. 27), as we reasoned thattumor-selective uptake of a putative SHMT2 inhibitor byPCFT (45) would be desirable. Because inhibition of mitochon-drial C1 metabolism (at SHMT2) should induce glycine auxot-rophy (22), and mitochondrial C1 metabolism generates glycineand C1 units (i.e., formate) critical for nucleotide biosynthesis,these experiments were performed in glycine- and nucleoside-freemedia, and the protective effects of added glycine (130 mmol/L)and/or adenosine (60 mmol/L) were determined. With the excep-tionofAGF147, these compounds inhibited cell proliferation and

Dekhne et al.





purine biosynthesis in R2/PCFT4 cells (reflected in adenosinerescue); however, glycine was not protective with or withoutadenosine addition, indicating a lack of mitochondrial C1 target-ing. Thus, it became imperative to design new analogues.

To engineer SHMT2 activity into our compound series,we merged structural features of our previous 5-substitutedpyrrolo[2,3-d]pyrimidine benzoyl and thienoyl compounds(43, 44) with those of 5,10-me-THF (SHMT2 product) and5-formyl-THF (SHMT2 inhibitor; ref. 24; Fig. 2). The resulting5-substituted pyrrolo[3,2-d]pyrimidine analogues included 3–5bridge carbons linked to benzoyl (i.e., AGF291, AGF300,and AGF299) or thienoyl (i.e., AGF331, AGF318, and AGF320)moieties. On the basis of the reported impact of 20 fluorinesubstitutions in increasing inhibitory potencies of pyrrolo[2,3-d]pyrimidine compounds (39), we designed 20-fluorinated ana-logues of AGF300 and AGF299, as well (AGF347 and AGF355,respectively; Fig. 2). Compounds were docked into humanSHMT2 (Supplementary Fig. S2) with docking scores (Supple-

mentary Table S1) better than for the pyrazolopyran inhibitorSHIN1 (23) (�5.58 kcal/mol) upon redocking.

Discovery of 5-substituted pyrrolo[3,2-d]pyrimidinecompounds that target mitochondrial C1 metabolism

We synthesized the novel analogues and screened these forinhibition of cell proliferation. To encompass themajormodes offacilitative transport, we initially assessed inhibition by thesecompounds (from 0 to 1,000 nmol/L) toward PCFT-expressingR2/PCFT4 CHO cells and an isogenic CHO subline engineered toexpress human RFC (PC43-10; ref. 26). Results were compared tothose for folate transporter-null R2CHOcells (25). IC50 values for"active" compounds are shown in Table 1 for the analogues, alongwith those for AGF94 (a "pure" GARFTase inhibitor; ref. 41) andPMX. Active compounds toward R2/PCFT4 and PC43-10 cellsincluded AGF291, AGF320, AGF331, and AGF347. These analo-gues were further tested with human tumor cell lines, includingH460 NSCLC, HCT116 colon cancer, and MIA PaCa-2 pancreatic

Figure 2.

Rational design of novel 5-substituted pyrrolo[3,2-d]pyrimidine benzoyl and thienoyl analogues. Structural features of the 5-substituted pyrrolo[2,3-d]pyrimidinesAGF127 and AGF136 (43, 44) were merged with the SHMT2 substrate 5,10-me-THF, and inhibitor 5-formyl-THF to generate the novel analogues inthis report. Key structural features for the various compounds are summarized including the bridge lengths (3 or 4 carbons), thienoyl or benzoyl side-chains, andfor the latter, the 20 ring substituent (F or H). NA, not applicable.






cancer cells, characterized by expression of PCFT and RFC but notFRa (refs. 46, 47; Supplementary Fig. S2). IC50 values for growthinhibition are in Table 1. Although there were differences ininhibitor sensitivities among the assorted tumor models,AGF291, AGF320, and AGF347were consistently the most activeof the series.

We again used glycine and nucleoside rescue studies (above) inH460, HCT116, and MIA PaCa-2 cells treated with AGF291,AGF320, or AGF347 to identify the targeted pathways. Resultswere compared with those for AGF94 and are shown in Fig. 3 forH460 cells. Adenosine (60 mmol/L) was completely protective upto 10 mmol/L AGF94, whereas glycine (130 mmol/L) had noeffect. We tested the protective effects of AICA (320 mmol/L),which is metabolized to AICAR (ZMP; AICARFTase substrate),thus circumventing GARFTase in de novo purine biosynthesis(ref. 41; Fig. 1). As AICA was completely protective (Supplemen-tary Fig. S4), GARFTase must be the intracellular target forAGF94 (41). For AGF291, AGF347, and AGF320, combinedadenosine and glycine was substantially protective (Fig. 3).These results strongly suggest that these compounds target bothmitochondrial C1 metabolism and cytosolic de novo purine bio-synthesis. Thymidine provided no protection from any of thecompounds and did not increase the extent of rescue by glycineand adenosine. For some analogues, notably AGF320, growthinhibition was modestly (and incompletely) reversed by AICA(with glycine), suggesting a secondary target (below). Analogousresults were obtained with HCT116 and MIA PaCa-2 tumorcells (Supplementary Fig. S4). These results establish thatAGF291, AGF320, and AGF347 likely target both mitochondrialand cytosolic C1 metabolism.

Identification of the mitochondrial enzyme target for AGF291,AGF320, and AGF347 by targeted metabolomics

To confirm the intracellular enzyme targets of the lead com-pounds AGF291, AGF320, and AGF347, we performed targetedmetabolomics with LC/MS using a [2,3,3-2H]serine tracer inHCT116, H460 and MIA PaCa-2 tumor cells. The cells wereprocessed for LC/MS analysis of total serine, and of isotopicallylabeled Mþ3, Mþ2, Mþ1, and Mþ0 serine (Mþn representsspecies with n deuterium atoms). Results for inhibitor-treatedcells were compared with those for untreated cells. For WTHCT116 cells, the results were compared with those for SHMT1KO, MTHFD2 KO, and SHMT2 KO cells (22). For H460 cells,controls included NTC and SHMT2 shRNA KD cells.

KO of cytosolic SHMT1 (SHMT1 KO) in HCT116 cells had noimpact on total serine pools. Serine pools were elevated approx-imately 10-fold in the HCT116 sublines with mitochondrial C1KOs (SHMT2 KO and MTHFD2 KO) relative to WT controls(Fig. 4B). This further supports the notion that SHMT2 ratherthan SHMT1 is the primary catabolic enzyme for serine inHCT116 cells (22). In AGF291-, AGF320- and AGF347-treatedHCT116 cells, serine also increased approximately 10-fold,indicating a profound loss of serine catabolism. Analogousresults were seen with the H460 and MIA PaCa-2 cell lines,including the H460 SHMT2 KD cells (Supplementary Fig. S5Aand S5B).

The flux of [2H]metabolites originating from [2,3,3-2H]serinetracer is depicted in Fig. 4A. In proliferating tumor cells, C1metabolism flows in a clockwise manner, with serine catabolizedin mitochondria (starting with SHMT2) and regenerated in thecytosol (via SHMT1; ref. 22). The reactions are reversible (e.g.,Ta

ble1.

IC50values

fora

ntiproliferativeactivities

towarden

ginee

redCHOan

dhu

man

tumorcelllines,and

cell-free

enzymeinhibitionby5-substituted

pyrrolo[3,2-d]pyrim

idineinhibitorsco

mpared

withAGF94an

dPMX

Proliferationinhibition[IC50(nmol/L)]

Invitroen

zymeassays

[Ki(m

mol/L)]

Compoun

dR2(null)

PC43-10

(RFC)

R2/PCFT4

(PCFT)

HCT116

HCT116

SHMT2

KO

H460

MIA

PaC

a-2

GARFTa

seAICARFTa

seSH

MT2

MTH

FD2

SHMT1

AGF94

136(3)

110(15)

3(1)

191(62)

295(59)a

166(59)

1,924

(539

)0.88(0.71)

NI

188.0

(91.8

)NI

112(57)

PMX

395(83)

124(39)

24(5)

293(81)

333(40)

165(59)

281(23)

5.19

(1.63)

0.88(0.56)

NI

NI

NI

AGF29

1NI

454

(87)

282(22)

2,26

6(450

)108(66)b

461(163)

3,664(721)

NI

13.31(3.16

)0.63(0.29)

NI

0.90(0.06)

AGF34

7NI

224(17)

479

(63)

437

(180)

42(19)c

214(88)

1,381(182)

3.13

(0.66)

3.72

(1.61)

2.19

(0.23)

NI

2.91(0.59)

AGF33

1NI

NI

661(62)

2,199(817)

969(254

)23

67(711)

6,681(604)

ND

ND

ND

ND

ND

AGF32

0NI

NI

694(56)

737(195)

117(60)a

573(144)

2,70

3(400)

0.33(0.22)

2.36

(1.99)

0.056

(0.020

)NI

0.12

6(0.030

)

NOTE:P

roliferationinhibitionassays

wereperform

edusingtheen

ginee

redCHOcelllines

R2(folate

tran

sporter-null),P

C43-10

(exp

resses

RFConly),and

R2/PCFT4(exp

resses

PCFTonly),and

human

tumorcelllines,

includ

ingHCT116(coloncancer),H460(lun

gcancer),an

dMIA

PaC

a-2(pan

crea

ticcancer).ForH

CT116,both

wild

-typ

e(W

T)a

ndCRISPR/C

as9SHMT2KOcells

(22,23

)weretested

.ForS

HMT2KOcells,glycine

was

added

tothemed

iato

circum

vent

theim

pactofloss

ofSHMT2,permitting

cellgrowth.R

esults

areshownas

mea

nIC

50va

lues

(�stan

darddev

iations),co

rrespond

ingto

theco

ncen

trations

that

inhibitgrowth

by50

%relative

tove

hicle-trea

tedco

ntrolcells,from

atleastfour

biologicalreplicates.a,P

<0.05;

b,P

<0.01;

c ,P<0.001in

HCT116SHMT2KOrelative

toWTcells.F

ortheinvitroen

zymeassays,K

iva

lues

arepresented

asmea

nva

lues

(�stan

darddev

iations)from

atleastthreereplicateexperim

ents."ND"den

otesassayno

tdone

."NI"den

otes"noinhibition"

at1,0

00nm

ol/LinhibitorfortheCHOcelllineproliferationexperim

entsan

dup

to20

0mm

ol/L

inhibitorforthein

vitroen

zymeassays.

Dekhne et al.





serine is resynthesized from formate in themitochondria; ref. 22).The compartmental metabolism of serine can be revealed byanalyzing the isotope tracing ("scrambling") patterns from[2,3,3-2H]serine in HCT116 cells (22, 23). These patterns includeboth unmetabolized (Mþ3) and metabolized (Mþ2 and Mþ1)forms of serine. The latter species are derived from the recombi-nation of unlabeled glycine with doubly labeled (Mþ2) 5,10-me-THF and the synthesis of serine from an unlabeled glycine and a

singly labeled 5,10-me-THF derived from a formate/10-formyl-THF precursor (Mþ1), respectively (Fig. 4A).

InWTHCT116 cells fed [2,3,3-2H]serine,most of the serinewascatabolized (only 10% Mþ3 serine remained; Fig. 4C). A largefraction of serine was incompletely labeled (scrambled; i.e., Mþ1or Mþ2), reflecting resynthesized serine from Mþ1 or Mþ2precursors. SHMT1 KO led to a drop (�60%) in theMþ1 fractionwithout a change in the Mþ2 serine fraction, consistent with

Figure 3.

In vitro antitumor efficacy and identification of targeted pathways and enzymes by novel pyrrolo[3,2-d]pyrimidine analogues in H460 tumor cells byproliferation assays. Dose–response growth inhibition curves are shown for AGF94, an established GARFTase inhibitor (41), and for AGF291, AGF320, andAGF347without additions, or in the presence of adenosine (60 mmol/L) and/or glycine (130 mmol/L). The results are mean values�standard deviations for threebiological replicates.

Figure 4.

Targeted metabolomics analysis to identify intracellular enzyme targets of AGF291, AGF320, and AGF347.A, A schematic of serine isotope scrambling and dTTPisotope analysis is shown. Heavy (2H) atoms in serine (red circles) enter the C1 flux in the cytosol through reversal of the SHMT1 reaction and TS leading to dTMPand dTTP (as Mþ2). In the mitochondria, 2H atoms in serine (blue circles) are metabolized via SHMT2, MTHFD2, and MTHFD1 to formate which is converted to10-formyl-THF and 5,10-me-THF, leading to dTMP and dTTP (as Mþ1). Most steps are reversible as noted. Adapted from ref. 22. Total serine pools (B) and thecorresponding serine isotope distributions (C) for WT and KO HCT116 sublines, and for inhibitor-treatedWT cells, are shown. Total GAR (D) and AICAR (E) poolsare shown forWT and SHMT2 KD H460 sublines, including inhibitor-treatedWT cells, with and without 1 mmol/L formate. F and G, Relative adenine nucleotidepools (AMP, ADP, and ATP) and dTTP pools with isotope distributions are shown for untreated and inhibitor-treated MIA PaCa-2 cells. For D and E, results forinhibitor-treated and SHMT2 KD cells were normalized to vehicle-treatedWT� formate or NTC� formate samples, as appropriate. Data are mean values� SDsfor three technical replicates. #, P < 0.10; � , P < 0.05; �� , P < 0.01; �� , P < 0.001; ^ or n are used in place of � to specify a significant increase or decrease,respectively. Statistical comparisons were with vehicle-treatedWT� formate or NTC� formate samples, as appropriate (ns, not significant).






SHMT1 in the cytosol being responsible for bulk synthesis ofserine from Mþ1 5,10-me-THF, and unlabeled (media) gly-cine (22). In SHMT2 KO cells, the Mþ3 fraction was >60% ofthe total serine (Fig. 4C), consistent with a profound loss of serinecatabolism. Reflecting this and the depletion of formate down-streamof SHMT2, there was approximately 90%decrease inMþ1serine (22). Mþ2 serine also decreased (�25%) compared withWTcells. Although similar resultswere seen for theMþ1andMþ3serine fractions with MTHFD2 KO cells (decreased �75% andincreased 3.5-fold, respectively, fromWT levels), Mþ2 serine wasincreased (�1.7-fold) rather than decreased (Fig. 4C). This reflectsaccumulation ofMþ25,10-me-THFwhenMTHFD2 is lost, whichdrives synthesis of Mþ2 serine from the largely unlabeled glycinepool. Thus, changes in serine isotope labeling from[2,3,3-2H]serine (particularly in the Mþ2 fraction) are diagnosticfor specific perturbations in folate metabolism, and also informupon the particular enzymatic step that is inhibited.

Treatment of WT HCT116 cells with AGF291, AGF320, orAGF347 closely recapitulated the effects of the SHMT2 KO,including a substantial increase in Mþ3 serine fraction(�55%–60% of total serine) and decreased Mþ2 serine (�2–3-fold) comparedwithWT cells, accompanied by nearly completeloss of Mþ1 serine (Fig. 4C). Similar results were obtained withSHMT2 KD H460 cells and inhibitor-treated WT H460 and MIAPaCa-2 cells (Supplementary Fig. S5C and S5D, respectively).These results identify SHMT2as the likelymitochondrial target forAGF291, AGF320, and AGF347.

Identification of the de novo purine biosynthetic pathway as adirect target for AGF291, AGF320, and AGF347

Both GARFTase and AICARFTase require 10-formyl-THFderived from formate, most of which is generated via mitochon-drial C1 metabolism from serine (ref. 22; Figs. 1 and 4A).Consistent with this, loss of SHMT2 in H460 SHMT2 KD cellsinduced significant increases in purine intermediates which aredependent on C1 pools (i.e., 10-formyl-THF), including GAR(GARFTase substrate; 21-fold; Fig. 4D) and AICAR (AICARFTasesubstrate; 65-fold; Fig. 4E). Likewise, treatment ofH460 cells withAGF291, AGF320, and AGF347 (10 mmol/L) increased GAR (10–2,300-fold) and AICAR (40–1,500-fold) relative to untreatedcontrols (Fig. 4D and E). Similar increases in GAR and AICARpools resulted in inhibitor-treated HCT116 andMIA PaCa-2 cells(Supplementary Fig. S5E–S5H). For the HCT116 sublines, theincreases in GAR and AICAR upon inhibitor treatments generallyexceeded those resulting from the SHMT2 KO.

To assess the possibility that AGF291, AGF320, and AGF347directly inhibit cytosolic enzyme targets in de novo purine bio-synthesis (i.e., GARFTase and/or AICARFTase), we treated theH460 cells with 1 mmol/L formate, to replenish the cytosolic C1pool while circumventing the mitochondrial C1 pathway. Wereasoned that formate treatment of SHMT2 KD cells shouldrestore levels of GAR and AICAR to those seen in NTC (WT) cells.However, if the cytosolic enzymeswere directly inhibited, formateshould not effectively reverseGARand/or AICARaccumulation. InH460 SHMT2 KD cells, treatment with formate completelyreversed elevated GAR (Fig. 4D) and AICAR (Fig. 4E) to NTClevels. However, for inhibitor-treated H460 cells, reversal byformate was incomplete, albeit to different extents for differentcompounds. These results implicate direct targeting of GARFTaseand/or AICARFTase by the novel inhibitors, in addition toSHMT2.

We measured the effects of the inhibitors on de novo purinebiosynthesis, as reflected in total cellular pools of AMP, ADP, andATP. In MIA PaCa-2 cells, all the adenine nucleotides werecomparably suppressed (�50%) with inhibitor treatments(Fig. 4F). In HCT116 cells (Supplementary Fig. S5I), AMP andADP pools were impacted more (50% suppression) than ATP(30% suppression), although SHMT2 KO had no effect. In H460cells (Supplementary Fig. S5J), SHMT2 KD decreased adeninenucleotide pools (20%–50%), yet in response to the inhibitors,AMP increased (�50%), ADP decreased (up to �60%), and ATPwas essentially unaffected. Thus, in spite of the requirementfor adenosine (with glycine) for complete rescue of all three celllines from the inhibitory effect of the pyrrolopyrimidine com-pounds (Fig. 3; Supplementary Fig. S4), the extent of adeninenucleotide depletion varied.

Identification of SHMT1 as a target for pyrrolo[3,2-d]pyrimidine analogues

As SHMT2 (UniProtKB: P34897) maintains 66% sequenceidentity (48) to SHMT1 (UniProtKB: 34896), we considered thepossibility that AGF291, AGF320, and AGF347 could also targetcytosolic SHMT1. By molecular modeling (SupplementaryFig. S2), these analogues bound to rabbit SHMT1 (UniProtKB:07511, 93% homology with human SHMT1; ref. 49) with dock-ing scores from �8.9 to �11.14 kcal/mol (SupplementaryTable S1), raising the possibility that they may directly inhibitthe human enzyme, as well.

To gauge potential cytosolic SHMT1 inhibition by our com-pounds, we traced [2,3,3-2H]serine into dTTP (dTMP) via C1transfer from 5,10-me-THF to dUMP by TS (ref. 22; Fig. 4A). Inall three cell lines, WT cells incubated with [2,3,3-2H]serinegenerated Mþ1 dTTP with little-to-no Mþ2 dTTP (Fig. 4G;Supplementary Fig. S5K and S5L). This confirms preferential[2,3,3-2H]serine metabolism through the mitochondrial C1pathway to [2H]formate and into dTTP (Mþ1) (22). KD ofSHMT2 in H460 cells or KO of SHMT2 in HCT116 cells induceda robust Mþ2 dTTP signal, reflecting serine-to-glycine fluxthrough SHMT1 in the cytosol (22). Consistent with SHMT2targeting, treatmentwithAGF291,AGF320, orAGF347 (10mmol/L)resulted in decreased Mþ1 and total dTTP in the cell lines(Fig. 4G; Supplementary Fig. S5K and S5L). However, only inMIA PaCa-2 cells were these changes accompanied by increas-ed Mþ2 dTTP (reflecting a SHMT1 compensation response;ref. 22); the other two cell lines showed no measurable Mþ2dTTP. These results likely reflect (i) sufficient residual fluxthrough SHMT2 to preserve the primary Mþ1 labeling of dTTP,and/or (ii) direct targeting of SHMT1, along with cell line–specific differences in SHMT1 activity and inhibition by ourcompounds. In spite of decreased dTTP pools, thymidine didnot rescue cells from effects of any of our inhibitors (Supple-mentary Fig. S4), a marked difference from TS-targeted inhibi-tors such as PMX (7). Thus, decreased dTTP likely reflects asecondary impact of targeting SHMT2/SHMT1 rather thandirectly targeting TS.

Further evidence for direct SHMT1 targeting by our inhibitorsinvolves the substantially increased antiproliferative effects ofAGF291, AGF320, and AGF347 toward HCT116 SHMT2 KO cellscompared withWT cells (21-fold for AGF291, 6-fold for AGF320,and 10-fold for AGF347; Table 1). Conversely, sensitivities toAGF94 and PMX were similar between WT and SHMT2 KO cells.These results mirror the increased potency toward SHMT2 KO

Dekhne et al.





cells relative to WT HCT116 cells seen with SHIN1 (23) andfurther suggest that our pyrrolo[3,2-d]pyrimidine compoundsinhibit SHMT1 in addition to SHMT2.

In vitro confirmation of enzyme targetsTo confirm the enzyme targets (SHMT2, GARFTase, AICARF-

Tase, and SHMT1) identified from our metabolomics experi-ments, we performed in vitro assays using purified recombinantenzymes. Our results demonstrate direct inhibition of SHMT2by AGF291, AGF320, and AGF347 (KIs of 0.63, 0.056 and2.19 mmol/L, respectively; Table 1). While the compounds inhib-ited AICARFTase within a 4–5-fold range, AGF320 was the mostpotent GARFTase inhibitor, corroborating the metabolomicsresults (Fig. 4D and E). For SHMT1, inhibitions paralleled thosefor SHMT2. None of the compounds inhibited MTHFD2 atconcentrations of up to 200 mmol/L of inhibitor (Table 1). Theseresults confirm that SHMT2, the purine biosynthetic enzymesGARFTase and AICARFTase, and SHMT1 are direct targets of ourlead pyrrolo[3,2-d]pyrimidine compounds. As a control, AGF94was tested up to 200 mmol/L inhibitor. AGF94 inhibited GARF-Tase (0.88mmol/L), however, therewas no inhibition forAICARF-Tase orMTHFD2andonlyweak inhibitionof SHMT2andSHMT1(Kis of 188 and 112 mmol/L, respectively). Notably, for PMX therewas no significant inhibition of MTHFD2, SHMT1, or SHMT2 upto 200 mmol/L; however, substantial inhibition was seen forGARFTase and AICARFTase (Table 1).

In vivo antitumor efficacy study of AGF347 with MIA PaCa-2tumor xenografts

We performed an in vivo efficacy trial with AGF347 in ahead-to-head comparison with gemcitabine toward MIA PaCa-2, an aggressive tumor model that reaches a euthanasia endpointwithin 2 weeks of implant. We initially used early-stage disease asa primary test of drug efficacy. Fourteen days before SCtumor implant, the mice were fed a folate-deficient dietto ensure that serum folate levels approximated those ofhumans (36, 38, 40, 41, 51). The mice were unselectivelydistributed to each group's treatment and control arms (5 miceper group), and dosed with AGF347 (15 mg/kg/injection every2 days� 8; total dose of 120mg/kg) or gemcitabine (120mg/kg/injection every 4 days � 4; total dose of 480 mg/kg) just belowtheir respective maximum tolerated doses (MTDs). Gemcitabinewas efficacious, with amedian tumor burden onday 16 of 420mg(range 284–552 mg) compared with 1,189 mg (range 601–1,711mg) for the control cohort;AGF347was distinctly superior,yielding a median tumor burden of 0 mg (range 0–276 mg).Median T/C values (see Materials and Methods for definition ofquantitative endpoints) on day 16were 35% for gemcitabine and0% for AGF347 (Fig. 5A). Tumor growth delays (median T-C toreach 1,000 mg) were 6 days for gemcitabine, compared with39 days for 4 of 5 mice treated with AGF347 and one tumor-freesurvivor (TFS) on day 136 (121 days post last treatment). Toconfirm whether immune factors contributed to this "curative"response, this mouse was "rechallenged" withMIA PaCa-2 tumor(bilateral SC implant). The tumor reached 1,000 mg 16 dayspost-implant, with comparable growth and doubling time to thecontrol group (median, 15.5 days; range, 13–20 days; Fig. 5A).Both drugs were well tolerated, with modest weight loss as theonly adverse symptom [for AGF347, 1% (�0.2 g) median nadironday17; for gemcitabine, 6.9%(�1.4 g)mediannadir onday3].After cessation of therapy, host recovery time was rapid for

AGF347 (1 day) compared with gemcitabine (14 days). Thus, atequitoxic dose levels,AGF347 showed superior antitumor efficacyover gemcitabine (>5 logs of cell kill compared with 1.1 logs forgemcitabine), with a 4-fold decreased dose requirement, no acuteor long-term toxicities other than reversibleweight loss, rapidhostrecovery time, and a tumor growth delay of >38 days, with 1 of 5TFS (Supplementary Table S2).

Studies were extended to upstage MIA PaCa-2 xenografts. Forthese experiments, after implantation, MIA PaCa-2 tumors wereallowed to grow for 7 days (100–150 mg), unselectively distrib-uted (5 mice/group), then treated as for the early-stage trial(Fig. 5B). AGF347 and gemcitabine were both efficacious, withmedian tumor burdens on day 21 of 733 mg (range 373–1,835 mg) and 524 mg (range 247–1,045 mg), respectively,compared with 1,736mg (range 1,117–2,046mg) for the controlcohort. AGF347 effected two partial remissions (PRs), includingone complete response (CR); no PRs or CRs were seen withgemcitabine. The PRs involved substantial decreases in tumorburden (688 to 220mg, and 400 to 0mg) and occurred on day 31(10 days post last dose). Overall tumor growth delays (median T-C to reach 1,000 mg) of 11 days (gemcitabine) and 7 days(AGF347) were recorded. For the PRs, AGF347 induced tumorgrowth delays of 26 and 52 days, as compared to 16 and 25 daysfor the best gemcitabine-responding mice. Both drugs were welltolerated,withmodestweight loss during treatment [AGF347, 1%(�0.2 g) median nadir on day 21 with full recovery on day 24;gemcitabine, 5.1% (�1.0 g) median nadir on day 20 with fullrecovery on day 24].

For matched cohorts of mice maintained on a standard diet(results in approximately 10-fold increased serum folate), anti-tumor activity of AGF347 was ablated (Supplementary Fig. S7).Thus, excessive levels of circulating folates largely abolish theantitumor activity of AGF347.

To confirm that the enzyme targets ofAGF347 identified in vitro(GARFTase, AICARFTase, and SHMT1/2) were inhibited in vivo,tumor samples from a separate cohort of AGF347-treated andcontrol mice (n ¼ 3 each) were harvested 6 hours after the sixthinjection of AGF347 (tumors > 1,000 mg). The tumors werehomogenized and metabolites analyzed by targeted metabolo-mics (Fig. 5C and D). In vivo administration of AGF347 inducedsubstantially increased GAR (200-fold) and AICAR (500-fold),consistent with direct inhibition of GARFTase and AICARFTase asseen in vitro (Fig. 4D and E); total serine trended up but did notdiffer significantly from the control, likely due to exchange withcirculating serine pools. We measured changes in purine nucle-otide pools in vivo, including 5-fold increased AMP, approximate-ly 2-fold decreased ADP and approximately 10-fold decreasedATP. This severe perturbation of energy charge resulted in acompensatory upregulation of cytochrome c oxidase activity(ref. 51; Supplementary Fig. S8), consistent with cytotoxic drugeffects over the chronic course of in vivo treatment.

DiscussionThe clinical utility of standard chemotherapy drugs is often

limited by toxicities toward normal tissues (reflecting a lack oftumor selectivity) and/or drug resistance. Discovery of new andpotent inhibitors of tumor-selective pathways remains a formi-dable challenge.

In this report, we identified novel inhibitors of C1 metabolisminmitochondria and cytosolwith in vivo antitumor activity. Serine






catabolism by mitochondrial C1 metabolism is the principalsource of C1 units for cytosolic de novo purine and thymidylatebiosynthesis, and a source of reducing equivalents andATP (2–5, 21). We targeted SHMT2, the first enzyme in themitochondrial C1 pathway, with 5-substituted pyrrolo[3,2-d]pyrimidine analogues, rationally designedbymolecularmodel-ing from the SHMT2 crystal structure (23) and based on structuralsimilarities to our previously reported 5-substituted pyrrolo[2,3-d]pyrimidine compounds (43, 44) and N-substituted THF meta-bolites. We identified lead compounds AGF291, AGF320, andAGF347 that inhibited in vitroproliferation of a broad spectrumoftumor subtypes including lung (H460), colon (HCT116), andpancreatic (MIA PaCa-2) cancers.

We assessed critical enzyme targets for our compoundsthrough glycine and nucleoside rescue, and by targeted meta-bolomics with a [2,3,3-2H]serine tracer, and identified SHMT2in mitochondria as an intracellular target, along with SHMT1,GARFTase, and AICARFTase in the cytosol. Inhibition of allthese targets was confirmed by in vitro assays with purifiedrecombinant enzymes. While the in vitro assays used "mono-glutamyl" inhibitors, greater inhibition should result followingmetabolism to polyglutamates (7, 9). Inhibition of SHMT1 by

our analogues prevents metabolic "compensation" by reversalof SHMT1 catalysis and synthesis of glycine in response to lossof SHMT2 activity (22, 23).

A number of pharmacodynamic factors could contribute to thein vitro antitumor effects of the novel analogues described herein.These include facilitated transport across the plasma membraneby PCFT and/or RFC and into mitochondria (potentially bySLC25A32; refs. 13, 14), and metabolism to drug polygluta-mates (7, 9). Variations in these parameters likely account fordifferences in relative antiproliferative activities toward the assort-ed tumor models. As AGF347 showed the most potent in vitroactivity of the series, this compound was tested in vivo with bothearly-stage and upstage MIA PaCa-2 xenograft models in SCIDmice. AGF347 exhibited potent in vivo efficacy superior to that ofgemcitabine, even at a 4-fold decreased total dose, resulting insustained tumor growth delay and a 20% complete response rate.For the upstage trial, for the PRs, substantial decreases in tumorburdenoccurred 10days after the last dose, likely reflecting a long-term impact of AGF347 polyglutamates, associated with targetinhibition.

Primary inhibition of SHMT2 in mitochondria results indecreased glycine and C1 units for cytosolic biosynthesis

Figure 5.

In vivo efficacy in early-stage and upstage trials and pharmacodynamics of AGF347 toward MIA PaCa-2 xenografts. A, Female NCr SCID mice (11 weeks old 20.2 gaverage body weight) were implanted bilaterally with human MIA PaCa-2 pancreatic cancer xenograft fragments. Treatment was initiated on day 1 followingsubcutaneous tumor implantation. Themice were dosed as follows: AGF347, Q2d�8 at 15 mg/kg/injection, total dose 120 mg/kg; and gemcitabine (GEM),Q4d�4 at 120mg/kg/i.v. injection, total dose 480mg/kg. Results are shown for individual mice. B, Female NCr SCID mice (10 weeks old; 19 g average bodyweight) were implanted bilaterally with human MIA PaCa-2 tumors. On day 7 following subcutaneous implantation, the mice were dosed as forA. C and D,Targeted metabolomics on tumor samples from upstage trial mice. GAR, AICAR, and serine (C) and adenine nucleotides (AMP, ADP, ATP; D) were quantified incontrol and AGF347-treated mice. Results for metabolomics (mean� SD) are normalized to control values and represent one tumor each taken from threedifferent control and AGF347-treated mice. The changes in GAR and AICAR were statistically significant (�� , P < 0.01) as were the changes in adenine nucleotides(�� , P < 0.01, #, P < 0.1), while the changes in serine did not reach statistical significance ("ns"¼ not significant), likely due to circulating serine pools.

Dekhne et al.





(4, 23), and complete ablation through genetic means leadsto defective mitochondrial respiration due to impaired syn-thesis of respiratory chain proteins (52, 53). Other effectsinclude decreased NADPH and glutathione with impacts onmitochondrial redox balance and reactive oxygen species(21), and decreased ATP synthesis due to impaired oxidativephosphorylation (52, 53) and C1-flux through MTHFD1L(54). These are exacerbated by direct inhibitions of the denovo purine biosynthetic enzymes GARFTase and AICARFTaseby our analogues. Although adenosine (with glycine) wasessential for in vitro rescue of the tumor cell lines from ourlead compounds, their effects on purine nucleotide pools werevariable in both magnitude and direction. This suggests thatdepletion of critical subcellular purine nucleotide pools maynot be reflected in measurements of total cellular purinenucleotides. The sustained treatment for AGF347 in vivoresulted in tumor killing with net energy depletion resultingin an increased AMP/ATP ratio. This was accompanied byactivation of cytochrome c oxidase, possibly reflecting post-translational modification (dephosphorylation; ref. 52).

Our results with active 5-substituted pyrrolo[3,2-d]pyrimidinecompounds expand upon earlier findings with nonfolate pyra-zolopyran inhibitors of human SHMT2 (23) and, to our knowl-edge, this series represents the first bona fide inhibitors of thisintracellular target with demonstrated in vivo antitumor efficacy.Thus, inhibition of SHMT2, coupled with direct inhibition ofmultiple C1-dependent targets including de novo purine biosyn-thesis and SHMT1, affords a valuable and exciting new platformfor future drug development.

Disclosure of Potential Conflicts of InterestJ.D. Rabinowitz reports receiving other commercial research support

from Raze, has ownership interest (including patents) in Raze andPrinceton Patent, and has consultant/advisory board relationship withRaze, Kadmon, and L.E.A.F. No potential conflicts of interest were disclosedby the other authors.

Authors' ContributionsConception and design: A.S. Dekhne, K. Shah, L. Polin, Z. Hou, C.E. Dann,A. Gangjee, L.H. MatherlyDevelopment of methodology: A.S. Dekhne, K. Shah, J.M. Katinas, J. Wong-Roushar, A. Doshi, X. Bao, L. Polin, J.D. Rabinowitz, J. Li, Z. Hou, C.E. Dann,A. Gangjee

Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A.S. Dekhne, G.S. Ducker, J.M. Katinas, J. Wong-Roushar, A. Doshi, C. Ning, X. Bao, S.H. Dzinic, K. White, M. H€uttemann,L. Polin, J. Li, C.E. DannAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A.S. Dekhne, K. Shah, G.S. Ducker, J.M. Katinas,J. Wong-Roushar, J. Wong-Roushar, M.J. Nayeen, A. Doshi, X. Bao, J. Liu,C. O'Connor, J. Kushner, S. Kim, M. H€uttemann, L. Polin, Z. Hou,C.E. Dann, A. Gangjee, L.H. MatherlyWriting, review, and/or revision of the manuscript: A.S. Dekhne, K. Shah,G.S. Ducker, J.M. Katinas, J. Wong-Roushar, A. Doshi, S.H. Dzinic,M. H€uttemann, L. Polin, J.D. Rabinowitz, J. Li, Z. Hou, C.E. Dann,A. Gangjee, L.H. MatherlyAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): A.S. Dekhne, K. Shah, J. Liu, A. Wallace-Povirk,C. O'Connor, K. White, J. Kushner, L. Polin, Z. Hou, C.E. Dann, A. Gangjee,L.H. MatherlyStudy supervision: A.S. Dekhne, L. Polin, Z. Hou, C.E. Dann, A. Gangjee,L.H. MatherlyOther (developed the synthetic methodology for the active molecules):K. ShahOther (performed research): M.J. NayeenOther (performed research and analyzed data): J. Fr€uhauf

AcknowledgmentsThis work was supported in part by R01 CA53535 (to L.H. Matherly and

Z. Hou), R01 CA152316 (to L.H. Matherly and A. Gangjee), K99 CA215307(to G.S. Ducker), DP1DK113643 (to J.D. Rabinowitz), R01 CA163591 (to J.D.Rabinowitz), and R01 CA166711 (to AGangjee, L.H. Matherly, and C.E. Dann)from theNIH, a pilot Small Molecule DrugDiscovery Grant from the KarmanosCancer Institute (to Z. Hou and L.H. Matherly), the Eunice and Milton RingEndowed Chair for Cancer Research (to L.H. Matherly), and the DuquesneUniversity Adrian Van Kaam Chair in Scholarly Excellence (to A. Gangjee). A.S.Dekhne was supported by T32 CA009531 (to L.H. Matherly) and F30CA228221. The Animal Model and Therapeutics Evaluation Core (to L. Polin,S.H. Dzinic, K. White, and J. Kushner), the Pharmacology Core (to X. Bao and J.Liu) and Biostatistics Core (to S. Kim) were supported, in part, by NIH Centergrant P30 CA022453 to the Karmanos Cancer Institute and the Wayne StateUniversity.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received January 9, 2019; revised April 5, 2019; accepted July 3, 2019;published first July 9, 2019.

References1. Pavlova NN, Thompson CB. The emerging hallmarks of cancer metabo-

lism. Cell Metab 2016;23:27–47.2. Newman AC, Maddocks ODK. One-carbon metabolism in cancer. Br J

Cancer 2017;116:1499–504.3. Yang M, Vousden KH. Serine and one-carbon metabolism in cancer.

Nat Rev Cancer 2016;16:650–62.4. Ducker GS, Rabinowitz JD. One-carbonmetabolism in health and disease.

Cell Metab 2017;25:27–42.5. Tibbetts AS, Appling DR. Compartmentalization of Mammalian folate-

mediated one-carbon metabolism. Annu Rev Nutr 2010;30:57–81.6. Field MS, Kamynina E, Chon J, Stover PJ. Nuclear folate metabolism.

Annu Rev Nutr 2018;38:219–43.7. Chattopadhyay S, Moran RG, Goldman ID. Pemetrexed: biochemical and

cellular pharmacology,mechanisms, and clinical applications.Mol CancerTher 2007;6:404–17.

8. Matherly LH, Wilson MR, Hou Z. The major facilitative folate transporterssolute carrier 19A1 and solute carrier 46A1: biology and role in antifolatechemotherapy of cancer. Drug Metab Dispos 2014;42:632–49.

9. Visentin M, Zhao R, Goldman ID. The antifolates. Hematol Oncol ClinNorth Am 2012;26:629–48.

10. Palmer AM, Kamynina E, Field MS, Stover PJ. Folate rescues vitaminB12 depletion-induced inhibition of nuclear thymidylate biosynthesisand genome instability. Proc Natl Acad Sci U S A 2017;114:E4095–E102.

11. Matherly LH, Hou Z, Deng Y. Human reduced folate carrier: translation ofbasic biology to cancer etiology and therapy. Cancer Metastasis Rev 2007;26:111–28.

12. ZhaoR,Diop-BoveN,VisentinM,Goldman ID.Mechanismsofmembranetransport of folates into cells and across epithelia. Annu Rev Nutr 2011;31:177–201.

13. McCarthy EA, Titus SA, Taylor SM, Jackson-CookC,Moran RG. Amutationinactivating themitochondrial innermembrane folate transporter creates aglycine requirement for survival of chinese hamster cells. J BiolChem2004;279:33829–36.

14. Lawrence SA, Hackett JC, Moran RG. Tetrahydrofolate recognition by themitochondrial folate transporter. J Biol Chem 2011;286:31480–9.






15. Lawrence SA, Titus SA, Ferguson J, Heineman AL, Taylor SM, Moran RG.Mammalian mitochondrial and cytosolic folylpolyglutamate synthetasemaintain the subcellular compartmentalization of folates. J Biol Chem2014;289:29386–96.

16. Jain M, Nilsson R, Sharma S, Madhusudhan N, Kitami T, Souza AL, et al.Metabolite profiling identifies a key role for glycine in rapid cancer cellproliferation. Science 2012;336:1040–4.

17. Kim D, Fiske BP, Birsoy K, Freinkman E, Kami K, Possemato R, et al.SHMT2 drives glioma cell survival in the tumor microenvironmentbut imposes a dependence on glycine clearance. Nature 2015;520:363–7.

18. Nilsson R, Jain M, Madhusudhan N, Sheppard NG, Strittmatter L,Kampf C, et al. Metabolic enzyme expression highlights a key role forMTHFD2 and the mitochondrial folate pathway in cancer.Nat Commun 2014;5:3128.

19. Zhang L, Chen Z, Xue D, Zhang Q, Liu X, Luh F, et al. Prognostic andtherapeutic value of mitochondrial serine hydroxyl-methyltransferase 2 asa breast cancer biomarker. Oncol Rep 2016;36:2489–500.

20. Loayza-Puch F, Rooijers K, Buil LC, Zijlstra J, Oude Vrielink JF, Lopes R,et al. Tumour-specific proline vulnerability uncovered by differentialribosome codon reading. Nature 2016;530:490–4.

21. Ye J, Fan J, Venneti S, Wan YW, Pawel BR, Zhang J, et al. Serine catabolismregulates mitochondrial redox control during hypoxia. Cancer Discov2014;4:1406–17.

22. Ducker GS, Chen L, Morscher RJ, Ghergurovich JM, Esposito M, Teng X,et al. Reversal of cytosolic one-carbon flux compensates for loss of themitochondrial folate pathway. Cell Metab 2016;23:1140–53.

23. Ducker GS, Ghergurovich JM, MainolfiN, Suri V, Jeong SK, Hsin-Jung Li S,et al. Human SHMT inhibitors reveal defective glycine import as a target-able metabolic vulnerability of diffuse large B-cell lymphoma. Proc NatlAcad Sci U S A 2017;114:11404–9.

24. Fu TF, Scarsdale JN, Kazanina G, Schirch V, Wright HT. Location of thepteroylpolyglutamate-binding site on rabbit cytosolic serine hydroxy-methyltransferase. J Biol Chem 2003;278:2645–53.

25. Flintoff WF, Davidson SV, Siminovitch L. Isolation and partial character-ization of three methotrexate-resistant phenotypes from Chinese hamsterovary cells. Somat Cell Mol Genet 1976;2:245–61.

26. Deng Y, Wang Y, Cherian C, Hou Z, Buck SA, Matherly LH, et al. Synthesisand discovery of high affinity folate receptor-specific glycinamide ribonu-cleotide formyltransferase inhibitors with antitumor activity. J Med Chem2008;51:5052–63.

27. Deng Y, Zhou X, Desmoulin SK, Wu J, Cherian C, Hou Z, et al. Synthesisand biological activity of a novel series of 6-substituted thieno[2,3-d]pyrimidine antifolate inhibitors of purine biosynthesis with selectivityfor high affinity folate receptors over the reduced folate carrier and proton-coupled folate transporter for cellular entry. J Med Chem 2009;52:2940–51.

28. Wong SC, Proefke SA, BhushanA,Matherly LH. Isolation of human cDNAsthat restore methotrexate sensitivity and reduced folate carrier activity inmethotrexate transport-defective Chinese hamster ovary cells. J Biol Chem1995;270:17468–75.

29. Hou Z, Gattoc L, O'Connor C, Yang S, Wallace-Povirk A, George C, et al.Dual targeting of epithelial ovarian cancer via folate receptor alpha and theproton-coupled folate transporter with 6-substituted pyrrolo[2,3-d]pyrimidine antifolates. Mol Cancer Ther 2017;16:814–30.

30. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Proteinmeasurement withthe Folin phenol reagent. J Biol Chem 1951;193:265–75.

31. Laemmli UK. Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature 1970;227:680–5.

32. Matsudaira P. Sequence from picomole quantities of proteins electro-blotted onto polyvinylidene difluoride membranes. J Biol Chem 1987;262:10035–8.

33. Lu W, Clasquin MF, Melamud E, Amador-Noguez D, Caudy AA, Rabino-witz JD. Metabolomic analysis via reversed-phase ion-pairing liquid chro-matography coupled to a stand alone orbitrap mass spectrometer.Anal Chem 2010;82:3212–21.

34. Clasquin MF, Melamud E, Rabinowitz JD. LC-MS data processing withMAVEN: a metabolomic analysis and visualization engine. Curr ProtocBioinformatics 2012;Chapter 14:Unit14.1.

35. Golani LK, Wallace-Povirk A, Deis SM, Wong JE, Ke J, Gu X, et al.Tumor targeting with novel 6-substituted pyrrolo [2,3-d] pyrimidine

antifolates with heteroatom bridge substitutions via cellular uptake byfolate receptor a and the proton-coupled folate transporter and inhi-bition of de novo purine nucleotide biosynthesis. J Med Chem 2016;59:7856–76.

36. Shih C, Chen VJ, Gossett LS, Gates SB, MacKellar WC, Habeck LL,et al. LY231514, a pyrrolo[2,3-d]pyrimidine-based antifolate thatinhibits multiple folate-requiring enzymes. Cancer Res 1997;57:1116–23.

37. Wang L, Wallace A, Raghavan S, Deis SM, Wilson MR, Yang S, et al.6-Substituted pyrrolo[2,3-d]pyrimidine thienoyl regioisomers astargeted antifolates for folate receptor alpha and the proton-coupled folate transporter in human tumors. J Med Chem 2015;58:6938–59.

38. Cherian C, Kugel Desmoulin S, Wang L, Polin L, White K, Kushner J, et al.Therapeutic targeting malignant mesothelioma with a novel 6-substitutedpyrrolo[2,3-d]pyrimidine thienoyl antifolate via its selective uptake by theproton-coupled folate transporter. Cancer Chemother Pharmacol 2013;71:999–1011.

39. Ravindra MP, Wilson MR, Tong N, O'Connor C, Karim MA, Polin L, et al.Fluorinated substituted pyrrolo[2,3-d]pyrimidine antifolates with tumor-targeting via cellular uptake by folate receptor a and the proton-coupledfolate transporter and inhibition of de novo purine nucleotide biosynthesis.J Med Chem 2018;61:4228–48.

40. Wang L, Cherian C, Desmoulin SK, Polin L, Deng Y, Wu J, et al. Synthesisand antitumor activity of a novel series of 6-substituted pyrrolo[2,3-d]pyrimidine thienoyl antifolate inhibitors of purine biosynthesis withselectivity for high affinity folate receptors and the proton-coupled folatetransporter over the reduced folate carrier for cellular entry. J Med Chem2010;53:1306–18.

41. Wang L, Desmoulin SK, Cherian C, Polin L, White K, Kushner J,et al. Synthesis, biological and antitumor activity of a highlypotent 6-substituted pyrrolo[2,3-d]pyrimidine thienoyl antifolateinhibitor with proton-coupled folate transporter and folate recep-tor selectivity over the reduced folate carrier that inhibits b--glycinamide ribonucleotide formyltransferase. J Med Chem 2011;54:7150–64.

42. Lee GY, Haverty PM, Li L, Kljavin NM, Bourgon R, Lee J, et al. Comparativeoncogenomics identifies PSMB4 and SHMT2 as potential cancer drivergenes. Cancer Res 2014;74:3114–26.

43. Wang Y, Cherian C, Orr S, Mitchell-Ryan S, Hou Z, Raghavan S, et al.Tumor-targeting with novel non-benzoyl 6-substituted straight chainpyrrolo[2,3-d]pyrimidine antifolates via cellular uptake by folate receptora and inhibition of de novo purine nucleotide biosynthesis. J Med Chem2013;56:8684–95.

44. Wang Y, Mitchell-Ryan S, Raghavan S, George C, Orr S, Hou Z, et al. Novel5-substituted pyrrolo[2,3-d]pyrimidines as dual inhibitors of glycinamideribonucleotide formyltransferase and 5-aminoimidazole-4-carboxamideribonucleotide formyltransferase and as potential antitumor agents. J MedChem 2015;58:1479–93.

45. Desmoulin SK, Hou Z, Gangjee A, Matherly LH. The human proton-coupled folate transporter: biology and therapeutic applications to cancer.Cancer Biol Ther 2012;13:1355–73.

46. Wilson MR, Hou Z, Yang S, Polin L, Kushner J, White K, et al. Targetingnonsquamous nonsmall cell lung cancer via the proton-coupled folatetransporter with 6-substituted pyrrolo[2,3-d]pyrimidine thienoyl antifo-lates. Mol Pharmacol 2016;89:425–34.

47. Kugel Desmoulin S, Wang L, Hales E, Polin L, White K, Kushner J, et al.Therapeutic targeting of a novel 6-substituted pyrrolo [2,3-d]pyrimidine thienoyl antifolate to human solid tumors based onselective uptake by the proton-coupled folate transporter.Mol Pharmacol 2011;80:1096–107.

48. Altschul SF, Wootton JC, Gertz EM, Agarwala R, Morgulis A, Schaffer AA,et al. Protein database searches using compositionally adjusted substitu-tion matrices. FEBS J 2005;272:5101–9.

49. Altschul SF, GishW,MillerW,Myers EW, LipmanDJ. Basic local alignmentsearch tool. J Mol Biol 1990;215:403–10.

50. Varela-Moreiras G, Selhub J. Long-term folate deficiency alters folatecontent and distribution differentially in rat tissues. J Nutr 1992;122:986–91.

51. Huttemann M, Lee I, Grossman LI, Doan JW, Sanderson TH. Phosphor-ylation of mammalian cytochrome c and cytochrome c oxidase in the

Dekhne et al.





regulation of cell destiny: respiration, apoptosis, and human disease.Adv Exp Med Biol 2012;748:237–64.

52. MintonDR,NamM,McLaughlinDJ, Shin J, Bayraktar EC, Alvarez SW, et al.Serine catabolism by SHMT2 is required for proper mitochondrial trans-lation initiation and maintenance of formylmethionyl-tRNAs. Mol Cell2018;69:610–21.

53. Morscher RJ, Ducker GS, Li SH, Mayer JA, Gitai Z, Sperl W, et al. Mito-chondrial translation requires folate-dependent tRNAmethylation.Nature2018;554:128–32.

54. Meiser J, Tumanov S, Maddocks O, Labuschagne CF, Athineos D, Van DenBroek N, et al. Serine one-carbon catabolism with formate overflow.Sci Adv 2016;2:e1601273.






2019;18:1787-1799. Published OnlineFirst July 9, 2019.Mol Cancer Ther Aamod S. Dekhne, Khushbu Shah, Gregory S. Ducker, et al. Antitumor Efficacyand Cytosolic One-carbon Metabolism with Broad-spectrum

]pyrimidine Compounds Target MitochondrialdNovel Pyrrolo[3,2-

Updated version

10.1158/1535-7163.MCT-19-0037doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2019/07/09/1535-7163.MCT-19-0037.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/18/10/1787.full#ref-list-1

This article cites 53 articles, 19 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/18/10/1787.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/18/10/1787To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-19-0037http://mct.aacrjournals.org/content/suppl/2019/07/09/1535-7163.MCT-19-0037.DC1http://mct.aacrjournals.org/content/18/10/1787.full#ref-list-1http://mct.aacrjournals.org/content/18/10/1787.full#related-urlshttp://mct.aacrjournals.org/cgi/alertsmailto:[email protected]://mct.aacrjournals.org/content/18/10/1787http://mct.aacrjournals.org/

/ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages false /GrayImageMinResolution 200 /GrayImageMinResolutionPolicy /Warning /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages false /MonoImageMinResolution 600 /MonoImageMinResolutionPolicy /Warning /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 900 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False

/CreateJDFFile false /Description > /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ > /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MarksOffset 18 /MarksWeight 0.250000 /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /NA /PageMarksFile /RomanDefault /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /LeaveUntagged /UseDocumentBleed false >> > ]>> setdistillerparams> setpagedevice

Home | Molecular Cancer Therapeutics - Novel Pyrrolo[3,2-d ......Small Molecule Therapeutics Novel...

Documents

Transcript of Home | Molecular Cancer Therapeutics - Novel Pyrrolo[3,2-d ......Small Molecule Therapeutics Novel...