Pharmacological blockade of ASCT2-dependent glutamine transport leads to antitumor efficacy in preclinical models

Abstract

The unique metabolic demands of cancer cells underscore potentially fruitful opportunities for drug discovery in the era of precision medicine. However, therapeutic targeting of cancer metabolism has led to surprisingly few new drugs to date. The neutral amino acid glutamine serves as a key intermediate in numerous metabolic processes leveraged by cancer cells, including biosynthesis, cell signaling, and oxidative protection. Herein we report the preclinical development of V-9302, a competitive small molecule antagonist of transmembrane glutamine flux that selectively and potently targets the amino acid transporter ASCT2. Pharmacological blockade of ASCT2 with V-9302 resulted in attenuated cancer cell growth and proliferation, increased cell death, and increased oxidative stress, which collectively contributed to antitumor responses in vitro and in vivo. This is the first study, to our knowledge, to demonstrate the utility of a pharmacological inhibitor of glutamine transport in oncology, representing a new class of targeted therapy and laying a framework for paradigm-shifting therapies targeting cancer cell metabolism.

Extended Data Fig. 1. Amino acid transporters and substrates used for selectivity screening

(A) Common transporters and amino acid substrates shown. Amino acids designated by one-letter codes. Transporters responsible for glutamine uptake shown in blue shading. Amino acid substrates utilized to evaluate the selectivity of V-9302 shown in red.

Extended Data Fig. 2. Drug Affinity Responsive Target Stability (DARTS) assay comparing the stabilization of ASCT2 and ASCT1 by V-9302

(A) Immunoblots for ASCT2 and ASCT2 illustrating concentration dependent stabilization of ASCT2 in the presence of the protease thermolysin by V-9302. ASCT1 is not similarly protected by V-9302. Uncroppped gel images are provided in Supplementrary Fig. 3. (B) Immunoblot densitometry analysis DARTS immunoblots for ASCT2 and ASCT1. Percent control (Ctrl) relative to ASCT2 or ASCT1 immunoreactivity, respectively, without thermolysin.

Extended Data Fig. 3. Summary and annotation of human cancer cell lines utilized in CellTiter Glo viability screen of V-9302 activity (graph shown in Figure 3)

Percent change in viability relative to vehicle control following 25 µM V-9302 exposure for 48 h (% △viability relative to vehicle); standard deviation (SD).

Extended Data Fig. 4. Follow-up viabiity screening of V-9302 activity in human CRC cells

In vitro multiplex assay (MultiTox Glo) of 15 colorectal cancer cell lines exposed to V-9302 (25 µM, 48 h). V-9302-dependent changes in the number of live cells (A, blue) and dead cells (B, red) relative to vehicle control; n = 3 independent experiments. (C) V-9302 Sulforhodamine B assay evaluating V-9302-dependent cell death in 9 colorectal cancer cell lines in vitro. Data for each assay relative to vehicle control; n = 3 independent experiments. Select mutational status (green squares) shown. All data is p < 0.05 as determined by Student’s t test relative to vehicle control unless otherwise indicated. n.s. = not statistically significant. Error bars represent ± std. dev.

Extended Data Fig. 5. Lack of correlation between ASCT2 protein levels and sensitivity to V-9302

(A) Membranous and (B) total ASCT2 levels evaluated by western blotting in thirteen human CRC cell lines. Uncropped gel images are provided in Supplementary Fig. 4. Neither membranous (C) nor total (D) ASCT2 levels quantified by densitometry correlated with V-9302 sensitivity. Immunoblots normalized to total protein gel loading control.

Extended Data Fig. 6. Effect of glutamine withdrawal (A/B) or combinatorial ASCT2 substrate withdrawal (C/D) on cellular viability and cell death in V-9302 sensitive human CRC cell lines

Cell lines propagated for 48 hrs in either glutamine-depleted or ASCT2-substrate-depleted media (without alanine, serine, cysteine, threonine, glutamine, asparagine, methionine, glycine, leucine, valine, glutamate); Multi-Tox Glo assay. (A) Cell viability and (B) cell death with glutamine withdrawal (C) Cell viability and (D) cell death with ASCT2-substrate withdrawal. Percent change relative to standard media control. n.s. = not statistically significant; ** = p < 0.01; **** = p < 0.0001 as determined by Student’s t test. Error bars represent ± std. dev.

Extended Data Fig. 7. Viability and activation of mouse T-cells following V-9302 exposure for 96 hrs

(A) Viability analyzed by propidium iodide flow cytometry. (B) Activation of T-cells as measured by flow cytometry for CD44 expressions; mean fluorescence intensity (MFI). n = 3 independent experiments. P values determined by Student’s t test

Extended Data Fig. 8. Effects of V-9302 exposure (25 µM, 48 h) on pS6 and pERK in HT29 cells

Uncropped gel images are provided in Supplementrary Fig. 5.

Extended Data Fig. 9. Validation of V-9302-dependent decrease in pERK levels in HCC1806 and HT29 cells; SureFire biochemical assay

Comparison of V-9302 and glutaminase inhibitor, CB839. (A) HCC1806 and (B) HT-29 cells. Drug concentrations shown; treatment duration 48 h; n = 3 independent experiments. P values determined by Student’s t test. n.s. = not statistically significant. Error bars represent ± std. dev.

Extended Data Fig. 10. Evaluation of oxidized glutathione (GSSG) levels in HT-29 cells with V-9302 or CB-839 exposure. (A) Oxidized glutathione (GSSG) and (B) reactive oxygen species (ROS)

Drug concentrations shown, cells treated for 48 hrs; n = 3 independent experiments. P values determined by Student’s t test. Error bars represent ± std. dev.

Extended Data Fig. 11. Analysis of autophagic flux with V-9302 exposure

(A) Autophagic vesicles in HCC1806 cells with exposure to increasing concentrations of V-9302; 8 h treatment duration, concentrations shown; n = 3 independent experiments. P values determined by Student’s t test. Vesicles induced by rapamycin (dotted line, 500 nM, positive control). Combination of lysosomal inhibitor chloroquine (10 µM) and V-9302. (B) or CB-839 (C) in HT29 cells. CLQ = chloroquine. n = 3 independent experiments. P values determined by Student’s t test. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max. Error bars represent ± std. dev.

Extended Data Fig. 12. Additive effect on cellular viability between V-9302 and chloroquine but not CB-839 and chloroquine

HT29 cells (A) and HCT-116 cells (B) exposed to V-9302 (concentrations shown), CB-839 (concentrations shown), chloroquine (CLQ, 10 µM), and the combinations thereof for 48 h; n = 3 independent experiments. P values determined by Student’s t test. n.s. = not statistically significant. Error bars represent ± std. dev.

Extended Data Fig. 13. Fluorescence lifetimes of NAD(P)H and FAD measured by optical spectroscopy were reduced with exposure to V-9302 (10 µM) in HCC1806 cells

Drug exposure 48 h, magnification 40×; n = 3 independent experiments. P values determined by Student’s t test. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max Error bars represent ± std. dev.

Extended Data Fig. 14. Viability of human CRC organoids A 007 (BRAF^V600E) (A) and A 008 (KRAS^G12V;p53^R248Q;PTEN^L140Y) (B) with exposure to CB-839 (35 µM, 96 h)

n = 3 independent experiments. P values determined by Student’s t test. n.s. = not statistically significant. Error bars represent ± std. dev.

Extended Data Fig. 15. Longitudinal assay of V-9302 concentration in whole blood following a single dose (75 mg/kg) in healthy C57BL/6 mice (n = 5 replicates at each time point)

Error bars represent ± std. dev.

Extended Data Fig. 16. Evaluation of acute (A/B) and chronic (C/D) V-9302 exposure on plasma glucose (A/C) and glutamine (B/D) in mice

Acute exposure consisted of a single dose (75 mg/kg) with metabolites assayed 4 h post-treatment. Chronic exposure analysis at the conclusion of 21 day treatment course (75 mg/kg daily), with metabolites assayed 4 h following the final V-9302 dose; n > 5 independent experiments. n.s. = not statistically significant. P values determined by Student’s t test.

Extended Data Fig. 17. Evaluation of V-9302 in vivo in HCC1806 cell line xenograft-bearing mice

(A) Tumor volumetric analysis of mice treated with vehicle or V-9302 (75 mg/kg per day) for 10 days. n = 5 mice per group. Treatment started 9 days post injection. P value at day 10 determined by Student’s t test. (B) Immunofluorescence analysis of tumor tissues harvested from vehicle- or V-9302-treated mice; LC3B and pAKT (Ser473) shown (pink). CD-31 positive vessels shown in green and nuclei in blue (DAPI). 20× magnification. P values determined by Student’s t test. (C) Effects of V-9302 treatment on pS6-positive cells and caspase 3-positive cells by immunohistochemistry. Quantitative analysis consisted of the mean counts of at least three representative fields from three vehicle and three V-9302-treated mice. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max. Error bars represent ± std. dev.

Extended Data Fig. 18. Evaluation of V-9302 in vivo in colo-205 cell line xenograft bearing mice

(A) Tumor volumetric analysis of mice treated with vehicle or V-9302 (75 mg/kg per day) for 10 days; n = 5 mice per group. Treatment started 6 days post injection. P value at day 10 determined by Student’s t test. (B) Immuno dot-blot assay of markers of cellular response to V-9302 in colo-205 cell line xenograft tumors; change in marker immunoreactivity relative to vehicle control. (C) Immnunofluorescence assay of LC3B-, pS6-, pAKT (Ser473)-, and BRDU-positive cells by immunofluorescence (pink). CD-31 positive vessels shown in green and nuclei in blue (DAPI). 20× magnification. Representative images shown. Quantitative analysis consisted of the mean counts of at least three representative fields from three vehicle and three V-9302 treated mice. P values determined by Student’s t test. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max. Error bars represent ± std. dev.

Extended Data Fig. 19

Longitudinal assessment of the mass of live athymic nude mice (n = 20 mice) treated daily with V-9302 (75 mg/kg per day) or vehicle over 21 days.

Extended Data Fig. 20. Representative photomicrographs of H&E stained liver sections from athymic nude mice chronically treated with 75 mg/kg per day V-9302 or vehicle over 21 days

Magnification 20× (A) and 40× (B) shown. An experienced GI pathologist (MKW) found no discernible difference between liver pathology in mice treated with V-9302 or vehicle (n = 5).

Figure 1. V-9302, an inhibitor of glutamine transport

(A) Chemical structure of V-9302. (B) Concentration-dependent inhibition of glutamine uptake in live HEK-293 cells; V-9302 (blue curve), GPNA (red curve); n = 3 independent experiments performed in triplicate. P < 0.001 at 10 µM by Student’s t test. Cellular glutamine accumulation normalized to vehicle control. Normalized amino acid uptake (relative to vehicle) in HEK-293 cells with V-9302 exposure at the IC₅₀ (10 µM, (C)) and 10X the IC₅₀ for glutamine inhibition (100 µM, (D)); n = 3 independent experiments. P < 0.001 by Student’s t test. Q=glutamine, Y=tyrosine, E=glutamic acid, D=aspartic acid, K=lysine, G=glycine, L=leucine. (E) Normalized uptake of ³H-labeled amino acids in HEK293 cells evaluated in the presence of increasing concentrations of V-9302; n = 3 independent experiments. Normalization relative to vehicle control. (F) Drug Affinity Responsive Target Stability (DARTS) assay visualized by immunoblot; tetracycline (TCN)-inducible ASCT2 HEK293 cells. ASCT2 is protected from proteolytic degradation by thermolysin (TLN) in the presence of increasing concentrations of V-9302 (veh = -, + = 50 µM, ++ = 100 µM, +++ = 200 µM. Uncropped gel images are provided in Supplementary Fig. 1. Error bars represent ± std. dev.

Figure 2. In silico modeling of V-9302 interactions with human ASCT2 (hASCT2)

(A) Homology model of hASCT2 (trimer shown) with V-9302 docked into the orthosteric binding site within the transmembrane region of the protein (extracellular membrane - red plane; intracellular membrane - blue plane). (B) Expanded view of residues proximal to V-9302 within the orthosteric binding site. Top scoring pose shown. (C) Overlay of V-9302 and ASCT2 substrate, glutamine, docked into the orthosteric binding site. (D) In silico alanine scan of the hASCT2 binding pocket. Positive values indicate alanine substitution interacts less favorably with V-9302 relative to the native residue. The total interface score is a weighted summation of the hydrogen bonding scores, repulsion penalties, solvation energies, and electrostatic potential. Glutamine and V-9302 were evaluated in a homology model of LAT1. (E) Ligand interaction diagram of glutamine or V-9302 in LAT1 visualized in the MOE molecular modeling and simulation package. Steric clash with the surrounding residues indicated in red (only seen with V-9302). (F) Docking scores for glutamine and V-9302 into LAT1; similar fit observed for glutamine in LAT1 and ASCT2, while V-9302 only fits the ASCT2 binding pocket. n = 100 top scoring poses per condition. P values determined by Student’s t test. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max.

Figure 3. In vitro Efficacy of V-9302

(A) A panel of 29 human cancer cell lines exposed to a single concentration of V-9302 (25 µM, 48 hrs); assay of ATP-dependent viability (CellTiter Glo). Select mutational status highlighted (green squares). Cell lines derived from lung cancer (white bars), breast cancer (pink bars), and colorectal cancer (blue bars) shown; n = 3 independent experiments. Specific cell lines representing a range of in vitro sensitivities prioritized for further evaluation in vivo indicated by arrows. Cell lines with P < 0.05 by Student’s t test relative to vehicle control below dotted line. (B) Direct comparison of V-9302 or CB-839 on the viability of human CRC cell lines. Drug incubated at concentrations shown for 48 hrs. Percent viability relative to vehicle control (MultiTox Glo assay); n = 3 independent experiments. P < 0.01 at 10 µM by Student’s t test. Error bars represent ± std. dev. *Estimated EC₅₀. Error bars represent ± std. dev.

Figure 4. Molecular determinants of ASCT2-antagonism in vitro

(A) Silencing ASCT2 (shRNA; HCC1806 cells; immunoblot shown) resulted in significantly attenuated pS6 and modestly decreased pERK. (B) V-9302 exposure (25 µM, 48 hrs) exhibited a similar inhibition profile to silencing ASCT2 with shRNA in HCC1806 cells. Uncropped gel images are provided in Supplementrary Fig. 2. (C) V-9302-dependent increase in oxidized glutathione (GSSG, left y-axis) and depletion of reduced glutathione (GSH, right y-axis) and (D) corresponding assay of reactive oxygen species (ROS); HCC1806 cells; n = 3 independent experiments. P < 0.01 by Student’s t test. (E) Effect of V-9302 exposure (25 µM, 48 hrs) on LC3B, a marker of autophagy, in HCC1806 cells. Immunofluorescence photomicrographs (left panels) showing cellular LC3B localization (pink fluorescence). Quantified numbers of LC3B-positive cells per field shown in right panel; n = 3 independent experiments. P < 0.01 by Student’s t test. Magnification 40×. (F) Effect of V-9302 exposure on optical redox-ratio ([FAD]/[NAD(P)H]) in HCC1806 cells. Representative photomicrographs (left) and quantification (right); concentrations shown, 48 hr exposure; n = 3 independent experiments. P values determined by Student’s t test. Magnification 40×. (G) Effects of V-9302 exposure (25 µM, 96 hrs) on the viability of two human colorectal cancer organoids (A007 - BRAF^V600E; A008 - KRAS^G12V;p53^R248Q;PTEN^L140Y.) Representative brightfield photomicrographs (left) and quantified organoid viability (right); n = 3 independent experiments. P values determined by Student’s t test. Error bars represent ± std. dev. For box plots, center line is plotted at the median; the box spans from the first quartile to the third quartile; whiskers represent min to max.

Figure 5. Evaluation of V-9302 in vivo

(A) Pharmacodynamic [¹⁸F]-4F-Gln PET imaging prior to and 4 h following a single administration of V-9302 (75 mg/kg) in HCC1806 cell line xenograft-bearing mice (arrows indicate xenograft tumor on right flank*). (B) Mean time activity curves (TACs) from tumor regions of interest (n = 4 measurements per condition); data prior to and 4 hrs following V-9302 administration. (C) P values determined by Student’s t test. Quantified tracer accumulation in xenograft tumors, muscle, and liver (n = 4 measurements per condition). Volumetric analysis over 21 day treatment regimen (Vehicle or V-9302; 75 mg/kg, daily) of HCT-116 (D) and HT29 (F) cell line xenografts propagated in athymic nude mice (n = 10 mice per group). Treatment started 12 days post tumor injection for HCT-116 and 4 days post injection for HT29. P values on day 21 determined by Student’s t test. Immunohistochemistry for pS6 and caspase 3 in vehicle-treated or V-9302-treated HCT-116 (E) and HT29 (G) xenografts. Representative photomicrographs and quantitation shown; magnification 20×. P values determined by Student’s t test. (H) Volumetric analysis over 31 day treatment regimen (Vehicle or V-9302; 75 mg/kg, daily) on athymic nude mice bearing patient-derived xenograft tumors (PDX A 008, F3 generation, treatment started 28 days post implantation, KRAS^G12V;p53^R248Q;PTEN^L140Y ; n = 10 mice per group) P value on day 31 determined by Student’s t test. (I) Photographs of A 008 PDX-bearing mice treated with V-9302 or vehicle; day 16 of 31. (Error bars represent ± std. dev. *Central photopenia observed.

Figure 6. Summary of cancer cell programs modulated by V-9302

(A) Global metabolomic analysis of HT-29 cell-line xenograft tumor-bearing mice treated with V-9302 or vehicle (n = 5 per condition). Individual statistically significant (p < 0.05 as determined by Welch’s two-sample t test) metabolites spanning seven distinct metabolic families highlighted. Select metabolites involved in glutamine-centric biological processes indicated (see text for details). (B) ASCT2 blockade with V-9302 results in attenuated cancer cell Growth and Proliferation, Cell Death, and Oxidative Stress. Arrows indicate V-9302-induced phenotypes relative to baseline homeostasis. *Additional substrates transported by ASCT2 include Alanine, Serine, Cysteine, Threonine, Leucine, and Asparagine.