Glutathione determines chronic myeloid leukemia vulnerability to an inhibitor of CMPK and TMPK

Abstract

Bcr-Abl transformation leads to chronic myeloid leukemia (CML). The acquirement of T315I mutation causes tyrosine kinase inhibitors (TKI) resistance. This study develops a compound, JMF4073, inhibiting thymidylate (TMP) and cytidylate (CMP) kinases, aiming for a new therapy against TKI-resistant CML. In vitro and in vivo treatment of JMF4073 eliminates WT-Bcr-Abl-32D CML cells. However, T315I-Bcr-Abl-32D cells are less vulnerable to JMF4073. Evidence is presented that ATF4-mediated upregulation of GSH causes T315I-Bcr-Abl-32D cells to be less sensitive to JMF4073. Reducing GSH biosynthesis generates replication stress in T315I-Bcr-Abl-32D cells that require dTTP/dCTP synthesis for survival, thus enabling JMF4073 susceptibility. It further shows that the levels of ATF4 and GSH in several human CML blast-crisis cell lines are inversely correlated with JMF4073 sensitivity, and the combinatory treatment of JMF4073 with GSH reducing agent leads to synthetic lethality in these CML blast-crisis lines. Altogether, the investigation indicates an alternative option in CML therapy.

Fig. 1. JMF4073 inhibits TMPK and CMPK.

a Schematic view of the synthesis of dTTP and dCTP. TMPK and CMPK mediate the synthesis of dTDP and dCDP, respectively, which subsequently are converted to dTTP and dCTP by nucleoside diphosphate kinase (NDPK). Thymidylate synthase (TS) converts dUMP to dTMP, which is inhibited by 5-FU. Ribonucleotide reductase (RNR) mediates the de novo synthesis of dCDP, dADP, dUDP, and dGDP. Salvage synthesis of dTMP and dCMP is mediated by thymidine kinase and deoxycytidine kinase. b The chemical structures, clogP, and IC50 of YMU1, JMF2977, and JMF4073. The clogP values were calculated by Chemsketch. The IC50 values against hTMPK and hCMPK were measured by NADH-coupled TMPK assay. c The measurement of dTTP and dCTP pools in untransformed 32D and HEK-293T cells after incubation with JMF4073 (10 μM) for 6 h. Data are represented as means ± S.D., n = 3 biological replicates. Asterisks denote *p < 0.05, **p < 0.01, or ***p < 0.001, from unpaired two-tailed Student’s t-test.

Fig. 2. JMF4073 eliminates WT-Bcr-Abl-transformed myeloid cells in vitro and in vivo.

a The comparison of overall pTyr in unstranformed, WT, and T315I Bcr-Abl -transformed 32D myeloid progenitor cells, After the treatment of imatinib (2 μM) and ponatinib (2 μM) for 4 h, cell lysates were analyzed by Western blot using phosphotyrosine (pY99) antibody. Cells were treated with JMF4073 at the indicated concentrations for (b) viability assays and (c) Western blot of tyrosine phosphorylation. d Cells were treated with 5-FU, and ara-C at the indicated concentrations for viability assays. Data are represented as means ± S.D., n = 3 biological replicates. e C3H/HeNCrNarl mice were intravenously injected with 5 × 10⁵ cells of WT-Bcr-Abl-32D cell. After 48 h of transplantation, mice were treated with vehicle (n = 7), or JMF4073 (5 mg/kg/time, n = 7) by intraperitoneal injection at 24 h intervals for 14 days. f C3H/HeNCrNarl mice were intravenously injected with 1 × 10⁶ cells of T315I-Bcr-Abl-32D cell. After 7 days of transplantation, mice were treated with vehicle (n = 7), or JMF4073 (5 mg/kg/time, n = 7) by intraperitoneal injection at 24 h intervals for 14 days. The Kaplan–Meier plot shows the survival of mice with treatment as indicated. Asterisks denote *p < 0.05, **p < 0.01, ***p < 0.001, or ****p < 0.0001 from unpaired two-tailed Student’s t-test or Log-rank (Mantel–Cox) test. The mice images are hand-drawn using the free Samsung PENUP software.

Fig. 3. The association of JMF4073 sensitivity with replication stress, low dTTP, and GSH pools.

Untransformed, WT-, and T315I-Bcr-Abl transformed 32D cells were subjected to (a) the measurement of four dNTP levels, b DNA fiber analysis to determine the replication fork speed in DNA replication speed (n = 200 fibers for each cell line) (Left). The workflow of DNA fiber labeling (Right). The representative labeled fibers are shown below. Replication speed in kb/min was calculated by the measured length of IdU (green) linking to CldU (red) in μm with a conversion factor of 0.34 μm/kb divided by the duration of the labeling pulse (Scale bar = 10 μm). The gray triangle indicates the boundary between red fluorescence and green fluorescence. c The measurement of intracellular ROS by the fluorescence intensity of CellROX Green staining using flow cytometric analysis. Data are presented as mean intensity relative to untransformed 32D cells. d, e The measurement of GSSG and GSH. Data are presented relative to untransformed 32D cells. f WT- and T315I-Bcr-Abl-32D cells after incubation with 4 mM U-¹³C glutamine medium for 2 h. Schematic illustration of the metabolic path U-¹³C glutamine incorporation into the GSH (left). Gray and white circles indicate ¹³C-label, and unlabeled ¹²C, respectively. The traced isotopologue abundance of GSH is shown (right). Data are represented as means ± S.D., n = 3 biological replicates. Asterisks denote *p < 0.05, **p < 0.01, ***p < 0.001, or ****p < 0.0001 from unpaired two-tailed Student’s t-test.

Fig. 4. Upregulation of ATF4-mediated network increases dTTP pool and GSH biosynthesis in T315I-Bcr-Abl 32D cells.

a The heat map of RNA-sequencing data for genes significantly altered in dTTP synthesis and glutathione metabolism, as determined by Ingenuity Pathway Analysis (IPA), sorted based on log2 fold changes >1 or < -1 with adjusted p < 0.05. b A schematic diagram illustrates genes linked to dTTP synthesis and GSH metabolism. Genes upregulated in T315I-Bcr-Abl-32D cells are shown in orange and downregulated in green. c The RT-qPCR analysis of genes indicated in (a). Data were normalized using GAPDH, and the expression levels were expressed relative to those in untransformed 32D cells. d Immunoblotting of ATF4 in untransformed, WT-, and T315I-Bcr-Abl transformed 32D cells. T315I-Bcr-Abl-32D cells infected with shLacZ and shATF4 lentivirus for 48 h were subjected to (e) immunoblotting and (f) RT-qPCR analysis to assess PSAT1 and SHMT2 levels. g RT-qPCR analysis of SHMT2 and the quantitation of dTTP level in T315I-Bcr-Abl-32D cells after infection with shLacZ and shSHMT2 lentivirus. h, i T315I-Bcr-Abl-32D cells after infection with shLacZ and shATF4 lentivirus for 48 h were subjected to (h) GSH level and (i) the DNA fiber analysis (n = 200, scale bar = 10 μm). The gray triangle indicates the boundary between red fluorescence and green fluorescence. Data are presented as means ± S.D, from three independent experiments. Asterisks denote **p < 0.01, ***p < 0.001, or ****p < 0.0001 from unpaired two-tailed Student’s t-test.

Fig. 5. GSH biosynthesis but not dTTP determine JMF4073 susceptibility.

T315I-Bcr-Abl-32D cells after infection with shLacZ and shSHMT2 lentivirus were examined for (a) DNA fiber analysis as described in the legend to Fig. 3b (scale bar = 10 μm), and (b) JMF4073 sensitivity analysis by viability assays. T315I-Bcr-Abl-32D cells after treatment with BSO (12.5 μM) for 16 h were subjected to (c) the measurement of GSH level, (d) DNA fiber analysis, and (e) JMF4073 sensitivity. WT-Bcr-Abl-32D cells after treatment with N-acetyl cysteine (NAC, 1 mM) for 16 h for (f) the measurement of GSH level, (g) DNA fiber analysis, and (h) JMF4073 sensitivity. The viability of cells treated with NAC in combination with JMF4073 (0.5 μM) for 72 h was determined. Data are presented as means ± S.D., n = 3 biological replicates. Asterisks denote *p < 0.05, **p < 0.01, or ****p < 0.0001 from unpaired two-tailed Student’s t-test.

Fig. 6. The combinatory treatment of JMF4073 and UK-5099 depletes dTTP/dCTP pools and eliminates T315I Bcr-Abl-32D cells in vivo.

T315I-Bcr-Abl-transformed and untransformed 32D cells were treated with UK-5099 (25 μM) for 16 h. These cells were subjected to (a) GSH measurement, and (b) DNA fiber assay. c JMF4073 sensitivity assay. d dTTP and dCTP pools measurement of T315I-Bcr-Abl-32D cells after treatment with vehicle, UK-5099 alone, and the combination of UK-5099 and JMF4073. e, f C3H/HeNCrNarl mice were intravenously injected with 1 × 10⁶ cells of T315I-Bcr-Abl-32D cell. After 7 days of transplantation, mice were treated vehicle (n = 7), UK-5099 (10 mg/kg/time, n = 7) (e) or JMF4073 (5 mg/kg/time) + UK-5099 (10 mg/kg/time) (n = 7) (f) by intraperitoneal injection at 24 h interval for 14 days. The Kaplan–Meier plot shows the survival of mice with treatment as indicated. g flow cytometry analysis of peripheral blood (n = 5) and (h) Spleen weights (n = 3) from T315I-Bcr-Abl-32D/EGFP+ bearing mice treated with vehicle, UK-5099, JMF4073, or JMF4073 combined with UK-5099. The number of EGFP+ cells in mice with T315I-Bcr-Abl-induced CML was determined on day 30 after transplantation. Data are represented as means ± S.D., n > 3 biological replicates. Asterisks denote *p < 0.05 from unpaired two-tailed Student’s t-test or Log-rank (Mantel–Cox) (f, g) test. The mice images are hand-drawn using the free Samsung PENUP software.

Fig. 7. Synthetic lethality by GSH blocker and JMF4073 in human CML-BC cells.

K562, TCCS, and KOPM28 were subjected to (a) GSH measurement, (b) immunoblotting of ATF4, and (c) DNA fiber analysis. d K562, TCCS and KOPM28 cells were treated with JMF4073. After 3 days, cells were analyzed by viability assays. GSH level in K562, TCCS, and KOPM28 cells after treated with (e) UK-5099 (50 μM) or (f) erastin (0.5 μM) for 16 h for GSH measurement. g After treatment with erastin (0.5 μM) for 16 h, DNA fiber analyses were performed in TCCS (n = 200) and KOPM28 cells (n = 200). h JMF4073 sensitivity in TCCS and KOPM28 cells after pretreatment with erastin (0.5 μM) for 16 h. All data are presented as means ± S.D. from 3 independent experiments. Asterisks denote **p < 0.01, or ****p < 0.0001 from unpaired two-tailed Student’s t test.