The advent of molecularly targeted therapeutic agents has opened a new era in cancer therapy. However, many tumors rely on nondruggable cancer-driving lesions. In addition, long-lasting clinical benefits from single-agent therapies rarely occur, as most of the tumors acquire resistance over time. The identification of targeted combination regimens interfering with signaling through oncogenically rewired pathways provides a promising approach to enhance efficacy of single-agent–targeted treatments. Moreover, combination drug therapies might overcome the emergence of drug resistance. Here, we performed a focused flow cytometry–based drug synergy screen and identified a novel synergistic interaction between GLUT1-mediated glucose transport and the cell-cycle checkpoint kinases ATR and CHK1. Combined inhibition of CHK1/GLUT1 or ATR/GLUT1 robustly induced apoptosis, particularly in RAS-mutant cancer cells. Mechanistically, combined inhibition of ATR/CHK1 and GLUT1 arrested sensitive cells in S-phase and led to the accumulation of genotoxic damage, particularly in S-phase. In vivo, simultaneous inhibition of ATR and GLUT1 significantly reduced tumor volume gain in an autochthonous mouse model of KrasG12D-driven soft tissue sarcoma. Taken together, these findings pave the way for combined inhibition of GLUT1 and ATR/CHK1 as a therapeutic approach for KRAS-driven cancers.

Significance:

Dual targeting of the DNA damage response and glucose transport synergistically induces apoptosis in KRAS-mutant cancer, suggesting this combination treatment for clinical validation in KRAS-stratified tumor patients.

Molecularly targeted small molecule inhibitors revolutionized cancer therapy throughout the last decade. In tumors that are driven by a single actionable genomic alteration, such as EGFR mutations or ALK and ROS1 rearrangements, these single-agent therapies often display remarkable clinical activity (1–3). However, many tumors rely on nondruggable cancer-driving lesions, such as somatic copy number gains of the MYC locus, disabling mutations of the tumor suppressor genes BRCA1 and -2 or activating mutations in KRAS (4–8). These genomic aberrations are typically associated with a massive rewiring of intracellular signaling circuits, affecting more than one downstream pathway. A well-studied example in this regard are oncogenic KRAS mutations, which can be detected in approximately 30% of all human tumors (9). Oncogenic KRAS mutations induce tonic signaling through the RAF/MEK/ERK/MAPK pathway, the PI3K/PDK1/AKT signaling cascade, and the RAL guanine nucleotide exchange factor pathway as their main downstream effectors (10–12). Emerging evidence suggests that KRAS-mutant tumor cells display alterations of glucose metabolism, leading to increased glucose uptake and turnover, channeling of glucose-derived metabolites into the pentose phosphate pathway, activation of hexosamine biosynthesis, as well as glutathione biosynthesis (13–17). Furthermore, copy-number gains of oncogenic KrasG12D were recently shown to drive additional metabolic alterations, such as shunting of glucose-derived metabolites into the tricarboxylic acid cycle and glutathione biosynthesis (13).

In addition to this metabolic rewiring, KRAS-mutant cells and tumors were previously shown to display signs of oncogene-induced genotoxic damage, likely as a result of replicative stress and elevated reactive oxygen species, which ultimately leads to constitutively active DNA damage response (DDR) signaling (18–24). Altogether, these rewired signaling cascades may offer opportunities for targeted therapeutic intervention strategies beyond classical MAPK, PI3K, and RAL signaling. However, despite numerous recent efforts to develop single-agent regimens to target KRAS-mutant tumors, no monotherapeutic approach has been clinically validated for this cancer genotype, to date. Thus, KRAS-driven tumors remain a clinical challenge that is difficult to treat with single-agent therapies or standard chemotherapy (6, 25).

Rational synergistic drug combination therapies might represent a viable strategy for therapeutic intervention particularly in tumors with nondruggable driver mutations. In this study, we focus on dual targeting of the DDR and glucose transport in KRAS-mutant cells, as well as an autochthonous mouse model of Kras-driven soft tissue sarcoma. In line with rewired glucose metabolism and a constitutively active DDR in KRAS-mutant cells, we find that simultaneous inhibition of the glucose transporter GLUT1 and the ATR/CHK1 cell-cycle checkpoint signaling axis synergistically induces apoptosis in KRAS-mutant cancer, in vitro and in vivo. Mechanistically, we show that combined inhibition of ATR/CHK1 and GLUT1 leads to the accumulation of genotoxic damage in sensitive cell lines, particularly in S-phase.

Cell lines and cell culture

Human cell lines were purchased from the ATCC (H1975: ATCC catalog no. CRL-5908, Calu1: ATCC catalog no. HTB-54, H647: ATCC catalog no. CRL-5834) or the German Resource Centre for Biological Material (DMSZ; MelJuso: catalog no. ACC-74, A498: catalog no. ACC-55, HCC44: catalog no. ACC-534, MFE296: catalog no. ACC-419, 639V: catalog no. ACC-413, A549: catalog no. ACC-107) and tested negatively for Mycoplasma contamination upon thawing using a PCR Mycoplasma Test Kit I/C (PromoKine catalog no. PK-CA91-1024). Experiments were performed between passage 4–15 upon purchase. All other human cell lines were provided by Roman Thomas (Department of Translational Genomics, University of Cologne, Cologne, Germany). If applicable, selected mutations were annotated using the Cancer Cell Line Encyclopedia (CCLE) database (26). Murine lung adenocarcinoma cell lines (KPL) were generated as described previously (27). Murine sarcoma cell lines (KPS) were a gift from David G. Kirsch (Duke University, Durham, NC). Cell lines were cultured in growth medium (10% FCS, 1% penicillin/streptomycin) and passaged using trypsin. See Supplementary Table S1 for entities, growth conditions, RRIDs, and primers used for genotyping.

Cell viability measurement

Cells were plated into 96-well plates (500 to 7,000 cells per well) in 100 μL complete medium and treated using a TECAN D300e digital dispenser (HP). After 72 hours of treatment, relative cell viability was assessed by measuring the ATP content in each well. Therefore, CellTiter-Glo Reagent (Promega, catalog no. G7573) was added 1:1. Upon 10 minutes of incubation, luminescence intensity was measured with a plate reader (Tecan Infinite M1000 Pro) and normalized to intensities of control wells.

Quantification of apoptosis by flow cytometry

Cell lines were seeded into 6-well plates at 40%–60% confluence. On the next day, medium was replaced with growth medium containing either single inhibitors or combination therapies as listed in Supplementary Table S1. Forty-eight hours later, cells were trypsinized, washed with ice-cold PBS, and incubated in antibody binding buffer [Annexin-V (BD Biosciences catalog no. 556420) and propidium iodide (PI; Carl Roth, catalog no. CN74, 5 μg/mL) in 2.5 mmol/L CaCl2, 10 mmol/L HEPES (pH 7.4), 140 mmol/L NaCl, 20% accutase solution (Sigma Aldrich, catalog no. A6964), 70% PBS]. Annexin V/PI staining was detected by flow cytometry (Beckman Coulter, Gallios Flow Cytometer). The fraction of apoptotic cells was quantified as the Annexin V, PI, and double positive populations using the Kaluza analysis software (Beckman Coulter). Treated samples were normalized to untreated control samples.

Xenograft model

All animal procedures of this study were approved by the local authorities and animal protection committee (LANUV NRW, 84-02.04.2016.A300). For each tumor, 3.5 × 106 (A549) or 3.0 × 106 cells (HCC44) were resuspended in 100 μL serum-free DMEM and injected subcutaneously into 6- to 7-week-old female nude mice (Rj:NMRI-Foxn1nu/nu, Janvier Labs). Treatment was started upon formation of palpable subcutaneous tumors. Animals were randomly allocated to one of the following treatment regimens for 14 consecutive days: Control group (vehicle solution), single PF477736 (15 mg/kg, i.p.), single CP724714 (50 mg/kg, orally), single WZB117 (30 mg/kg, i.p.), or the combination therapies consisting of PF477736/WZB117 (15 mg/kg or 30 mg/kg, respectively) or PF477736/CP724714 (15 mg/kg or 50 mg/kg, respectively). Compounds were obtained from Selleckchem, dissolved in DMSO, and kept at −80°C. Prior to use, components of vehicle solution [5% DMSO, 20% PEG400 (Carl Roth, catalog no. 0144.1), 5% Tween 80 (Carl Roth, catalog no. 4859.1), 65% PBS] were sequentially added. Tumor size was monitored every other day by measurement of perpendicular diameters using an external caliper and calculated using the hemi-ellipsoid formula:

formula

Gene set enrichment analysis

Gene-centric RMA-normalized mRNA expression data v. 2012-09-29 was downloaded from the CCLE (https://portals.broadinstitute.org/ccle, accessed January 31, 2017; ref. 26). Gene set enrichment analysis (GSEA) was performed with GSEA v2.2.3 (28, 29) using the Hallmark and Gene Ontology gene sets from the Molecular Signatures Database v6.0 (28) and the synergy scores for the 24 cell lines in Fig. 1C of the CHK1/GLUT1 and ATR/GLUT1 combination as phenotype. For the comparison of the mean MTORC1 expression levels, the HALLMARK_MTORC1_SIGNALING gene set was downloaded from the Molecular Signatures Database v6.0 and used to select the corresponding genes in the CCLE expression data. The cell lines were divided into three groups: CCLE cell lines not included in our study, cell lines that show synergy (≥15) after inhibition of both ATR/GLUT1 and CHK1/GLUT1, and cell lines that do not show synergy in both settings.

Figure 1.

A combination screen identifies a synergistic interaction between ATR/CHK1 and GLUT1 inhibition. A, Schematic representation of the synergy screening experiments. Cells were exposed to single compounds or combined compound treatment for 48 hours, harvested, and incubated with AnnexinV-FITC and PI. Apoptotic cells were quantified using flow cytometry. The expected apoptotic effect was calculated by the Bliss independence model using observed apoptotic effects for each single compound. The synergy score was calculated as the difference between the observed and expected apoptotic effect. B, Pie chart showing the representation of different tumor entities covered by the 24 cell lines used for the synergy screen. C, Heatmap showing synergy scores in percent of 23 compound combinations across 24 cancer cell lines. Combinations were arranged by their synergy frequency. Cell lines were arranged by their genomic background for mutations in KRAS, NRAS, EGFR, and TP53 genes. For concentrations of single compounds and apoptotic effects, see Supplementary Tables S1 and S4 and Supplementary Fig. S3. D, GSEA plot showing the significant positive correlation between the synergy score (combined CHK1/GLUT1 inhibition) of the cell lines in C and the expression of genes included in the HALLMARK_MTORC1_SIGNALING gene set. The green line shows the running enrichment score; the black bars show the genes included in the signature, which are concentrated on the left of the ranked list of all considered genes (gray waterfall plot). See Supplementary Fig. S6A for the GSEA plot of combined ATR/GLUT1 inhibition.

Figure 1.

A combination screen identifies a synergistic interaction between ATR/CHK1 and GLUT1 inhibition. A, Schematic representation of the synergy screening experiments. Cells were exposed to single compounds or combined compound treatment for 48 hours, harvested, and incubated with AnnexinV-FITC and PI. Apoptotic cells were quantified using flow cytometry. The expected apoptotic effect was calculated by the Bliss independence model using observed apoptotic effects for each single compound. The synergy score was calculated as the difference between the observed and expected apoptotic effect. B, Pie chart showing the representation of different tumor entities covered by the 24 cell lines used for the synergy screen. C, Heatmap showing synergy scores in percent of 23 compound combinations across 24 cancer cell lines. Combinations were arranged by their synergy frequency. Cell lines were arranged by their genomic background for mutations in KRAS, NRAS, EGFR, and TP53 genes. For concentrations of single compounds and apoptotic effects, see Supplementary Tables S1 and S4 and Supplementary Fig. S3. D, GSEA plot showing the significant positive correlation between the synergy score (combined CHK1/GLUT1 inhibition) of the cell lines in C and the expression of genes included in the HALLMARK_MTORC1_SIGNALING gene set. The green line shows the running enrichment score; the black bars show the genes included in the signature, which are concentrated on the left of the ranked list of all considered genes (gray waterfall plot). See Supplementary Fig. S6A for the GSEA plot of combined ATR/GLUT1 inhibition.

Close modal

Crystal violet stainings

Cells were seeded into 6-well plates at 20% confluence and incubated overnight. Cells were exposed to treatment for 72 hours and subsequently incubated in normal medium until 100% confluence was reached in control wells. Cells were fixed with methanol, stained with 0.5% crystal violet (Sigma Aldrich, catalog no. HT90132), and subjected to microscopy on a Carl Zeiss Stemi 2000-C Stereo microscope equipped with a CCD camera (Zeiss). Three images per well were taken and the percentage of area covered by crystal violet–stained cell colonies was determined (ImageJ 1.50i, Fiji).

siRNA-mediated knockdown of GLUT1, ATR, and CHK1

Cells were transfected with ON-TARGETplus siRNA pools targeting either GLUT1 (SLC2A1), ATR or CHK1, or a nontargeting control (Dharmacon, catalog no. L-007509-02-0010, #L-003202-00-0010, #L-003255-00-0010, #D-001810-10-05) using Lipofectamine RNAiMAX (Thermo Fisher Scientific, catalog no. 13778150). Knockdown efficacy was determined by Western blotting and qRT-PCR. For viability assays, cells were seeded into 96-well plates and transfected on the next day. After 48 hours, compound was added using a TECAN D300e digital dispenser (HP). Cell viability was determined 96 hours after transfection by measuring the ATP content (CellTiter-Glo assay).

Cell-cycle analysis and DNA damage

Cells were seeded at 40%–60% confluence in 6-cm dishes. Upon 48 hours of treatment, cells were harvested by trypsination, washed twice in PBS, and fixed in methanol (−20°C). Cells were then permeabilized (0.25% TritonX-100 in PBS), blocked (1% BSA in PBS), and stained with antibodies against cleaved caspase-3 (CC3,) and γH2AX (Ser-139) for 4 hours. Upon incubation with secondary antibodies for 1 hour and washing, cells were incubated with RNAse A (1 μL per sample, Sigma catalog no. R6148) and PI (Carl Roth, catalog no. CN74, 5 μg/mL) for 20 minutes. See Supplementary Table S1 for catalog numbers of antibodies and RRIDs. Samples were analyzed by flow cytometry. We used the Kaluza Analysis Software to quantify the distribution of the cellular DNA content as well as the population of γH2AX and CC3-positive cells. Cell doublets were excluded from analysis by plotting the PI signal integral against the PI signal peak.

Immunofluorescence

Cells were seeded into 96-well imaging plates. Upon 24 hours of treatment, cells were fixed in 4% formaldehyde. For permeabilization and extraction of nonchromatin-bound proteins, cells were incubated in cytoskeleton buffer [10 mmol/L PIPES (pH 6.8), 100 mmol/L NaCl, 300 mmol/L sucrose, 3 mmol/L MgCl2, 1 mmol/L EGTA, 0.5% Triton X-100) for 10 minutes, followed by incubation in cytoskeleton stripping buffer [10 mmol/L Tris-HCl (pH 7.4), 10 mmol/L NaCl, 3 mmol/L MgCl2, 2% Tween 20, 0.5% sodium deoxycholate] for 10 minutes. After 30-minute blocking [5% BSA, 2% normal goat serum, 0.01% Triton X-100 in PBS (PBS-B)], cells were incubated with γH2AX and RPA70 and secondary antibodies as listed in Supplementary Table S1. After three final washes, wells were filled with PBS and stored at 4°C until scanning.

Autochthonous model

Soft tissue sarcomas were induced by intramuscular injection of adenovirus expressing Cre recombinase (Ad-Cre) in the extremities of KrasLSL.G12D/wt;Trp53fl/fl (KP) mice as described previously (30). Ad-Cre virus (2–5 × 107 PFU/mL, UI Viral Vector Core Web, catalog no. VVC-U of Iowa-5) was applied as a mixture of 6.6% Ad-Cre virus, 92.8% Opti-MEM, and 0.005% 2 mol/L CaCl2. Following virus injection, mice were palpated every day to detect formation of tumor nodules as soon as possible, while the first MRI scan was procured 30 days after virus injection. Subsequently, MR imaging was performed once a week until tumor formation was detected. Mice were randomly allocated to one of the following treatment regimens for 7 consecutive days: Vehicle, single VE822 (40 mg/kg), single WZB117 (50 mg/kg), combination treatment (VE822: 40 mg/kg and WZB117: 50 mg/kg). WZB117 (Merck Millipore, catalog no. 400036, dissolved in DMSO/PBS) was administered daily by intraperitoneal injection. VE822 (Selleckchem, dissolved in 70% PBS, 30% polyethylene glycol, and 0.5% Tween) was administered daily by oral gavage.

Combined inhibition of the ATR/CHK1 axis and GLUT1 displays synergistic toxicity

To search for compound combinations that display synergistic cytotoxicity, we performed a focused cell line–based synergy screen. To assess cytotoxicity, we employed flow cytometry–based apoptosis measurements, using Annexin V/PI colabeling (Fig. 1A). In essence, we treated 24 distinct cancer cell lines derived from various entities, including non–small cell lung cancer (Calu1, A549, H1299, H1792, H1975, H2087, H647, H661, HCC15, HCC44), small-cell lung cancer (SBC5), colorectal cancer (Colo320, HCT116, HCT15), skin cancer (SKMEL2, MelJuso), endometrial cancer (HEC1A, MFE296), renal cell carcinoma (A498, Caki2), bladder cancer (639V), pancreatic ductal adenocarcinoma (Panc1), salivary gland cancer (A253), and ovarian carcinoma (SKOV3; Fig. 1B), with a series of 18 distinct drugs. These compounds included inhibitors targeting BCL2 (ABT199), EGFR/HER2 (afatinib), MEK1 (AZD6244), the proximal DDR kinase ATR (VE822), as well as its downstream effector kinase CHK1 (PF477736), FGFR (BGJ398), SRC/ABL (bosutinib), PDK1 (BX912), GSK3 (CHIR99021), HER2/ERBB2 (CP724714), AMPK (dorsomorphine), topoisomerase-II (etoposide), ALK (LDK378), WEE1 (MK1775), PARP1/2 (olaparib), IGF1R (OSI906), the multi-tyrosine kinase inhibitor sorafenib and the glucose transporter GLUT1 (WZB117). We determined the IC50 for each compound in a series of preparatory experiments (Supplementary Fig. S1; Supplementary Table S2). On the basis of these profiles, we determined three representative concentrations for each compound, that reflected the low-activity spectrum. Using these drug concentrations, we performed a prescreen (4 × 4 drug concentration matrix), assessing cell viability through ATP measurements on a panel consisting of eight distinct cell lines (HCC44, 639V, A549, Calu1, MelJuso, MFE296, H1975, and A498; Supplementary Fig. S2; Supplementary Table S3). The final drug concentrations used in our apoptosis screen were determined on the basis of drug synergy observed in our prescreen (Fig. 1C; Supplementary Fig. S2). To benchmark our results, we purposefully included four drug combinations that were previously reported to display synergistic cytotoxicity. These benchmarks consisted of inhibitors targeting WEE1/CHK1, WEE1/SRC, BCL2/CHK1, as well as the multi-tyrosine kinase inhibitor sorafenib plus the MEK inhibitor AZD6244 (31–39).

In our apoptosis screen, we specifically assessed the synergistic cytotoxic potential of 23 distinct compound combinations depicted in Fig. 1C. Drug synergy was determined using the Bliss independence algorithm, which assumes that the actions of the individual compounds under investigation are similar and independent (40). In essence, a synergy score >0 formally indicates drug synergy. For the data presented in this article, we defined a score of >15 as indicative of a synergistic compound interaction. This restrictive synergy calling strategy was employed to reduce the probability of false positives. As shown in Fig. 1C, Supplementary Fig. S3, and Supplementary Table S4, the majority of drug combinations displayed no, or only mild synergistic cytotoxicity. However, we found eight distinct drug combinations that displayed a robust synergistic interaction in at least nine distinct cell lines. These were VE822/WZB117, WZB117/PF477736, CP724714/PF477736, MK1775/PF477736, MK1775/bosutinib, LDK378/BGJ398, LDK378/afatinib, and BX912/LDK378 (Fig. 1C). Importantly, the previously described synergistic combinations consisting of inhibitors targeting WEE1/CHK1 and WEE1/SRC also scored as synergistic in our screen (Fig. 1C; refs. 31–37). Of note, the combination of sorafenib plus the MEK inhibitor AZD6244 and the combined inhibition of BCL2/CHK1 also displayed synergy in our screen, albeit only in six and five distinct cell lines, respectively (Fig. 1C).

Interestingly, we could detect robust synergistic cytotoxicity in cell lines treated with inhibitors targeting CHK1/HER2 and ALK/EGFR. These two drug combinations have been reported to be effective in HER2-positive gastric cancer (CHK1/HER2) and ALK-rearranged lung cancer (ALK/EGFR); however, they have not been explored for synergism to date (41–43).

Overall, we identified eight drug combinations with a substantial synergistic potential for cytotoxicity in at least nine distinct cancer cell lines (Fig. 1C).

To assess the validity of our screening results, we reassessed three combinations by further in vitro and in vivo experiments (Supplementary Figs. S4A–S4C and S5A–S5C). We specifically verified drug synergy of the combination regimens consisting of VE822/WZB117, WZB117/PF477736 (Supplementary Fig. S4A) and CP724714/PF477736 (Supplementary Fig. S5A) on an extended 6 × 6 dose concentration matrix by intracellular ATP measurements, as a surrogate for viability. These experiments confirmed the synergistic drug interactions between these compounds (Supplementary Figs. S4A and S5A). Moreover, we also performed a series of in vivo xenograft experiments to further validate drug synergy between WZB117/PF477736 and CP724714/PF477736. For this purpose, we subcutaneously injected A549 and HCC44 cells into Rj:NMRI-Foxn1nu/nu immunocompromised mice. Upon tumor formation, animals were treated with vehicle solution, single-agent PF477736, WZB117, or CP724714, as well as drug combinations consisting of WZB117/PF477736 and CP724714/PF477736 for 2 weeks. As shown in Supplementary Figures S4B and S5B single-agent treatment with PF477736, WZB117, or CP724714 resulted in continuous tumor volume growth. In contrast, and fully in line with the synergistic drug interactions that we had observed in vitro, combined treatment with CP724714/PF477736 resulted in tumor volume stabilization in A549- and HCC44-derived xenograft tumors. The combination therapy consisting of WZB117 and PF477736 induced a significant volume shrinkage of A549-derived xenograft tumors and tumor volume stabilization of HCC44-derived xenograft tumors (Supplementary Figs. S4B and S5B). Of note, both single-agent treatments, as well as the combination regimens were well-tolerated by the animals and did not have a significant effect on body weight of the mice throughout the 14-day treatment course (Supplementary Figs. S4C and S5C).

The combinations consisting of the GLUT1 inhibitor WZB117 with the proximal DDR kinase ATR or its downstream target CHK1 emerged as highly synergistic from our screen. Particularly, the observation that different components of the same DDR signaling cascade, namely ATR/CHK1, displayed drug synergy with the GLUT1 inhibitor WZB117 cross-validated a potential synergy between compounds targeting glucose transport and DDR signaling. We therefore performed a series of experiments to further characterize these synergistic drug interactions. To this end, we initially asked whether specific genomic or transcriptomic aberrations in our cell line panel were enriched in the cell lines displaying the strongest synergy between compounds targeting the ATR/CHK1 axis and GLUT1. We specifically interrogated the mutational profile of our cell line panel, using the CCLE dataset (26). This analysis revealed that oncogenic mutations in KRAS and NRAS were highly prevalent among the cell lines that displayed synergy between ATR/CHK1 and GLUT1 inhibitors, although this enrichment in RAS pathway–mutant cells was not statistically significant, likely due to the overall high frequency of cells harboring RAS pathway alterations in our cell line panel (Fig. 1C). We next employed GSEA (28, 29) to ask whether the synergistic interaction between inhibitors targeting the ATR/CHK1 axis and GLUT1 inhibitors correlates with specific transcriptome profiles. This analysis revealed a significant positive correlation between the expression of genes within the Molecular Signatures Database HALLMARK_MTORC1_SIGNALING gene set and the level of synergism measured in the cell lines, indicating that synergistic cell lines tend to have higher expression of genes activated by the mTORC1 complex relative to nonsynergistic cell lines (Fig. 1D; Supplementary Fig. S6A; Supplementary Table S5). Moreover, we compared the mean expression of genes within the MTORC1 signature of our cell line panel to all CCLE cell lines. Cell lines that display synergistic effects for both combinations show significantly higher mean MTORC1 signature expression, than nonsynergistic cell lines (Supplementary Fig. S6B). We note that the mTOR pathway is one of the dominant signaling cascades activated by oncogenic KRAS signaling (44). Thus, the observation that synergistic cell lines display a high frequency of KRAS and NRAS mutations, as well as higher expression of genes downstream of mTORC1, may suggest that activation of the RAS/mTOR pathway determines sensitivity against combined ATR/CHK1 and GLUT1 inhibition. To further validate the sensitivity of Kras-mutant cells against combined ATR/CHK1 and GLUT1 inhibition, we next profiled the cytotoxic activity of PF477736, VE822, and WZB117 on a panel of three murine lung adenocarcinoma cell lines and three murine sarcoma cell lines that were derived from KrasG12D/wt;Trp53−/− tumors. For that purpose, we employed flow cytometry–based apoptosis measurements. As shown in Supplementary Fig. S6C and S6D, both combined VE822/WZB117 and PF477736/WZB117 were highly synergistic in these Kras-driven murine cancer cell lines. The presence of an oncogenic KrasG12D mutation as well as a homozygous deletion of Trp53 was verified by genotyping PCR in these cell lines (Supplementary Fig. S6E and S6F). We note that KPS3 cells only harbored a heterozygous Trp53 deletion.

Altogether, our focused cell line–based drug synergy screen revealed a series of previously described and novel synergistic drug interactions. In particular, the synergistic interaction between compounds targeting the ATR/CHK1 axis and GLUT1 in RAS pathway–altered and mTORC1 signature–expressing cell lines has not been reported previously, leading us to further pursue these interactions.

Combined inhibition of the ATR/CHK1 axis and GLUT1 synergistically induces apoptosis

To further confirm the data derived from our screen, we reassessed the single-agent and combination therapy–induced cytotoxicity of VE822 (ATRi), PF477736 (CHK1i), and WZB117 (GLUT1i) in a validation cell line panel consisting of the synergistic cell lines HCC44 (KRASmut, TP53mut), A549 (KRASmut, TP53mut), as well as the nonsynergistic cell lines H1975 (KRASwt, NRASwt) and A498 (KRASwt, NRASwt). Of note, the presence of oncogenic KRAS p.G12C and p.G12S mutations in HCC44 and A549 cells, respectively, as well as a p.L858R and p.T790M EGFR mutation in H1975 cells and a deletion c.426-429 in VHL in A498 cells were verified by Sanger sequencing, prior to these experiments (Supplementary Fig. S7A). In both synergistic cell lines, the combinations consisting of PF477736/WZB117 (CHK1i/GLUT1i) and VE822/WZB117 (ATRi/GLUT1i) induced significantly higher levels of apoptosis than the respective single agents (Fig. 2A and B), as measured by flow cytometry. In contrast, the combination regimens did not induce higher levels of apoptosis than single treatments in the control cell line H1975, apart from slightly increased levels of apoptosis for combined inhibition of ATR/GLUT1, compared with single treatment with VE822 (ATRi) (Fig. 2C).

Figure 2.

Combined inhibition of the ATR/CHK1 axis and GLUT11 synergistically induces apoptosis in vitro. A–D, Cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents, or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 48 hours. Cells were then co-labeled with Annexin V/PI and analyzed by flow cytometry. Error bars, SD. Each data point represents an independent experiment. E, Synergy scores were calculated for all independent experiments using the Bliss independence model. Mean synergy scores ±SD are shown, revealing robust synergistic effects (>15%) for both combination regimens in the sensitive cell lines A549 and HCC44. Displayed colors resemble the synergy score heatmap in Fig. 1C. F–I, Cells were exposed to DMSO (left), 500 nmol/L PF477736 (CHK1i; middle), 250 nmol/L VE822 (ATRi; right) in combination with (+GLUT1i; bottom rows) or without 25 μmol/L WZB117 (-GLUT1i; top rows) for 72 hours. Left panels show representative pictures. Scale bar, 200 μm. Right, quantification of crystal violet–positive cells is shown. Values were normalized to control wells. Error bars, SD; four independent experiments were performed. Significance in A–D and F–I was tested by one-way ANOVA with Tukey multiple comparison test and is indicated by asterisks. *, P < 0.05; **, P < 0.01; ***, P < 0.001 for CHK1i versus CHK1i/GLUT1, GLUT1i versus CHK1i/GLUT1, ATRi versus ATRi/GLUT1i, and GLUT1i versus ATRi/GLUT1i. Nonsignificant levels are not labeled. See Supplementary Fig. S7B and S7C for all significance levels. J, Graphs show the total number of population doublings (population doubling level). Mean doubling times of all passages were tested for significant differences by Student t test.

Figure 2.

Combined inhibition of the ATR/CHK1 axis and GLUT11 synergistically induces apoptosis in vitro. A–D, Cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents, or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 48 hours. Cells were then co-labeled with Annexin V/PI and analyzed by flow cytometry. Error bars, SD. Each data point represents an independent experiment. E, Synergy scores were calculated for all independent experiments using the Bliss independence model. Mean synergy scores ±SD are shown, revealing robust synergistic effects (>15%) for both combination regimens in the sensitive cell lines A549 and HCC44. Displayed colors resemble the synergy score heatmap in Fig. 1C. F–I, Cells were exposed to DMSO (left), 500 nmol/L PF477736 (CHK1i; middle), 250 nmol/L VE822 (ATRi; right) in combination with (+GLUT1i; bottom rows) or without 25 μmol/L WZB117 (-GLUT1i; top rows) for 72 hours. Left panels show representative pictures. Scale bar, 200 μm. Right, quantification of crystal violet–positive cells is shown. Values were normalized to control wells. Error bars, SD; four independent experiments were performed. Significance in A–D and F–I was tested by one-way ANOVA with Tukey multiple comparison test and is indicated by asterisks. *, P < 0.05; **, P < 0.01; ***, P < 0.001 for CHK1i versus CHK1i/GLUT1, GLUT1i versus CHK1i/GLUT1, ATRi versus ATRi/GLUT1i, and GLUT1i versus ATRi/GLUT1i. Nonsignificant levels are not labeled. See Supplementary Fig. S7B and S7C for all significance levels. J, Graphs show the total number of population doublings (population doubling level). Mean doubling times of all passages were tested for significant differences by Student t test.

Close modal

In A498 cells, WZB117 did not induce substantial levels of apoptosis, whereas both PF477736 and VE822 displayed marked single-agent activity, inducing 50.27% ± 13.99% and 50.10% ± 10.55% apoptosis, respectively (Fig. 2D). We note that significance in Fig. 2A–D was determined using ANOVA with Tukey multiple comparison test (Fig. 2A–D; Supplementary Fig. S7B). The Bliss independence test was used to determine drug synergy in every single experiment. Both combination therapies robustly scored as synergistic in the sensitive cell lines. A mean synergy score of 33.5% ± 12.5% and 22.9% ± 11.5% was calculated for dual inhibition of CHK1 and GLUT1, combined inhibition of ATR and GLUT1 synergistically induced 30.4% ± 9.1% and 32.4% ± 6.8% apoptosis in the sensitive cell lines A549 and HCC44, respectively (Fig. 2E). In contrast, we did not detect synergism for combination regimens, neither in H1975 (synergy scores of −0.8% ± 5.9% and 0.7% ± 5.8% apoptosis for combined inhibition of CHK1/GLUT1 or ATR/GLUT1, respectively; Fig. 2E) nor A498 cells (synergy scores of 9.0% ± 7.9% and 12.5% ± 7.5% apoptosis for combined inhibition of CHK1/GLUT1 or ATR/GLUT1, respectively; Fig. 2E).

To further substantiate these data, we next performed clonogenic survival assays. In brief, cells were seeded at 20% confluency and subsequently exposed to either single-agent VE822, PF477736, and WZB117, as well as the VE822/WZB117 or PF477736/WZB117 combination for 72 hours. Cells were then allowed to proliferate until DMSO-treated control dishes reached confluency. As shown in Fig. 2F–I, the synergistic cell lines displayed a significantly reduced colony formation when treated with the combination regimens compared with the respective single agents. In contrast, addition of WZB117 to either VE822 or PF477736 did not lead to a further reduction of surviving colonies compared with single-agent VE822 or PF477736 in the nonsynergistic H1975 and A498 cells (Fig. 2F–I; Supplementary Fig. S7C).

As specifically the cell-cycle checkpoint kinases ATR and CHK1 are critically important for proper S-phase progression and ATR and CHK1 inhibitors have previously been shown to induce S-phase–specific genotoxic damage, we next asked whether synergistic cell lines may differ in their proliferation kinetics compared with nonsynergistic cells lines. For that purpose, we performed population doubling assays. As shown in Fig. 2J, the mean doubling time of the cell lines for all passages (26.9 hours for HCC44, 26.2 hours for A549, 28.1 hours for H1975, 32.7 hours for A498) displayed no significant difference. Of note, the ATRi/CHKi monosensitive control cell line A498 exhibited the slowest growth rate, making proliferation kinetics as a sensitivity-mediating mechanism for the synergistic interaction unlikely.

As both VE822 and PF477736 are ATP-competitive kinase inhibitors, we next aimed to solidify the specificity of our observations. First, we performed RNA interference–mediated genetic loss-of-function experiments. We specifically used siRNA pools targeting ATR, CHK1, or GLUT1 and confirmed their knockdown efficiency using qRT-PCR and immunoblotting (Supplementary Fig. S8A). By combining these siRNA pools with WZB117, PF477736, or VE822, respectively, we confirmed the synergistic effect obtained by pharmacologic inhibition of these targets in sensitive A549 and HCC44 cells, using viability measurements (Supplementary Fig. S8B). In contrast, no synergy could be detected in the control cell line H1975 (Supplementary Fig. S8B). As a further validation step, we performed a series of flow cytometry–based apoptosis measurements on our validation cell line panel. In these experiments, we replaced the CHK1 inhibitor PF477736 with the structurally distinct CHK1 inhibitor AZD7762 and exchanged the ATR inhibitor VE822 for the structurally distinct ATR inhibitor AZD6738 (Supplementary Fig. S9A and S9B). These validation experiments confirmed that addition of WZB117 to AZD7762 or AZD6738 in the sensitive cell lines HCC44 and A549 led to a synergistic increase in the percentage of apoptotic cells compared with single agents (Supplementary Fig. S9A and S9B). Moreover, and fully in line with our previous results, treatment of H1975 or A498 cells with combined WZB117/AZD7762 or WZB117/AZD6738 did not synergistically increase levels of apoptosis, when compared with single-agent treatments (Supplementary Fig. S9A and S9B).

To verify that PF477736 and VE822 led to target inhibition at the concentrations employed in our experiments, we performed immunoblot experiments, to monitor ATR-mediated CHK1 phosphorylation on Ser-345, as well as the CHK1 autophosphorylation site Ser-296. As shown in Supplementary Fig. S9C, hydroxyurea exposure of HCC44 and H1975 cells induced robust Ser-296 and Ser-345 phosphorylation of CHK1. More importantly, pretreatment with PF477736 (CHK1i) significantly decreased Ser-296 phosphorylation, whereas VE822 (ATRi) completely repressed Ser-345 phosphorylation, thus indicating that both compounds induce proper target inhibition (Supplementary Fig. S9C). In line with a compensatory upregulation of ATR and possibly ATM, as well as DNA-PKcs activity, following CHK1 inhibition, we observed a PF477736-induced hyperphosphorylation of CHK1 on Ser-345 in HCC44 and H1975 cells (Supplementary Fig. S9C). Altogether, these pharmacologic and genetic validation experiments confirm the synergistic interaction between ATR/CHK1 and GLUT1 inhibition.

Combined inhibition of the ATR/CHK1 axis and Glut1 synergistically induces S-phase–specific genotoxic stress

The ATR/CHK1 pathway plays a central role in establishing an intra-S-phase cell-cycle checkpoint in response to the exposure of single-stranded DNA and stalled or collapsed replication DNA forks (45). Furthermore, CHK1 deletion in murine mammary epithelial cells has previously been shown to induce unscheduled replication firing and the induction of apoptosis (46). Moreover, DNA synthesis during S-phase is associated with a substantially increased energy demand (47), which may rationalize an increased demand for glucose uptake in this setting. On the basis of these considerations, we hypothesized that combined inhibition of the ATR/CHK1 axis and GLUT1 may cause S-phase–specific genotoxic stress and apoptosis induction in synergistic cell lines. To directly assess whether combination treatment induced S-phase–specific DNA damage, we treated our validation cell line panel with DMSO, PF477736, VE822, WZB117, PF477736/WZB117, and VE822/WZB117 for 48 hours and subjected the cells to flow cytometry–based γH2AX (a marker for genotoxic damage; refs. 27, 48) and cleaved-caspase-3 staining (a marker for apoptosis; refs. 27, 48). Cells were counterstained with PI to quantify DNA content (Fig. 3A–D; Supplementary Fig. S10A–S10D). None of the single agents induced substantial γH2AX- or cleaved caspase-3 positivity in the synergistic lines (Fig. 3A and B; Supplementary Fig. S10A and S10B). In contrast, populations of γH2AX- and cleaved caspase-3–positive cells were significantly increased following treatment with the combination regimen consisting of VE822/WZB117, compared with single-agent treatment (Fig. 3A and B; Supplementary Fig. S10A, S10B, S10E, and S10F). The populations of γH2AX-positive cells were also significantly increased upon treatment with PF477736/WZB117 in both cell lines, while the populations of cleaved caspase-3–positive cells were only significantly increased in A549 cells (Fig. 3A and B; Supplementary Fig. S10A, S10B, S10E, and S10F). However, the induction of apoptosis (cleaved caspase-3–positive cells) upon treatment with both combination regimens scored as synergistic in both A549 and HCC44 cells (Fig. 3B; Supplementary Fig. S10B). A further inspection of the flow cytometry profiles revealed that the combination treatment induced an intra-S-phase arrest, which coincided with a massive accumulation of γH2AX- and cleaved caspase-3–positive cells in S-phase (Fig. 3A and B; Supplementary Fig. S10A, S10B, S10G, and S10H). In line with the induction of apoptosis, we observed the occurrence of a sub-G1 population, specifically in HCC44 and A549 cells, following exposure to the combination regimens (Fig. 3A; Supplementary Fig. S10A). In marked contrast to the results obtained in the synergistic cell lines, no substantial alterations in the cell-cycle profiles of H1975 cells could be detected following treatment with the single agents or the combination regimens (Fig. 3C). Moreover, neither of these treatments induced significant increases in the percentage of γH2AX-positive or cleaved caspase-3–positive cells in total or in S-phase in H1975 cells (Fig. 3C and D; Supplementary Fig. S10E–S10H), apart from significantly increased caspase-3–positive cell populations upon combined treatment with VE822/WZB117, compared with WZB117 alone. However, there was no synergism detectable (Fig. 3D). In line with our previous observation that ATR and CHK1 inhibition induces single-agent toxicity in A498 cells, the flow cytometry experiment revealed that single-agent PF477736 and VE822 induced γH2AX- and cleaved caspase-3–positivity in these cells. In contrast to the ATR and CHK1 inhibitors, the GLUT1 inhibitor WZB117 did not induce γH2AX- or cleaved caspase-3 positivity beyond that observed for vehicle control. Combination treatment with PF477736/WZB117 or VE822/WZB117 did not lead to a synergistic increase in cleaved caspase-3–positive cells. (Supplementary Fig. S10D).

Figure 3.

Combined inhibition of the ATR/CHK1 axis and GLUT1 synergistically induces S-phase–specific genotoxic stress, resulting in apoptosis. HCC44 (A and B) and H1975 (C and D) cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 48 hours. Cells were stained for γH2AX and cleaved caspase-3 (CC3) and incubated with PI to quantify the cellular DNA content. Samples were analyzed by flow cytometry. A and C, Representative original data are shown. Dashed rectangles show the relative numbers of γH2AX- or CC3-positive cells in percent. B and D, γH2AX- and CC3-positive cells were quantified in total (top plots) and for cells in G1-, S-, and G2 phase (middle and bottom plots). Bliss synergy scores were calculated using the apoptotic effects (CC3-positive cells) of single inhibitors and combination regimens. Synergy levels are annotated (, 15%–29.9%; , 30%–45%; n.syn., synergy score of <15%). Error bars, SD. At least three independent experiments were performed. See Supplementary Fig. S10 for additional cell lines and significance levels.

Figure 3.

Combined inhibition of the ATR/CHK1 axis and GLUT1 synergistically induces S-phase–specific genotoxic stress, resulting in apoptosis. HCC44 (A and B) and H1975 (C and D) cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 48 hours. Cells were stained for γH2AX and cleaved caspase-3 (CC3) and incubated with PI to quantify the cellular DNA content. Samples were analyzed by flow cytometry. A and C, Representative original data are shown. Dashed rectangles show the relative numbers of γH2AX- or CC3-positive cells in percent. B and D, γH2AX- and CC3-positive cells were quantified in total (top plots) and for cells in G1-, S-, and G2 phase (middle and bottom plots). Bliss synergy scores were calculated using the apoptotic effects (CC3-positive cells) of single inhibitors and combination regimens. Synergy levels are annotated (, 15%–29.9%; , 30%–45%; n.syn., synergy score of <15%). Error bars, SD. At least three independent experiments were performed. See Supplementary Fig. S10 for additional cell lines and significance levels.

Close modal

To further solidify the data derived from our flow cytometry–based γH2AX quantification, we next performed immunoblot experiments in our validation cell line panel. Cells were treated with vehicle, PF477736, VE822, WZB117, PF477736/WZB117, and VE822/WZB117 for 24 hours, lysed, and subjected to immunoblot analysis. As shown in Supplementary Fig. S11, combination treatment consisting of PF477736/WZB117 and VE822/WZB117 significantly induced H2AX Ser-139 phosphorylation in HCC44 and A549 cells compared with single agents, whereas single-agent treatment only led to a mild signal increase compared with vehicle control (Supplementary Fig. S11A and S11B). Further in line with our flow cytometry data, neither single agent nor combination regimen induced any substantial H2AX Ser-139 phosphorylation in H1975 cells (Supplementary Fig. S11C), whereas single-agent ATR and CHK1 inhibition produced a strong γH2AX signal in A498 cells, which did not substantially increase when the combination regimens consisting of ATR/GLUT1 or CHK1/GLUT1 inhibitors were applied (Supplementary Fig. S11D). Altogether, these data indicate that combined inhibition of the ATR/CHK1 axis and GLUT1-induced genotoxic damage and apoptosis in synergistic cell lines, specifically in S-phase.

To further assess the genotoxic lesions induced by combined ATR/CHK1 and GLUT1 inhibition, we next performed high-content screening microscopy (Fig. 4A and B). We specifically treated our validation cell line panel with vehicle, PF477736, VE822, WZB117, PF477736/WZB117, and VE822/WZB117 for 24 hours. Upon completion of drug exposure, cells were fixed and subjected to indirect immunofluorescence staining with antibodies detecting γH2AX and RPA (a marker for single-stranded DNA). Cells were counterstained with DAPI. We particularly focused on the quantification of RPA-coated single-stranded DNA, as activity of the ATR/CHK1 kinase branch has been associated with replication fork stabilization in certain cellular systems (49). We suspected that ATR or CHK1 inhibition may lead to replication fork collapse, which exposes large stretches of single-stranded DNA, which is subsequently coated by RPA (45). Somewhat surprisingly, we found that neither PF477736 nor VE822 treatment induced any substantial nuclear RPA foci in HCC44 or A549 cells. However, single WZB117 treatment significantly increased the amount of RPA foci–positive cells in the synergistic cell lines (Fig. 4C; Supplementary Fig. S12; Supplementary Table S6). Simultaneously, we saw a lack of nuclear γH2AX staining in response to all three distinct single agents (Fig. 4C; Supplementary Fig. S12), whereas combined inhibition of either ATR/GLUT1 or CHK1/GLUT1 in HCC44 and A549 cells led to the induction of large populations of RPA foci–positive cells, which similarly coincided with γH2AX positivity (Fig. 4C; Supplementary Fig. S12; Supplementary Table S6). In contrast to the synergistic lines, neither single agents, nor combination therapies, induced any significant increase in the population of γH2AX- or RPA foci–positive H1975 cells (Fig. 4C; Supplementary Fig. S12; Supplementary Table S6). In A498 cells, single-agent PF477736 and VE822 induced not only a robust increase in γH2AX staining intensity, but also a mild, yet consistent increase in the population of RPA-positive cells (Fig. 4C; Supplementary Fig. S12; Supplementary Table S6). However, neither the γH2AX nor the RPA staining intensity increased significantly in A498 cells, when cells were treated with the combination regimens compared with single-agent VE822 or PF477736 exposure (Fig. 4C; Supplementary Fig. S12; Supplementary Table S6). These data suggest that inhibition of GLUT1 induces replicative stress in the sensitive cell lines, leading to the exposure of single-stranded DNA. Because the single treatment with WZB117 did not induce increased γH2AX-levels, cells seem to have repair mechanisms that can avert the induction of apoptosis. This is fully in line with the observation of reduced colony formation upon single treatment with WZB117 (Fig. 2F and G), which, however, did not induce H2AX phosphorylation or apoptosis in any other experiment we performed (Figs. 2C and 3C and D; Supplementary Fig. S11). In contrast, combined inhibition of GLUT1 and CHK1/ATR significantly increased both γH2AX- and RPA foci–positive cell populations (Fig. 4; Supplementary Fig. S12) and subsequently led to a massive induction of apoptosis in sensitive cell lines (Fig. 3; Supplementary Fig. S10).

Figure 4.

Combined inhibition of ATR/CHK1 and GLUT1 induces S-phase–specific genotoxic damage. Cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 24 hours. A and B, Representative images of γH2AX-immunostained cells are shown. Green or red encircled cells indicate inclusion or exclusion of objects for quantification. Insets in B illustrate representative cells with low or high γH2AX intensity, respectively. Scale bar, 100 μm. C, Bar graphs show cells per field (third row), numbers of analyzed cells per condition (bottom plots), and the quantification of mean γH2AX and RPA70 intensities normalized to the vehicle control, respectively. Values are means ± SEM, n = 4. See Supplementary Fig. S12 for original data and Supplementary Table S6 for P values. Significance of mean γH2AX and RPA70 intensities was tested by one-way ANOVA with Tukey honest significant difference test.

Figure 4.

Combined inhibition of ATR/CHK1 and GLUT1 induces S-phase–specific genotoxic damage. Cells were treated with DMSO, VE822 (250 nmol/L; ATRi), PF477736 (500 nmol/L; CHK1i), WZB117 (25 μmol/L; GLUT1) as single agents or with PF477736/WZB117 (500 nmol/L/25 μmol/L; CHK1i/GLUT1i) or VE822/WZB117 (250 nmol/L/25 μmol/L; CHK1i/GLUT1i) in combination for 24 hours. A and B, Representative images of γH2AX-immunostained cells are shown. Green or red encircled cells indicate inclusion or exclusion of objects for quantification. Insets in B illustrate representative cells with low or high γH2AX intensity, respectively. Scale bar, 100 μm. C, Bar graphs show cells per field (third row), numbers of analyzed cells per condition (bottom plots), and the quantification of mean γH2AX and RPA70 intensities normalized to the vehicle control, respectively. Values are means ± SEM, n = 4. See Supplementary Fig. S12 for original data and Supplementary Table S6 for P values. Significance of mean γH2AX and RPA70 intensities was tested by one-way ANOVA with Tukey honest significant difference test.

Close modal

Thus, altogether, our data indicate that combined ATR/CHK1 and GLUT1 inhibition is selectively cytotoxic in KRAS-mutant HCC44 and A549 cells. This synergistic cytotoxicity was associated with the induction of S-phase–specific genotoxic damage following exposure to both combination regimens. Moreover, this S-phase–specific DNA damage was associated with the induction of apoptotic cell death in S-phase.

Combined ATR/GLUT1 inhibition displays synergistic toxicity in a KrasG12D-driven autochthonous mouse model of soft tissue sarcoma

Given that the majority of cell lines that displayed a synergistic drug interaction between the ATR and GLUT1 inhibitors in our primary screen harbored oncogenic KRAS mutations, we next aimed to assess the preclinical efficacy of a combination regimen consisting of VE822 and WZB117 in an autochthonous mouse model of KrasG12D-driven soft tissue sarcoma. For that purpose, we generated KrasLSL.G12D/wt;Trp53fl/fl animals. Intramuscular Adeno-CMV-Cre injection leads to deletion of a transcriptional/translational STOP cassette from the endogenous Kras locus, allowing the transcription of an oncogenic KrasG12D allele. In parallel, Cre recombinase activity mediates the deletion of both Trp53 alleles in virally infected cells (30). Following Adeno-CMV-Cre injection into the thighs of KrasLSL.G12D/wt;Trp53fl/fl mice, sarcoma development was monitored by weekly MRI scans (Supplementary Fig. S13A). Upon tumor detection, mice were treated daily with vehicle, VE822 (40 mg/kg), WZB117 (50 mg/kg) or VE822/WZB117 (40 mg/kg and 50 mg/kg; Fig. 5A; Supplementary Fig. S13A). While sarcomas treated with vehicle (n = 24), VE822 (n = 8), or WZB117 (n = 15) displayed continuous growth throughout the 7-day treatment period, sarcoma-bearing mice that were treated with the combination regimen (n = 21) showed a significantly reduced tumor volume gain compared with single-agent treatments (Fig. 5B and C; Supplementary Fig. S13B). Of note, 5 of 21 sarcomas in KrasLSL.G12D/wt;Trp53fl/fl mice treated with combined ATR/GLUT1 inhibition displayed tumor volume shrinkage (Fig. 5C). In general, both the single-agent treatments, as well as the combination treatment were well-tolerated by the animals and did not have a significant effect on body weight of the mice throughout the 7-day treatment course (Supplementary Fig. S13C).

Figure 5.

Combined ATR/GLUT1 inhibition displays synergistic toxicity in a KrasG12D-driven sarcoma model in vivo. A,KrasLSL.G12D/wt;Trp53fl/fl mice were divided into four different treatment cohorts. Thirty days after injection of Ad-Cre (green bar), mice were imaged weekly to detect formation of tumor nodules via MRI (yellow bar). Upon tumor detection (day 0), mice were treated for 7 consecutive days (red bar). A final MRI scan was performed upon completion of therapy (day 7). B, Representative MRI images of tumors before and after treatment are shown. Tumor nodules are encircled in the transversal scans, green and red lines mark two separate tumors. C, Tumor volumes were calculated for each lesion and normalized to the respective volumes at the beginning of therapy, yielding fold changes (y-axis). Error bars, SEM. Fold changes in all four cohorts were compared against one another by Kruskal–Wallis test with Dunn multiple comparisons test. Significance levels are denoted by asterisks. *, P < 0.05; **, P < 0.01; nonsignificant levels are not labeled.

Figure 5.

Combined ATR/GLUT1 inhibition displays synergistic toxicity in a KrasG12D-driven sarcoma model in vivo. A,KrasLSL.G12D/wt;Trp53fl/fl mice were divided into four different treatment cohorts. Thirty days after injection of Ad-Cre (green bar), mice were imaged weekly to detect formation of tumor nodules via MRI (yellow bar). Upon tumor detection (day 0), mice were treated for 7 consecutive days (red bar). A final MRI scan was performed upon completion of therapy (day 7). B, Representative MRI images of tumors before and after treatment are shown. Tumor nodules are encircled in the transversal scans, green and red lines mark two separate tumors. C, Tumor volumes were calculated for each lesion and normalized to the respective volumes at the beginning of therapy, yielding fold changes (y-axis). Error bars, SEM. Fold changes in all four cohorts were compared against one another by Kruskal–Wallis test with Dunn multiple comparisons test. Significance levels are denoted by asterisks. *, P < 0.05; **, P < 0.01; nonsignificant levels are not labeled.

Close modal

These data derived from an autochthonous Kras-driven soft tissue sarcoma model indicate that combined ATR/GLUT1 inhibition displays substantial preclinical activity in this difficult to treat entity.

Through the advent of novel sequencing technologies, the genomic underpinnings of the process of malignant transformation have been unraveled in an unprecedentedly fine-grained manner. Numerous cancer-specific genomic aberrations have been identified that drive tumorigenesis. On the basis of these genomic discoveries, single-agent activity of a wide spectrum of drug compounds has recently been profiled across broad panels of genomically-characterized cancer cell lines (26, 50). These approaches have led to the discovery and validation of numerous cancer genotype-specific molecular liabilities, such as oncogenic EGFR- and BRAF mutations, as well as BCR-ABL rearrangements (26, 50). However, while some cancer-driving mutations and rearrangements, particularly kinase oncogenes, encode for pharmacologically actionable gene products (e.g., EGFR, ROS1, ALK; refs. 1–3), others are not amenable to direct pharmacologic inhibition. More recently, focused drug combination screens with the specific goal of overcoming resistance to single-agent drugs, such as BRAF inhibitors, have been conducted on melanoma-restricted cell line panels (51, 52). Moreover, a recent screening effort characterized synergistic effects between 18 classical genotoxic agents, but only included few target-specific small-molecule inhibitors (53). Here, we conducted a focused flow cytometry–based drug synergy screen, using a series of small-molecule compounds interfering with cancer-relevant pathways. We employed Bliss independence–based drug synergy scoring and identified eight distinct drug combinations that displayed a synergistic interaction in at least nine cancer cell lines. Of note, our approach also validated four synergistic drug interactions within our cell line panel that had been described previously. This observation indicates that our synergy screening approach is valid. The synergy data obtained in our initial screen were confirmed in the context of validation experiments (Fig. 2A–E). We note that these validation experiments uniformly returned synergistic drug interactions between VE822/PF477736 with WZB117 in the sensitive lines A549 and HCC44. However, only eight of nine synergy indices that we determined for the PF477736/WZB117 combination in A549 cells achieved a Bliss synergy score >15, while one experiment revealed a synergy index <10 (Fig. 2A and E). Similarly, in HCC44 cells, the PF477736/WZB117 combination revealed a synergy index >15 in eight of ten experiments, while synergy indices <15 were measured in two of ten experiments (Fig. 2B and E). Thus, while the experiments are highly consistent, the synergy scores display a relatively high degree of variation, which may explain the relatively high SDs in these experiments.

Intriguingly, KRAS-mutant cells were recently shown to display sensitivity against combined inhibition of the direct downstream effectors MEK and PI3K (54). These data suggest that the plethora of pathways rewired by oncogenic KRAS encode further vulnerabilities, amenable to combination drug regimens. For instance, several lines of evidence suggest that KRAS mutations induce an extensive rewiring of glucose metabolism and GLUT1 overexpression. Furthermore, increased 18FDG avidity has been reported, specifically in KRAS- or BRAF-mutant cells (14, 16, 55–58). Moreover, oncogenic KRAS mutations were shown to induce a profound rewiring of metabolic circuits in transformed cells, including increased glucose uptake and lactate production, shifting of the glycolytic flux into the nonoxidative arm of the pentose phosphate pathway, as well as activation of the hexosamine biosynthesis pathway (14–17). These effects appear to be gene dosage–dependent, as Kras copy number gains were recently shown to drive additional metabolic alterations, such as shunting of glucose-derived metabolites into the tricarboxylic acid cycle and glutathione biosynthesis (13). Furthermore, in KrasG12D-driven murine pancreatic ductal adenocarcinomas, acute withdrawal of oncogenic KrasG12D substantially decreased levels of glucose-6-phosphate, fructose-6-phosphate and fructose-1,6-bisphosphate within 24 hours (15). These changes were accompanied by reduced glucose uptake and lactate production, indicating that KrasG12D drives glycolytic flux in this model (15). In addition, expression levels of the glycolytic genes Glut1, Hk1, Hk2, Pfkl, and Ldha markedly decreased in response to KrasG12D withdrawal, further confirming that KrasG12D is critical for resetting glucose uptake and metabolism (15). Moreover, it was shown that KrasG12D-driven pancreatic ductal adenocarcinoma displays shunting of the glycolytic flux into the nonoxidative arm of the pentose phosphate pathway, to generate ribose-5-phosphate for DNA/RNA biosynthesis (14, 15). Altogether, these data indicate that KRAS-mutant cells and tumors display a marked rewiring of glucose metabolism. These metabolic alterations may partially explain why GLUT1 inhibition in combination with blockade of the ATR/CHK1 kinase axis results in toxicity in KRAS-mutant cells (Fig. 1) and a Kras-driven murine sarcoma model (Fig. 5).

In addition to these KrasG12D-dependent alterations of glucose metabolism, oncogenic KRAS has been shown to induce genotoxic stress (19, 24, 59). This KRAS-induced DNA damage is likely the result of increased production of reactive oxygen species, which subsequently causes oxidative damage to the DNA (20–23). Furthermore, oncogenic KRAS has been shown to induce replication stress through unscheduled replication firing and alterations in fork progression, leading to genotoxic damage at the sites of active DNA replication (18, 60, 61). This oncogene-induced genotoxic stress has been shown to lead to a constitutive activation of the DDR network (19, 61), likely to cope with the detrimental effects of replication-induced DNA damage in KRAS-mutant tumor cells. Intriguingly, this tonic DNA damage signaling has recently been shown to be critical for the survival of KRAS-mutant cells (19, 62). Moreover, the ATR/CHK1 kinase branch is primarily active in response to single-stranded DNA stretches and stalled replication forks (45). This may rationalize the efficacy of ATR/CHK1–targeting compounds as part of combination regimens including GLUT1 inhibitors in KRAS-mutant settings, in which oncogene-induced replication stress is prevalent.

Fitting with a role of the ATR/CHK1 pathway in safeguarding S-phase progression and fitting with an increased energy demand during S-phase, we found that combined inhibition of GLUT1 and the ATR/CHK1 axis arrested RAS-altered cells in S-phase. Paralleling this S-phase stalling, we observed the occurrence of genotoxic damage in S-phase. Altogether, these data may indicate that combined repression of GLUT1 and the ATR/CHK1 axis synergistically intercepts two signaling nodes, which are critical for the survival of KRAS-driven cancer. On the basis of our results, we recommend these combination regimens for clinical validation in KRAS-stratified tumor patients.

H.C. Reinhardt reports receiving a commercial research grant from Gilead Pharmaceuticals, has received speakers bureau honoraria from Abbvie and AstraZeneca, and has unpaid consultant/advisory board relationship with Vertex Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J. Erber, J.D. Steiner, T. Persigehl, H.C. Reinhardt

Development of methodology: J. Erber, J.D. Steiner, F. Beleggia, R.W.J. Kaiser, H.C. Reinhardt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Erber, J.D. Steiner, J. Isensee, L.A. Lobbes, A. Toschka, A. Schmitt, R.W.J. Kaiser, F. Siedek, T. Persigehl, T. Hucho

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Erber, J.D. Steiner, J. Isensee, A. Toschka, F. Beleggia, R.W.J. Kaiser, T. Persigehl, T. Hucho, H.C. Reinhardt

Writing, review, and/or revision of the manuscript: J. Erber, J.D. Steiner, F. Beleggia, F. Siedek, T. Persigehl, H.C. Reinhardt

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Erber, A. Schmitt, F. Siedek, T. Persigehl, H.C. Reinhardt

Study supervision: H.C. Reinhardt

Other (high content screening microscopy): J. Isensee

This work was supported by the Deutsche Forschungsgemeinschaft (KFO-286/RP2, RE 2246/2-1, RE 2246/7-1 CECAD, SFB-829 to H.C. Reinhardt), the German federal state North Rhine Westphalia (NRW) as part of the EFRE initiative (grant LS-1-1-030a to H.C. Reinhardt), the Else Kröner-Fresenius Stiftung (EKFS-2014-A06 to H.C. Reinhardt), the Deutsche Krebshilfe (111724 to H.C. Reinhardt; Mildred-Scheel-Doktorandenprogramm to J. Erber), and Boehringer Ingelheim Fonds (MD fellowship to R.W.J. Kaiser).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Shaw
AT
,
Ou
SH
,
Bang
YJ
,
Camidge
DR
,
Solomon
BJ
,
Salgia
R
, et al
Crizotinib in ROS1-rearranged non-small-cell lung cancer
.
N Engl J Med
2014
;
371
:
1963
71
.
2.
Mok
TS
,
Wu
YL
,
Thongprasert
S
,
Yang
CH
,
Chu
DT
,
Saijo
N
, et al
Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma
.
N Engl J Med
2009
;
361
:
947
57
.
3.
Shaw
AT
,
Kim
DW
,
Nakagawa
K
,
Seto
T
,
Crino
L
,
Ahn
MJ
, et al
Crizotinib versus chemotherapy in advanced ALK-positive lung cancer
.
N Engl J Med
2013
;
368
:
2385
94
.
4.
Pao
W
,
Girard
N
. 
New driver mutations in non-small-cell lung cancer
.
Lancet Oncol
2011
;
12
:
175
80
.
5.
Karnoub
AE
,
Weinberg
RA
. 
Ras oncogenes: split personalities
.
Nat Rev Mol Cell Biol
2008
;
9
:
517
31
.
6.
Cox
AD
,
Fesik
SW
,
Kimmelman
AC
,
Luo
J
,
Der
CJ
. 
Drugging the undruggable RAS: mission possible?
Nat Rev Drug Discov
2014
;
13
:
828
51
.
7.
Meyer
N
,
Penn
LZ
. 
Reflecting on 25 years with MYC
.
Nat Rev Cancer
2008
;
8
:
976
90
.
8.
Lord
CJ
,
Ashworth
A
. 
BRCAness revisited
.
Nat Rev Cancer
2016
;
16
:
110
20
.
9.
Baines
AT
,
Xu
D
,
Der
CJ
. 
Inhibition of Ras for cancer treatment: the search continues
.
Future Med Chem
2011
;
3
:
1787
808
.
10.
Hunter
T
. 
Oncoprotein networks
.
Cell
1997
;
88
:
333
46
.
11.
Campbell
SL
,
Khosravi-Far
R
,
Rossman
KL
,
Clark
GJ
,
Der
CJ
. 
Increasing complexity of Ras signaling
.
Oncogene
1998
;
17
:
1395
413
.
12.
Vigil
D
,
Cherfils
J
,
Rossman
KL
,
Der
CJ
. 
Ras superfamily GEFs and GAPs: validated and tractable targets for cancer therapy?
Nat Rev Cancer
2010
;
10
:
842
57
.
13.
Kerr
EM
,
Gaude
E
,
Turrell
FK
,
Frezza
C
,
Martins
CP
. 
Mutant Kras copy number defines metabolic reprogramming and therapeutic susceptibilities
.
Nature
2016
;
531
:
110
3
.
14.
Yun
J
,
Mullarky
E
,
Lu
C
,
Bosch
KN
,
Kavalier
A
,
Rivera
K
, et al
Vitamin C selectively kills KRAS and BRAF mutant colorectal cancer cells by targeting GAPDH
.
Science
2015
;
350
:
1391
6
.
15.
Ying
H
,
Kimmelman
AC
,
Lyssiotis
CA
,
Hua
S
,
Chu
GC
,
Fletcher-Sananikone
E
, et al
Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose metabolism
.
Cell
2012
;
149
:
656
70
.
16.
Yun
J
,
Rago
C
,
Cheong
I
,
Pagliarini
R
,
Angenendt
P
,
Rajagopalan
H
, et al
Glucose deprivation contributes to the development of KRAS pathway mutations in tumor cells
.
Science
2009
;
325
:
1555
9
.
17.
Racker
E
,
Resnick
RJ
,
Feldman
R
. 
Glycolysis and methylaminoisobutyrate uptake in rat-1 cells transfected with ras or myc oncogenes
.
Proc Natl Acad Sci U S A
1985
;
82
:
3535
8
.
18.
Grabocka
E
,
Commisso
C
,
Bar-Sagi
D
. 
Molecular pathways: targeting the dependence of mutant RAS cancers on the DNA damage response
.
Clin Cancer Res
2015
;
21
:
1243
7
.
19.
Dietlein
F
,
Kalb
B
,
Jokic
M
,
Noll
EM
,
Strong
A
,
Tharun
L
, et al
A synergistic interaction between Chk1- and MK2 inhibitors in KRAS-mutant cancer
.
Cell
2015
;
162
:
146
59
.
20.
Weyemi
U
,
Lagente-Chevallier
O
,
Boufraqech
M
,
Prenois
F
,
Courtin
F
,
Caillou
B
, et al
ROS-generating NADPH oxidase NOX4 is a critical mediator in oncogenic H-Ras-induced DNA damage and subsequent senescence
.
Oncogene
2012
;
31
:
1117
29
.
21.
Lee
AC
,
Fenster
BE
,
Ito
H
,
Takeda
K
,
Bae
NS
,
Hirai
T
, et al
Ras proteins induce senescence by altering the intracellular levels of reactive oxygen species
.
J Biol Chem
1999
;
274
:
7936
40
.
22.
Irani
K
,
Xia
Y
,
Zweier
JL
,
Sollott
SJ
,
Der
CJ
,
Fearon
ER
, et al
Mitogenic signaling mediated by oxidants in Ras-transformed fibroblasts
.
Science
1997
;
275
:
1649
52
.
23.
Ogrunc
M
,
Di Micco
R
,
Liontos
M
,
Bombardelli
L
,
Mione
M
,
Fumagalli
M
, et al
Oncogene-induced reactive oxygen species fuel hyperproliferation and DNA damage response activation
.
Cell Death Differ
2014
;
21
:
998
1012
.
24.
Di Micco
R
,
Fumagalli
M
,
Cicalese
A
,
Piccinin
S
,
Gasparini
P
,
Luise
C
, et al
Oncogene-induced senescence is a DNA damage response triggered by DNA hyper-replication
.
Nature
2006
;
444
:
638
42
.
25.
Macerelli
M
,
Caramella
C
,
Faivre
L
,
Besse
B
,
Planchard
D
,
Polo
V
, et al
Does KRAS mutational status predict chemoresistance in advanced non-small cell lung cancer (NSCLC)?
Lung Cancer
2014
;
83
:
383
8
.
26.
Barretina
J
,
Caponigro
G
,
Stransky
N
,
Venkatesan
K
,
Margolin
AA
,
Kim
S
, et al
The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity
.
Nature
2012
;
483
:
603
7
.
27.
Schmitt
A
,
Knittel
G
,
Welcker
D
,
Yang
TP
,
George
J
,
Nowak
M
, et al
ATM deficiency is associated with sensitivity to PARP1 and ATR inhibitors in lung adenocarcinoma
.
Cancer Res
2017
;
77
:
3040
56
.
28.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
29.
Mootha
VK
,
Lindgren
CM
,
Eriksson
KF
,
Subramanian
A
,
Sihag
S
,
Lehar
J
, et al
PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes
.
Nat Genet
2003
;
34
:
267
73
.
30.
Kirsch
DG
,
Dinulescu
DM
,
Miller
JB
,
Grimm
J
,
Santiago
PM
,
Young
NP
, et al
A spatially and temporally restricted mouse model of soft tissue sarcoma
.
Nat Med
2007
;
13
:
992
7
.
31.
Carrassa
L
,
Chila
R
,
Lupi
M
,
Ricci
F
,
Celenza
C
,
Mazzoletti
M
, et al
Combined inhibition of Chk1 and Wee1: in vitro synergistic effect translates to tumor growth inhibition in vivo
.
Cell Cycle
2012
;
11
:
2507
17
.
32.
Chaudhuri
L
,
Vincelette
ND
,
Koh
BD
,
Naylor
RM
,
Flatten
KS
,
Peterson
KL
, et al
CHK1 and WEE1 inhibition combine synergistically to enhance therapeutic efficacy in acute myeloid leukemia ex vivo
.
Haematologica
2014
;
99
:
688
96
.
33.
Chila
R
,
Basana
A
,
Lupi
M
,
Guffanti
F
,
Gaudio
E
,
Rinaldi
A
, et al
Combined inhibition of Chk1 and Wee1 as a new therapeutic strategy for mantle cell lymphoma
.
Oncotarget
2015
;
6
:
3394
408
.
34.
Guertin
AD
,
Martin
MM
,
Roberts
B
,
Hurd
M
,
Qu
X
,
Miselis
NR
, et al
Unique functions of CHK1 and WEE1 underlie synergistic anti-tumor activity upon pharmacologic inhibition
.
Cancer Cell Int
2012
;
12
:
45
.
35.
Magnussen
GI
,
Emilsen
E
,
Giller Fleten
K
,
Engesaeter
B
,
Nahse-Kumpf
V
,
Fjaer
R
, et al
Combined inhibition of the cell cycle related proteins Wee1 and Chk1/2 induces synergistic anti-cancer effect in melanoma
.
BMC Cancer
2015
;
15
:
462
.
36.
Davies
KD
,
Cable
PL
,
Garrus
JE
,
Sullivan
FX
,
von Carlowitz
I
,
Huerou
YL
, et al
Chk1 inhibition and Wee1 inhibition combine synergistically to impede cellular proliferation
.
Cancer Biol Ther
2011
;
12
:
788
96
.
37.
Cozzi
M
,
Giorgi
F
,
Marcelli
E
,
Pentimalli
F
,
Forte
IM
,
Schenone
S
, et al
Antitumor activity of new pyrazolo[3,4-d]pyrimidine SRC kinase inhibitors in Burkitt lymphoma cell lines and its enhancement by WEE1 inhibition
.
Cell Cycle
2012
;
11
:
1029
39
.
38.
Zhao
J
,
Niu
X
,
Li
X
,
Edwards
H
,
Wang
G
,
Wang
Y
, et al
Inhibition of CHK1 enhances cell death induced by the Bcl-2-selective inhibitor ABT-199 in acute myeloid leukemia cells
.
Oncotarget
2016
;
7
:
34785
99
.
39.
Wang
C
,
Jin
H
,
Gao
D
,
Lieftink
C
,
Evers
B
,
Jin
G
, et al
Phospho-ERK is a biomarker of response to a synthetic lethal drug combination of sorafenib and MEK inhibition in liver cancer
.
J Hepatol
2018
;
69
:
1057
65
.
40.
Bliss
CI
. 
The toxicity of poisons applied jointly
.
Ann Appl Biol
1939
;
26
:
585
615
.
41.
Bai
M
,
Song
N
,
Che
X
,
Wang
X
,
Qu
X
,
Liu
Y
. 
Chk1 activation attenuates sensitivity of lapatinib in HER2-positive gastric cancer
.
Cell Biol Int
2018
;
42
:
781
93
.
42.
Miyawaki
M
,
Yasuda
H
,
Tani
T
,
Hamamoto
J
,
Arai
D
,
Ishioka
K
, et al
Overcoming EGFR bypass signal-induced acquired resistance to ALK tyrosine kinase inhibitors in ALK-translocated lung cancer
.
Mol Cancer Res
2017
;
15
:
106
14
.
43.
Yamaguchi
N
,
Lucena-Araujo
AR
,
Nakayama
S
,
de Figueiredo-Pontes
LL
,
Gonzalez
DA
,
Yasuda
H
, et al
Dual ALK and EGFR inhibition targets a mechanism of acquired resistance to the tyrosine kinase inhibitor crizotinib in ALK rearranged lung cancer
.
Lung Cancer
2014
;
83
:
37
43
.
44.
Vasan
N
,
Boyer
JL
,
Herbst
RS
. 
A RAS renaissance: emerging targeted therapies for KRAS-mutated non-small cell lung cancer
.
Clin Cancer Res
2014
;
20
:
3921
30
.
45.
Saldivar
JC
,
Cortez
D
,
Cimprich
KA
. 
The essential kinase ATR: ensuring faithful duplication of a challenging genome
.
Nat Rev Mol Cell Biol
2017
;
18
:
622
36
.
46.
Lam
MH
,
Liu
Q
,
Elledge
SJ
,
Rosen
JM
. 
Chk1 is haploinsufficient for multiple functions critical to tumor suppression
.
Cancer Cell
2004
;
6
:
45
59
.
47.
Solaki
M
,
Ewald
JC
. 
Fueling the cycle: CDKs in carbon and energy metabolism
.
Front Cell Dev Biol
2018
;
6
:
93
.
48.
Doerr
F
,
George
J
,
Schmitt
A
,
Beleggia
F
,
Rehkamper
T
,
Hermann
S
, et al
Targeting a non-oncogene addiction to the ATR/CHK1 axis for the treatment of small cell lung cancer
.
Sci Rep
2017
;
7
:
15511
.
49.
Couch
FB
,
Bansbach
CE
,
Driscoll
R
,
Luzwick
JW
,
Glick
GG
,
Betous
R
, et al
ATR phosphorylates SMARCAL1 to prevent replication fork collapse
.
Genes Dev
2013
;
27
:
1610
23
.
50.
Garnett
MJ
,
Edelman
EJ
,
Heidorn
SJ
,
Greenman
CD
,
Dastur
A
,
Lau
KW
, et al
Systematic identification of genomic markers of drug sensitivity in cancer cells
.
Nature
2012
;
483
:
570
5
.
51.
Held
MA
,
Langdon
CG
,
Platt
JT
,
Graham-Steed
T
,
Liu
Z
,
Chakraborty
A
, et al
Genotype-selective combination therapies for melanoma identified by high-throughput drug screening
.
Cancer Discov
2013
;
3
:
52
67
.
52.
Friedman
AA
,
Amzallag
A
,
Pruteanu-Malinici
I
,
Baniya
S
,
Cooper
ZA
,
Piris
A
, et al
Landscape of targeted anti-cancer drug synergies in melanoma identifies a novel BRAF-VEGFR/PDGFR combination treatment
.
PLoS One
2015
;
10
:
e0140310
.
53.
O'Neil
J
,
Benita
Y
,
Feldman
I
,
Chenard
M
,
Roberts
B
,
Liu
Y
, et al
An unbiased oncology compound screen to identify novel combination strategies
.
Mol Cancer Ther
2016
;
15
:
1155
62
.
54.
Engelman
JA
,
Chen
L
,
Tan
X
,
Crosby
K
,
Guimaraes
AR
,
Upadhyay
R
, et al
Effective use of PI3K and MEK inhibitors to treat mutant Kras G12D and PIK3CA H1047R murine lung cancers
.
Nat Med
2008
;
14
:
1351
6
.
55.
Sasaki
H
,
Shitara
M
,
Yokota
K
,
Hikosaka
Y
,
Moriyama
S
,
Yano
M
, et al
Overexpression of GLUT1 correlates with Kras mutations in lung carcinomas
.
Mol Med Rep
2012
;
5
:
599
602
.
56.
Kawada
K
,
Nakamoto
Y
,
Kawada
M
,
Hida
K
,
Matsumoto
T
,
Murakami
T
, et al
Relationship between 18F-fluorodeoxyglucose accumulation and KRAS/BRAF mutations in colorectal cancer
.
Clin Cancer Res
2012
;
18
:
1696
703
.
57.
Iwamoto
M
,
Kawada
K
,
Nakamoto
Y
,
Itatani
Y
,
Inamoto
S
,
Toda
K
, et al
Regulation of 18F-FDG accumulation in colorectal cancer cells with mutated KRAS
.
J Nucl Med
2014
;
55
:
2038
44
.
58.
Sheu
JJ
,
Guan
B
,
Tsai
FJ
,
Hsiao
EY
,
Chen
CM
,
Seruca
R
, et al
Mutant BRAF induces DNA strand breaks, activates DNA damage response pathway, and up-regulates glucose transporter-1 in nontransformed epithelial cells
.
Am J Pathol
2012
;
180
:
1179
88
.
59.
Pylayeva-Gupta
Y
,
Grabocka
E
,
Bar-Sagi
D
. 
RAS oncogenes: weaving a tumorigenic web
.
Nat Rev Cancer
2011
;
11
:
761
74
.
60.
Maya-Mendoza
A
,
Ostrakova
J
,
Kosar
M
,
Hall
A
,
Duskova
P
,
Mistrik
M
, et al
Myc and Ras oncogenes engage different energy metabolism programs and evoke distinct patterns of oxidative and DNA replication stress
.
Mol Oncol
2015
;
9
:
601
16
.
61.
Fikaris
AJ
,
Lewis
AE
,
Abulaiti
A
,
Tsygankova
OM
,
Meinkoth
JL
. 
Ras triggers ataxia-telangiectasia-mutated and Rad-3-related activation and apoptosis through sustained mitogenic signaling
.
J Biol Chem
2006
;
281
:
34759
67
.
62.
Grabocka
E
,
Pylayeva-Gupta
Y
,
Jones
MJ
,
Lubkov
V
,
Yemanaberhan
E
,
Taylor
L
, et al
Wild-type H- and N-Ras promote mutant K-Ras-driven tumorigenesis by modulating the DNA damage response
.
Cancer Cell
2014
;
25
:
243
56
.

Supplementary data