mTORC1/2 and Protein Translation Regulate Levels of CHK1 and the Sensitivity to CHK1 Inhibitors in Ewing Sarcoma Cells

Ewing sarcoma is a highly aggressive bone and soft tissue cancer that is caused by the EWS-FLI1 fusion protein (1). EWS-FLI1 is required for tumor growth and survival (2). However, directly targeting EWS-FLI1 with drugs has been challenging and an alternative therapeutic approach is to identify unique vulnerabilities incurred by the oncoprotein (3). In previous work, we used a human embryonic stem cell model to identify that Ewing sarcoma cells are vulnerable to gemcitabine, hydroxyurea, ciclopirox, and other drugs that inhibit ribonucleotide reductase (RNR), which is the rate-limiting enzyme in the synthesis of deoxyribonucleotides (4, 5).

More recently, we found that the inhibition of checkpoint kinase 1 (CHK1), a downstream effector of Ataxia Telangiectasia and Rad3-Related Protein (ATR) and a critical mediator of the response to DNA replication stress, increases the sensitivity of Ewing sarcoma cells to RNR inhibitors in vitro and in vivo (5). Ewing sarcoma cells also exhibit sensitivity to CHK1 and ATR inhibitors as single agents (6–8). However, Nieto-Soler and colleagues have shown that Ewing sarcoma cells exhibit high levels of the CHK1 protein, which may reflect a response to elevated levels of endogenous DNA replication stress (6). In addition, Wayne and colleagues have shown that the complete and sustained inhibition of CHK1 is necessary to inhibit the growth of sensitive cell lines (9). Consequently, we began this work with the aim of determining the effect of CHK1 protein levels on drug sensitivity and identifying the mechanisms that regulate CHK1 levels in Ewing sarcoma cells.

In this report, we show that the high levels of the CHK1 protein in Ewing sarcoma cells limit the efficacy of CHK1 inhibitors. Notably, multiple mechanisms are known to regulate CHK1 protein levels and activity in cancer cells, but recent work has identified that the mTOR pathway can regulate levels of CHK1 via both transcriptional and translational mechanisms (10–15). The mTOR pathway is also known to have a critical role in Ewing sarcoma tumors, and mTOR inhibitors reduce the growth of Ewing sarcoma cells in vitro and in xenograft experiments (16–18). Moreover, in addition to CHK1, the mTOR pathway has also been reported to regulate the abundance of the RRM1 and RRM2 subunits of RNR (19, 20). Consequently, based on our previous work demonstrating synergy between RNR and CHK1 inhibitors in Ewing sarcoma, we initially focused our investigation on the role of mTOR signaling in the regulation of CHK1 levels in Ewing sarcoma cells (5).

We identified that the inhibition of mTORC1/2, but not mTORC1, activates the protein translation repressor 4E-BP1, reduces protein synthesis, and decreases levels of the CHK1 protein in Ewing sarcoma cells. In addition, prexasertib, a catalytic CHK1 inhibitor that also inhibits the mTOR pathway, activates 4E-BP1, inhibits protein synthesis, and reduces CHK1 protein levels in Ewing sarcoma cells. Moreover, the combination of prexasertib and gemcitabine was synergistic in vitro and significantly prolonged mouse survival in a xenograft experiment. Overall, our results provide insight into Ewing sarcoma biology and identify a candidate pathway that can be targeted in Ewing sarcoma tumors.

Cell lines were maintained at 37°C in a 5% CO₂ atmosphere. The A673, TC32, TC71, SK-NEP, TTC466 and EW8 cell lines were kindly provided by Dr. Kimberly Stegmaier (Dana-Farber Cancer Institute, Boston, MA). The BJ-tert, HEK-293T, HT1080, RPE-tert, HeLa, and U2OS cell lines were obtained from ATCC. The RH30, RD, and SAOS cell lines were provided by Dr. Munir Tanas (University of Iowa, Iowa City, IA). The Panc-1 and PaCa-2 cell lines were provided by Dr. Garry Buettner (University of Iowa, Iowa City, IA). Cells were cultured as described previously (4, 5). Cell lines were authenticated by DNA fingerprinting using the short tandem repeat method and used within 8–10 passages of thawing.

Chemical compounds were purchased from Sigma (gemcitabine and temsirolimus), Selleckchem (prexasertib, LY2603618, U0126, and TAK-228), APExBio (VE-822), Thermo Fisher Scientific (puromycin), and MedChemExpress (Torin2, MK-8776, olaparib, and 4EGI-1).

Protein synthesis was assessed using puromycin labeling (SUnSET technique), as described previously (21). For in vitro labeling, puromycin (2 μg/mL, Thermo Fisher Scientific) was added to cells at a 1:400 dilution. The cells were then incubated with the puromycin for 1 hour before cell lysates were collected, as described in the Immunoblotting section. Protein loading for the immunoblots was normalized using cell number.

Cell proliferation was measured using the resazurin (AlamarBlue) fluorescence assay as described previously (5). The combination index (CI) as a measure of drug synergy was determined using the method of Chou and Talalay with five drug concentrations at a fixed dose ratio (22). The data were analyzed using the CompuSyn software (http://www.combosyn.com/).

Cells (1.5–3 × 10⁵) were plated one day prior to transfection in 6-well plates. Cells were transfected with siRNA using Lipofectamine RNAiMax (Thermo Fisher Scientific) as described previously (5). siCHK1.1, siCHK1.2, and siCHK1.3 were obtained from IDT. siCHK1.pool was a SMARTpool ON-TARGETplus reagent (GE Dharmacon) and siRRM2 was described previously (4). siControl was purchased from Cell Signaling Technology (#6568). siEWS-FLI1 (5′-GCAGAACCCUUCUUAUGACUU-3′) was synthesized by IDT (23).

A shERWOOD UltramiR Lentiviral Inducible shRNA plasmid targeting CHK1 (ULTRA342152) was obtained from Transomic Technologies. Lentivirus was produced by transfecting HEK-293T cells with the shRNA plasmid and packaging plasmids (psPAX2 and pMD2.G) according to the FuGENE 6 (Roche) protocol. For the lentiviral transduction, Ewing sarcoma cells were incubated with 2 mL of virus and 6 μg/mL of polybrene (Sigma-Aldrich) for 12–16 hours. Cells were selected in 1 μg/mL puromycin 48 hours after transduction.

The Institutional Animal Care and Usage Committee at the University of Iowa approved the animal studies. The studies were conducted in adherence with the NIH Guide for the Care and Use of Laboratory Animals. Approximately 1.0 × 10⁶ TC71 cells were mixed with 30% Matrigel and injected subcutaneously into the flanks of 6-week old, female NCr mice. When tumors were palpable (∼100–200 mm³), the mice were divided into cohorts (8–9 mice per cohort) and treated with drug or vehicle. Mice cohorts were treated with vehicle, gemcitabine (150 mg/kg, intraperitoneal, day 1), prexasertib (10 mg/kg, subcutaneous, twice daily, days 1, 2, 3 and 14, 15, 16), and the combination of gemcitabine and prexasertib (dosing as for the single agents). Tumor volumes were measured periodically using calipers (volume = 0.5 × length × width²). Animals were sacrificed when a tumor reached 20 mm in any dimension. GraphPad Prism was used to generate survival curves. For the in vivo puromycin labeling experiment, NCr mice were subcutaneously injected with two Ewing sarcoma cell lines (TC71 and EW8) and allowed to develop measurable tumors. The mice were then divided into four cohorts, with 2 mice per cohort, and treated with drugs, as described above. On day 2, the mice were injected with puromycin (0.5 mg) and then sacrificed 90 minutes later (24).

Immunoblots were performed as described previously (4, 5). Protein loading for the immunoblots was normalized using cell number. Antibodies to the following proteins were used in the immunoblots: phospho-Histone-139 H2A.X (Cell Signaling Technology, #9718, 1:1,000), phospho-Chk1-345 (Cell Signaling Technology, #2348, 1:1000), phospho-Chk1-296 (Cell Signaling Technology, #12302, 1:1,000), Chk1 (Cell Signaling Technology, #2360, 1:1,000), c-MYC (Cell Signaling Technology, #5605, 1:1,000), puromycin (Millipore, #AF488, 1:2,000), 4E-BP1 (Cell Signaling Technology, #9644, 1:1,000), PARP (Cell Signaling Technology, #9532, 1:1,000), cleaved caspase-3 (Cell Signaling Technology, #9664, 1:1,000), p-4E-BP1-37/46 (Cell Signaling Technology, #2855, 1:1,000), RRM1 (Cell Signaling Technology, #8637, 1:1,000), RRM2 (Santa Cruz Biotechnology, #398294, 1:500), CHK2 (Cell Signaling Technology, #6334, 1:1,000), S6 (Abcam, #40820, 1:1,000), P-S6-235/236 (Cell Signaling Technology, #4858, 1:1,000), and tubulin (Proteintech, 66031-1, 1:2,000).

Tumor xenografts were fixed in formalin for IHC staining. Immunostaining for Ki67 was conducted using a rabbit anti-human Ki67 antibody (Cell Signaling Technology, #9027, 1:400) on tumor sections. Staining was performed using the Vectastain ABC Kit (Vector Laboratories), according to the manufacturer's instructions.

Data for 4E-BP1 expression were obtained from the Cancer Cell Line Encyclopedia (Broad Institute, Cambridge, MA; https://portals.broadinstitute.org/ccle; ref. 25). A Mann–Whitney test was performed to compare the Ewing sarcoma cell lines to the other cell lines.

Cells (500 cells/well) were plated in triplicate in a 6-well plate and allowed to adhere overnight. The cells were then treated with prexasertib, or vehicle, for 6 hours. Prexasertib was then removed and the cells were washed three times with PBS. The cells were allowed to grow for approximately 10–14 days and then fixed and counted as described previously (5).

Cells (3 × 10⁵ cells/well) were plated in a 6-well plate and allowed to adhere overnight. The cells were then treated with drugs, or vehicle, for 6 hours. Flow cytometry for γH2AX was then performed as described previously (5).

Annexin V was measured by flow cytometry using a FITC Annexin V/Dead Cell Apoptosis Kit (Thermo Fisher Scientific; ref. 5).

Total RNA was extracted from cells using a RNeasy Kit (Qiagen) following the manufacturer's instructions. One microgram of total RNA was reverse-transcribed into first-strand cDNA using random hexamer primers and the SuperScript III Reverse Transcriptase (Invitrogen). qRT-PCR was performed on the ViiA 7 Real-Time PCR System (Life Technologies) using SYBR Select Master Mix (Thermo Fisher Scientific). Reactions were performed in triplicate and gene expression was normalized to GAPDH. The PCR primer sequences for CHEK1 and GAPDH were 5′-ATATGAAGCGTGCCGTAGACT-3′ and 5′-CTGGGCTACACTGAGCACC-3′, respectively.

Student t test was used to calculate P values for the comparison of two groups and analyses for more than two groups were conducted with a one-way ANOVA followed by Dunnett multiple comparisons test. P values for the tumor volume measurements in the xenograft experiment were determined by two-way ANOVA and the log-rank (Mantel–Cox) test was used to calculate P values comparing the mouse survival curves. Statistical analyses were conducted using GraphPad Prism 5.0.

Ewing sarcoma cell exhibit high levels of CHK1 protein (6). We used multiple siRNAs to knockdown CHK1 in Ewing sarcoma cells (Fig. 1A; Supplementary Fig. S1). However, the approximately 70% reduction in CHK1 protein caused by the siRNAs did not reduce the growth of the Ewing sarcoma cells, which is consistent with our previous data and work from Wayne and colleagues showing that the complete and sustained inhibition of CHK1 is necessary to inhibit the growth of sensitive cell lines (Fig. 1B; refs. 5, 9). Similar results were obtained using a doxycycline-inducible shRNA that targets CHK1 (Fig. 1C and D). Next, we used immunoblotting for the CHK1 auto-phosphorylation site Ser-296 to test the ability of three CHK1 inhibitors, LY2603618, MK-8776, and prexasertib, to inhibit CHK1 function (26–29). Figure 1E shows that the three CHK1 inhibitors caused a dose-dependent inhibition of CHK1-296 autophosphorylation when the cells were also treated with gemcitabine, which inhibits ribonucleotide reductase and activates CHK1 (5). Notably, prexasertib was significantly more potent at inhibiting CHK1 than either LY2603618 or MK-8776 and inhibited CHK1 activity at drug concentrations <1 nmol/L.

Inhibition of CHK1, in addition to impairing the response to DNA replication stress, also causes DNA replication stress by altering origin licensing, which results in the phosphorylation of CHK1 itself at Ser-345 via ATR (30). Figure 1F and G show that treating Ewing sarcoma cells with prexasertib resulted in the phosphorylation and activation of CHK1, which was blocked by cotreatment with the ATR inhibitor VE-822 (31). We then tested whether reducing CHK1 levels, using siRNA knockdown, would increase the sensitivity of Ewing sarcoma cells to the CHK1 inhibitors. Figure 1H and I show that reducing CHK1 levels in Ewing sarcoma cells using either siRNA or shRNA significantly sensitized the cells to concentrations of CHK1 inhibitors that blocked CHK1 activity (Fig. 1E), but did not impair cell growth as single agents. This reduction in cell growth was associated with cleavage of PARP and caspase-3, markers of apoptosis, in multiple Ewing sarcoma cell lines (Fig. 1J and K).

The mTOR pathway, which has been reported to regulate CHK1 levels, is known to be active in Ewing sarcoma cells (16, 17). To test whether the mTOR pathway regulates levels of CHK1 in Ewing sarcoma cells, we treated cell lines with a mTORC1/2 inhibitor (TAK-228) that was previously shown by Slotkin and colleagues to inhibit the growth of Ewing sarcoma cells in vitro and in xenograft experiments (17). Figure 2A shows that the treatment of three Ewing sarcoma cell lines, A673, EW8, and TC71, with TAK-228 caused a dose-dependent reduction in cell growth. Treatment of Ewing sarcoma cells with TAK-228 reduced protein levels of CHK1, but not the related protein CHK2 (Fig. 2B). On the basis of reports in the literature, as well as the sensitivity of Ewing sarcoma cells to inhibitors of RNR, we also tested whether inhibition of the mTOR pathway reduces levels of RRM1 and RRM2 (4, 19, 20). Figure 2B shows that inhibition of mTOR reduced levels of both RRM1 and RRM2, although the effect was larger for RRM2 than RRM1. TAK-228 also reduced CHK1 levels in three non-Ewing sarcoma cell lines, HT1080, RD, and U2OS, but did not alter CHK1 levels in HEK-293T cells (Fig. 2C).

The mTOR pathway regulates protein synthesis via multiple mechanisms (32). Consequently, we used puromycin labeling (SUnSET technique) to directly detect the synthesis of nascent proteins in Ewing sarcoma cells in the presence and absence of mTOR inhibition (21). Figure 2D shows that TAK-228 reduced protein synthesis in four Ewing sarcoma cell lines, similar to results obtained with the protein translation inhibitor cycloheximide (Supplementary Fig. S2). TAK-228 also reduced protein synthesis in several non-Ewing sarcoma cell lines with the exception of HEK-293T cells, similar to the results for CHK1 protein levels (Fig. 2E).

To delineate the impact of inhibition of mTORC1 versus mTORC1/2 we also tested an mTORC1 inhibitor (temsirolimus) and an additional ATP-competitive, mTORC1/2 inhibitor (Torin2; refs. 33, 34). Figure 2F shows that the mTORC1/2 inhibitors (TAK-228 and Torin2), but not the mTORC1 inhibitor temsirolimus, suppressed protein synthesis in the EW8 and TC71 Ewing sarcoma cell lines. We then evaluated the effect of the TAK-228 on the phosphorylation of two mTOR targets, 4E-BP1 and ribosomal protein S6 (S6), which are critical regulators of protein translation (32). Figure 2G shows that temsirolimus reduced phosphorylation of ribosomal protein S6, but not 4E-BP1. In contrast, the mTORC1/2 inhibitors TAK-228 and Torin2 inhibited the phosphorylation of both 4E-BP1 and S6. Similarly, the mTORC1/2 inhibitors TAK-228 and Torin2, but not temsirolimus, also reduced levels of CHK1. These findings are consistent with work that has shown that the ATP-competitive mTORC1/2 inhibitors suppress cap-dependent protein translation through 4E-BP1 protein dephosphorylation, while the effects of temsirolimus on this pathway are more variable (35–38).

Analysis of data generated by the Cancer Cell Line Encyclopedia project (CCLE, Broad Institute) demonstrate that Ewing sarcoma cell lines express high levels of 4E-BP1 mRNA, total 4E-BP1 protein, and phospho-4E-BP1 protein relative to other cancer types (Supplementary Fig. S3; ref. 25). Active (dephosphorylated) 4E-BP1 protein negatively regulates protein synthesis by binding the cap-binding protein eIF4E, inhibiting eIF4G binding, and preventing assembly of the eIF4F complex (39). To further evaluate the role of cap-dependent protein translation in the regulation of levels of CHK1, we treated Ewing sarcoma cells with 4EGI-1, a drug that disrupts the eIF4F complex and inhibits cap-dependent translation (40). Figure 2H shows that 4EGI-1 reduced total protein synthesis, as assessed using puromycin labeling, and specifically reduced levels of CHK1 in Ewing sarcoma and osteosarcoma cells. 4E-BP1 is known to regulate the translation of c-Myc, which has a well-described role in Ewing sarcoma tumorigenesis, and Fig. 2I shows that TAK-228 caused a dose-dependent reduction in the c-Myc protein (39). Finally, to test whether a gene transcription mechanism may also contribute to the regulation of CHK1 protein levels, we isolated mRNA from EW8 and TC71 cells treated with TAK-228 and performed RT-qPCR to measure CHEK1 mRNA. Figure 2J demonstrates that TAK-228 caused a small, approximately 20% reduction in CHEK1 mRNA in both EW8 and TC71 cells, suggesting that inhibition of protein synthesis is the primary mechanism regulating CHK1 levels.

TAK-228 was previously shown by Slotkin and colleagues to inhibit the growth of CHP100 Ewing sarcoma cells in vivo in xenograft experiments (17). Consequently, to identify additional drugs that regulate CHK1 levels in Ewing sarcoma, we turned our focus to the CHK1 inhibitor prexasertib that was reported by Sen and colleagues to also target the mTOR and protein synthesis pathways in small-cell lung cancer cells (27, 41). Figure 3A shows that treatment of Ewing sarcoma cells with prexasertib caused a decrease in phosphorylation (activation) of 4E-BP1 and a depletion of CHK1 protein in two Ewing sarcoma cell lines. Figure 3B and C show that prexasertib also caused a dose-dependent inhibition of protein synthesis in the Ewing sarcoma and osteosarcoma cells, respectively. We then evaluated the effect of prexasertib on the growth of Ewing sarcoma cells, as well as additional transformed (RH30, SAOS, HEK-293T, and HT1080) and nontransformed (BJ-tert and RPE-tert) cell lines. In a dose–response assay, Ewing sarcoma cells were sensitive to treatment with prexasertib (Fig. 3D) with IC₅₀ values in the low nanomolar range. In comparison, the other cancer cell lines showed variable sensitivity, while the nontransformed cells lines were relatively resistant to prexasertib (Fig. 3E). Prexasertib also significantly inhibited the growth of Ewing sarcoma cells, but not nontransformed cell lines, in long-term clonogenic assays (Fig. 3F; Supplementary Fig. S4). Similarly, treatment of Ewing sarcoma cells with prexasertib resulted in cleavage of PARP and an increase in cells positive for Annexin-V and propidium iodide, both markers of apoptosis (Supplementary Fig. S5). These results are consistent with work by Lowery and colleagues, and others, that has shown that a wide range of adult and pediatric cancer cell lines, including Ewing sarcoma, are sensitive to prexasertib (8, 27, 42).

Next, we tested whether prexasertib would increase the sensitivity of Ewing sarcoma cells to gemcitabine, which impairs DNA replication and is synergistic with the CHK1 inhibitor LY2603618 in Ewing sarcoma cells (5). Figure 3G shows that prexasertib (1 nmol/L) increased the sensitivity of EW8 Ewing sarcoma cells to gemcitabine, with an approximately 4-fold reduction in the IC₅₀ value. We also used the method of Chou and Talalay to calculate a combination index (CI) to test whether the combination of gemcitabine and prexasertib was synergistic (22). The combination of gemcitabine and prexasertib demonstrated synergism (CI < 0.9) in two Ewing sarcoma cell lines (EW8 and TC71) with CI values ranging from 0.25 to 0.80 (Fig. 3H). Consistent with this cell viability data, the combination of prexasertib (1 nmol/L) with gemcitabine (10 nmol/L) caused robust phosphorylation of H2AX, a marker of DNA damage and double strand breaks (Fig. 3I; ref. 43).

We then evaluated whether gemcitabine and prexasertib could inhibit the growth of tumor cells in a mouse xenograft experiment. NCr mice were subcutaneously injected with Ewing sarcoma (TC71) cells and allowed to develop measurable tumors. The mice were then divided into four cohorts and treated with either vehicle, gemcitabine, prexasertib, or the combination of gemcitabine and prexasertib. Figure 4A and B shows that there was a statistically significant difference in tumor volumes between the control and drug-treated groups. Notably, both prexasertib and the combination of gemcitabine and prexasertib resulted in tumor regression over the first 10 days of the experiment without apparent systemic toxicity (Fig. 4B; Supplementary Fig. S6). However, the effect of gemcitabine in combination with prexasertib resulted in a more sustained inhibition of tumor growth and prolongation of mouse survival compared with prexasertib alone (Fig. 4C). The median survival of the mice treated with vehicle, gemcitabine, prexasertib, and the combination of gemcitabine and prexasertib was 21, 27, 34, and 99 days, respectively.

Next, we used in vivo puromycin labeling to assess the effect of gemcitabine, prexasertib, and the combination of the drugs on protein synthesis in vivo in a xenograft experiment (24). NCr mice were subcutaneously injected with two Ewing sarcoma cell lines (TC71 and EW8) and allowed to develop measurable tumors. The mice were then divided into four cohorts and treated with drugs, as described above. On day 2, the mice were injected with puromycin and then sacrificed 90 minutes later. Figure 4D shows that the combination of prexasertib and gemcitabine reduced protein synthesis in vivo in the tumor xenografts. In addition, gemcitabine, and to a lesser extent prexasertib, also reduced protein synthesis in vivo. Similarly, Fig. 4E shows that the combination of prexasertib and gemcitabine caused phosphorylation of CHK1 and gH2AX, as well as cleavage of PARP and caspase-3, in vivo in the xenografts. The combination of gemcitabine and prexasertib also reduced expression Ki-67, a marker of cell proliferation (Fig. 4F).

On the basis of the unexpected finding that gemcitabine reduced protein synthesis in Ewing sarcoma cells in vivo, we then treated Ewing sarcoma (TC71) and osteosarcoma (U2OS) cells with gemcitabine in vitro and assessed protein synthesis using puromycin labeling. Figure 5A shows that treatment of Ewing sarcoma cells, but not osteosarcoma cells, with gemcitabine (100 nmol/L) reduced protein synthesis. Additional Ewing sarcoma cell lines (A673, TC32, and EW8) also showed a reduction in protein synthesis after treatment with gemcitabine (Fig. 5B). The effect of gemcitabine on protein synthesis in Ewing sarcoma cells was dose-dependent (Fig. 5C) and evident by the end of the 6 hour drug incubation (Fig. 5D). We then tested additional non-Ewing sarcoma cell lines, including BJ-tert, RPE-tert, and HEK-293T, and did not observe a reduction in protein synthesis after treatment with gemcitabine (Fig. 5E; Supplementary Fig. S7). To evaluate for a dose dependency effect in the non-Ewing sarcoma cells lines, we treated U2OS and RPE-tert cells with 10-fold and 100-fold higher concentrations of gemcitabine, but did not observe any reduction in protein synthesis (Fig. 5F). Finally, we used siRNA to knockdown the EWS-FLI oncogene in Ewing sarcoma A673 cells, which tolerate a reduction in EWS-FLI1 with a decrease in tumorigenicity but not viability, and observed that knockdown of EWS-FLI rescued protein synthesis in cells treated with gemcitabine (Fig. 5G; ref. 44). In addition, based on the xenograft results, we also tested the effect of the combination of gemcitabine and prexasertib on protein synthesis. Figure 5H shows that gemcitabine (10 nmol/L) and prexasertib (1 nmol/L or 2 nmol/L) also suppressed protein synthesis, in contrast to either drug as a single agent (at the specified doses; Figs. 3B and 5C).

On the basis of the high levels of expression of 4E-BP1 in Ewing sarcoma cells, as well as the activation of 4E-BP1 by mTORC1/2 inhibitors and prexasertib, we then tested whether gemcitabine results in activation of 4E-BP1 in Ewing sarcoma cells. Figure 6A shows that treatment of Ewing sarcoma cell lines (A673, EW8, TC32, and TC71) with gemcitabine resulted in an increase in the active (unphosphorylated) form of 4E-BP1, which suppresses protein translation. However, in contrast to the Ewing sarcoma cells, treatment of other cell types with gemcitabine did not alter 4E-BP1 phosphorylation, with the exception of a modest increase in the unphosphorylated form of 4E-BP1 in BJ-tert cells (Fig. 6B; Supplementary Fig. S8). Next, we tested whether inhibition of protein synthesis by gemcitabine would also reduce CHK1 levels. Figure 6C shows that treatment of Ewing sarcoma cells with gemcitabine caused phosphorylation of CHK1 at 6 hours, but a loss of total CHK1 protein at the 24-hour time point. In contrast, treatment of U2OS osteosarcoma cells with gemcitabine resulted in phosphorylation of CHK1 at 6 hours, but no change in total CHK1 protein levels. Figure 6D shows that gemcitabine caused a dose-dependent activation of 4E-BP1 and reduction in both CHK1 and c-Myc protein levels. In addition, the knockdown of EWS-FLI1 in A673 cells reversed the effect of gemcitabine on 4E-BP1 phosphorylation (Fig. 6E). Knockdown of EWS-FLI1 also caused a reduction in total 4E-BP1 levels, which may provide an explanation for the high levels of 4E-BP1 in Ewing sarcoma cells (Supplementary Fig. S3). In an approach complementary to gemcitabine treatment, we also used siRNA (siRRM2) to knockdown RRM2 in two Ewing sarcoma cell lines and identified a reduction in protein synthesis (Fig. 6F), as well as a loss of CHK1 (Fig. 6G) and c-Myc (Fig. 6H) proteins (4). Finally, to evaluate whether the inhibition of protein synthesis is a general stress response in Ewing sarcoma cells, we treated two cell lines with olaparib (PARP inhibitor) and U0126 (MEK inhibitor), drugs that are known to cause toxicity in Ewing sarcoma (45). Neither of these drugs, though, had an effect on protein synthesis or 4E-BP1 phosphorylation (Fig. 6I).

The treatment of Ewing sarcoma, in particular metastatic and recurrent disease, is challenging and requires highly intensive, cytotoxic chemotherapy in combination with surgery and/or radiation (1). We previously used a human embryonic stem cell model of Ewing sarcoma to identify that the combination of gemcitabine, an irreversible inhibitor of RNR, and a checkpoint kinase 1 (CHK1) inhibitor is synergistic in vitro and significantly prolongs mouse survival in xenograft experiments (4, 5, 46). In this work, we identified that high levels of the CHK1 protein in Ewing sarcoma cells limit the efficacy of CHK1 inhibitors. However, from a mechanistic standpoint, we discovered that inhibition of mTORC1/2 decreases levels of CHK1 protein via a reduction in protein synthesis. Inhibition of mTORC1/2 also reduced levels of the subunits of RNR, RRM1, and RRM2, in Ewing sarcoma cells. This reduction of RRM1, RRM2, and CHK1 levels by inhibition of mTORC1/2 may provide an explanation, in part, for the toxicity of mTOR inhibitors toward Ewing sarcoma cells, although other targets and mechanisms likely contribute as well (16, 17).

We identified a novel pharmacologic approach to modulate CHK1 levels in Ewing sarcoma cells using prexasertib. Prexasertib is an ATP-competitive, catalytic inhibitor of CHK1 that also reduces CHK1 protein levels by inhibiting protein synthesis in Ewing sarcoma cells (27). Although the decrease in CHK1 levels caused by prexasertib contributes to the synergy of this drug with gemcitabine, we also expect that the effect of prexasertib on overall protein synthesis is a critical factor in modulating the drug sensitivity and synergy. Notably, the combination of prexasertib and gemcitabine was synergistic in vitro, decreased protein synthesis in vivo in Ewing sarcoma xenografts, and significantly prolonged mouse survival in the xenograft experiment. We do note that gemcitabine is a nucleoside analogue that causes DNA chain termination, in addition to inhibiting the RRM1 subunit of RNR, and that this effect of the drug on cellular proliferation may also contribute to its mechanism of action.

Prexasertib is currently being tested in clinical trials in children and adults (47). Although the combination of prexasertib and gemcitabine has not been tested in a clinical trial, the combination of gemcitabine and other CHK1 inhibitors has been tested in several clinical trials (clinicaltrials.gov). The combination of prexasertib with other targeted agents, including WEE1, ATR and PARP inhibitors, may also offer a novel therapeutic approach for treating Ewing sarcoma tumors. For example, Sen and colleagues recently demonstrated that prexasertib improves the response to the PARP inhibitor olaparib in small-cell lung cancer xenograft models (41). Several groups have identified that Ewing sarcoma cells are sensitive to PARP inhibitors, which suggests that prexasertib in combination with a PARP inhibitor may warrant investigation (48).

We also identified that multiple drugs, including TAK-228, prexasertib, and gemcitabine, activate 4E-BP1 and inhibit protein synthesis in Ewing sarcoma cells. TAK-228 is an ATP-competitive inhibitor of mTORC1/2 and known activator of 4E-BP1 (17). Similarly, prexasertib was reported by Sen and colleagues to block mTOR signaling via direct inhibition of p90RSK (27, 41). However, the finding that gemcitabine and/or knockdown of RRM2 activated 4E-BP1 and inhibited protein synthesis in Ewing sarcoma cells was unexpected. But, other groups have reported similar findings. For example, Williams and colleagues demonstrated that gemcitabine activates 4E-BP1 in non–small cell lung cancer cells and Jacobson and colleagues showed that the toxicity of gemcitabine is enhanced by concurrent blockade of cap-dependent translation (49, 50). Similarly, Tee and colleagues identified that the treatment of cells with multiple DNA-damaging agents, including etoposide, mitomycin-C, and cisplatin, can activate 4E-BP1 and reduce protein synthesis (51). Overall, based on this published work and our data demonstrating that activation of 4E-BP1 is not a general response to gemcitabine in all cell types, we believe that the regulation of 4E-BP1 is likely a complex interplay between cell backgrounds, drug mechanisms, and oncogenic signaling pathways. Work is currently underway in our laboratory investigating how these factors regulate 4E-BP1 in Ewing sarcoma cells and, in particular, whether the high levels of 4E-BP1 mRNA and protein in Ewing sarcoma contribute to this phenotype.

We have not yet identified the upstream pathway that mediates the effects of gemcitabine on protein synthesis and 4E-BP1 activation in Ewing sarcoma cells, but inhibition of mTOR signaling can occur via multiple mechanisms (32). Interestingly, Zhou and colleagues recently showed that ciclopirox, which inhibits RNR and the growth of Ewing sarcoma cells in vitro and in vivo, inhibits mTOR signaling by activation of AMP-activated protein kinase (AMPK; ref. 52). In addition, Kruiswijk and colleagues identified that genotoxic stress can activate AMPK and inhibit protein synthesis via regulation of eukaryotic elongation factor 2 kinase (eEFK2; ref. 53). Consequently, we are currently investigating whether RNR inhibitors, including ciclopirox and gemcitabine, activate the AMPK/tuberous sclerosis/raptor pathways in Ewing sarcoma cells. In addition, although mTOR is the best known upstream regulator of 4E-BP1 phosphorylation, other kinases have also been reported to phosphorylate 4E-BP1, including CDK1, MAP Kinase p38, GSK3beta, PIM2, and ATM (54).

Similarly, alterations in multiple proteins and pathways that regulate protein synthesis have also been described in Ewing sarcoma tumors. Consequently, there may be underlying cellular abnormalities driven by EWS-FLI1 that predispose Ewing sarcoma cells to inhibition of protein synthesis. For example, Erkizan and colleagues first reported a critical role for RNA helicase A (RHA), which functions in mRNA translation, in Ewing sarcoma tumors (55). Similarly, Wilky and colleagues demonstrated that Ewing sarcoma cells express high levels of the RNA helicase DDX3, which is critical for ribosome assembly and the translation of mRNAs with complex 5′ untranslated regions (56). In addition, Y-box binding protein 1 (YB-1), which regulates the translation of specific mRNAs, is also overexpressed in Ewing sarcoma cells (57). Overall, the pleiotropic effects of the EWS-FLI1 oncogene may generate a cell state that is particularly prone to inhibition of protein synthesis in the setting of DNA replication stress, DNA damage, or other stressors.

In summary, we have identified that protein translation is a critical regulator of CHK1 levels and the response to DNA replication stress in Ewing sarcoma cells. Overall, we believe that our preclinical data provides a rationale testing the combination of gemcitabine and ATR–CHK1 pathway inhibitors in patients with relapsed and refractory Ewing sarcoma.

No potential conflicts of interest were disclosed.

Conception and design: S.L. Koppenhafer, W.W. Terry, D.J. Gordon

Development of methodology: D.J. Gordon

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.L. Koppenhafer, K.L. Goss, D.J. Gordon

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.L. Koppenhafer, D.J. Gordon

Writing, review, and/or revision of the manuscript: W.W. Terry, D.J. Gordon

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.L. Koppenhafer, K.L. Goss

Study supervision: D.J. Gordon

We thank Mohammed Milhem, Varun Monga, Munir Tanas, Rebecca Dodd, and Yasmeen Rose for helpful discussions. D.J. Gordon is supported by a University of Iowa Dance Marathon Award, a Holden Comprehensive Cancer Center Sarcoma Multidisciplinary Oncology Group Seed Grant, a University of Iowa Oberley Seed Grant, a University of Iowa Stead Family Research Award, The Matt Morrell and Natalie Sanchez Pediatric Cancer Research Foundation, and NIH grant R37-CA217910.

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.