Endogenous Replication Stress Marks Melanomas Sensitive to CHEK1 Inhibitors In Vivo

Cell-cycle checkpoints are potential therapeutic targets in melanoma (1). CHEK1 is a cell-cycle checkpoint effector that is activated in response to DNA damage and replication stress to block cell-cycle progression and apoptosis, and regulate DNA replication and repair (2, 3). It is activated in response to many chemotherapeutic agents, and inhibiting CHEK1 activity strongly enhances the potency of a range of chemotherapeutic drugs, particularly drugs such as gemcitabine that enhance replication stress (4). CHEK1 inhibitors (CHEK1i) are being investigated in clinical trials to enhance the efficacy of chemotherapeutic drugs, primarily gemcitabine (5, 6).

CHEK1i have been reported not to have broad activity as single agents, although they have been shown to be potent anticancer agents in melanoma, Myc-driven lymphoma, and a small number of other cancer cell lines that have been reported to have high levels of endogenous replication stress (7–10). CHEK1i sensitivity is strongly linked to replication stress and replication catastrophe (7, 11). Replication stress is a common feature of cancer, a product of oncogene-induced senescence bypass and mechanisms that drive replication through damaged DNA (12, 13).

CHEK1i are thought to act as chemosensitizing agents by abrogating the G₂–M phase checkpoint resulting in mitotic catastrophe and cancer cell death (2, 14). However, considering the numerous roles that CHEK1 plays in cell cycle and damage response pathways, this mechanism alone may not represent the entirety of sensitivity to CHEK1i as chemosensitizers or as single agents.

Inhibition of CHEK1 in conditions of high replication stress induced by gemcitabine or millimolar concentrations of thymidine or hydroxyurea, which deplete the dNTP pools, results in a massive increase in the numbers of replication origins fired, RPA exhaustion, and DNA strand breaks (11, 15, 16). While high levels of replication stress induced by drugs such as hydroxyurea and gemcitabine are well characterized, there is a lack of clarity as to what defines high levels of endogenous replication stress, and how this might contribute to CHEK1i sensitivity as a single agent. Here we have investigated the molecular basis of sensitivity to CHEK1i as a single agent in a panel of melanoma cell lines hypersensitive to single-agent CHEK1i.

A panel of melanoma cell lines, A02, A04, A15, A2058, BL, C002, C003, C32, C011, C012, C013, C025, C038, C045, C052, C054, D04, D20, D22, D24, D25, D28, D35, D41, HT144, MM127, MM170, MM200, MM329, MM370, MM383, MM415, MM426, MM466, MM576, MM603, MM604, MM648, MM96L, SKMEL13 and SKMEL28, primary neonatal foreskin fibroblasts (NFF), primary adult human melanocytes (HEMa; Life Technologies) and melanoblasts (QF1597, QF1610, QF1618 and QF1619) were assessed in this study. All melanoma cell lines were kindly provided by Prof. Nick Hayward (QIMR Berghofer Medical Research Institute, Brisbane, Queensland, Australia) and A, C, and D series lines were originally sourced from the Australasian Biospecimen Network (Oncology).

All melanoma cell lines were cultured as described previously (17). NFF cells were cultured in DMEM (Sigma Aldrich) containing 10% FBS (Bovogen). Human melanoblasts were cultured as described (18). HEMa cells were cultured in Medium 254 supplemented with HMGS-2 (Life Technologies). The identity of all cell lines was confirmed by short tandem repeat fingerprinting, and used within 3 months of thawing and were confirmed mycoplasma free. Tumorspheres (TS) were grown in suspension culture in tissue culture flasks coated with the biocompatible hydrogel pHEMA (poly-hydroxyethylmethacrylate), in media described previously (19) without β-mercaptoethanol. Nucleoside supplementation used EmbryoMax Nucleosides (Merck). CHEK1 inhibitors, GNE-323 (7) and GDC-0575 (20) were provided by Genentech. Dose–response experiments were performed as described previously (7). For TS dose–response experiments, cells were seeded and allowed to form spheres for 24 hours before drug treatment for 72 hours. Cell viability values for TS were determined by CellTiter Glo 3D cell viability assay in triplicate. IC₅₀ values were calculated for each of the cell line. The surviving fraction was calculated from the 10 μmol/L drug treatment. The data are the mean and SD of triplicate determinations.

Cell pellets were lysed and immunoblotted as described previously (7). Membranes were probed with antibodies against RPA2, Cyclin E1, and RRM2 (Santa Cruz Biotechnology), pRPA2 Ser4/8, pRPA2 Ser33 (Bethyl Laboratories), cleaved PARP, pCHEK1 Ser317, γH2AX, E2F1, (Cell Signaling Technology), CDC25A (Abcam), PCNA (DAKO), and α-tubulin (Sigma Aldrich). Proteins were visualized using chemiluminescence detection.

Cells were transduced using the lentiviral plvEIG (pLV411) vector expressing CDC25A, Cyclin E1, or the empty vector control as described previously (21). The transduced cells were cultured for 2 weeks then sorted by their GFP fluorescence. Stably transduced cells were investigated for sensitivity to GNE-323 by dose–response experiments as above. For 5-ethynyl-2-deoxyuridine (EdU), pRPA2 Ser/8, or γH2AX staining, cells were treated with 2 mmol/L hydroxyurea, 0.5 μmol/L camptothecin or 1 μmol/L GNE-323 for 24 hours and then labeled for 2 hours with EdU prior to fixation and staining of incorporated EdU as described in ref. 22. Plates were scanned on InCell 2000 (GE), images were then analyzed using the Cell Profiler analysis software (23), and data processed in R Studio (https://cran.r-project.org/).

For immunofluorescence imaging of TS and xenograft sections, these were formalin fixed and paraffin embedded, then sectioned and dewaxed and rehydrated as per standard protocols. Antigen retrieval was performed in a decloaking chamber (Biacore) using 10 mmol/L tri-sodium citrate buffer (pH 6.0) at 125°C for 5 minutes. Sections were probed with mouse Ki67 and rabbit pRPA2 Ser4/8 antibodies and appropriate secondary antibodies, and DAPI for DNA. The stained sections were imaged on an Olympus FLUOVIEW 3000 Confocal microscope and the images were then analyzed using the Cell Profiler (23) and R Studio.

Cells were treated with either 1 μmol/L GNE-323 or control (DMSO) and then images were taken using Zeiss Axiovert 200M Live Cell Imager with a 37°C incubator and 5% CO₂. Images were captured at 20-minute intervals with a minimum of 100 cells per condition per cell line analyzed. The number and timing of apoptosis relative to mitosis was quantified by visual inspection and was scored as described previously (24).

All animal experiments were performed according to the guidelines of the Australian and New Zealand Council for the Care and Use of Animals in Research and was approved by the University of Queensland Animal Ethics number UQDI/049/14/CA. Female nude BALB/c mice were injected with 2–3 × 10⁶ melanoma cells in Matrigel (BD Biosciences) by subcutaneous injection on the hind flank. Once tumors reached approximately 100 mm³, mice were treated with CHEK1i or vehicle (0.5% w/v methylcellulose and 0.2%v/v Tween 80) by oral gavage for 3 cycles where one cycle is three consecutive days of treatment followed by four rest days. Tumor size was measured three times per week using calipers. Mice were sacrificed at up to 6 weeks after terminating the treatment or when tumor size measured >1 cm³.

We have previously reported that melanoma cells were sensitive to killing by the CHEK1i GNE-323 (7), and now increased the panel to 45 cell lines including four normal melanocyte/melanoblast lines. CHEK1i were cytotoxic in the lines with IC₅₀ ≤ 150 nmol/L and surviving fractions <20% after 72-hour treatment with 10 μmol/L GNE-323, and cytostatic in cell lines (including normal melanocytes, melanoblasts, and fibroblasts) with IC₅₀ > 700 nmol/L, with >40% surviving fraction (Fig. 1A). We generated TS cultures from 20 melanoma cell lines and tested them for sensitivity to the CHEK1i, GDC-0575 (20), a clinical lead compound that has a toxicity profile similar to GNE-323 (Supplementary Fig. S1). Most of the cell lines with IC₅₀ ≤ 150 nmol/L to GNE-323 remained sensitive to cytotoxic effects of GDC-0575 when grown as TS (IC₅₀ ≤ 250 nmol/L, surviving fraction ≤25%), but 13 of 20 lines grown as TS were relatively insensitive with IC₅₀ > 1 μmol/L and surviving fraction of >40% (Fig. 1B). None of the major melanoma mutations were selectively enriched in the hypersensitive lines, nor was there any enrichment for defects in the p53 pathway previously suggested to be a sensitizing factor for CHEK1i (ref. 2; Supplementary Table S1). The cell lines with low IC₅₀ and low surviving fractions in each culture condition will be referred to as hypersensitive, and the other lines are defined as insensitive in this article.

CHEK1i sensitivity of TS in vitro was found to translate in vivo. Lines hypersensitive as TS retained sensitivity to the drugs when grown as xenografts in immunocompromised mice and treated with 50 mg/kg GNE-323 (D20 and MM96L) or 50 mg/kg GDC-0575 (C002) or initially 25 mg/kg and then 50 mg/kg GDC-0575 (C045; Fig. 1C). The C045 data demonstrated that GDC-0575 was active at 25 mg/kg as a single agent, but the efficacy was improved at the higher drug dose (Fig. 1C). The drugs effectively blocked tumor growth in the D20 and C002 xenografts, and the effect was maintained for at least 10 days after the final dose was administered. MM96L xenografts were the least responsive, although they still showed significant growth reduction with treatment. Lines insensitive as TS retained insensitivity to the drugs when grown as xenografts in immunocompromised mice and treated with 50 mg/kg GNE-323 (MM370) or 50 mg/kg GDC-0575 (A2058) using the same dosing schedule as above (Fig. 1D). Increased γH2AX-stained nuclei were found in all drug-treated tumors, indicating the drug penetrated all tumors (Fig. 1E). Therapy was well tolerated and there was no effect on mouse body weights (Supplementary Fig. S2A), and CHEK1i treatment did noticeably reduce the level of Ki67 staining (Supplementary Fig. S2B). Together, these data indicate that melanoma lines hypersensitive as TS retain sensitivity to GNE-323/GDC-0575 as a single agent in vivo.

Several studies have demonstrated that CHEK1i drives cells into an aberrant mitosis, and this is thought to trigger cell death (10, 14, 25, 26). We have previously reported that CHEK1i-sensitive melanomas accumulate with S-phase DNA content (7), which was a common feature of the hypersensitive cell lines. In contrast, insensitive melanomas and melanoblasts had a relatively unaffected cell-cycle distribution with drug treatment (Fig. 2A). Using time-lapse microscopy, the fate of cells and the timing of cell death relative to mitosis were assessed over 160-hour treatment with 1 μmol/L GNE-323. Cell death was readily detected by membrane blebbing, a feature of apoptosis (Fig. 2B). Addition of GNE-323 to hypersensitive melanoma cell lines resulted in delayed progression into mitosis, and cell death before, during or after mitosis (Fig. 2B). In the less sensitive cell lines, a proportion of cells either failed to enter mitosis or remained viable after mitosis (Supplementary Fig. S3). In the hypersensitive cell lines, the majority of cells died prior to mitosis, whereas in the insensitive cell lines, the majority of cells died either during or after mitosis (Fig. 2C). Inhibiting entry into mitosis using the CDK1 inhibitor RO-3306 had little effect on GNE-323–induced death in the hypersensitive cell lines, but effectively inhibited death in the insensitive cell lines (Fig. 2D). This indicates that the hypersensitive cells do not require transit through mitosis to trigger apoptosis. Thus, melanomas hypersensitive to GNE-323 die without entering mitosis or undergoing mitotic catastrophe, suggesting that the S-phase effects of CHEK1is are sufficient to trigger cell death.

GNE-323 treatment promoted apoptosis in the hypersensitive lines, indicated by the increased cleaved PARP1, a marker of apoptosis. CHEK1 inhibition or depletion in cells treated with hydroxyurea or DNA-damaging agents to promote high levels of replication stress triggers hyperphosphorylation of the RPA2 subunit of the single-stranded DNA (ssDNA) binding RPA complex by DNA-PK, ATM, and ATR (15, 27–29). The accumulation was less prominent in the insensitive cell lines, including primary fibroblasts and melanoblasts. Phosphorylation of the DNA-PK–dependent RPA2 Ser4/8 sites was observed in the untreated hypersensitive lines and accumulated to higher levels in the hypersensitive lines with CHEK1i treatment. The ATR-dependent RPA2 Ser33 site accumulated in the hyperphosphorylated form in the CHEK1i-treated cells, paralleling the accumulation of the pRPA2 Ser4/8, being more prominent in the hypersensitive lines, MM96L was the exception. The slower migrating hyperphosphorylated form of RPA2 (the band migrating with apparent molecular size of 36 kDa; Fig. 3A) accumulated after treatment with GNE-323.

Robustly increased γH2AX levels were found with GNE-323 treatment in the hypersensitive lines, with a weaker accumulation in the insensitive lines. The accumulation of γH2AX corresponded to the increase in cleaved PARP (Fig. 3A). The same differential effects on RPA2 hyperphosphorylation and γH2AX levels between the hypersensitive and insensitive lines were also found in the corresponding TS (Fig. 3B).

The ribonucleotide reductase subunit RRM2 is downregulated in response to CHEK1i (30, 31), which was thought to be the mechanism by which CHEK1i generates the same level of replication stress as treatment with 2 mmol/L hydroxyurea that rapidly depletes dNTP pools. Hydroxyurea treatment increased and GNE-323 treatment reduced RRM2 levels in both hypersensitive and insensitive cell lines including the primary melanoblasts (Fig. 4A). One of the regulators of RRM2 levels is E2F1, which is upregulated in response to replication stress in a CHEK1-dependent manner (32). All cell lines showed some decrease in E2F1 levels with GNE-323 treatment (Fig. 4A). Similar changes in RRM2 levels were observed in the TS cultures, although the changes in E2F1 levels were more muted (Fig. 4B).

The reduced RRM2 levels did not appear to affect replication in the insensitive cell lines, as the percentage of cells actively replicating and the fluorescence intensity of EdU incorporated into the replicating cells was not significantly changed from the controls. In contrast, there was a reduction of the proportion of cells actively replicating and/or efficiency of replication in the GNE-323–treated hypersensitive cell lines, indicated by the proportion of EdU-positive cells and fluorescence intensity of EdU in these cells. The total EdU incorporation (calculated as the product of the proportion of EdU-labeled cells and intensity of EdU labeling) was to approximately 50% of control levels in GNE-323–treated hypersensitive cells (Fig. 4C).

The reduced EdU incorporation in the hypersensitive cell lines with GNE-323 treatment suggested that the cytotoxic effects of GNE-323 was simply due to the loss of RRM2 reducing dNTP levels in the hypersensitive lines below levels that were sufficient for viability; the insensitive lines being more tolerant of the low dNTP levels. However, while addition of nucleosides was sufficient to reduce the level of CHEK1 activation and hyperphosphorylated RPA2 and γH2AX in 2 mmol/L hydroxyurea-treated cells, indicating that nucleoside supplementation effectively reversed the effects of dNTP depletion, it had only a minimal effect on these measures of replication stress or PARP cleavage in GNE-323–treated cells (Fig. 4D). Supplementation with up to 5-fold more nucleosides was no more effective (data not shown). This indicated that depletion of the dNTP pools was also only a minor contributing factor to the replication stress induced by GNE-323 treatment, but had little influence on the cell killing observed.

The elevated levels of replication stress markers in the CHEK1i-hypersensitive lines without evidence of reduction in DNA replication suggested that the S-phase cell-cycle checkpoint mechanisms normally triggered by replication stress might be defective. The S-phase checkpoint was triggered in cell lines using 2 mmol/L hydroxyurea or the TOPOI inhibitor Camptothecin (33), and checkpoint activation and arrest assessed using EdU incorporation and the presence of RPA foci. We observed RPA foci in all hydroxyurea-treated cell lines, and replication was inhibited in CHEK1i-insensitive cell lines indicated by the loss of EdU incorporation. In contrast, in CHEK1i-hypersensitive cell lines EdU incorporation continued in the presence of hydroxyurea (Supplementary Fig. S4). EdU incorporation in hydroxyurea and camptothecin-treated cells was quantified in a panel of melanoma cell lines using high-content imaging. Hydroxyurea and camptothecin treatment resulted in the expected reduction in the proportion of actively replicating cells only in the CHEK1i-insensitve cell lines, demonstrated by the reduced percentage (0%–7.5%) of EdU⁺ cells; Fig. 5A and B). In contrast, CHEK1i-hypersensitive cell lines continued incorporating EdU with hydroxyurea and camptothecin treatment, although the intensity of labeling was decreased (Fig. 5A and B). The reduced labeling intensity was a consequence of the depletion of dNTP pools with hydroxyurea treatment and slower replication with camptothecin treatment. The lack of a checkpoint arrest was not due to failure of checkpoint activation, indicated by the RPA foci (Supplementary Fig. S4) and robust activation of CHEK1 in all cell lines assessed (Supplementary Fig. S5). Loss of the checkpoint is also indicated by the enhanced ability of low-dose hydroxyurea (0.2 mmol/L), to promote micronucleus formation in the CHEK1i-hypersensitive cells, but had no effect on the proportion of cells with micronuclei in the insensitive lines (Fig. 5C). Micronucleus formation is a feature of cells with under-replicated DNA (34, 35). The significant increase micronuclei in the S-phase–defective hypersensitive cells demonstrates that they can progress into mitosis without completing replication, whereas checkpoint functional insensitive cells complete replication prior to mitosis.

Loss of the S-phase checkpoint arrest correlated with CHEK1i hypersensitivity. To determine whether to loss of checkpoint arrest was responsible for the hypersensitivity, inhibition of WEE1 which should mimic loss of the checkpoint was assessed. WEE1 inhibitors have been reported to increase sensitivity to CHEK1i (36). The WEE1 inhibitor MK-1775 synergized with GNE-323 in a small panel of CHEK1i-insensitive melanoma lines. Interestingly, the degree of synergy appeared to be dependent on the initial sensitivity of the cell line to GNE-323 (Fig. 5D).

Stable overexpression of CDC25A or the downstream drivers of S-phase progression, Cyclin E (37) in the CHEK1i-insensitive D28 line produced a modest increase in sensitivity to GNE-323 IC₅₀ reducing from 550 to 205 and 380 nmol/L, respectively (Supplementary Fig. S6A). The level of overexpression of either CDC25A or Cyclin E1 was modest, and insufficient to overcome the S-phase checkpoint arrest (Supplementary Fig. S6B and S6C). CDC25A overexpression modestly increased the level of pRPA2 Ser4/8 and γH2AX with GNE-323 treatment. A similar modest effect and failure to overcome the S-phase checkpoint was observed with CDC25A and Cyclin E1 overexpression in another insensitive line C013 (data not shown). Thus, dysregulated expression of the S-phase checkpoint regulators in CDC25A pathway appear to be insufficient for overcoming the S-phase, and have only a minor effect on sensitivity to CHEK1i.

In vivo GNE-323 treatment elevated pRPA2 Ser4/8 levels in hypersensitive and insensitive xenografts harvested 4 hours after the final drug treatment (Supplementary Fig. S7A) as we have seen in vitro (Fig. 3). Inspection of the pRPA2 Ser4/8 staining indicated that untreated hypersensitive xenografts had high numbers of foci compared with the insensitive MM370 xenograft (Fig. 6A), suggesting this might be a marker of high level of replication stress and correlated with sensitivity to CHEK1i in vivo. Quantitative imaging of sections of untreated xenografts stained for pRPA Ser4/8, and Ki67 as a marker of proliferation, using immunofluorescence, was performed. This revealed the hypersensitive xenografts had increased numbers of pRPA2 foci (Fig. 6B) and intensity of foci staining (Supplementary Fig. S7B) than the insensitive MM370. A similar effect was observed in the untreated TS, with the hypersensitive lines having higher numbers of pRPA foci and/or more proliferating cells (Ki67 positive) cells than the insensitive TS lines (Fig. 6C). The lower proliferative rate may in part explain changes in CHEK1i sensitivity in TS culture conditions.

Here we have investigated the hypersensitivity of a panel of melanoma cell lines to the selective and potent CHEK1i GNE-323 and GDC-0575. Comparison of a panel of CHEK1i demonstrated that GDC-0575 was significantly more potent in promoting DNA damage, replication stress and cell death than V158411, LY2603618, and MK-8776 (26). We have previously demonstrated that the selectivity of GNE-323 is based on its inhibition of CHEK1 rather than any significant off-target effect (7). The hypersensitivity we observed in vitro using both conventional cell culture and a modified tumor sphere culture model corresponds to in vivo sensitivity to GNE-323 and GDC-0575 as single agents.

It is currently thought that CHEK1 inhibition results in S-phase DNA damage, and also overcomes the cell-cycle arrest triggered by the this damage to promote mitotic catastrophe, which is the driver of cell death with CHEK1i treatment (10, 11, 14, 25, 26, 38, 39). In contrast, we found that melanoma cells hypersensitive to GNE-323 died in S-phase, whereas it was the less sensitive cell lines that required transit through mitosis for efficient killing. The latter was likely to be through a mitotic catastrophe-dependent mechanism as reported previously. Thus, the hypersensitive melanoma cell lines represent a subset of melanomas that are critically dependent on CHEK1 for survival during S-phase progression. The effect of CHEK1i in the hypersensitive cells is similar to the effect of CHEK1i treatment with exogenously imposed replication stress such as high-dose hydroxyurea. In both cases, this triggers RPA exhaustion, indicated by the strong accumulation of hyperphosphorylated RPA2 marking the chromatin associated RPA complex, and accompanied by increased γH2AX (15). It is likely that CHEK1 is essential for the viability of hypersensitive cells through its role in regulating the activation of new replication factories. In cells with modestly increased replication stress imposed with low-dose hydroxyurea, CHEK1 is essential in limiting the number of replication origins firing and thereby blocking RPA exhaustion (40). The question is what promotes this increased endogenous replication stress? CHEK1i treatment itself can promote replication stress by downregulation of RRM2 (31), through either downregulated expression due to E2F6-dependent repression (32), or direct CDK2-mediated destabilization of RRM2 (31, 41), thereby depleting dNTP pools. However, similarly reduced levels of RRM2 were found in both hypersensitive and insensitive cell lines with CHEK1i treatment, although the effect on replication was only observed in hypersensitive lines. Supplementation with nucleosides did not significantly reduce sensitivity to CHEK1i, indicating that reduced dNTP pools is not a major component of the cytotoxic action of CHEK1i in the hypersensitive melanomas. Reduction in RRM2 levels and dNTP pools has also been reported to promote senescence (42); however, there was no evidence of senescence with CHEK1i treatment.

Several groups have demonstrated that the primary effect of CHEK1i, either in cells with high levels of endogenous or exogenously applied replication stress is through a CDC25A-CDK2–dependent mechanism to promote replication fork collapse and S-phase DNA damage (10, 14), DNA double-strand breaks produced by action of MRE11 and MUS81 (16, 43). Here we have also found that overexpression of CDC25A and Cyclin E modestly increase CHEK1i sensitivity, as did inhibition of WEE1. The enhanced CHEK1i sensitivity with WEE1 inhibitor or overexpression of S-phase checkpoint regulators indicates that promoting S-phase progression increases sensitivity to these drugs. However, this by itself did not appear to promote hypersensitivity to CHEK1i. We have found that the CHEK1i hypersensitive melanomas have a defective S-phase checkpoint arrest, and loss of the checkpoint function was strongly correlated with increased replication stress. This is likely to be through the failure to cell-cycle arrest when replication defects are encountered, demonstrated by the increased micronucleus formation when the checkpoint-defective cells were grown under conditions of modestly increased stress. It is the increased replication stress rather than loss of the checkpoint itself that is responsible for the increased sensitivity to CHEK1i. It was surprising that overexpression of CDC25A or Cyclin E1 did not overcome the S-phase checkpoint, suggesting that simple dysregulation of this pathway may not be sufficient for the defective checkpoint. Likewise, CHEK1 was normally activated and RPA foci formed with exogenously applied replication stress, indicating that normal S-phase checkpoint activation occurs, but cells are unable to maintain the checkpoint arrest. The molecular defects underlying the defective S-phase checkpoint remain elusive.

In summary, the CHEK1 inhibitors GNE-323 and GDC-0575 have activity against a subset of melanomas with elevated levels of endogenous replication stress. This hypersensitivity translates to sensitivity in xenograft models, suggesting that CHEK1i may have clinical utility for patient's tumors with elevated levels of replication stress. In melanomas, the elevated replication stress was strongly correlated with a defective S-phase checkpoint, the defective checkpoint likely to be responsible for the elevated replication stress. Inhibiting CHEK1 in these cells enhances replication stress, but importantly triggers excessive replication origin firing, RPA exhaustion, DNA damage, and cell death (Fig. 6D). Insensitive lines are protected by the relative absence of endogenous replication stress. CHEK1i have shown activity as single agents in early-stage trials (6, 44), but the identification of the patients with elevated replication stress and most likely to benefit from this class of drug, is still required.

No potential conflicts of interest were disclosed.

Conception and design: Z.Y. Oo, A.J. Stevenson, N.K. Haass, B. Gabrielli

Development of methodology: Z.Y. Oo, A.J. Stevenson, D. Škalamera, S.A. Ainger

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z.Y. Oo, A.J. Stevenson, M. Proctor, S.M. Daignault, S. Walpole, C. Lanagan, J. Chen, S.A. Ainger, R.A. Sturm, B. Gabrielli

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z.Y. Oo, A.J. Stevenson, M. Proctor, S.M. Daignault, L. Spoerri, N.K. Haass, B. Gabrielli

Writing, review, and/or revision of the manuscript: Z.Y. Oo, A.J. Stevenson, M. Proctor, S.M. Daignault, S. Walpole, D. Škalamera, L. Spoerri, S.A. Ainger, R.A. Sturm, N.K. Haass, B. Gabrielli

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z.Y. Oo, A.J. Stevenson, M. Proctor, C. Lanagan

Study supervision: A.J. Stevenson, D. Škalamera, L. Spoerri, N.K. Haass, B. Gabrielli

The authors acknowledge the assistance of Dr. Sandrine Roy (TRI Microscopy facility). This work was funded by grants to B. Gabrielli from Cancer Australia and Cancer Council Queensland. B. Gabrielli, M. Proctor, and D. Škalamera are supported by funding from Smiling for Smiddy and the Mater Foundation.

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.