Inhibition of ATR-Dependent Signaling by Protoapigenone and Its Derivative Sensitizes Cancer Cells to Interstrand Cross-link–Generating Agents In Vitro and In Vivo

Ataxia telangiectasia-mutated (ATM) and ATM and Rad3-related (ATR) are 2 members of the phosphoinositide 3-kinase–related protein kinase family that play a central role in DNA damage response (DDR) coordination; they also function in the signaling machinery of cell-cycle arrest, DNA repair and transcription, and cell death. Although ATM is predominantly activated in response to DNA strand breaks, ATR is activated in response to damage arising from UV rays or replication block; both kinases activate signaling cascades that involve 2 checkpoint kinase effectors, Chk1 and Chk2, whose roles were previously suggested to be redundant (1). In contrast to ATM, ATR has been reported to be indispensible for cell growth and for life. ATR knockout mouse embryos died early due to mitotic catastrophe characterized by incomplete DNA replication and chromosomal fragmentation (2, 3). Moreover, ATR gene mutations are rarely found in humans. The only mutation variants that can survive are heterozygous or hypomorphic variants. Furthermore, cells derived from patients with Seckel syndrome exhibit cellular features associated with ATR signaling defects. Consistent with this phenotype, Seckel-like mouse embryonic cells showed accelerated aging due to replicative stress, exhibiting an accumulation of lethal chromosomal breaks (4, 5). However, with regard to its role in regulating the replication checkpoint, ATR is activated by most cancer chemotherapeutic agents that target DNA in replicating cells. Therefore, inhibition of ATR signaling is a valid and promising strategy that can improve efficiency of chemotherapy or radiotherapy (6, 7). Thus far, several inhibitors of DDR-related kinases, including Chk1 and Chk2, have been successfully used alone or in combination with each other in clinical trials (8, 9). Recently, several chemicals that inhibit ATR kinase activity in vitro were used to support the hypothesis that ATR kinase can be targeted to improve cancer therapy (10–14). Because most of these studies are in their initial stages, it is imperative to focus more efforts toward investigating strategies to inhibit ATR signaling.

WYC02 (Fig. 1A) is a flavonoid that we previously isolated and identified from the whole plant extract of Thelypteris torresiana, a fern species native to Taiwan. This compound was originally screened for cytotoxicity function; using a colorimetric cytotoxicity assay, WYC02 showed therapeutic effects and was a lead compound for potential anticancer drug development (15–17). In previous studies, WYC02 and its more potent analog WYC0209 (Fig. 1A) were shown to induce oxidative stress, consequently activating the p38 and c-jun-NH₂-kinase (JNK)1/2 mitogen-activated protein kinase (MAPK) pathways following cell-cycle arrest and apoptosis in several cancer cell types. These compounds were also found to reduce the size of tumor xenografts in nude mice without exerting toxic effects on the recipient (18–22). Recently, WYCs were found to induce chromosomal breakage through oxidative stress (18, 20), implicating a role for WYCs in interfering with DNA metabolism. To date, the biomolecular actions and implications of this WYC-mediated interference are mostly undetermined. Here, we found that WYCs are capable of inhibiting DNA damage–induced activation of ATR targets Chk1 and Fanconi anemia group D2 protein (FANCD2), which then sensitize tumor cells to chemotherapy, and finally result in tumor size reduction in mice. Our results propose that these new flavonoid compounds are of noteworthy potential to treat cancers by inducing replication stress via the inhibition of ATR signaling.

MDA-MB-231 [breast adenocarcinoma; ATCC HTB-26, BCRC 60425] and A549 (lung adenocarcinoma; ATCC CCL-185, BCRC 60074) human cell lines were purchased from Bioresource Collection and Research Center (BCRC) and were authenticated by American Type Culture Collection (ATCC). U2OS (osteosarcoma), HeLa (cervical adenocarcinoma), and HEK 293T (embryonic kidney cells) human cell lines were provided by Dr. Sheau-Yann Shieh (Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan). Cells were maintained in Dulbecco's Modified Eagle's Medium (Sigma-Aldrich) supplemented with 10% FBS (Gibco). For DDR induction, freshly diluted H₂O₂ (Merck) was added to the culture medium 1 hour before the cells were harvested. For UV irradiation treatment, the cells were irradiated for 10 J/m² by a cross-linker (UVP CL-1000) 1 hour before analysis. WYC02 and WYC0209 were isolated and synthesized as described previously (15–17).

To evaluate in vitro chemosensitization, cells were seeded in 6-well plates 1 day before the experiment at a density of 100 to 400 cells per well. The drugs were incubated with the cells for 6 hours, after which the medium was replaced with fresh drug-free FBS-containing medium. The colonies became visible and were counted 7 to 10 days later with 0.1% crystal violet staining following image capture by a CCD camera (LAS 4000 mini; Fujifilm).

To evaluate the effect of DNA damage checkpoint activation on cell-cycle distribution, the cells were harvested at indicated time points and fixed with methanol for at least 2 hours. The DNA was then stained with a solution containing propidium iodide (PI) and RNase A (Sigma-Aldrich). Fluorescently labeled cells were subsequently analyzed by the flow cytometer (LSR II; BD Biosciences) with a suitable selection of excitation and emission wavelengths. The percentages of different fluorescent cell populations were analyzed by WinMDI version 2.9 (The Scripps Research Institute).

In brief, 5 × 10⁵ Chinese hamster ovary (CHO) cells were seeded in 60-mm dishes 1 day before the experiment. WYC02-induced structural chromosomal changes after 20 hours were compared with that of the cells cultured in 2 μmol/L mitomycin C (MMC). At 18 hours after WYC02, 0.1 μg/mL colchicine was added for 2 hours, and metaphase cells were collected by shaking them off the dishes. Mitotic cells were treated with 0.5% KCl for 10 minutes and fixed with a 3:1 mixture of methanol:glacial acetic acid. The cells were then spread on slides for chromosome staining with 5% Giemsa solution. We then analyzed the chromosome structure of 200 well-spread metaphase cells (100 metaphase cells per experiment) under a ×100 oil immersion objective.

The plasmids ATR, ATRIP, and claspin were kindly provided by Dr. X. Wu (The Scripps Research Institute, La Jolla, CA), and TopBP1 was provided by Dr. J. Chen (University of Texas MD Anderson Cancer Center, Houston, Texas). The siRNA sequences of the target ATM (5′-AAGCGCCTGATTCGAGATCCT-3′), ATR (5′-CCTCCGTGATGTTGCTTGATT-3′), DNA-PKcs (5′-GATCGCACCTTACTCTGTTGA-3′), and the random sequence that served as the control (5′-AAGTCAATATGCGACTGATGG-3′) were synthesized by Sigma-Proligo (23, 24). All transfections in HEK293T cells were carried out by the calcium phosphate precipitation method.

Cell lysate preparations, gel electrophoresis, and immunoblotting were carried out as previously described (23). The binding of primary antibodies were detected by horseradish peroxidase–coupled secondary antibodies (Jackson ImmunoResearch) followed by enhanced chemiluminescence (Millipore). The images of nonsaturated bands were captured by a luminescent image analyzer (LAS-4000 mini; Fujifilm). The antibodies used in this study are listed in Supplementary Materials.

DNA constructs of the recombination substrate pHPRT-DRGFP, in which the I-SceI site lies within 1 copy of 2 mutated tandem repeated GFP genes, and the I-SceI endonuclease expression vector pCBASceI, were originally constructed by Dr. M. Jasin (25). In brief, we generated a stable pHPRT-DRGFP construct in HeLa cells and evaluated the chromosomal breaks generated by I-SceI endonuclease expression. Six hours after pCBASceI was delivered into the cells, complete medium with or without WYC02 or WYC0209 was replaced onto the cells. Forty-eight hours after delivery, the efficiency of chromosomal homologous recombination repair (HRR) was obtained as the percentage of GFP-positive cells, which was assessed by flow cytometry.

Human breast cancer MAD-MB-231 cells were harvested from the culture, resuspended in medium without serum at 1 × 10⁸ cells/mL, and 0.1 mL of this suspension was subcutaneously injected into the right flank of female nude mice (BALB/cAnN-Foxn1nu/Crl Narl; purchased from the National Science Council Animal Center). Tumor-injected mice that successfully developed tumors that grew to approximately 50 to 100 mm³ in volume were randomly sorted into groups and used for the experiments. Control vehicle or 2 mg/kg of cisplatin with or without 0.2 mg/kg of WYC0209 was administered intraperitoneally every 4 days throughout the experiment. The drug formulations are described in Supplementary Materials.

Previously, WYC02 and WYC0209 were shown to cause DNA strand breaks and apoptosis in lung and prostate cancers (18, 20), suggesting that inducing DNA damage may be the potential mechanism underlying the anticancer effect of WYCs. To test this hypothesis, we investigated the cytogenetic effect of WYC02 on CHO cells (Fig. 1B). Low WYC02 concentrations produced dose-dependent increases in chromosomal structural changes, such as breakages, radials, and chromosomal polyploidy, similar to the effects seen with MMC treatment; however, the complete mitotic chromosome could not be obtained upon high-dose WYC02 treatment. Because MMC can induce DDR in many cancers, we investigated what kind of DDR signaling was activated by WYC02. Surprisingly, high doses of WYC02 in HEK293T cells did not induce noticeable changes in the putative DDR signaling, which we measured by analyzing the phosphorylation of the ATM-dependent Chk2 Thr⁶⁸ residue and the ATR-dependent Chk1 Ser³⁴⁵ residue (Fig. 1C). We did observe that WYC02 treatment caused slight accumulation of the p53 protein, which could have been the result of several posttranslational modifications. However, phosphorylation of the p53 Ser¹⁵ residue did not contribute to this WYC02-induced p53 protein accumulation, suggesting that WYC02 does not directly damage DNA because DNA damage normally stimulates ATM/ATR-dependent p53 Ser¹⁵ phosphorylation. Our result is similar to previous reports that p38 MAPK is activated by WYC02 (19, 21, 22), as its downstream target MAPKAPK2 was found to be phosphorylated starting as early as 2 hours after WYC02 exposure (Fig. 1C). We repeated the WYCs experiment on lung and breast carcinoma cell lines A549 and MDA-MB-231 cells, respectively, and obtained similar results. Consistently, no marked changes in Chk1 and Chk2 phosphorylation signaling were detected even at high doses of either drug for as long as 8 hours after drug treatment (Supplementary Fig. S1A). The cytotoxic effect by WYCs on cancer cells was determined by MTT assay at 48 hours of incubation (Supplementary Fig. S1B); our data indicated that the IC₅₀ value range for cytotoxicity was similar to those in previous reports, confirming that WYCs are stable compounds that do not directly cause DNA damage.

Understanding the mechanism by which the WYCs compounds cause chromosomal breakages and other abnormalities might aid in identifying their targets. We hypothesized that genes with functions associated with DNA damage checkpoints and/or DNA repair might be targeted by WYCs. To test this hypothesis, we assessed the effects of WYCs on DDR induced by H₂O₂. WYC02 was found to inhibit Chk1, but promote Chk2 phosphorylation in A594 cells treated with 0.1 mmol/L H₂O₂ for 2 hours; however, ATM autophosphorylation was not affected (Fig. 2A). Pretreatment of cells with okadaic acid (OA; a phosphatase inhibitor) or MG132 (a proteasome inhibitor) could not reverse the WYC02-induced inhibition of Chk1 phosphorylation, indicating that the inhibition does not occur due to phosphatase activation or proteasome degradation by other regulatory factors (Fig. 2B). Furthermore, we investigated other sources of DNA stimuli specific for ATR activation; our results show that UV-induced Chk1 phosphorylation was dose dependently inhibited by WYCs within different cells (Fig. 2C). In response to DNA double-strand breaks (DSB), FANCD2 is known to be monoubiquitinated on K561 (FANCD2-Ub) in an ATR-dependent manner to stimulate repair (26, 27). We showed that FANCD2-Ub was also inhibited by WYCs (Fig. 2A and C); furthermore, ATR inhibition by WYCs was also observed in cells treated with currently prescribed chemotherapeutic agents (Supplementary Fig. S2). Collectively, these findings indicate that WYCs can modify ATR signaling after various types of DNA damage. Interestingly, WYC0209 was more potent than WYC02 in inhibiting Chk1 phosphorylation and cytotoxicity (Fig. 2C and Supplementary Fig. S1B). We speculate that the replacement of 2 hydroxyl groups on WYC02 with an additional benzene ring contributes positively to this ATR inhibition; however, the definite pharmacophores need to be further investigated when the ATR protein structure is resolved.

To elucidate the specificity of the WYC inhibition on ATR-mediated signaling, we compared the change between cells treated with WYCs or the ATM-specific inhibitor KU55933 (Fig. 1A) before the induction of DDR. After H₂O₂ damage, ATM is thought to be the principal responder, and KU55933 treatment strongly inhibited ATM-mediated Chk2 phosphorylation specifically, but its effect on ATR-mediated Chk1 phosphorylation was small (Fig. 3A, top). In contrast, after hydroxyurea (HU; a replication blocker) damage, ATR is thought to be the principal responder, and WYCs treatment significantly inhibited Chk1 phosphorylation, but only slightly inhibited Chk2 phosphorylation (Fig. 3A, bottom). Using these pharmacologic methods, we showed that the specificity of DDR inhibition between WYCs and KU55933 was completely different. To strengthen the argument that WYCs specifically inhibit ATR signaling, small inhibitory RNAs against ATM, ATR, and the catalytic subunit of DNA protein kinase (DNA-PKcs) were introduced into HEK293T cells before exposure to UV or H₂O₂. Our results showed that WYCs completely inhibited UV- or H₂O₂-induced Chk1 phosphorylation in a manner similar to siRNA knockdown of ATR, but not ATM or DNA-PKcs (Fig. 3B and Supplementary Fig. S3). The siRNAs against ATM and DNA-PKcs decreased the UV- or H₂O₂-induced Chk2 phosphorylation, which were not altered by the addition of WYC02, but were increased by WYC0209 treatment. Interestingly, neither siRNA targeted to ATM or ATR nor DNA-PKcs affected the WYC0209-mediated increase in Chk2 phosphorylation. Because a high dose of WYC0209 itself slightly induces Chk2 activation (Supplementary Fig. S1), the increased Chk2 phosphorylation was likely a synergistic effect due to DNA damage. To further identify the specific mediator that contributes to the effect of WYC02 on the initiation of ATR kinase activation, we tested whether TopBP1, ATRIP, and claspin were involved, as they have been identified as mediators of ATR kinase activation (28). Our results showed that overexpression of ATRIP or TopBP1 did not reverse the inhibitory effect of WYC02 on Chk1 phosphorylation, whereas overexpression of claspin or ATR did (Fig. 3C), suggesting that WYC02 might affect the function of ATR or claspin contributing to ATR signaling inhibition.

Previously, it has been shown that S–M and G₂–M checkpoints are activated by ATR in response to different types of DNA damage (6, 7). Of these, the G₂–M checkpoint involves ATM and ATR in collaboration, whereas the S–M checkpoint is mediated solely by ATR. To maintain genetic integrity, ATR can prevent premature mitotic entry in the event of incomplete DNA replication or unrepaired DNA damage (6). To evaluate the effect of WYCs on these ATR-associated DNA damage checkpoints, we observed the effect of WYCs on mitotic entry following HU or cisplatin treatment. In MDA-MB-231 cells, HU and cisplatin significantly decreased the number of mitotic cells, indicating that the S–M and G₂–M checkpoints are intact in MDA-MB-231 cells (Fig. 4A). WYCs or KU55933 treatment increased the percentage of mitotic cells in cisplatin-treated cells, suggesting that all of these compounds inhibited the damage-induced G₂–M checkpoint. However, WYCs, but not KU55933, significantly increased the HU-induced mitotic entry that is specific for the S–M checkpoint, indicating that WYCs specifically impaired this distinctive checkpoint mediated solely by ATR (Fig. 4A).

ATR function is also linked to DNA repair via its coupled targets (29). To examine the effect of WYCs treatment on DNA repair, we carried out an HRR assay in HeLa cells. Our result showed that chromosomal breaks normally repaired by HRR were dose dependently inhibited by WYC02 at low concentrations. WYC0209 produced similar effects at doses that were 10-fold lower than that of WYC02 (Fig. 4B). From these results, we assumed that the cells carrying unrepaired DNA would enter into mitosis following DNA damage. To verify this assumption, we analyzed the DNA-damage marker γH2AX on mitotic cells using immunofluorescence staining. As expected, the numbers of large γH2AX foci were increased upon addition of WYCs in both unperturbed and perturbed mitotic cells (Fig. 4C and Supplementary Fig. S8), suggesting that WYCs increase DNA damage in mitotic cells. The chromosomes became flat and aggregated after WYC treatment, differing from the 3-dimensional and hair-like appearance of normal chromosomes at metaphase.

Inhibition of the checkpoint and repair mechanisms leads to chemosensitization in cancers. We questioned whether WYCs could function as sensitizers for the chemotherapeutic drug cisplatin that has been shown to induce ATR activation as well as FANCD2 monoubiquitination, which is the vital step for DNA crosslink repair (26, 30, 31). We found that WYCs treatment decreased the cisplatin-induced Chk1 phosphorylation and FANCD2 monoubiquitination in A549, MDA-MB-231, and U2OS cells (Fig. 5A). Using the same doses, WYC0209 not only inhibits monoubiquitination of FANCD2 but also affects FANCD2 protein stability; these data emphasize that WYC0209 has more potent inhibitory effects as than WYC02. We further treated individual cells with cisplatin in combination with several varying doses of WYCs and counted survival colonies to determine their ability to survive after cisplatin-induced damage. Our results showed that WYCs effectively decreased the clonogenic survival in cisplatin-treated MDA-MB-231 and A549 cells in the nanomolar dose range (Fig. 5B). To investigate the chemosensitization effect of low-dose WYCs in vivo, we established a tumor xenograft in nude mice with human MDA-MB-231 tumor cells, which are considered to be more resistant to cisplatin and are also sensitive to treatment with WYCs, at least as compared with A549 cells (Supplementary Fig. S1B and S4). When the mice were treated with 0.2 mg/kg WYC0209 in combination with 2 mg/kg cisplatin, the tumor inhibitory effect was greater than that of cisplatin treatment alone (Fig. 5C). However, WYC02 unexpectedly did not affect the cisplatin sensitivity of MDA-MB-231 tumors when a higher dose of 2 mg/kg was used in our experiments (data not shown). The pharmacokinetic data of WYC02 and WYC0209 needs to be compared in future studies to determine the differences in the chemical effects of these 2 compounds in vitro and in vivo.

Many chemotherapeutic agents kill cancers by interfering with DNA metabolism and lethally damaging the DNA. ATM and ATR are defined as the key checkpoint sensors in the DDR. Before the recent discovery of ATR inhibitors (10–14), ATM and Chk1/2 inhibitors were proven to efficiently enhance the sensitivity of cancers to IR therapy or chemical agents (8). Because the effects of flavonoids on ATR signaling have not yet been investigated, we consider the findings of our study to be novel. The synthetic derivative of WYC02, WYC0209, was found to have similar, yet more potent, inhibitory effects on ATR signaling that further validate our findings.

Low doses of WYCs significantly slowed cancer growth and caused S-phase delay and inhibition of DNA synthesis (Supplementary Fig. S5); these events are similar to previously reported ATR defects (32). WYCs treatment exhibited selective inhibition of ATR signaling, which we determined experimentally by different types of DNA-damage inducers. We clearly indicated that KU55933 specifically inhibits the ATM-mediated phosphorylation of Chk2 and WYCs specifically inhibit the ATR-mediated phosphorylation of Chk1 (Figs. 2A and 3A). Because the ATM activation that normally follows replication fork stalling or UV-induced irradiation damage depends on ATR kinase activity (33), the WYCs-mediated reduction in ATM activation upon HU, but not H₂O₂, treatment might result from ATR kinase dysfunction (Figs. 2A and 3A). This specificity was also supported by our observation that the WYC-mediated inhibition of DNA damage-induced Chk1 phosphorylation was dependent on ATR but not on ATM and DNA-PKcs (Fig. 3B and Supplementary Fig. S3). Furthermore, radial changes in chromosome structure are characteristics of cells derived from Fanconi anemia patients but not from patients with Ataxia telangiectasia (AT) or AT-like disease (34, 35); therefore, the WYC02-mediated inhibition of ATR-dependent FANCD2 activation links this effect to radial changes in chromosome structure from FA patients (Fig. 1B).

UV-induced Chk1 phosphorylation can be restored by overexpression of claspin and ATR but not TopBP1 or ATRIP (Fig. 3C). These data suggest that WYC02 might block the function of ATR kinase itself or the ATR–claspin–Chk1 interactions that then affect Chk1 activation. However, we could not exclude the possibility that WYCs target other mediators involved in ATR kinase activation. For example, FANCM has been shown to be necessary for both ATR activation and FANCD2 monoubiquitination (36). It remains unclear as to whether WYCs actually inhibit the ATR kinase enzyme activity or just abort the process of activation, and further investigations are required to clarify the precise mechanism(s) of action by WYCs. We screened a human kinome set for WYC02 involving 58 diverse in vitro kinase assays (Kinome Diversity Screen; MDS Pharma Services) and found that only PLK1 and SGK2 kinases were moderately inhibited by a high dose of WYC02 (Supplementary Fig. S6); therefore, it is clear that WYC02 does not target any of the kinases tested in this assay.

Our results supported 2 possible mechanisms by which WYCs could enhance cancer cell sensitivity to cisplatin: first, inhibition of the ATR checkpoint could lead cells that carry erroneous DNA into the next generation; second, inhibition of the FA repair pathway could result in elevated DNA DSB formation and error accumulation. We showed that WYCs compromise the cisplatin-induced checkpoint and lead to DSB accumulation within mitotic cells (Fig. 4A and C), suggesting that mitotic catastrophe might be induced in WYC-treated cells, whose DNA replication was not completed before entering into mitosis (37). Although ATM inhibition by KU55933 treatment rapidly accumulates damage in the late S and G₂-phase cells rather than in the M-phase cells (38), and induces cell death via apoptosis (39), WYC02 treatment mainly causes cells to undergo necrosis-like death rather than apoptosis in our system (Supplementary Fig. S7). These results indicate that the inhibition of ATM or ATR differentially cause accumulation of DNA lesions in cells undergoing diverse fates via different death pathways.

Previous data suggest that ATR and ATM play different roles in preventing DSB formation through HRR (40). ATR and its targets, such as Chk1, FANCD2, and WRN, are required to repair cells damaged during the S-phase via HRR (26, 29, 40, 41). We found that low doses of WYCs almost completely inhibited the HRR-mediated repair of chromosomal breaks, even in cells that possessed ATM function (Fig. 4B). Although ATM-deficient cells did not show reduced frequency of HRR to repair introduced DSBs (42), ATM-dependent FANCD2 Ser²²² phosphorylation is required for the intra-S checkpoint, whereas the ATR-dependent FANCD2 Lys⁵⁶¹ monoubiquitination is necessary for HRR (41). In the specific case of recovering from a stalled replication fork, we propose that ATM and ATR collaborate to induce checkpoint activation and recruit repair proteins for the initial step of HRR, whereas ATR signaling perhaps plays an indispensable role in the entire HRR process.

Our data revealed that WYCs could not only modify cancer cell responses to cisplatin but also to other therapeutic agents that induce Chk1 phosphorylation (Figs. 3A, 5A, and Supplementary Fig. S2), suggesting that WYCs may potentially function as cancer cell sensitizers for chemotherapeutic agents that activate the ATR checkpoint. Inhibition of DNA damage checkpoints and repair factors, such as ATM and BRCA1, are associated with promoting carcinogenesis as a consequence of genomic instability (43). However, there is no known cancer associated with ATR gene alterations. Specific ATR inhibition by expression of kinase-inactive ATR specifically in the skin prevents UV-induced skin carcinogenesis (44), suggesting that ATR inhibition has chemopreventive effects on nontransformed cells. Conversely, ATR has been proposed to play a critical role in directing cell death by senescence in unperturbed cells, where ATR is activated by overexpression of the oncoprotein Mos or the ATR activator TopBP1, respectively (45, 46). One plausible explanation proposed by Kawasumi and colleagues is that chemoprevention by ATR inhibition only works on precancerous cells in which oncogenes are not yet activated (44). In DNA-perturbed cells, the difference in the lesions induced by various DNA-damaging agents could also explain the different consequences of ATR/Chk1 inhibition, such as whether this inhibition promotes or suppresses apoptosis (47). However, while ATR activation clearly protects cells from death via suppression of apoptosis when DNA damage is caused by oncogene-induced replication stress or DNA cross-linking, limited ATR activity selectively kills these cells (48–50), probably mediated by an ATM-dependent mitotic catastrophe (51). Taken together, the genetic and pharmacologic evidence suggests that ATR inhibition potentially plays a dual role by achieving both chemoprevention and chemotherapy. Given that the efficiency of ATR inhibition causes the opposite effect of synthetic lethality or tumorigenesis on Ras-transformed cells (52), inhibition of ATR function for anticancer therapy should be carefully evaluated by considering the genetic background and responsiveness of the tumor to this type of therapy; furthermore, the efficiency of ATR inhibition should be cautiously monitored during treatment.

No potential conflicts of interest were disclosed.

The authors thank Drs. Chen-Yang Shen and Sheau-Yann Shieh for their support in providing some of their research material for this work and for their critical review of the manuscript.

This work was supported by grants NSC-97-2314-B-037-051 and NSC-99-2320-B-037-008 (H.-C. Wang), NSC 98-2311-B-037-001-MY3 (A.Y-L. Lee), NSC-99-2323-B-039-002 (Y.-C. Wu), and NSC 100-2325-B-037-002 (C.-C Wu) from the National Science Council of Taiwan; grant CA-100-PP-08 (A.Y-L. Lee) from the National Health Research Institutes of Taiwan; and grant KMU-Q098021 (H.-C. Wang) from the Kaohsiung Medical University Research Foundation. A. Hunyadi was supported by the Zoltán Magyary postdoctoral fellowship (EEA and Norway grants, 104/2009).

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.