Abstract
The transient opening and closing of tumor vasculature result in periods of severe oxygen deprivation (hypoxia) followed by reoxygenation. This exerts a positive selective pressure for cells that have lost their sensitivity to killing by reduced oxygen levels. These cells are effectively resistant to hypoxia-induced apoptosis and conventional therapeutic approaches. Here we show hypoxia-induced S-phase arrest results in regions of single-stranded DNA in stalled replication forks and signals the activation of ATR. S-phase cells represent the most sensitive phase of the cell cycle to the stress of hypoxia/reoxygenation. Loss of ATR or inhibition of ATR kinase activity results in a further loss of reproductive viability in S-phase cells when exposed to hypoxic conditions followed by reoxygenation but has little effect on the inhibition of DNA synthesis. This is, at least in part, mediated via Chk1 signaling because loss of Chk1 also results in increased sensitivity to hypoxia/reoxygenation. The observed decrease in reproductive survival is in part because of the accumulation of DNA damage in S-phase cells during hypoxia exposure in the absence of full ATR activity. Therefore, ATR acts to protect stalled replication forks during hypoxia exposure. In conclusion, ATR and Chk1 play critical roles in the cellular response to hypoxia/reoxygenation, and inhibitors of ATR and Chk1 represent new hypoxic cell cytotoxins.
INTRODUCTION
Regions of hypoxia occur within solid tumors because of an inadequate, inefficient, and generally poorly formed blood supply. The level of hypoxia within a tumor correlates with a poor patient prognosis because these tumors are resistant to treatment and have more aggressive genotypes (1, 2, 3, 4) . Unlike normal vasculature, those vessels found in tumors often contain dead ends, arterial to venous shunts, and temporary blockages. Therefore, during tumor expansion, hypoxia within tumors is inevitably followed by periods of reoxygenation (5) . The range of oxygen concentrations within tumors varies from ∼8% for those cells closest to the capillary to anoxia (0%) for those cells >150 μm from the capillary. Exposure of tumor cells to anoxic levels results in the phosphorylation of p53 and histone H2AX by the ATM- and Rad-3–related kinase ATR (6) . Both of these molecules are phosphorylated in the hypoxic regions of tumors, consistent with previous findings that oxygen levels within tumors can reach these low levels (7) . ATR has been found to respond by relocalizing in the cell nucleus to form foci in response to agents that cause a replication stress (e.g., hydroxyurea, aphidicolin, and hypoxia; ref. 8 ). Of these stresses, hypoxia is undoubtedly the most physiologically relevant because it occurs during tumorigenesis and normal embryonic development (3 , 9) . In contrast, ATM is activated primarily in the presence of DNA breaks and is mutated in people with ataxia telangiectasia syndrome.
Exposure to oxygen levels close to anoxia (0.02%) causes a rapid S-phase arrest, which leads to ATR foci formation and the phosphorylation of downstream target molecules, such as p53, histone H2AX, and Chk1. These phosphorylation events are not altered or delayed by the absence of ATM but are inhibited by the loss of ATR. Whereas hypoxia-mediated S-phase arrest occurs in the absence of DNA damage that can be detected by comet assay, cells that are reoxygenated accumulate significant amounts of damage.
In this study, we show the presence of single-stranded regions of DNA in the nuclei of cells exposed to hypoxia and propose that these act as an initiating signal for ATR activity. Loss or inhibition of ATR or Chk1 activity results in increased sensitivity to hypoxia/reoxygenation as assessed by apoptosis and clonogenic assays. Cells exposed to severe hypoxia do not sustain comet-detectable DNA damage unless they also are reoxygenated. However, when ATR activity is inhibited, through the overexpression of an ATR dominant negative molecule, S-phase cells accumulate DNA damage in response to hypoxia and in the absence of any reoxygenation. These findings indicate the importance of the ATR and Chk1 kinases in the response to the physiologically relevant stress of hypoxia/reoxygenation.
MATERIALS AND METHODS
Cell Lines and Transfections.
The U2OS cell line with inducible ATRkd (clone GK41) was grown in DMEM supplemented with 10% FCS, 200 μg/mL geneticin, and 50 μg/mL hygromycin. To induce ATRkd expression, doxycycline was added to a final concentration of 1 μg/mL for between 36 and 48 hours. All of the derivatives of the HCT-116 cell lines were maintained in McCoy’s media supplemented with 10% FCS. The RKO cell line was maintained in DMEM with 10% FCS (10) . Transfections with small interfering RNA (siRNA) were carried out using the oligonucleotide to Chk1 described previously, sequence AAU CGU GAG CGU UUG UUG AAC (Dharmacon, Lafayette, CO; ref. 11 ). Oligofectamine was used to transfect the siRNA as described by the manufacturers. Cells were used 48 hours after transfection.
Hypoxia Treatment.
All of the hypoxia treatments were carried out in an anaerobic chamber (<0.02% O2; Sheldon Corp., Cornelius, OR). Cells were plated in glass dishes.
Precipitation of [3H]-Thymidine–Labeled DNA with Trichloroacetic Acid (TCA).
This assay was carried out as described previously (12) . All of the experiments were carried out in triplicate.
Colony Survival Assay.
Between 300 and 500 cells were plated in 4-cm glass dishes and allowed to adhere for 1 to 2 hours. Cells then were exposed to hypoxia, after which they were replaced in normal tissue culture incubators. The dishes were left untouched for 10 to 14 days. Colonies were stained with crystal violet and counted. Each condition was carried out in triplicate.
Comet Assay.
Comet assays were carried out as described previously (12 , 13) . For experiments using synchronized U2OS cells, a mitotic shake was carried out 12 hours before the start of hypoxia treatment to generate S-phase cells or 3 hours before treatment to isolate G1-phase populations.
Immunoblot Analysis.
Cells were lysed in buffer UTB [9 mol/L urea, 75 mmol/L Tris-HCl (pH 7.5), and 0.15 mol/L β-mercaptoethanol]. To disrupt the DNA, cells were sonicated briefly. After quantification, 50 μg of total cell protein were electrophoresed on 7.5% polyacrylamide gels. The primary antibodies used were p53 D0–1, Chk1 sc-8408, Chk2 sc-17748, Rad-17 sc-5613 (Santa Cruz Biotechnology, Santa Cruz, CA) and phospho-p53 Ser-15 #9286 (Cell Signaling Technology, Beverly, MA).
Detection of Single-Stranded DNA.
RKO cells were grown on chamber slides in normal tissue culture media with 30 μmol/L BrdUrd for 25 hours. During this time, cells were kept entirely in the dark. Cells were treated as indicated. To visualize incorporated BrdUrd, cells were fixed in methanol (20°C) for 10 minutes and blocked for 30 minutes at room temperature in 20% heat-inactivated goat serum, 0.1% BSA, and 0.1% sodium azide in PBS. The primary antibody, Anti-BrdU (Becton Dickinson, Franklin Lakes, NJ), was used as described by the manufacturer. To visualize BrdUrd incorporated into single-stranded regions only, cells were not denatured (i.e., the acid denaturation step common to most BrdUrd protocols was omitted; ref. 14 ).
RESULTS
Single-Stranded Regions of DNA in Cells Exposed to Hypoxia.
The recent report that ATR was found at regions of single-stranded DNA (ssDNA) at stalled replication forks along with replication protein A (RPA) and ATR interacting protein (ATRIP) suggested to us that hypoxia-induced replication arrest could result in regions of ssDNA (15) . To investigate this, RKO cells were grown on slides in the presence of BrdUrd for 24 hours in complete darkness. During this time, each cell underwent cell division and so incorporated BrdUrd. Cells were treated with hypoxia or hydroxyurea to inhibit DNA synthesis (16 , 17) . Fixation of hypoxic samples was carried out under hypoxic conditions with equilibrated solutions. Treated and untreated cells then were stained with an anti-BrdUrd antibody. In these experiments, the DNA was not denatured; therefore, only ssDNA would be visualized. A denatured sample also is shown (Fig. 1) ⇓ . In the normoxic controls, the cells exhibited little positive staining. Any staining that was apparent was cytoplasmic and caused by either labeled RNA or mitochondrial DNA. In contrast, cells treated with either hypoxia or hydroxyurea showed dense nuclear punctate staining. This indicated that like hydroxyurea, hypoxia gave rise to regions of ssDNA that are likely to be at the sites of stalled replication forks. These data tell us not only about the potential ATR trigger in response to hypoxia but also about the mechanism of arrest. It has been unclear whether hypoxia-induced replication arrest was the result of a failure of origins to fire or the stalling of previously active replication forks. These data indicate that the replication arrest is caused by failure of replication forks to progress. Unlike hydroxyurea, this was not found to be because of a lack of ribonucleotides (18) .
Hypoxia-mediated replication arrest results in the accumulation of single-stranded regions of nuclear DNA. RKO cells were grown on chamber slides under normal incubation conditions in the presence of 30 μmol/L BrdUrd in complete darkness. After 24 hours, cells were exposed to hypoxia or 1 mmol/L Hu for 12 hours. Cells then were fixed and stained with anti-BrdUrd antibody without a DNA denaturation step unless otherwise indicated.
The replication arrest induced by severe hypoxia suggests that S-phase cells may be more sensitive to hypoxia treatment than cells in other phases of the cell cycle. To investigate this, we carried out colony survival assays after hypoxia treatment with cells enriched for either G1- or S-phase cells. Because of the replication arrest induced by severe hypoxia, cells will not form colonies under hypoxia treatment. Therefore, to analyze the sensitivity to hypoxia, cells must be reoxygenated after treatment. G1- and S-phase populations were generated from an asynchronous population of RKO cells using the mitotic shake-off technique (12) . Immediately after shake off at least 75% of the cells were doublets, indicating they had completed mitosis and were in the G1 phase of the cell cycle. The relative survival of both cells in G1 and S phase after hypoxia treatment is shown (Fig. 2A) ⇓ . After exposure to hypoxia, S-phase cells were significantly more sensitive than the G1 cells. We have shown previously that cells exposed to severe hypoxia do not accumulate DNA damage detectable by alkaline comet assay. However, when hypoxic cells are reoxygenated, they acquire significant DNA damage. We hypothesized that S-phase cells might accumulate more reoxygenation-induced DNA damage. To address this, G1- and S-phase cells were exposed to hypoxia, reoxygenated, and comet assays were carried out (Fig. 2B) ⇓ . Comet assays show that the S-phase cells accumulate more damage than G1-phase cells, indicating that they are more sensitive to the reoxygenation event. These findings, combined with the discovery of increased ssDNA in hypoxia-treated cell nuclei, support the hypothesis that severe hypoxic/reoxygenation stress is primarily cytotoxic to S-phase cells.
S-phase cells are more sensitive to hypoxia/reoxygenation. A, The mitotic shake-off technique was used to generate populations of RKO cells enriched for either G1- or S-phase cells. In each case, the cells were plated at ∼300 cells/5-cm glass dish and allowed to adhere for a 2-hour period before being exposed to hypoxia (0.02%) for the periods indicated. After hypoxia exposure, the cells were returned to normal incubation conditions. This experiment was carried out in triplicate. B, The RKO cell line again was used to generate S- and G1-phase populations. Cells were exposed to hypoxia for 12 hours. Some were harvested under entirely hypoxic conditions, whereas others were reoxygenated, and comet assays were carried out. Without reoxygenation, neither population had undergone DNA damage. In contrast, after reoxygenation both populations showed increased levels of damage, but the S-phase cells were significantly more damaged than the G1-phase cells.
Loss of ATR Activity Decreases Survival after Hypoxia/Reoxygenation.
Because of the lethality of homozygous ATR null cells, research into ATR-mediated signaling has had to rely on either systems with decreased levels of ATR or the overexpression of an ATR dominant negative molecule (8) . We have made use of both of these alternatives. Because of the presence of one intact ATR allele in the HCT-116 ATR−/flox cell line, the levels of ATR protein were found to be diminished when compared with the wild-type cell line. The levels of ATR and p53 protein are shown in genetic variants of the HCT-116 cell line; the levels of Rad-17 protein in each cell line also are shown as loading control (Fig. 3A) ⇓ . We investigated whether the heterozygote ATR cell line might exhibit hypoxia-induced S-phase arrest with slower kinetics. Using thymidine incorporation as a means to assess DNA synthesis, the ATR heterozygote cell line arrested with similar kinetics to the wild-type version (Fig. 3B) ⇓ . These experiments also were carried out with BrdUrd and gave similar results (data not shown). These findings suggested that ATR was not required for the replication arrest induced by hypoxia, although the lack of truly ATR null cells makes this difficult to determine definitively. Similar experiments showed that ATM also was not required for hypoxia-induced replication arrest (data not shown).
Inhibition of ATR leads to increased sensitivity to hypoxia/reoxygenation. A, Western blot showing the levels of ATR and p53 protein in untreated HCT-116 wild-type (wt), HCT-116 ATR−/flox, and HCT-116 p53−/−. B, Thymidine incorporation assay, HCT-116 wt, and HCT-116 ATR−/flox were exposed to hypoxia for the periods indicated. This experiment was carried out in triplicate. C, HCT-116 wt, HCT-116 ATR−/flox, and HCT-116 p53−/− were plated at a density of 300 cells/5-cm glass dish, and a colony survival assay was carried out in triplicate after hypoxia/reoxygenation. D, HCT-116 wt, HCT-116 ATR−/flox, and HCT-116 p53−/− were used for a colony survival assay after ionizing radiation. E, Cells taken from a person with Seckel syndrome show increased sensitivity to hypoxia/reoxygenation. Primary fibroblasts were taken from an individual with Seckel syndrome and a normal matched control. The cells were plated at a density of 150/5-cm glass dish and treated as indicated. F, Overexpression of the ATRkd molecule during hypoxia/reoxygenation leads to increased sensitivity. Doxycycline was added to the cells for a 36-hour period before hypoxia exposure.
To investigate the consequences of hypoxia/reoxygenation treatment, we analyzed cell survival by the colony formation assay. Three genetic variations of the HCT-116 cell line (ATR−/flox, p53−/−, and wild type) were exposed to hypoxia/reoxygenation, and colonies were allowed to form during a 10-day period (Fig. 3C) ⇓ . In contrast to the p53 null cell line, which showed no increased sensitivity to hypoxia/reoxygenation over the wild-type cell line, the ATR heterozygotes were consistently more sensitive to the treatment. We did not use adenoviral Cre to remove the remaining ATR allele from the HCT-116 ATR−/flox cell line for colony forming ability after exposure to hypoxia/reoxygenation because removing ATR completely leads to cell death of cells in culture (19) . To establish that this finding was specific to hypoxia/reoxygenation stress, the same cell lines were exposed to ionizing radiation, and a colony formation assay was carried out. In this case, the decreased amount of ATR observed in the ATR heterozygote cell line had no effect on the dose-survival relationship seen after treatment with ionizing radiation (Fig. 3D) ⇓ . To support this novel finding, we used other systems in which ATR levels were either compromised or inhibited. Patients with Seckel syndrome recently were identified to have a splicing mutation in ATR, leading to a dramatic reduction of the ATR protein level within the cells of affected individuals (20) . The relative survival after hypoxia/reoxygenation was compared in primary fibroblasts from a Seckel patient or a normal matched control (Fig. 3E) ⇓ . The Seckel syndrome fibroblasts were significantly more sensitive to hypoxia/reoxygenation than the normal controls. This again shows that one of the consequences of decreased ATR protein levels is increased sensitivity to hypoxia/reoxygenation. Finally, we used a stable U2OS cell line (U2OS-GK41), which in response to doxycycline expresses an ATR dominant negative molecule (ATRkd; ref. 21 ). U2OS-GK41 cells were treated with doxycycline to induce ATRkd before plated for a colony survival assay. Cells expressing ATRkd were more sensitive to hypoxia/reoxygenation than the wild-type controls (Fig. 3F) ⇓ . These data not only support our previous findings but also indicate that it is specifically the loss of ATR kinase activity that leads to increased sensitivity to hypoxia/reoxygenation.
The decreased survival after hypoxia/reoxygenation observed in cell lines lacking ATR activity could be caused by an increase in the apoptosis rate. HCT-116 cells were exposed to hypoxia for 16 hours, followed by reoxygenation for the periods indicated before the rate of apoptosis was determined morphologically by Hoechst/propidium iodide staining (Fig. 4) ⇓ . In our study, the parental HCT-116 cell line has been found resistant to p53-dependent hypoxia-induced apoptosis during a 24-hour period despite having wild-type p53. Immediately after hypoxia treatment, the cell lines (wild-type, p53−/−, and ATR−/flox) showed similar amounts of apoptosis that were not increased over background levels. However, after reoxygenation, the ATR−/flox cells showed significantly more apoptosis than either the wild-type or p53−/− cell lines. We observed similar results when the cells were treated for shorter periods of hypoxia (8 hours; data not shown). Interestingly, we did observe a small amount (15%) of p53-dependent apoptosis after reoxygenation.
Loss of one ATR allele increases apoptosis induced by hypoxia/reoxygenation. The HCT-116 wild-type (wt), HCT-116 ATR−/flox, and HCT-116 p53−/− cell lines were exposed to hypoxia for 16 hours, after which they were returned to normal culture conditions for the periods indicated. Apoptosis was determined by nuclear morphology using Hoechst and propidium iodide.
Loss of ATR Activity Leads to DNA Damage during Hypoxia.
On the basis of our findings, we predicted that the lack of ATR during exposure to extreme hypoxia might result in DNA damage. To test our hypothesis, we again used the U2OS-GK41 cell line with inducible ATRkd. Cells were exposed to hypoxia, and comet assays were carried out. Care was taken to ensure that cells were not reoxygenated during preparation. The median tail moments are shown in Fig. 5A ⇓ . Cells exposed to ionizing irradiation were used as a reference. As reported previously, cells exposed to hypoxia showed no increase in tail moment (7) . We find that this was irrespective of the overexpression of ATRkd. However, a more detailed examination of the results indicated that in hypoxic samples with the presence of ATRkd, the distribution of tail moments was increased. This indicated that these populations represented a mixture of cells, some damaged and some not. On the basis of earlier survival curves, we hypothesized that the S-phase cells might be accumulating damage, whereas the others did not. To investigate this further, we again used the mitotic shake-off technique to generate synchronized populations of cells in G1 and S phase. These cells were exposed to hypoxia, and comet assays were carried out. Synchronizing cells in G1 and S phase made it possible to detect DNA damage specific to S-phase cells overexpressing ATRkd (Fig. 5B) ⇓ . The S-phase cells expressing ATRkd accumulated DNA damage in response to hypoxic treatment, whereas those expressing only wild-type ATR or those in the G1 phase of the cell cycle did not. Significantly, we also observed a higher incidence of apoptosis in hypoxia-treated S-phase cells in the presence of ATRkd (26%) compared with hypoxia-treated S-phase cells in the absence of ATRkd (11%). These data support the hypothesis that if ATR is compromised or decreased during hypoxia, DNA damage results. These findings and those of other investigators suggest a critical role for ATR in the preservation of stalled replication forks.
Inhibition of ATR activity by overexpression of ATRkd leads to DNA damage in S-phase cells. A, Asynchronous populations of U2OS-GK41 were treated with doxycycline to induce ATRkd and then exposed to hypoxia for 16 hours or 6 Gy ionizing radiation. Comet assays were carried without reoxygenation in the case of the hypoxia-treated samples. B, S-phase and G1-phase populations were generated and exposed to hypoxia for 16 hours. Comet assays then were carried out without reoxygenation.
Loss of Chk1 Also Leads to Increased Sensitivity to Hypoxia/Reoxygenation.
We investigated whether given the increased sensitivity to hypoxia/reoxygenation in the presence of reduced ATR activity, some of the downstream targets of ATR also would show an increased sensitivity. We chose to study Chk1 because it is phosphorylated in response to hypoxia in an ATR-dependent manner and is an essential gene (12 , 22) . To study the role of Chk1 in hypoxia, we used siRNA to reduce the levels of Chk1 within HCT-116 cells. Cells were transfected and allowed to recover, and the levels of Chk1 were assessed by Western blot analysis at the times indicated; a representative Western blot is shown in Fig. 6A ⇓ . Chk1 protein levels were reduced by at least 60% at 24 hours after transfection and remained reduced at this level for at least 72 hours. The protein levels of the related kinase Chk2 and glyceraldehyde-3-phosphate dehydrogenase remained unchanged throughout the experiment, indicating that the Chk1 siRNA is specific in action. Chk1 recently has been shown to play a role in S-phase arrest after treatment with ionizing radiation (11) . We have determined that RKO cells with levels of Chk1 reduced but not abolished by siRNA underwent hypoxia-induced S-phase arrest with identical kinetics to untransfected cells (Fig. 6B) ⇓ . HCT-116 cells transfected with Chk1 siRNA were used to assess sensitivity to hypoxia/reoxygenation by colony formation assay (Fig. 6C) ⇓ . Reducing the level of Chk1 protein significantly increased the sensitivity of HCT-116 cells to hypoxia/reoxygenation treatment. We also observed this effect in the RKO cell line (data not shown). This increased sensitivity was apparent even after relatively short treatment times; for example, after 4 hours of hypoxia, the viability of the untransfected cells had decreased by 30%, whereas those treated with Chk1 siRNA showed a 40 to 50% loss of viability (Fig. 6D) ⇓ . Having seen an increased rate of apoptosis in the HCT-116ATR/flox cell line after hypoxia/reoxygenation, we investigated the rate of apoptosis in cells with reduced levels of Chk1 after hypoxia/reoxygenation. HCT-116 wild-type cells were transfected with Chk1 siRNA and exposed to 12 hours of hypoxia, followed by 7 hours of reoxygenation. The rate of apoptosis was determined by Hoechst/propidium iodide staining in cells exposed to hypoxia alone, hypoxia and reoxygenation, and normoxia (Fig. 6E) ⇓ . We found a significant increase in the amount of apoptosis in cells with reduced Chk1 protein levels after hypoxia alone compared with wild-type cells, and this increased with reoxygenation. We have shown previously that p53 is not phosphorylated at residue 20 in response to hypoxia, indicating that p53 is not a Chk1 target during hypoxia exposure (18 , 23) . As expected, inhibiting Chk1 with specific siRNAs also led to an accumulation of cdc25A (24) . However, the levels of cdc25A protein did not appear to change during hypoxia exposure (data not shown). These data show that the ATR-mediated protective effect against the stress of hypoxia/reoxygenation is at least in part because of signaling via Chk1.
Reduction of Chk1 protein levels by siRNA results in increased sensitivity and apoptosis in response to hypoxia/reoxygenation. A, Western blot showing the levels of Chk1, Chk2, and as a loading control, glyceraldehyde-3-phosphate dehydrogenase, in HCT-116 wild-type (wt) cells after transfection with Chk1 siRNAs. Cells were transfected using Oligofectamine and harvested at the time points shown. B, Thymidine incorporation assay showed that cells transfected with Chk1 siRNA still underwent hypoxia-induced replication arrest, the kinetics of which did not significantly differ from the untreated cells. C, HCT-116 wt cells were treated with Chk1 siRNA, allowed to recover, and then plated for a colony survival assay. This was carried out in triplicate. D, examples of the colonies seen after 4 hours of hypoxia, followed by reoxygenation in HCT-116 wt cells and HCT-116 wt cells transfected with Chk1 siRNA. E, HCT-116 wt cells were transfected with Chk1 siRNA, allowed to recover, and then plated along with untreated cells as a control. Cells then were exposed to hypoxia for 12 hours or 12 hours of hypoxia followed by 7 hours reoxygenation before the amount of apoptosis was determined by Hoechst/propidium iodide staining.
DISCUSSION
The mechanism by which S-phase cells arrest in response to severe hypoxia remains unclear and is most likely multifactorial. We have shown previously that hypoxia-induced arrest of DNA synthesis does not occur as a result of lack of ribonucleotides (18) . We also have shown conclusively that ATR responds to severe hypoxia and that there is a complete correlation between ATR activity and oxygen levels low enough to induce replication arrest. However, before this study, we have not been able to determine what feature of hypoxia-induced replication arrest leads to the relocalization and activation of ATR. This problem was compounded by our finding that there is no DNA damage associated with hypoxia-induced arrest, at least that can be detected by comet assay. Here we show that regions of ssDNA are clearly visible in the nuclei of hypoxia-treated cells. The presence of ssDNA in hypoxia-treated cell nuclei gives novel insight into the mechanism by which ATR is activated in response to severe hypoxia and also the replication arrest itself. Hypoxia-induced replication arrest previously has been attributed to a stalling of replication forks and a failure to initiate replication from origins (25 , 26) . If hypoxia-induced replication arrest were entirely caused by the inhibition of firing from replication origins, regions of ssDNA would not be generated. The formation of regions of ssDNA during hypoxia strongly suggests the presence of stalled replication forks. Although this does not, however, rule out an additional and separate hypoxia-induced inhibitory effect on replication origins, this finding is supported by those of Zou and Elledge (15) , who showed the binding of ssDNA by RPA and the ATR binding partner, ATRIP. We predict that further studies will show a colocalization of RPA/ATRIP and ATR with the nuclear BrdUrd punctuate staining seen in hypoxia-treated S-phase cells. We propose that the increased sensitivity of S-phase cells to hypoxia/reoxygenation and the concomitant increased accumulation of DNA damage after reoxygenation are attributable to the presence of these stalled replication forks.
The finding that mice homozygous null for the ATR locus are embryonic lethal has had a profound impact on the study of ATR and has highlighted the diversity between ATM and ATR function (19) . This study clearly shows the importance of ATR-mediated signaling in the response to hypoxia-induced replication arrest. This and recent reports from the study of yeast and mammalian cells have led to the conclusion that ATR functions during normal replication (21 , 27, 28, 29) . These recent reports showed that ATR could protect and stabilize stalled replication forks, as well as initiate a cell cycle checkpoint in response to incomplete replication, for example, in the area of fragile sites. Reduction of ATR activity either because of decreased levels of ATR protein, in a heterozygous or Seckel line, or the overexpression of a dominant negative molecule resulted in an increased loss of viability in response to hypoxia/reoxygenation. Significantly, we did not observe an increased sensitivity to ionizing radiation in the HCT-116 ATR−/flox cell line, again showing the disparity between ATR and ATM because loss of ATM leads to classic radiosensitivity (30 , 31) . It seems unlikely that this increased sensitivity to hypoxia/reoxygenation is because of a failure in a checkpoint response (32 , 33) ; therefore, we suggest that during hypoxia exposure ATR acts to protect and stabilize stalled replication forks from collapse and hence DNA damage and that this is at least in part mediated via Chk1 activity. The consequences of damaged replication forks are numerous and include genomic instability and reproductive failure. It is certain that even partially reducing ATR levels or activity sensitizes cells exposed to hypoxia/reoxygenation. This has significant therapeutic implications because cells undergoing cycles of hypoxia/reoxygenation within a solid tumor represent the most aggressive population with the worst prognosis for effective treatment (5 , 34) . The use of therapeutics to completely inhibit the action of either ATR or Chk1 is undesirable because complete inhibition of ATR-mediated signaling would kill all of the cells, normal and cancerous (22 , 35) . However, it would be advantageous to partially inhibit ATR and/or Chk1 to specifically sensitize tumor cells exposed to cycles of hypoxia/reoxygenation. We have only considered ATR and Chk1 during these studies, but it seems probable that other components of the damage response pathway also could lead to hypoxic cytotoxicity. For example, BRCA2 recently has been shown to have an important role in the stabilization of stalled replication forks during normal division (36 , 37) . Rad-17 also is an attractive target because loss of Rad-17 is embryonic lethal and essential for chromosomal stability (38 , 39) . Future studies, particularly in animal tumor models, will show whether pharmacologic inhibition of these critical effector molecules decreases cell viability, particularly when combined with chemo/radiation therapy of solid tumors. The finding that cells from an individual with Seckel syndrome are sensitive to hypoxia/reoxygenation also may shed some light on the physical characteristics of this disease. For example, it is possible that there is a link between the microcephaly seen in patients with Seckel syndrome and the finding that the head and brain are particularly hypoxic areas during embryonic development (9 , 20 , 40 , 41) .
Cycles of hypoxia followed by reoxygenation are not only common to most if not all of the solid tumors but also have a significantly detrimental effect on prognosis. Tumor cells insensitive to hypoxia/reoxygenation (e.g., because of the loss of active p53) are positively selected and lead to a more aggressive and harder to manage tumor. Sensitizing these cells to hypoxia/reoxygenation without relying on, for example, p53 status represents a real and exciting potential therapeutic strategy.
Acknowledgments
We thank Drs. Mark O’Driscoll and Penny Jeggo for the Seckel syndrome cell lines and Dr. Paul Nghiem for the U2OS-GK41 cell line.
Footnotes
-
Grant support: NIH grant CA88480 to A. J. Giaccia.
-
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.
-
Requests for reprints: Amato J. Giaccia, Center for Clinical Sciences Research, Department of Radiation Oncology, Stanford University, Stanford, CA, 94305. Phone: 650-723-7366; Fax: 650-723-7382; E-mail: giaccia{at}stanford.edu
- Received May 4, 2004.
- Revision received June 23, 2004.
- Accepted July 13, 2004.
- ©2004 American Association for Cancer Research.
References
Volume 64, Issue 18
