Methoxyamine potentiates iododeoxyuridine-induced radiosensitization by altering cell cycle kinetics and enhancing senescence Free

After existing for a century, radiation therapy is still a major treatment component in the curative management of most human solid tumors. In many solid tumors, chemotherapy drugs, such as fluoropyrimidine analogues and platinum analogues, and newer targeted drugs are used during radiation therapy to improve the therapeutic index (1, 2). Based on these clinical data, it can be argued that a better understanding of the cellular and molecular mechanisms underlying the observed radiosensitization by these combined treatments may lead to further gains in the therapeutic index.

Iododeoxyuridine (IUdR) is a halogenated pyrimidine that can be incorporated into DNA in place of thymidine through DNA replication. This halogenated pyrimidine is considered a clinical radiosensitizer, where the extent of radiosensitization correlates directly with the level of halogenated pyrimidine-DNA incorporation (3–5). Therefore, treatment strategies that selectively enhance the incorporation of IUdR into tumor cell DNA would be expected to enhance tumor radiosensitivity and the therapeutic gain of tumor to normal tissues. However, the cellular and molecular mechanisms of IUdR radiosensitization are not completely understood. It is believed that IUdR sensitizes cells to ionizing radiation through enhancing the formation of ionizing radiation–induced double-strand breaks (DSB; refs. 6–9).

Our recent studies have shown that IUdR incorporated into DNA (particularly a G-IU mispair) can be recognized by mismatch repair (10) and base excision repair (BER) proteins (11, 12). Therefore, the cellular status of these two repair systems in tumor cells may affect the degree of IUdR-DNA incorporation and subsequently IUdR-mediated radiosensitization. Methoxyamine is a specific chemical inhibitor of BER mediated by tight binding to AP sites generated by cleavage of BER glycosylases and rendering the phosphodiester bonds adjacent to the AP site refractory to the catalytic activity of AP endonuclease (APE1; ref. 13). A previous study by our group found that methoxyamine enhanced IUdR-DNA incorporation and potentiated IUdR-induced radiosensitization in human tumor cells (12). In this study, we explore the possible mechanisms underlying the enhanced radiosensitivity using methoxyamine in combination with IUdR. We found that the combined treatment of IUdR and methoxyamine altered cell cycle kinetics and led to a stringent G₁-S checkpoint and an insufficient G₂-M checkpoint, resulting in a prolonged G₁ arrest of 2C and 4C tumor cell populations. These cell cycle changes were associated with a suppression of apoptotic cell death but an enhancement of stress-induced premature senescence following ionizing radiation.

RKO is a human colorectal carcinoma cell line that is p53 proficient but deficient in mismatch repair (hMLH1⁻ because of hypermethylation of the gene promoter region; ref. 14). The cell population doubling time is 20 ± 2 hours. The RKO cells are grown in DMEM (Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 2 mmol/L glutamine, and 0.1 mmol/L nonessential amino acids (Invitrogen, Carlsbad, CA) in 10% CO₂ at 37°C. The drug-ionizing radiation treatment protocol is as follows. First, RKO cells were seeded at 5 × 10⁵ per 10-cm dish in complete medium. Next, the medium was changed 18 hours later to medium supplemented with dialyzed serum, IUdR (3 μmol/L; Sigma-Aldrich, St Louis, MO), and methoxyamine (3 mmol/L; Sigma-Aldrich), both drugs or no drugs. The medium containing dialyzed serum and drugs was changed every 24 hours for a total of 48 hours. The cells were then irradiated (model 109 137Cs irradiator at a dose rate of 3.7 Gy/min) in serum-free medium. After ionizing radiation, cells were incubated in complete medium without drugs until being harvested for cell cycle, Western blotting, and cell death analyses.

Immediately following ionizing radiation, the control and drug-treated RKO cells were trypsinized and counted. The cells were then plated into triplicate 60-mm dishes at the appropriate cell numbers to ensure the formation of ∼50 to 100 colonies per dish. Cells were allowed to grow for 10 to 12 days. The colonies were stained with 0.5% crystal violet in methanol/acetic acid (3:1; Fisher Scientific, Pittsburgh, PA), and those colonies containing ≥50 cells were scored. Plating efficiency was determined by dividing the number of colonies by the number of cells seeded. The surviving fraction was calculated by dividing the plating efficiency for irradiated cells by the plating efficiency for unirradiated cells. The experiments were done six times, each in triplicate.

Western blotting analyses were carried out as described previously (15). The antibodies used were as follows: β-actin and α-tubulin (Sigma-Aldrich); pcdc2(Y15), pChk1(S317), pChk2(T68), Rb, pRb(S780), and p-histone H3(S10) (Cell Signaling, Beverly, MA); γ-H2AX (Upstate, Charlottesville, VA); PARP p85 (Promega, Madison, WI); cyclin A, cyclin B1, p53, p21, p27, and secondary antibody IgG-horseradish peroxidase conjugates (Santa Cruz Biotechnology, Santa Cruz, CA). All experiments were done at least twice.

Determination of the cell cycle profile was described previously (16). For dual-variable flow cytometry, 1 × 10⁶ fixed cells were incubated with cyclin B1/FITC (PharMingen, San Diego, CA) before the DNA was stained with propidium iodide (Sigma-Aldrich; 33 μg propidium iodide/mL in PBS, including 1 mg/mL RNase A, 0.5 mmol/L EDTA, and 0.2% NP40). The flow cytometric data were analyzed with Modfit 3.0 and Winlist 5.0 software (Verity Software, Topsham, ME). The cells containing a 2× DNA content (G₂, M, and 4CG1) will be referred to as 4C (here, C is DNA content). All experiments were done at least twice.

RKO cells were treated with drugs and ionizing radiation as previous described. The control and drug-treated cells were analyzed immediately following the 2-day drug exposure as well as at 24, 48, and 72 hours following ionizing radiation. The differently treated cell populations were washed with PBS to remove floating cells and then trypsinized. The cells were allowed to recover from trypsinization in serum-containing medium on ice for 30 minutes. One hundred microliters of the cell suspension (2–5 × 10⁶ cells per mL in PBS) was then incubated with 5 μL of propidium iodide solution (250 μg/mL) for 10 minutes on ice in the dark. The cell samples were analyzed by flow cytometry immediately after adding 0.6 mL PBS. Only necrotic cells were propidium iodide positive due to the loss of membrane integrity. The data were analyzed with Winlist 5.0 software. The experiment was done thrice.

Cells were seeded and treated with IUdR, methoxyamine, or both as above and then stained at days 5 to 8 after ionizing radiation with a senescence detection kit (Calbiochem, San Diego, CA), according to manufacturer's instruction, with the exception of extending the incubation time in the staining solution to 48 hours. We found that this modification allowed for an enhanced blue color, although the cells were fixed before staining. For quantitation, we found that direct counting on the slides was not accurate due to repopulation of surviving cells. Therefore, we carried out experiments similar to a clonogenic survival assay. However, only single large viable cells were counted at days 10 to 12 after ionizing radiation, as we found that these cells were senescence-activated β-galactosidase (β-gal) positive. The data were normalized against the cell number seeded and the plating efficiency. The experiment was done thrice.

The data, where applicable, represent the means ± SE. Data were analyzed using the Student's t test. Bar and line graphs of flow cytometry data were usually derived from representative experiments and therefore are presented without error bars. The reason for this is that although we observed the same trends in changes repeatedly, the fraction of cells arrested in each phase varied from one experiment to another, presumably because asynchronous cells were used.

We first confirmed, in RKO cells, our previous finding in HCT116 cells (12) that methoxyamine enhances IUdR-DNA incorporation and potentiates IUdR-induced radiosensitization. As shown in Fig. 1A, IUdR treatment alone (3 μmol/L) for 48 hours results in 9% of thymidine replacement, whereas treatment with IUdR (3 μmol/L)/methoxyamine (3 mmol/L) for 48 hours increases the thymidine replacement to 15%, which, assuming thymidine represents 25% of total DNA, is equivalent to 2.25% and 3.75% IUdR in the total DNA, respectively. Using clonogenic survival as an end point (Fig. 1B), IUdR pretreatment significantly sensitized cells to ionizing radiation, and the radiosensitization was further increased when cells were pretreated with IUdR/methoxyamine. The sensitizer enhancement ratio is 2.37 ± 0.58 for 2.5 Gy and 2.12 ± 0.59 for 5 Gy, respectively, for IUdR/methoxyamine–pretreated cells compared with IUdR-pretreated cells. No difference in clonogenic survival was seen with methoxyamine pretreatment alone.

Figure 1C shows the effect of IUdR/methoxyamine pretreatment on cell proliferation following ionizing radiation. The % cell number changes are shown as the total number of viable cells for up to 72 hours following ionizing radiation over the total number of viable cells before ionizing radiation. The combined treatment with IUdR/methoxyamine for 48 hours before ionizing radiation reduced cell growth to a greater extent (∼30% decrease compared with the control) than IUdR or methoxyamine pretreatment alone (5–15% decrease). The growth inhibition was reversible after removal of the drugs. As shown in Fig. 1C, the cell numbers in the control and methoxyamine-pretreated groups are increased significantly following ionizing radiation, but the cell number changes in the IUdR- and IUdR/methoxyamine–pretreated cells are minimal. Therefore, we chose to compare the effects of IUdR pretreatment with IUdR/methoxyamine pretreatment in the following cell cycle and cell death experiments and used the responses of methoxyamine pretreatment and/or no drug treatment as baseline controls.

To determine what cellular and molecular events might underlie the enhanced ionizing radiation cell killing following the combined treatment of IUdR/methoxyamine, we examined cell cycle profiles. As shown in Fig. 2A and B, IUdR/methoxyamine treatment for 48 hours leads to an increased G₁ and a reduced S phase (time point 0). In contrast, the cell cycle profiles were not changed significantly by a 48-hour incubation with IUdR alone or methoxyamine alone. Using nocodazole trapping, which blocks cell cycle progression in the mitotic phase, we confirm that IUdR/methoxyamine–treated cells progress more slowly through G₁ phase than the other treatment groups (Fig. 2C). The time course of early cell cycle responses following 5 Gy ionizing radiation (time points 1–6 hours in Fig. 2A and B) shows that the differences in G₁ and S phases found before ionizing radiation between IUdR/methoxyamine–pretreated cells and the other groups are sustained after ionizing radiation. Furthermore, although cells from all treatment groups respond to ionizing radiation damage with a decrease in G₁ and an increase in S and G₂ phases, the changes in cell cycle kinetics in IUdR/methoxyamine–pretreated cells are slower than the other groups (Fig. 2A and B). These results indicate that IUdR/methoxyamine pretreatment not only alters cell cycle kinetics before ionizing radiation but also changes the early cell cycle response for at least up to 6 hours following ionizing radiation.

Supporting a strong G₁-S checkpoint, we found that the levels of protein expression of two cyclin-dependent kinase 2 inhibitors (p21 and p27) are significantly higher in the IUdR/methoxyamine–pretreated cells before and after ionizing radiation than in the other groups (Fig. 2D; data of the methoxyamine group are not shown). p21 and p27 were also increased after IUdR treatment alone for 48 hours (time point 0) but not to the extent found in IUdR/methoxyamine–treated cells. Furthermore, an S-phase marker (cyclin A) and a G₂-phase marker [cyclin B1/pcdc2(Y15), phosphorylated cdc2 at Tyr¹⁵] are only increased slightly in response to ionizing radiation in the IUdR/methoxyamine–pretreated cells compared with the increased levels in the other treatment groups, indicating a reduced activation of S and G₂-M checkpoints by this combined drug treatment.

We further evaluated the DNA damage responsive proteins H2AX, Chk1, and Chk2. Histone H2AX is phosphorylated at Ser¹³⁹ (γ-H2AX) in response to DNA DSBs (17). IUdR-pretreated cells show a higher level of ionizing radiation–induced γ-H2AX as expected (Fig. 3), because it is known that IUdR increases ionizing radiation–induced DSB (3–9). Considering that methoxyamine enhances IUdR-DNA incorporation (Fig. 1A), it was expected that IUdR/methoxyamine–pretreated cells would show an even higher level of ionizing radiation–induced γ-H2AX than IUdR-pretreated cells. However, ionizing radiation–induced γ-H2AX in IUdR/methoxyamine–pretreated cells neither exceed that in IUdR-pretreated cells nor that in the control cells. Notably, drug treatment alone for 48 hours results in an increase in γ-H2AX in both IUdR- and IUdR/methoxyamine–treated cells, but it is more evident in IUdR-treated cells than in IUdR/methoxyamine–treated cells (Fig. 3, time point 0). These results suggest that either the formation of ionizing radiation–induced DSBs may not be enhanced in IUdR/methoxyamine–pretreated cells, or γ-H2AX signaling may be impaired in IUdR/methoxyamine–pretreated cells.

ATM/Chk2 and ATR/Chk1 DNA damage signaling pathways are known to be activated in response to ionizing radiation (18, 19). As expected, both phosphorylated Chk1 at Ser³¹⁷ [pChk1(S317)] and Chk2 at Thr⁶⁸ [pChk2(T68)] were increased dramatically in all groups in response to 5 Gy ionizing radiation (Fig. 3; data of the methoxyamine group are not shown). Ionizing radiation–induced pChk1(S317) and pChk2(T68) levels are greater in IUdR-pretreated cells than in the control cells as expected. However, both the highest level of ionizing radiation–induced pChk2(T68) and the lowest level of ionizing radiation–induced pChk1(S317) are found in the IUdR/methoxyamine–pretreated cells. IUdR/methoxyamine pretreatment itself does not seem to generate an appreciable damage signal before ionizing radiation (Fig. 3, time point 0). These data support our speculation that there may be a stringent G₁ checkpoint and insufficient S and G₂-M checkpoints in IUdR/methoxyamine–pretreated cells because it has been recognized that pChk1(S317) plays an important role in a G₂ arrest, whereas pChk2(T68) plays an important role in a G₁ arrest (19–21).

Having documented these cell cycle and DNA damage signaling changes (Figs. 2 and 3), we next questioned what type of ionizing radiation–induced cell death is enhanced by IUdR/methoxyamine. Figure 4A shows a representative histogram from a flow cytometry assay measuring the apoptotic cell population following 5 Gy ionizing radiation. A large sub-G₁ fraction is found in the control and methoxyamine-pretreated cells at 48 and 72 hours following ionizing radiation (the medium was changed every 24 hours). However, this ionizing radiation–induced sub-G₁ population was reduced in IUdR- and IUdR/methoxyamine–pretreated cells. These flow cytometry data were confirmed with 4′,6-diamidino-2-phenylindole staining of the nucleus (Fig. 4B), where condensed chromatin was found readily in control and methoxyamine-pretreated cells but not in IUdR- and IUdR/methoxyamine–pretreated cells. Figure 4C provides data on an additional apoptosis marker (i.e., cleaved PARP p85 fragment), which is increased significantly at 72 hours following ionizing radiation in the control cells but only moderately in IUdR- and IUdR/methoxyamine–pretreated cells.

Together, it seems that both IUdR and IUdR/methoxyamine pretreatments partially suppress ionizing radiation–induced apoptotic cell death. It is likely that a selection was made to favor cell cycle checkpoint arrest over apoptosis because a sustained ionizing radiation–induced 4C DNA cell peak (containing G₂, M, and 4CG1 cells; Fig. 4A) occurred concomitantly with reduced apoptosis in IUdR- and IUdR/methoxyamine–pretreated cells. Furthermore, no significant differences in apoptotic cell death were found between the IUdR- and IUdR/methoxyamine–pretreated cells, suggesting that IUdR/methoxyamine does not enhance IUdR-induced radiosensitization by promoting apoptotic cell death.

We next examined necrotic cell death following ionizing radiation. Figure 5A shows the results of a flow cytometry assay using propidium iodide staining for necrotic cells. The permeability of the cytoplasmic membrane is increased following ionizing radiation in all treatment groups, and the percentage of heavily propidium iodide–stained cells (membrane-compromised cells) were similar (12–14%) over a 72-hour period after ionizing radiation among all treatment groups, indicating that ionizing radiation–induced necrotic cell death is not enhanced by IUdR/methoxyamine pretreatment. The use of a trypan blue exclusion assay for necrotic cells gave similar results (data not shown).

Autophagy, a vacuolar process of cytoplasmic degradation, is implicated as a form of programmed cell death distinct from apoptosis (22). We checked the levels of autophagosome-associated protein LC3-II as a marker of autophagic death in IUdR/methoxyamine–pretreated cells. In this death pathway, LC3-II is formed from LC3-I after lipid modification, and the amount of LC3-II is directly correlated with the extent of autophagosome formation. We found that LC3-II is strongly induced following ionizing radiation (Fig. 5B), indicating that the autophagy signaling pathway is activated following ionizing radiation in these RKO cells. However, no significant differences were found in the level of LC3-II among the different groups (data of methoxyamine group not shown). Thus, autophagy does not seem to explain the enhanced radiosensitization following IUdR/methoxyamine pretreatment.

Ionizing radiation is also known to induce a senescence-like irreversible growth arrest (23). During the course of the experiments, we noted that after ionizing radiation, IUdR- and IUdR/methoxyamine–pretreated cells exhibited senescence-like morphologic features (i.e., enlarged and flattened cells with increased adhesion to the tissue culture dishes). We, therefore, examined senescence-activated β-gal activity following these treatments. Senescence-activated β-gal activity was examined on days 5 to 8 following ionizing radiation and was found to be positive in IUdR- and IUdR/methoxyamine–pretreated cells (Fig. 6A). Spot senescence-activated β-gal positive cells were also found in the control and methoxyamine-pretreated cells; however, the majority of these cell populations have undergone cell division at the time of staining. Because almost all the single large cells were senescence-activated β-gal positive, we counted the single large viable cells at day 10 following ionizing radiation. According to the initial seeding cell number and the plating efficiency, it was determined (Fig. 6B) that IUdR/methoxyamine pretreatment leads to a greater percentage of ionizing radiation–induced senescent cells (33%) than IUdR pretreatment (18%). Together, these data indicate that IUdR/methoxyamine pretreatment enhances radiosensitization through enhancing IUdR/ionizing radiation–induced stress-induced premature senescence. Of note, the majority of the large cells contain an intact, homologous, and mononucleated nucleus, suggesting that mitotic catastrophe is not a major death pathway.

To verify the cell cycle phases where senescent cells may derive, we examined cell cycle status for longer times following ionizing radiation. As shown in Fig. 4A, both IUdR and IUdR/methoxyamine pretreatment result in a persistent 4C DNA cell peak following ionizing radiation. Using a dual-variable flow cytometry assay with propidium iodide and cyclin B1 double staining, we found that this 4C cell fraction contains two populations: G₂ cells that express high levels of cyclin B1 and 4CG1 cells that do not express cyclin B1 (Fig. 7A). Within the 4C cell population, IUdR/methoxyamine–pretreated cells contain the highest percentage of 4CG1 and the lowest percentage of G₂ cells for up to 72 hours following ionizing radiation compared with IUdR-pretreated cells as illustrated in the bar graphs in Fig. 7A, which are derived from the histograms (Fig. 7A). Thus, IUdR/methoxyamine–pretreated cells contain a higher total G₁ (2CG1 + 4CG1) population than IUdR-pretreated cells (Fig. 7B), indicating that the senescent cells may be derived from prolonged G₁-arrested cells. Furthermore, the low levels of pcdc2(Y15) in Fig. 7C confirm that the G₂-M checkpoint in the IUdR/methoxyamine–pretreated cells is not sufficiently activated. Moreover, the level of phosphorylated histone H3 at Ser¹⁰, a marker of M phase, indicates that the IUdR/methoxyamine–pretreated cells have not arrested in M phase either before or after ionizing radiation (Fig. 7C; data of methoxyamine group not shown).

Using the flow cytometry variable, forward scatter, which increases when cell size increases, we found that after ionizing radiation, the IUdR/methoxyamine–pretreated cells show a more gradual increase in cell size, which persists longer than IUdR-pretreated cells (Fig. 7D). At 168 hours (7 days) after ionizing radiation, when most cells in the control group and half of the cells in IUdR-pretreated group seemed normal in size, the majority of the IUdR/methoxyamine–pretreated cells still remain enlarged. This result is consistent with the observation that a higher percentage of ionizing radiation–induced senescent cells are found in IUdR/methoxyamine–pretreated group.

We further examined molecular characteristics of senescence in IUdR/methoxyamine–pretreated cells. Following ionizing radiation, two important senescence-related proteins, p53 and Rb (24, 25), were activated to a greater extent in IUdR/methoxyamine–pretreated cells than in IUdR-pretreated cells (Fig. 8). Indeed, both accumulation of p53 and induction of p21, the major mediator of p53-dependent senescence, increase more significantly in IUdR/methoxyamine–pretreated cells at both early (Fig. 2D) and later (Fig. 8) times following ionizing radiation. Furthermore, IUdR/methoxyamine–pretreated cells show reduced hyperphosphorylated Rb(S780) (inactive form) but greater hypophosphorylated Rb (active form) compared with IUdR-pretreated cells (Fig. 8). These data suggest that senescence signaling may be activated more efficiently following ionizing radiation in IUdR/methoxyamine–pretreated cells than in IUdR-pretreated cells.

It remains unclear how methoxyamine enhances IUdR-DNA incorporation. We previously proposed that methoxyamine may inhibit short-patch BER by blocking APE activity, therefore enhancing long-patch BER to remove methoxyamine-AP (12), which would result in increased IUdR-DNA incorporation. We further suspect other possibilities, including (a) when the template for DNA synthesis contains a methoxyamine-AP site, the chance that an incorrect base (including IUdR) will be inserted may be higher; (b) DNA structure distortions from methoxyamine-AP may initiate nucleotide excision repair, which increases IUdR-DNA incorporation through repair synthesis; and (c) the overall removal of IUdR from DNA may slow down significantly due to structural twists induced by methoxyamine-AP. IUdR is known as a radiosensitizer that increases ionizing radiation–induced DNA DSB (3–9). However, IUdR/methoxyamine seems to enhance IUdR-induced radiosensitization through different mechanisms other than increasing ionizing radiation–induced DSB. Examination of γ-H2AX, as a generally accepted DSB marker, provides us with two interesting implications in this regard (Fig. 3).

First, without ionizing radiation, γ-H2AX is increased moderately after IUdR and IUdR/methoxyamine treatment with a higher level found in IUdR-treated cells, indicating that DSBs are formed with IUdR treatment per se, and methoxyamine does not enhance DSB formation. It is likely that these DSBs are formed from single-strand breaks generated during BER (11, 12). Because methoxyamine stabilizes AP sites and blocks cleavage of the phosphodiester bond, it reduces the BER intermediate. Second, although γ-H2AX is increased significantly following ionizing radiation in all groups, the level in IUdR/methoxyamine–pretreated cells is similar to that in the control cells and does not exceed that in IUdR-pretreated cells. This result was unexpected because IUdR/methoxyamine–pretreated cells have a higher level of IUdR-DNA incorporation (Fig. 1A). One possible explanation is that in IUdR/methoxyamine–pretreated cells, ionizing radiation–induced single-strand breaks are increased, but they do not form recognizable DSB due to a twisted DNA structure. It is also plausible that methoxyamine-AP alters the overall state of chromatin into a configuration that inhibits the activation of a γH2AX DNA damage response. Finally, it is difficult to understand but cannot be excluded that methoxyamine-AP complex formation may lead to less ionizing radiation–induced DSB.

We found that IUdR/methoxyamine results in a reduced G₁-S transition, as evidenced by the results of nocodazole trapping (Fig. 2C) and the increased p21 and p27 proteins (Fig. 2D, time 0). This shift in the G₁-S transition may be a key factor determining radiosensitization by IUdR/methoxyamine. It is known that the efficiency of DNA repair is cell cycle dependent (26, 27). S-phase cells, which are dominated by homologous recombination repair in repairing DSB, are most resistant to ionizing radiation, whereas G₁ cells, which have nonhomologous end joining as a dominant DSB repair process, are sensitive to ionizing radiation (next to the most sensitive mitotic phase). Thus, the increased ionizing radiation sensitivity found in the IUdR/methoxyamine–pretreated cells can be correlated to their cell cycle profile of an increased G₁ phase and a decreased S phase. Although this report focuses on RKO cells, similar results were found with HCT116 cells, another human colon cancer cell line (data not shown). Furthermore, IUdR/methoxyamine pretreatment not only sets up a cell cycle stage before ionizing radiation that favors ionizing radiation sensitization but also alters cell cycle responses following ionizing radiation that may affect the ultimate decision making between apoptosis and senescence. It seems that IUdR/methoxyamine pretreatment renders the cells with a stringent G₁ checkpoint but an insufficient G₂ checkpoint, which result in a persistent ionizing radiation–induced G₁ (2CG1 + 4CG1) arrest (Fig. 7B), where senescent cells may reside.

For the past decade, a great deal of attention in cancer research has been given to apoptotic cell death. However, many human tumors carry mutations that inactivate apoptotic pathways; in addition, it is known that in most solid tumors, apoptosis is not a major cell death pathway following ionizing radiation damage (28). Tumor suppression is a complex process including different cell death pathways. In our case, apoptosis, necrosis, autophagy, and senescence were all found to contribute to overall ionizing radiation–induced cell death. We further observed that whereas RKO control cells showed the highest clonogenic survival following ionizing radiation, they also showed the highest apoptotic cell death. In contrast, IUdR and IUdR/methoxyamine pretreatment partially suppresses ionizing radiation–induced apoptosis but shows reduced survival following ionizing radiation. This partial suppression of ionizing radiation–induced apoptosis by IUdR and IUdR/methoxyamine occurs concomitantly with an increase in stress-induced premature senescence following ionizing radiation. Therefore, senescence that promotes loss of proliferative capacity seems more decisive in the decreased clonogenic survival in IUdR/ionizing radiation– and IUdR/methoxyamine/ionizing radiation–treated cells.

Senescence as an alternative cell death pathway following DNA damage is an emerging concept (25, 29, 30). It is recognized that tumor cells may harbor an innate senescence program that can be activated upon DNA damage (31). Senescence does not lead to immediate cell death but also does not enhance mutational changes. Because rapid repopulation of tumor cells during conventional radiotherapy is a factor that significantly impairs tumor response in some human cancers, a therapeutic strategy that enhances senescence may improve the therapeutic gain in certain human cancers. Our study suggests one such combinational treatment strategy. Although methoxyamine at 3 mmol/L is used in this in vitro study, it has been reported recently that methoxyamine at lower concentrations showed enhancement of methylating drug cytotoxicity in vivo (32, 33). We are currently assessing the in vivo therapeutic gain of methoxyamine enhancement of IUdR-mediated radiosensitization in human tumor xenografts.

It is evident that both IUdR and IUdR/methoxyamine pretreatments before ionizing radiation suppress apoptosis but enhance senescence. However, there seems to be a mechanistic difference in how senescence is enhanced. IUdR pretreatment seems to increase ionizing radiation–induced DSB and trigger a strong G₂-M checkpoint after ionizing radiation, whereas IUdR/methoxyamine seems to alter cell cycle kinetics and trigger a stringent G₁-S checkpoint but insufficient G₂-M checkpoints before and after ionizing radiation. There are many unanswered questions regarding these results that are under investigation in our laboratory. Currently, we are attempting to answer several pertinent issues, including how methoxyamine enhances IUdR-DNA incorporation; why γH2AX is not induced significantly in IUdR/methoxyamine pretreated cells following ionizing radiation; which signaling component is targeted by IUdR/methoxyamine treatment that can lead to an alteration in cell cycle kinetics; how the cellular decision between apoptosis and senescence is made in IUdR/methoxyamine/ionizing radiation treated cells; and whether the methoxyamine potentiation of IUdR/ionizing radiation–induced cell killing could be found in p53 mutant tumor cells.

In summary, pretreatment of cells with IUdR/methoxyamine before ionizing radiation alters cell cycle kinetics, particularly at the G₁-S transition, and enhances tumor radiosensitization. IUdR/methoxyamine pretreatment suppresses ionizing radiation–induced apoptotic cell death while enhancing the stress-induced premature senescence pathway. This newly recognized ionizing radiation–induced tumor suppression pathway may be effectively targeted by IUdR/methoxyamine, resulting in an improved therapeutic index for ionizing radiation.

We thank Dr. Tamotsu Yoshimori (Department of Cell Genetics, National Institute of Genetics, Shizuoka-ken, 411-8450 Japan) for providing us with anti-LC3 antiserum.