Pharmacologic Disruption of Base Excision Repair Sensitizes Mismatch Repair-deficient and -proficient Colon Cancer Cells to Methylating Agents

INTRODUCTION

In DNA repair-competent cells, DNA adducts formed by methylating agents may be repaired efficiently or be sites of both mutagenic and cytotoxic damage. In this process, the cellular response is specific for each of the DNA adducts formed. Perhaps the best studied is the response to O⁶mG.3 This adduct may be repaired in a single step reaction by O⁶-alkylguanine-DNA (AGT); however, saturation of this protein by an excess of adducts or inhibition by BG results in residual adducts that are both cytotoxic and mutagenic (1) . Cytotoxicity results from recognition of this adduct by components of the MMR system, a five- or six-protein complex that recognizes O⁶mG:thymine base pairs formed by DNA replication past O⁶mG, and excises thymine and surrounding bases, resulting in DNA strand breaks. However, a thymine is preferentially reincorporated opposite the persisting O⁶mG, triggering MMR function again. It has been hypothesized that this repetitive aberrant repair process increases DNA double-strand breaks and acts as a trigger of apoptosis (2) .

MMR deficiency results in inability to process the O⁶mG:T mispair; consequently cells replicate DNA past O⁶mG lesions without cell cycle arrest, chromosomal aberrations, or apoptosis and survive in the face of persistent DNA damage (3, 4, 5, 6, 7) . The presence of MMR deficiency in a number of colon cancer cell lines allowed us the opportunity to evaluate the relative contribution of this DNA repair defect in resistance to the methylating chemotherapeutic agent, TMZ. We found that MMR deficiency resulted in 35–60-fold resistance to TMZ in cells defective in either MLH1 or MLH6 even after inhibition of AGT by BG (7) .

Although O⁶mG is the best studied cytotoxic DNA adduct, it is not the most abundant. TMZ, like other methylating agents, also forms N⁷mG and N³mA DNA adducts at frequencies 11 and 1.5 times that of O⁶mG.4 These DNA adducts are efficiently removed by BER and appear to contribute little to cytotoxicity. In the first step of BER, a series of glycosylases recognize abnormal bases such as N³mA and N⁷mG (8 , 9) , the T:G mismatch (10) , and deaminated bases such as hypoxanthine/oxidized 8-oxo-7,8-dihydroguanine or uracil:A (11, 12, 13, 14) . After enzymatic or spontaneous hydrolysis of the N-glycosidic bond and release of the abnormal base, AP endonuclease hydrolyzes the phosphodiester backbone 5′ to the lesion and dRpase (a DNA deoxyribophosphodiesterase with activity associated with polymerase β) excises the residual 2-deoxyribose-5-phosphate, generating a gap of one nucleotide. DNA polymerase β fills the gap, and DNA ligase seals the nick. This pathway is called short-patch BER. An alternative pathway for BER involves DNA synthesis to fill a gap of 2 to ≤ 13 nucleotides. This long-patch repair requires proliferating cell nuclear antigen and proliferating cell nuclear antigen-dependent DNA polymerase (15) .

PARP acts as a nick sensor of DNA strand breaks by itself or interaction with XRCC1 and involves in BER. PARP binds damaged DNA, resulting in autoribosylation. The modified protein then releases and allows other proteins to access and repair DNA strand breaks (15, 16, 17) . Therefore, PARP participates in BER after nick formation in both short- and long-patch repair. It appears most active in the alternative pathway for BER.

BER as a therapeutic target to increase the cytotoxicity of methylating agents has been documented. Cells deficient in DNA polymerase β or blocked in expression of AP endonuclease by antisense oligonucleotides are sensitized to methylating agents (18 , 19) . In addition, mice deficient in N³mA DNA glycosylase exhibited increased sensitivity to alkylating drugs such as BCNU and mitomycin C (20) . On the other hand, overexpression of the N³mA DNA glycosylase, which increases the number of AP sites formed, also increases the cytotoxicity of methylating agents (21) . Finally, cells lacking PARP activity are more sensitive to alkylating agents, with increased apoptosis and chromosomal instability (22 , 23) . These data suggest that balanced expression of proteins in the BER complex is important to the efficient processing of lesions. BER is an important mechanism of resistance to therapeutic methylating agents.

We examined two classes of agents that could inhibit the BER pathway to determine whether they would increase the cytotoxicity of methylating agents in colon cancer cells, particularly in cells deficient in MMR. Because MMR-deficient cells are tolerant to O⁶mG formed by TMZ, any change in cytotoxicity observed after use of a BER inhibitor would be due to interruption in repair of N⁷mG and N³mA DNA adducts. Our first strategy was to combine MX with TMZ. MX has been shown to react with the free aldehyde formed at the abasic site exposed by glycosylases and to reduce cleavage at AP sites in mammalian cells, suggesting that the MX-bound abasic site is not a substrate for AP endonuclease (24) . In the regard that AP sites modified by MX are relatively stable and must be converted to cytotoxic lesions, we hypothesized that MX would interrupt BER in cells and potentiate the cytotoxic effects of TMZ, even in MMR-defective cells. The second strategy we used was to inhibit PARP with PD128763, 3-AB, or 6-AN and to subsequently treat cells with TMZ. We hypothesize that inactivated PARP would affect short- and long-patch BER, destabilize strand breaks, reduce interaction with other proteins during repair of methylated DNA adducts, and lead to cell death, again in both MMR-proficient and -deficient cells.

MATERIALS AND METHODS

Chemicals and Reagents.

BG was generously provided by Dr. Robert Moschel (Frederick Cancer Research and Development Center, National Cancer Institute, Frederick, MD). Stock solution was made in DMSO. TMZ and BCNU were obtained from the Drug Synthesis and Chemistry Branch, Drug Therapeutic Program, National Cancer Center Institute (Rockville, MD). PD128763 was a gift from Park-Davis Pharmaceutical Division (Ann Arbor, MI). 6-AN, 3-AB, MX, and MMS were purchased from Sigma Chemical Co. (St. Louis, MO). Stock solutions of PD128763, 3-AB, and 6-AN were prepared by dissolving in DMSO and added to cell culture at a final concentration of < 1% DMSO when cells were treated with these compounds. MX was dissolved in sterilized water (pH 7.0). All stock solutions were kept at −20°C. BCNU was prepared fresh in 0.5 ml of 100% ethanol, diluted in PBS, and used within 10 min.

Colony Survival Assay.

SW480 cells were obtained from the American Type Culture Collection, Rockville, MD. HCT116 was obtained from R. Boland, University of Michigan Medical Center (Ann Arbor, MI). All cell lines were cultured in appropriate growth media.

Cells (2000/dish) were plated, adhered for 18 h, and treated with TMZ or MMS plus or minus variable modifiers such as BG, MX, 6-AN, 3-AB, or PD128763, according to experimental protocol. After treatment, cells were washed and fresh medium was added. The cells were grown for a further 7 days prior to staining with methylene blue for determination of colonies containing more than 50 cells. Comparisons of drug-induced cytotoxicity consisted of a calculation of the DMF, defined as the ratio of the IC₅₀ of either TMZ or MMS in the absence of indicated modifier(s) to that in the presence of indicated modifier(s), i.e., DMF = IC₅₀ for TMZ alone/IC₅₀ for TMZ plus modifier(s). The DMF indicates the degree of potentiation of cytotoxic agents by modulator.

Median Effect Analysis.

Median effect analysis was used to determine the dose-response interactions between TMZ and either MX or PD128763. Drugs were combined at the ratio of the IC₅₀ values for either TMZ and MX or TMZ and PD128763 as determined by survival/concentration curves. The combination was compared with the cytotoxicity of each drug alone in every experiment. The combination index was determined from colony-forming assays at increasing levels of cell killing, using an analysis of multiple drug interaction program (Biosoft, Cambridge, United Kingdom) developed based on the method of Chou and Talalay (25) . Combination index values of less than or greater than 1 indicate synergy and antagonism, respectively, whereas a combination index value of 1 indicates additivity of the drugs.

Flow Cytometry for Cell Cycle Distribution Analysis.

For cell cycle analysis, 10⁶ cells were plated in 100-mm tissue culture dishes and exposed to MX (6 mm)/PD128763 (100 μm) or MX (6 mm)/PD128763 (100 μm) plus TMZ (300 μm) at 37°C. After 24–72 h of culture, cells were fixed in 80% ethanol and DNA was stained with 20 μg/ml propidium iodide. The DNA fluorescence of propidium iodide-stained cells was measured with an Elite ESP flow cytometer/cell sorter (Coulter, Miami, FL). Cell cycle distribution was analyzed with the Modfit 5.2 program (Verity Software, Topsham, MA) with at least 10,000 cells per data point.

Western Blotting for PARP Cleavage Detection.

Cell extracts were resolved by SDS-PAGE (8% polyacrylamide) in a Bio-Rad minigel apparatus at 150 V for 1 h. Proteins were transferred onto PVDF membranes, using a Bio-Rad mini Trans-Blot cell for 1 h at 100 V. The blotted membranes were blocked with 5% dry milk in Tris-buffered saline and then probed for 2 h with anti-PARP antibody C2-10 (Trevigen, Gaithersburg, MD). After three 5-min washes with Tris-buffered saline-Tween 20 (0.05%), the blots were incubated with secondary antibody, antimouse horseradish peroxidase-anti-IgG for 1 h (Amersham Life Science, Arlington Heights, IL). Antibody binding was visualized by the ECL method, according to manufacturer’s instructions (Amersham).

RESULTS

MX Potentiates Cytotoxicity of TMZ.

We previously reported the comparative cytotoxicity of TMZ and BG in the SW480 and HCT116 cell lines (7) . To test whether MX would alter TMZ cytotoxicity, we treated SW480 and HCT116 with 6 mm MX (IC₅₀ for MX alone was 50 μm in SW480 and 28 μm in HCT116 cells) plus TMZ (0–1500 μm) for 2 h, with or without BG to abolish AGT-mediated removal of O⁶mG DNA adducts. SW480 cells were moderately resistant to TMZ, with an IC₅₀ of 395 μm, which was reduced 14-fold to 28 μm by BG pretreatment. Greater resistance to TMZ was observed in MLH1-defective HCT116 cells, even after inhibition of AGT by BG (TMZ IC₅₀, 950 μm). In both cell lines, MX potentiated the cytotoxic effect of TMZ (Fig. 1)<$REFLINK> with a DMF of 2.3 ± 0.12 (P = 0.0002) in SW480 and 3.1 ± 0.16 (P < 0.0001) in HCT116. In SW480 cells, additive effects of MX and BG were noted, (IC₅₀ was reduced from 395 μm to 6 μm, and the DMF was 65.8), whereas with HCT116 cells, no effect of BG was seen in the presence or absence of MX.

To further decipher the role of N³mA and N⁷mG DNA adducts in the relative absence of O⁶mG, we evaluated the effect of MX on MMS-mediated cytotoxicity. MMS is a methylating agent that produces far fewer O⁶mG adducts (0.3%) and a greater proportion of N³mA (10%) and N⁷mG adducts (87%) than TMZ (25) . The IC₅₀ of MMS was 0.82 mm in SW480 and 1.4 mm in HCT116 cells. This difference is smaller than the difference in the TMZ IC₅₀ (395 versus 950 μm) between these cell lines, probably because the low concentration of O⁶mG adducts formed by MMS increases the impact of other DNA adducts. After cells were treated with MMS (0–3 mm) plus 6 mm MX for 1 h, the IC₅₀ DMFs, compared with MMS alone, were 2.0 ± 0.14 (P < 0.002) in SW480 and 2.3 ± 0.17 (P = 0.002) in HCT116 (Fig. 2)<$REFLINK> . These DMFs were similar to that observed with TMZ. Compared with treatment of SW480 with BG plus TMZ (DMF of 14), BG plus MMS induced less enhancement of cytotoxicity (DMF of 6). This is perhaps due to fewer O⁶mG adducts formed by MMS; however, even a small number of O⁶mG adducts contribute to cytotoxicity in MMR-proficient cells. When MMS was combined with BG and MX, greater than 10-fold potentiation of cytotoxicity at the IC₅₀ for TMZ alone was observed in SW480, whereas no increased toxicity over that of the combination of BG, MX, and MMS was seen in HCT116 cells. From these data, we infer that MX had equal ability to interrupt BER in these two cell lines.

Inhibitors of PARP Modulate the Sensitivity of Cells to TMZ.

Because inhibitors of PARP may interrupt BER and increase sensitivity to methylating agents, we examined whether inhibitors of PARP sensitize cells to TMZ. Figs. 3<$REFLINK> and 4<$REFLINK> display survival after combined treatment of TMZ with PD128763, 3-AB, or 6-AN in both SW480 and HCT116 cells. In the SW480 cell line, 100 μm PD128763 (IC₅₀ for PD128763 alone, 625 μm) sensitized cells to TMZ with a DMF of 3.1 ± 0.12 (P < 0.0002). The combination of PD128763, BG, and TMZ was even more toxic, with a DMF of 36 (Fig. 3A)<$REFLINK> . In HCT116 cells, the DMF for PD128763 and TMZ compared with TMZ alone was 4.7 ± 0.2 (P < 0.0001). However, the combination of PD128763, BG, and TMZ had no greater effect than PD128763 and TMZ (Fig. 4A)<$REFLINK> , indicating that persistent O⁶mG had no effect on cytotoxicity in this MMR-defective cell line. Potentiation of TMZ cytotoxicity was also observed in both cell lines treated with two other PARP inhibitors, 3-AB (Figs. 3B<$REFLINK> and 4B)<$REFLINK> and 6-AN (Figs. 3C<$REFLINK> and 4C)<$REFLINK> . Although the specific activity of these agents varied considerably, DMF values of 3 to 4 were observed for both 3-AB and 6-AN when combined with TMZ compared to TMZ alone.

Synergistic Interaction between TMZ and MX or PD128763.

We investigated the nature of the interaction between TMZ and MX in these two cell lines. These cells were incubated in the presence of a range of concentrations of TMZ (37.5–750 μm) and MX (0.75–15.0 mm) and a constant molar ratio mixture of TMZ and MX (1: 20), based on the relative IC₅₀ for 2 h. Cells were also exposed to TMZ (18.8–750 μm) and PD128763 (15.6–625 μm) alone and at the fixed dose ratio of the combination of (1:0.83) for 2 h to analyze synergism. As shown in Fig. 5<$REFLINK> , synergistic interaction (C I≪ 1; P < 0.001) was found in both SW480 and HCT116 cells for the combination of TMZ with either MX or PD128763 despite the fact that the HCT116 cells were TMZ resistant. This synergistic interaction was observed even at very low concentrations, which were absolutely nontoxic when each drug was used alone, indicating that BER inhibitors significantly synergize methylating agent cytotoxicity in both MMR-deficient and -proficient colon cancer cells.

Effect of BER Inhibitors on BCNU Cytotoxicity.

To test whether MX is also able to sensitize colon cancer cells to chloroethylating agents, these two cell lines were pretreated with 6 mm MX for 2 h, followed by BCNU. No enhancement of BCNU cytotoxicity by MX was observed (Fig. 6)<$REFLINK> ; the BCNU IC₅₀ was 45 μm in HCT116 cells (Fig. 6A)<$REFLINK> and 27–29 μm, respectively, in SW480 cells treated with BCNU alone or BCNU plus MX (Fig. 6B)<$REFLINK> . A greater sensitization to BCNU was observed in these two cell lines when cells were treated with MX plus BG and BCNU; the BCNU IC₅₀ for both cell lines was 5 μm under these conditions. However, most of the effect was potentiation due to BG, which increased BCNU cytotoxicity by 3–4-fold. As shown in Fig. 7<$REFLINK> , no sensitization to BCNU cytotoxicity was seen after treatment with addition of PD 128763.

Effect of Inhibitors of BER on Cell Cycle Distribution and PARP Cleavage.

The cell cycle and apoptosis responses of SW480 and HCT116 cells were examined at various times after treatment with TMZ (300 μm) alone or with MX (6 mm), PD 128763 (100 μm), or BG (25 μm). After treatment, cells were divided into two aliquots for analysis of cell cycle/apoptosis on days 1 and 3, and for detection of PARP cleavage (see below). Cell cycle distribution was measured by flow cytometry according to DNA content, and estimation of the duration of the G₁, S, and G₂-M phases was based on untreated, exponentially growing, asynchronous cells. MX and PD128763 alone did not affect the distribution of cell cycle in these two cell lines (data not shown). At 24 h, 75–90% of SW480 cells accumulated in the S and G₂ phases after treatment with TMZ alone, and this S-G₂ phase arrest was more pronounced in cells pretreated with either MX or PD 128763 (Fig. 8A)<$REFLINK> . S-G₂ phase arrest was still present after 3 days in cells treated with the combination of MX or PD 128763 and TMZ (in both instances, 13–20% of cells were apoptotic). In SW480 cells treated with TMZ alone, the S-G₂ phase block was less obvious at day 3, with only 8% of cells showing evidence of apoptosis. In contrast, HCT116 cells had a normal cell cycle distribution after treatment with TMZ alone, and no effect was seen with BG and TMZ. However, accumulation in the S phase was observed (Fig. 8B)<$REFLINK> 24 h after treatment with PD 128763 plus TMZ. At 72 h, HCT116 cells had moved through the S phase, and thereafter, a significant portion of cells (90%) remained arrested in the G₂ phase with apoptosis present in 14% of cells. A similar but less striking result was observed with MX and TMZ in HCT116 cells. By 72 h, 60% of cells were still arrested in the S and G₂ phases and 10% of cells were apoptotic.

Finally, as a marker of apoptosis-induced cell death, we examined PARP cleavage after cells were treated with these drug combinations at 3 days (Fig. 9)<$REFLINK> . PARP cleavage was observed in SW480 cells after exposure to TMZ alone and TMZ plus BG but was not seen in HCT116 cells with the same treatment, indicating that the apoptotic process is triggered when O⁶mG lesions are repaired by the MMR system. However, PARP cleavage was detected in MMR-proficient and -deficient cells treated with TMZ plus either MX or PARP inhibitors.

DISCUSSION

Because MMR defective cell lines are remarkably resistant to methylating agents yet accumulate high levels of three methylating DNA adducts, O⁶mG, N³mA, and N⁷mG, we reasoned that the interruption of repair of N³mA and N⁷mG adducts by the BER process would sensitize cells to methylating agents. To address this issue, we studied the effect of MX on potentiation of TMZ cytotoxicity. MX interacts specifically with the tautomeric open-ring form of deoxyribose generated from the removal of an abnormal base by glycosylase. The MX-modified AP site is relatively stable (26 , 27) and inhibits the cleavage of AP sites in DNA by AP endonuclease in mammalian cells. This has been shown to protect cells from cytotoxicity, mutagenicity induced by SN1-type ethylating agents, such as ethylnitrosourea, but not SN2 alkylating agents, such as diethyl sulfate and MMS (24 , 28) . Moreover, the protection was strictly time dependent and was limited to the short period (30 min) after exposure to the alkylating agents (28) . In our studies, we had also observed that MX reduced cleavage at AP sites and decreased BER in human colon cancer cell extracts.5 However, we did not see protection of these two cell lines from ethylnitrosourea cytotoxicity when longer exposures to MX were used. The short duration of MX studied previously may not have the same impact on BER inhibition as does longer exposure to MX. Our results showed that MX synergistically increased TMZ-induced cytotoxicity in human colon cancer cell lines in both MMR-proficient and -deficient cells. A similar degree of enhanced cytotoxicity was observed with MX and MMS and with TMZ as well. The effect of BG inhibition of AGT was additive to the effect of MX only in the MMR-proficient SW480 cell line but not in the MMR-defective HCT116 cell line. These data suggest that O⁶mG DNA adducts do not contribute to the enhanced cytotoxic effect of TMZ by MX. Furthermore, a similar degree of enhanced cytotoxicity was observed with MX and MMS as with TMZ, again implicating N³mA- and N⁷mG-induced abasic sites as the major targets for MX. In our recent studies, a prolonged exposure to low-dose MX results in even greater potentiation of TMZ cytotoxicity.

The mechanisms of MX-enhanced cytotoxicity of methylating agents in colon cancer cells have not been fully understood. It is fair to suggest that MX enhanced the cytotoxic effect of TMZ because (a) the MX-AP site complex is able to block the AP endonucleolytic step of the BER pathway; (b) the persistence of abasic sites may increase topoisomerase II-mediated DNA cleavage (29) ; and (c) AP sites inhibit DNA replication and trigger programmed cell death (30) .

Under normal circumstances, TMZ produces strand breaks during BER-mediated repair of N⁷mG and N³mA adducts that are repaired efficiently and do not contribute to cytotoxicity until high concentrations of adducts are achieved. When DNA strand breaks are present, one component of the response is recognition, binding, and activation of PARP. Activated PARP leads to autoribosylation, and this in turn facilitates access of repair enzymes to DNA damage (15 , 31) and appears to enhance processing of stand breaks and religation by polymerase-β and ligase I (32) . In the alternative BER pathway, PARP interacts with XRCC1 to facilitate repair (17) . It seems likely that PARP plays an important role in communication between repair proteins and the stability of the repair complex involved in BER. This suggests that inhibition of PARP leads to an impaired ability to rejoin DNA strand breaks, which can initiate both apoptotic and nonapoptotic cell death cascades and thereby increase cytotoxicity of TMZ (33) . Our results support this hypothesis. Potentiation of cytotoxicity of methylating agent with PARP inhibitors was observed with a marked increase in apoptosis and PARP cleavage.

The results of BCNU combined with either MX or PD128763 are in sharp contrast to TMZ: little if any potentiation is observed in the absence of BG in either cell line. This suggests that although BER appears to process BCNU-induced cross-links, inhibition of BER in this manner has little impact on BCNU toxicity. One of the best studied BCNU-induced lethal lesions is the N³-cytosine-N¹-guanine interstrand cross-link formed after initial chloroethyl monoadducts at O⁶-guanine and cyclic rearrangement to N¹,O⁶-ethanoguanine (34) . However, treatment of cells with BCNU also produces alkylated bases that may be labile and spontaneously result in breakage or nicking of the phosphoribosyl backbone (35) . Because PARP has been shown to bind to BCNU-induced DNA nicks in vitro (36) , it is reasonable to assume that PD128763 might increase BCNU cytotoxicity. However, our studies showed only minor enhancement of toxicity in HCT116 cells and no enhancement in SW480 cells. Although methyladenine DNA glycosylase has been implicated in BCNU cross-link repair and its absence sensitizes cells to BCNU (37) , we did not observe sensitization to BCNU by treatment with MX in the absence of BG. In the presence of BG, MX potentiated BCNU toxicity, indicating that MX may interfere with DNA cross-link repair pathway and suggesting that BER may be involved in repair of the N¹,O⁶-ethanoguanine cross-link, which is not formed if AGT reacts with the O⁶-chloroethylguanine adduct. Taken together, these data suggest a different reaction of MX with damaged DNA induced by BCNU compared with TMZ. With TMZ, MX-increased cytotoxicity is associated with AP sites generated from repair of N⁷mG and N³mA DNA adducts formed by methylating agent; however, with BCNU, it might be the O⁶ lesion-induced cross-link that controls BCNU toxicity.

It appeared that apoptosis mediates both MX and PD128763-enhanced cytotoxicity of TMZ. Increased apoptosis was observed in MMR wild-type SW480 cells but not in MMR-deficient HCT116 cells after treatment with BG and TMZ. This suggests that MMR processing of O⁶mG is a potent apoptosis-inducing event (38) . Although the biological and functional consequences of PARP and its cleavage in apoptosis still remain to be further identified, it has been demonstrated that PARP is rapidly and specifically cleaved during apoptosis (39 , 40) . PARP cleavage was observed in both SW480 and HCT116 cells after treatment with either MX or one of the PARP inhibitors and TMZ, confirming activation of apoptotic pathways.

We noted that arrest at cell cycle checkpoints paralleled the cellular response to DNA damage and that these were dependent on MMR and BER pathways. MMR wild-type SW480 cells were sensitive to TMZ alone with arrest in the S and G₂ phases (2) . The S and G₂ phase arrests were potentiated by MX or by PD128763 despite the fact that SW480 is a p53 mutant cell line. In contrast, even high levels of DNA adducts formed by TMZ in the MMR-deficient HCT116 cells did not induce cell cycle checkpoint arrest despite the fact that p53 is wild type in this cell line. This dysregulation of damage-induced cell cycle checkpoint control appeared because of failure of processing O⁶mG lesions in MMR-deficient cells. However, after combined treatment with TMZ and either MX or PD128763, HCT116 cells showed S-G₂ phase arrest and apoptosis. These results are consistent with previous studies of cell cycle changes after MMS exposure or other compounds that produce 90% N³mA (41, 42, 43) and the prolonged G₂ phase arrest observed in PARP knockout mice or derived cell lines (22) following DNA damage. These data indicate that both SW480 and HCT116 cells have a similar response to persistent N⁷mG and N³mA lesions, following interruption of BER.

In summary, we have shown that disrupted BER processing of non-O⁶mG, most likely N⁷mG and N³mA, DNA adducts formed by TMZ is cytotoxic to colon cancer cell lines. This may be particularly important in MMR-deficient cells, which are resistant to TMZ alone because of the failure to recognize O⁶mG DNA adducts. These studies provide evidence that disrupting repair of N⁷mG and N³mA by inhibiting BER or PARP may improve the therapeutic efficacy of methylating agents.