Purpose: Defining the cytotoxicity of individual adducts in DNA is necessary for mechanistic understanding of human brain tumor resistance to therapeutic alkylating agents and for design of DNA repair-related antiresistance strategies. Our purpose is to characterize the sensitivity of human glioma cells to methyl-lexitropsin (Me-lex), a sequence-specific alkylator that produces 3-methyladenine (3-meA) as the predominant (>90%) DNA lesion.

Experimental Design: We quantitated the Me-lex cytotoxicity of 10 human glioma cell lines that differ in O6-methylguanine (O6-meG)-DNA methyltransferase (MGMT) and mismatch repair activity. We used antisense suppression of alkyladenine DNA glycosylase (AAG) and Ape1 to assess the contribution of 3-meA and abasic sites to lethality and measured abasic sites.

Results: (a) The LD10 for Me-lex varied widely among the cell lines. (b) MGMT-proficient lines were more resistant than MGMT-deficient lines, an unexpected finding because Me-lex produces very little O6-meG. (c) Suppression of AAG increased Me-lex killing and reduced abasic site content. (d) Suppression of Ape1 increased Me-lex killing and increased abasic site content. (e) Ablation of MGMT had no effect on Me-lex cytotoxicity.

Conclusions: (a) Me-lex is cytotoxic in human glioma cells and AAG promotes resistance, indicating that 3-meA is a lethal lesion in these cells. (b) Abasic sites resulting from 3-meA repair are cytotoxic and Ape1 promotes resistance to these derivative lesions. (c) A factor(s) associated with MGMT expression, other than repair of O6-meG, contributes to Me-lex resistance. (d) Me-lex may have clinical utility in the adjuvant therapy of gliomas. (e) AAG and Ape1 inhibitors may be useful in targeting alkylating agent resistance.

Many alkylating agents undergo chemical decomposition to produce highly reactive electrophiles (e.g., diazonium and carbonium ions) that readily transfer their alkyl moieties to multiple nucleophilic sites on purines and pyrimidines in DNA (reviewed in ref. 1). In dsDNA, reaction occurs preferentially at ring nitrogens of purines. The majority of methylation, including that produced by the chemotherapeutic agents procarbazine and temozolomide, occurs at N7 of guanine (∼70-80%) and N3 of adenine (∼9%). Other N-methylpurines (e.g., 3-methylguanine, 1-methyladenine, and 7-methyladenine) constitute between 0.5% and 2.0% of total base adducts (1, 2). Work with repair-deficient bacteria (3, 4) indicates that 7-methylguanine is biologically innocuous and implicates 3-methyladenine (3-meA) as an important cytotoxic lesion. However, definitive elucidation of the contribution of individual N-methylpurines to cell killing has been hampered by the diversity of lesions produced by methylators, by the large biological effects of some quantitatively minor products [e.g., O6-methylguanine (O6-meG); refs. 5, 6], and by the instability of potentially cytotoxic lesions to manipulation in vitro. A novel approach that circumvents these problems exploited the sequence-specific binding properties of the antibiotics distamycin and netropsin, whose N-methylpyrrolecarboxamide moieties preferentially bind at A-T rich regions in the minor groove of DNA (reviewed in ref. 7). Gold et al. synthesized a series of compounds, designated lexitropsins, consisting of N-methylpyrrolecarboxamide subunits appended with methylating functionalities (8, 9). One derivative appended with an O-methyl sulfonate ester [1-methyl-4-(1-methyl-4-(3-methoxysulfonyl-propanamido)pyrrole-2-carboxamido) propane; methyl-lexitropsin (Me-lex)] methylates DNA almost exclusively at the N3 atom of adenine (8). Quantitation of Me-lex-induced adducts revealed that >90% of base lesions are 3-meA (8, 9).

The cytotoxicity of Me-lex has been documented in bacteria, yeast, and mammalian cells, strongly indicating that 3-meA is directly lethal or is converted into a lethal lesion (7). Available evidence indicates that cell killing by 3-meA is due, at least in part, to blockage of DNA replication as evidenced by arrest of S-phase progression in murine cells treated with Me-lex and by arrest of in vitro DNA synthesis by bacterial and phage DNA polymerases (reviewed in ref. 10). X-ray structures of bacterial and phage and also mammalian DNA polymerases bound to DNA have revealed contact between template-binding amino acid side chains and the N3 position of adenine. This interaction, which is sterically blocked by the methyl group in 3-meA, is apparently necessary for DNA synthesis. In light of the conservation of structure and mechanism among DNA polymerase families (11), the interaction is likely required for synthesis by mammalian polymerases as well, although direct demonstration is currently lacking. The lethality of 3-meA may thus reflect a crucial role of the N3 atom of adenine in stabilizing contact between DNA polymerases and the template strand required for insertion of the complementary deoxynucleotide triphosphate.

Base excision repair is the primary defense against 3-meA in both prokaryotes and eukaryotes (12). In mammalian short patch base excision repair, removal of 3-meA is initiated by alkyladenine DNA glycosylase (AAG; also methylpurine-DNA glycosylase and 3-meA-DNA glycosylase), which cleaves the glycosylic linkage between the adducted base and deoxyribose, producing an abasic site (13). AAG remains bound to its abasic site product until displaced by Ape1, an apurinic endonuclease (Ap endo) that incises the phosphodiester backbone 5′ to the baseless residue. The terminal 5′-deoxyribose phosphate is then excised by the 5′-deoxyribose phosphodiesterase activity of the 8-kDa NH2-terminal domain of DNA polymerase β. The resulting single nucleotide gap, with 3′-OH and 5′-phosphate at its margins, is filled and ligated by a complex of DNA polymerase β and DNA ligase III.

Bacterial and yeast mutants lacking DNA glycosylases homologous to AAG are hypersensitive to methylating agents (13), including Me-lex (1416). In accord, Samson et al. have shown that embryonic stem cell lines derived from Aag null mice (17) are hypersensitive to killing by Me-lex (10). Deficiency of AAG resulted in elevated rates of sister chromatid exchange and chromosome aberration, induction of p53, and prolonged S-phase arrest with eventual exit into apoptosis. This constellation of findings, together with detection of 3-meA in Aag null but in not wild-type cells (18), represents the first documentation of the deleterious biological effects of unrepaired 3-meA in mammalian cells. Of significance is the observation that deletion of Ap endo in yeast potentiates Me-lex sensitivity, suggesting that abasic sites arising from disrupted repair of 3-meA can lead to cell killing (15, 16).

The methylating agents procarbazine and temozolomide are used extensively in the adjuvant chemotherapy of gliomas (19, 20). These agents produce roughly equal amounts of 3-meA and O6-meG, another documented cytotoxic lesion (5). The importance of O6-meG in killing by methylators and the contribution of repair of O6-meG to methylator resistance have been intensively examined in glioma cells (e.g., refs. 21, 22), whereas comparable studies for 3-meA are lacking. To address this deficiency, we have characterized the Me-lex sensitivity of 10 human glioma-derived cell lines that differ in their ability to repair or tolerate O6-meG. Our findings indicate that 3-meA is cytotoxic in human glioma cells and suggest that Me-lex may have clinical utility.

Cell culture. Cell lines were grown as adherent cultures at 37°C in 95%/5% humidified air/CO2 in DMEM/F12 supplemented with 2% iron-supplemented bovine calf serum (HyClone, Logan, UT), 100 units/mL penicillin, 100 mg/mL streptomycin, and 0.25 mg/mL amphotericin B (Invitrogen, Carlsbad, CA). Alkylating agent resistance and dependence of resistance on O6-meG-DNA methyltransferase (MGMT) of the glioblastoma-derived lines SNB19, SF763, SF767, T98G, UW18, and UW456 (21, 22) as well as of A1235 and its mismatch repair (MMR)-deficient derivative A1235-MR4 (hereafter MR4; ref. 23) have been described previously. Res256 was derived from a pediatric medulloblastoma (24). The MMR-deficient variant Res256MSH6 was isolated by serial selection for resistance to temozolomide; increased resistance was accompanied by loss of expression of MSH6 mRNA and MSH6 protein as revealed by reverse transcription-PCR and Western blotting, respectively.6

6

M. Bobola and J.R. Silber, unpublished data.

AAG and Ap endo activity. Subconfluent cultures were harvested by trypsinization, washed twice with PBS, flash frozen in liquid N2, and stored at −80°C. Extracts were prepared as described previously (25) and stored as multiple small aliquots. Aliquots could be flash frozen and thawed thrice without detectable loss of AAG or Ap endo activity. Activity was normalized to cell number by using DNA content of crude cell lysates estimated by the method of Burton (26), assuming a DNA content of 5 pg for a human cell.

AAG activity (pmol 3-meA released/106 cells/hour) was measured in an established assay (27, 28) that quantitates release of 3H-labeled N-methylpurines from a DNA substrate methylated with [3H]methylnitrosourea (13-20 Ci/mmol; Amersham, Indianapolis, IN). Each determination comprised measurement of radioactivity released as a function of time by three different amounts of extract. Rates for each amount were determined by regression analysis of points on the initial, linear portion of the curve. TLC of the released bases showed that ∼90% of the counts comigrated with a 3-meA standard and 10% with a 7-methylguanine standard (data not shown), in accord with the observed substrate specificity of human AAG (27).

Ap endo activity (fmol abasic sites incised/cell/minute) was measured in a standard, highly sensitive assay that measures the conversion of plasmid DNA from supercoiled to relaxed form caused by incision at an abasic site (29). A detailed description of the procedure has been presented previously (30). Each determination comprised measurement of four different amounts of extract, each assayed in triplicate, and yielded activity calculated by regression analysis of points on the linear portion of the curve.

Antisense suppression of AAG and Ape1. Two antisense oligonucleotides (ASOs; Operon, Huntsville, AL) directed against human AAG, 5′-GTCTGCCGGCCACCCCTTGAGT-3′, and 5′-ATTACCTGTCCCAGAAATGCC-3′ were used. The first targets nucleotides 405 to 426 of the mRNA (accession NM 002434.2), and the second targets the second exon-intron junction (nucleotides 2,435-2,456 of the genomic sequence; accession AF499437) and includes the TCCC motif that occurs in half of ASOs that suppress expression of tumor necrosis factor-β (31). Both have low potential for self-association. As a control for specificity, cells were concurrently transfected with oligomers complementary to the ASOs [i.e., sense oligonucleotides (SO)]. The sequences of ASOs targeting Ape1 have been described previously (32). ASOs were introduced into cells by cationic lipid-mediated transfection as described previously (32). After 72 h of incubation, cells were plated for enzymatic assay and cytotoxicity determination and for abasic site measurement.

Me-lex survival. Me-lex (8) was dissolved and diluted in DMSO immediately before use so that a constant volume was added for all doses. No-drug controls received an equivalent volume of solvent. The final concentration of DMSO was 1%. Colony-forming ability following a 2-h exposure to Me-lex was determined by clonogenic assay. Drug sensitivity was quantitated by analysis of survival curves (log surviving fraction versus dose) using standard methods we have described previously in detail (e.g., ref. 21). The survival variables LD10 (the dose required to reduce overall survival to 10%), DT (the threshold dose below which cells are insensitive to alkylator cytotoxicity), and D37 (the rate of killing as indicated by the slope of the linear portion of the survival curve) were derived by regression analysis of the linear portion of a composite kill curve derived from the replicate assays. The three survival variables are presented as mean ± SD and provide a complete description of alkylator sensitivity. Unless otherwise stated, each survival curve represents three to five independent experiments as indicated, in which every dose was assayed in triplicate (i.e., 9 to 15 determinations per drug concentration).

Quantitation of abasic sites. Abasic sites were quantitated by using a slot-blot technique (33) as we have described previously in detail (32), except that chemiluminescent bands were measured using phosphorimage analysis rather than autoradiography. Abasic site levels in cellular DNA were determined by comparison with a standard curve constructed with reference DNA containing 1,080 abasic sites per 106 nucleotides, prepared as described previously (32).

Statistical analysis. Data analysis and statistical procedures were done using Microsoft Excel. Comparison of means was by Student's t test assuming unequal variances. Relationships between continuous variables were assessed by regression analysis. Statistically significant relationships were determined at the 95% confidence level.

Me-lex sensitivity of human glioma cells. In this work, we assessed the effect of 3-meA in DNA on survival of human glioma cells. To do so, we measured the cytotoxicity of Me-lex, a sequence-specific alkylator that introduces 3-meA as the predominant adduct (>90% of base lesions; ref. 9). We quantitated Me-lex cytotoxicity by clonogenic assay and analysis of survival curves in 10 human glioma lines, including lines that are either proficient or deficient in MGMT, the repair protein that removes cytotoxic O6-meG adducts from DNA (6). The survival variables DT, D37, and LD10 that characterize Me-lex sensitivity of the lines are summarized in Table 1; examples of survival curves from which these variables were derived are shown in Fig. 1. The lines differed widely in sensitivity as evidenced by a 43-fold range in LD10, a 40-fold range in D37, and a >31-fold difference in DT. Incubating two of the lines (SF767 and SNB19) with hydrolyzed Me-lex did not reduce survival (data not shown), confirming that an active alkylating moiety is required for cell killing (10). Simultaneous incubation with Me-lex and a 5-fold molar excess of netropsin, an innocuous natural product that competes with Me-lex for binding to the minor groove in A-T rich regions of DNA (14), decreased cell killing 4-fold (data not shown). These control experiments indicate that Me-lex cytotoxicity was associated with methylation at minor groove A-T rich sequences in glioma cells.

Table 1.

Me-lex sensitivity of human brain tumor lines

LineDT (μmol/L)D37 (μmol/L)LD10 (μmol/L)MGMT*AAGAp endo
MGMT+       
    SF763 185 ± 10 429 ± 18 119 ± 19 2.8 ± 0.51 0.63 ± 0.13 
    UW456 44 ± 4.7 103 ± 8.9 225 ± 21 144 ± 22 4.5 ± 1.2 0.38 ± 0.08 
    UW18 37 ± 4.4 89 ± 4.2 194 ± 11 74 ± 20 3.5 ± 0.40 0.39 ± 0.07 
    T98G 19 ± 3.9 65 ± 3.9 144 ± 8 104 ± 14 3.0 ± 0.62 0.20 ± 0.05 
    SF767 1.4 ± 0.3 4.6 ± 0.5 10 ± 1.2 61 ± 12 3.5 ± 0.59 0.36 ± 0.04 
MGMT       
    A1235 9.4 ± 1.9 15 ± 3.2 32 ± 6.0 <0.25 2.7 ± 0.44 0.45 ± 0.07 
    Res256 7.5 ± 0.5 13 ± 0.7 28 ± 2 <0.25 3.1 ± 0.58 0.28 ± 0.05 
    SNB19 2.8 ± 0.9 8.7 ± 1.9 19 ± 4.2 <0.25 3.8 ± 0.31 0.51 ± 0.14 
MGMT MMR       
    MR4 9.0 ± 2.0 15 ± 1.6 32 ± 2.5 <0.25 4.5 ± 0.82 0.85 ± 0.24 
    Res256MSH6 6.2 ± 1.1 11 ± 1.1 24 ± 2.3 <0.25 4.7 ± 0.97 0.30 ± 0.6 
LineDT (μmol/L)D37 (μmol/L)LD10 (μmol/L)MGMT*AAGAp endo
MGMT+       
    SF763 185 ± 10 429 ± 18 119 ± 19 2.8 ± 0.51 0.63 ± 0.13 
    UW456 44 ± 4.7 103 ± 8.9 225 ± 21 144 ± 22 4.5 ± 1.2 0.38 ± 0.08 
    UW18 37 ± 4.4 89 ± 4.2 194 ± 11 74 ± 20 3.5 ± 0.40 0.39 ± 0.07 
    T98G 19 ± 3.9 65 ± 3.9 144 ± 8 104 ± 14 3.0 ± 0.62 0.20 ± 0.05 
    SF767 1.4 ± 0.3 4.6 ± 0.5 10 ± 1.2 61 ± 12 3.5 ± 0.59 0.36 ± 0.04 
MGMT       
    A1235 9.4 ± 1.9 15 ± 3.2 32 ± 6.0 <0.25 2.7 ± 0.44 0.45 ± 0.07 
    Res256 7.5 ± 0.5 13 ± 0.7 28 ± 2 <0.25 3.1 ± 0.58 0.28 ± 0.05 
    SNB19 2.8 ± 0.9 8.7 ± 1.9 19 ± 4.2 <0.25 3.8 ± 0.31 0.51 ± 0.14 
MGMT MMR       
    MR4 9.0 ± 2.0 15 ± 1.6 32 ± 2.5 <0.25 4.5 ± 0.82 0.85 ± 0.24 
    Res256MSH6 6.2 ± 1.1 11 ± 1.1 24 ± 2.3 <0.25 4.7 ± 0.97 0.30 ± 0.6 
*

fmol [3H]CH3 groups removed/106 cells: the lower limit of detection is 0.25.

pmol 3-meA released/106 cells/hour.

fmol abasic sites incised/cell/minute.

Fig. 1.

Me-lex cytotoxicity in human glioma-derived cell lines. Me-lex sensitivity was measured by clonogenic assay in the MGMT+ MMR+ cell lines SF763 (▪), UW456 (▴), UW18 (•), T98G (⧫; A), and SF767 (□; B); the MGMT MMR+ lines A1235 (×) and SNB19 (◊; B); and the MGMT MMR line MR4 (○; B). Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion of the curves was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. The survival variables LD10, DT, and D37 derived from the curves are compiled in Table 1.

Fig. 1.

Me-lex cytotoxicity in human glioma-derived cell lines. Me-lex sensitivity was measured by clonogenic assay in the MGMT+ MMR+ cell lines SF763 (▪), UW456 (▴), UW18 (•), T98G (⧫; A), and SF767 (□; B); the MGMT MMR+ lines A1235 (×) and SNB19 (◊; B); and the MGMT MMR line MR4 (○; B). Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion of the curves was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. The survival variables LD10, DT, and D37 derived from the curves are compiled in Table 1.

Close modal

AAG, Ap endo, and MGMT activities. We measured the activity of AAG, the base excision repair enzyme that excises 3-meA from DNA, as well as Ap endo activity, which cleaves the DNA backbone adjacent to the resulting abasic site (Table 1). In contrast to Me-lex cytotoxicity, AAG and Ap endo activity varied little among the cell lines, suggesting that additional mechanisms may contribute to sensitivity. Regression analysis revealed no significant relationship between the two activities. In addition, univariate and multivariate regression analyses showed no association between AAG or Ap endo activity and either LD10, DT, or D37 for Me-lex. Nevertheless, because correlative data cannot establish the presence or absence of biochemical function, we used antisense suppression to directly examine the contribution to Me-lex resistance of both AAG and Ape1, the predominant Ap endonuclease activity in human cells (34), as presented below.

We also measured the activity of MGMT. We observed that, on average, MGMT-proficient lines were more resistant to Me-lex than MGMT-deficient lines as judged by a 7-fold greater LD10 (Table 1). We did not expect this result because the yield of O6-meG adducts introduced by Me-lex is very low (0.07%; ref. 9) and, we believe, insufficient to confer demonstrable cytotoxicity. We examine this result in relation to additional findings in a later section.

Suppression of AAG activity increases Me-lex cytotoxicity. AAG is the only known enzyme in mammalian cells that catalyzes excision of 3-meA (12, 13), and Aag null mouse embryonic stem cells and fibroblasts are hypersensitive to Me-lex (17, 27). To examine the contribution of AAG to Me-lex resistance in human glioma cells, we used ASOs to suppress activity in SF767 (MGMT+MMR+), SNB19 (MGMTMMR+), and MR4 (MGMTMMR). As illustrated in Fig. 2, ASO treatment reduced AAG activity and resistance to Me-lex in the three lines. Resistance was uniformly reduced within each of the ASO-treated lines as indicated by the constant slope of the final linear portion of the survival curves, suggesting that all cells were similarly affected. As summarized in Table 2, ASO treatment reduced AAG activity 1.6- to 1.8-fold, and LD10 1.4- to 1.5-fold, relative to SO-treated cells, differences that were statistically significant in every case. Greater sensitivity to Me-lex reflected statistically significant reduction in both DT and D37. The enhancement in glioma cell killing accompanying suppression of AAG was comparable in magnitude with the difference in sensitivity between wild-type and Aag−/− mouse embryonic cells (17, 27).

Fig. 2.

ASOs against AAG reduce excision of 3-meA (A-C) and increase Me-lex cytotoxicity (D-F) in MGMT+ MMR+ (SF767), MGMT MMR+ (SNB19), and MGMT MMR (MR4) human glioma cell lines. AAG activity in ASO-treated (×) and SO-treated (○) cells was measured by biochemical assay: representative time courses are shown. Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. Mean AAG activities, and survival variables derived from the curves, are compiled in Table 2. C, inset, AAG protein content in 15 μg of whole-cell extract prepared from SO- and ASO-treated MR4 detected by Western blotting as described previously (25) using a goat anti-human AAG antibody (Imgenex, San Diego, CA). Densitometry revealed that ASO treatment reduced signal intensity 2-fold comparable with the 1.8-fold mean reduction in activity (Table 2). Std, 100 ng of hexahistidine-tagged human recombinant AAG (a generous gift of Dr. Cheng-Yao Chen, University of Washington, Seattle, WA).

Fig. 2.

ASOs against AAG reduce excision of 3-meA (A-C) and increase Me-lex cytotoxicity (D-F) in MGMT+ MMR+ (SF767), MGMT MMR+ (SNB19), and MGMT MMR (MR4) human glioma cell lines. AAG activity in ASO-treated (×) and SO-treated (○) cells was measured by biochemical assay: representative time courses are shown. Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. Mean AAG activities, and survival variables derived from the curves, are compiled in Table 2. C, inset, AAG protein content in 15 μg of whole-cell extract prepared from SO- and ASO-treated MR4 detected by Western blotting as described previously (25) using a goat anti-human AAG antibody (Imgenex, San Diego, CA). Densitometry revealed that ASO treatment reduced signal intensity 2-fold comparable with the 1.8-fold mean reduction in activity (Table 2). Std, 100 ng of hexahistidine-tagged human recombinant AAG (a generous gift of Dr. Cheng-Yao Chen, University of Washington, Seattle, WA).

Close modal
Table 2.

Antisense suppression of AAG and Ape1 increases Me-lex cytotoxicity in human glioma cells

LineAAG
Ape1
SenseASOSense/ASOPSenseASOSense/ASOP
SF767         
    Activity* 3.3 ± 0.9 1.8 ± 0.6 1.8 ≤0.015 0.30 ± 0.06 0.16 ± 0.2 1.9 ≤0.020 
    LD10 (μmol/L) 10 ± 1.3 6.6 ± 0.9 1.5 ≤0.0002 10 ± 1.3 7.0 ± 0.7 1.5 ≤0.001 
    DT (μmol/L) 1.4 ± 0.3 0.5 ± 0.2 2.8 ≤0.0006 1.6 ± 0.3 1.0 ± 0.4 1.6 ≤0.05 
    D37 (μmol/L) 4.5 ± 0.5 2.8 ± 0.7 1.6 ≤0.0004 4.7 ± 0.2 3.3 ± 0.2 1.6 ≤0.0001 
SNB19         
    Activity 3.9 ± 0.3 2.4 ± 0.1 1.6 ≤0.002 0.38 ± 0.12 0.20 ± 0.04 1.9 ≤0.05 
    LD10 (μmol/L) 22 ± 2.0 15 ± 1.4 1.5 ≤0.0002 26 ± 5.1 16 ± 1.6 1.5 ≤0.01 
    DT (μmol/L) 3.6 ± 1.1 1.0 ± 1.0 3.6 ≤0.006 3.8 ± 1.1 2.0 ± 1.3 1.9 ≤0.05 
    D37 (μmol/L) 9.9 ± 0.8 6.7 ± 0.7 1.5 ≤0.0002 12 ± 3.4 7.3 ± 0.7 1.6 ≤0.04 
MR4         
    Activity 4.2 ± 0.8 2.3 ± 0.6 1.8 ≤0.012 0.36 ± 0.19 0.19 ± 0.06 1.9 ≤0.05 
    LD10 (μmol/L) 31 ± 0.9 22 ± 0.2 1.4 ≤0.003 39 ± 1.2 28 ± 1.4 1.4 ≤0.0001 
    DT (μmol/L) 3.7 ± 0.5 2.9 ± 0.2 1.3 ≤0.028 7.7 ± 1.4 4.2 ± 1.4 1.8 ≤0.015 
    D37 (μmol/L) 14 ± 2.8 9.7 ± 0.2 1.4 ≤0.0001 18 ± 2.8 13 ± 0.5 1.4 ≤0.0001 
LineAAG
Ape1
SenseASOSense/ASOPSenseASOSense/ASOP
SF767         
    Activity* 3.3 ± 0.9 1.8 ± 0.6 1.8 ≤0.015 0.30 ± 0.06 0.16 ± 0.2 1.9 ≤0.020 
    LD10 (μmol/L) 10 ± 1.3 6.6 ± 0.9 1.5 ≤0.0002 10 ± 1.3 7.0 ± 0.7 1.5 ≤0.001 
    DT (μmol/L) 1.4 ± 0.3 0.5 ± 0.2 2.8 ≤0.0006 1.6 ± 0.3 1.0 ± 0.4 1.6 ≤0.05 
    D37 (μmol/L) 4.5 ± 0.5 2.8 ± 0.7 1.6 ≤0.0004 4.7 ± 0.2 3.3 ± 0.2 1.6 ≤0.0001 
SNB19         
    Activity 3.9 ± 0.3 2.4 ± 0.1 1.6 ≤0.002 0.38 ± 0.12 0.20 ± 0.04 1.9 ≤0.05 
    LD10 (μmol/L) 22 ± 2.0 15 ± 1.4 1.5 ≤0.0002 26 ± 5.1 16 ± 1.6 1.5 ≤0.01 
    DT (μmol/L) 3.6 ± 1.1 1.0 ± 1.0 3.6 ≤0.006 3.8 ± 1.1 2.0 ± 1.3 1.9 ≤0.05 
    D37 (μmol/L) 9.9 ± 0.8 6.7 ± 0.7 1.5 ≤0.0002 12 ± 3.4 7.3 ± 0.7 1.6 ≤0.04 
MR4         
    Activity 4.2 ± 0.8 2.3 ± 0.6 1.8 ≤0.012 0.36 ± 0.19 0.19 ± 0.06 1.9 ≤0.05 
    LD10 (μmol/L) 31 ± 0.9 22 ± 0.2 1.4 ≤0.003 39 ± 1.2 28 ± 1.4 1.4 ≤0.0001 
    DT (μmol/L) 3.7 ± 0.5 2.9 ± 0.2 1.3 ≤0.028 7.7 ± 1.4 4.2 ± 1.4 1.8 ≤0.015 
    D37 (μmol/L) 14 ± 2.8 9.7 ± 0.2 1.4 ≤0.0001 18 ± 2.8 13 ± 0.5 1.4 ≤0.0001 
*

AAG activity is expressed as pmol 3-meA released/106 cells/hour and Ape1 as fmol abasic sites incised/cell/minute.

Multiple control experiments provide evidence that increased Me-lex sensitivity reflected reduction of AAG activity rather than nonspecific effects of transfection or ASOs on cell physiology or DNA repair capacity. First, ASO- and SO-treated cells did not differ in plating efficiency (SF767, 28 ± 6% versus 30 ± 7%; SNB19, 27 ± 5% versus 29 ± 5%; MR4, 27 ± 10% versus 30 ± 10%). Second, treatment with cationic lipid alone had no effect on alkylator sensitivity in SNB19 (32) or on sensitivity to hydroxyurea and 4-NQO in SF767 (25). Third, the AAG activity and Me-lex sensitivity of SO-treated cells did not differ from that of untransfected cells (data not shown; Table 1). Fourth, ASOs and SOs had no effect on Ap endo activity in the three lines or on MGMT activity in SF767.6 Fifth, ASOs did not increase the sensitivity of SNB19 to camptothecin or cisplatin or of SF767 to camptothecin or 4-NQO,6 agents that do not produce 3-meA in DNA.

Suppression of AAG reduces Me-lex-induced abasic sites. Excision of 3-meA by AAG produces abasic sites that can be detected in a slot-blot protocol using an aldehyde reactive probe that reacts with the ring-open isomer of deoxyribose in DNA (33). We expected that suppression of AAG would result in reduced abasic site content in Me-lex-treated cells. To test this prediction, we measured abasic site levels in ASO- and SO-treated SF767 immediately following exposure to 8 μmol/L Me-lex (see Fig. 2). As shown in Fig. 3A, abasic site abundance was lower in DNA from ASO-treated cells, in accord with our expectation. In the experiment shown, a 1.9-fold reduction in AAG activity (2.3 ± 0.4 versus 1.2 ± 0.09 pmol 3-meA excised/106 cells/hour) was accompanied by a 1.4-fold reduction in abasic site content (26 ± 1.0 versus 18 ± 4.0 Ap sites/106 nucleotides). In three different determinations, the abasic site level was, on average, 1.6-fold lower in ASO-treated than in SO-treated cells (P ≤ 0.017). Preliminary experiments revealed comparable reduction in Me-lex-induced abasic site content in SNB19 and MR4 treated with anti-AAG ASO (data not shown). These findings provide additional evidence that increased Me-lex sensitivity in ASO-treated glioma cells reflects reduction of AAG activity and that Me-lex cytotoxicity is mediated by unrepaired 3-meA.

Fig. 3.

Abasic site content of Me-lex-treated glioma cells following antisense-mediated suppression of AAG (A) and Ape1 (B) activity. SF767 cells were treated with 8 μmol/L Me-lex (A) and SNB19 cells with 40 μmol/L (B). Abasic sites in ASO-treated (x-x) and SO-treated (o-o) cells were measured in a slot-blot assay using an aldehyde-reactive probe (33). B, inset, example of a standard curve constructed with calf thymus DNA containing increasing amounts of abasic sites prepared as described in Silber et al. (32).

Fig. 3.

Abasic site content of Me-lex-treated glioma cells following antisense-mediated suppression of AAG (A) and Ape1 (B) activity. SF767 cells were treated with 8 μmol/L Me-lex (A) and SNB19 cells with 40 μmol/L (B). Abasic sites in ASO-treated (x-x) and SO-treated (o-o) cells were measured in a slot-blot assay using an aldehyde-reactive probe (33). B, inset, example of a standard curve constructed with calf thymus DNA containing increasing amounts of abasic sites prepared as described in Silber et al. (32).

Close modal

Suppression of Ape1 increases Me-lex cytotoxicity. Concurrent deletion of the genes for the two known Ap endo activities in the yeast Saccharomyces cerevisiae increases Me-lex killing (15, 16). To examine the contribution of Ap endo activity to Me-lex resistance in human glioma cells, we used ASO against Ape1 in SF767, SNB19, and MR4. Ape1 is the predominant Ap endo activity in human cells, catalyzing >95% of abasic site cleavage (34). As illustrated in Fig. 4, ASO treatment reduced Ap endo activity and resistance to Me-lex in the three lines. In all cases, resistance was uniformly reduced in the ASO-treated lines as indicated by the constant slope of the final linear portion of the survival curves, suggesting that all cells in each line were similarly affected. As summarized in Table 2, ASO treatment reduced Ap endo activity 1.9-fold, and LD10 1.4- to 1.5-fold, relative to SO-treated cells, differences that were statistically significant in every case. Greater sensitivity to Me-lex reflected significant reduction in both DT and D37. We have previously shown that suppression of the Ap endo activity of Ape1 increases the level of temozolomide-induced and 1,3-bis(2-chloroethyl)-1-nitrosourea–induced abasic sites in SNB19 (32). In accord, abasic site content was 1.6-fold greater in ASO-treated compared with SO-treated SNB19 (68 ± 5.4 versus 41 ± 2.4 Ap sites/106 nucleotides) 2 h after exposure to 40 μmol/L Me-lex (Fig. 3B). The 1.5-fold reduction in LD10 that accompanied the increase in abasic sites suggests that unrepaired abasic sites derived from AAG-catalyzed excision of 3-meA mediate Me-lex cytotoxicity.

Fig. 4.

ASOs against Ape1 reduce Ap endo activity (A-C) and increase Me-lex cytotoxicity (D-F) in MGMT+ MMR+ (SF767), MGMT MMR+ (SNB19), and MGMT MMR (MR4) human glioma cell lines. Ap endo activity and Me-lex sensitivity in ASO-treated (×) and SO-treated (○) cells were measured by biochemical and clonogenic assay, respectively. Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. Mean Ap endo activities, and survival variables derived from the curves, are compiled in Table 2. We have previously shown that ASO-mediated suppression of Ap endo activity is accompanied by a comparable reduction of Ape1 protein content (e.g., ref. 32).

Fig. 4.

ASOs against Ape1 reduce Ap endo activity (A-C) and increase Me-lex cytotoxicity (D-F) in MGMT+ MMR+ (SF767), MGMT MMR+ (SNB19), and MGMT MMR (MR4) human glioma cell lines. Ap endo activity and Me-lex sensitivity in ASO-treated (×) and SO-treated (○) cells were measured by biochemical and clonogenic assay, respectively. Points, mean of 9 to 15 determinations (i.e., triplicate determinations at each drug dose in three to five independent experiments); bars, SD. The final linear portion was constructed by linear regression analysis. Where error bars are not shown, the SD is too small to be indicated. Mean Ap endo activities, and survival variables derived from the curves, are compiled in Table 2. We have previously shown that ASO-mediated suppression of Ap endo activity is accompanied by a comparable reduction of Ape1 protein content (e.g., ref. 32).

Close modal

A factor(s) associated with expression of the MGMT gene promotes Me-lex resistance. Five of the 10 lines we examined are MGMT+ and another 3 are MGMT; as indicated in Table 1, these 8 lines are proficient in MMR. We also measured Me-lex cytotoxicity in MMR-deficient derivatives of two of the MGMT lines. Loss of MMR in MGMT cells permits tolerance of O6-meG lesions without lethality (5).

We have previously shown that, among the eight MMR+ lines, MGMT proficiency is accompanied by greater resistance to clinical and laboratory methylators, such as temozolomide and N-methyl-N′-nitro-N-nitrosoguanidine (data not shown; refs. 21, 22). We did not anticipate a comparable result for Me-lex because the yield of O6-meG adducts is ∼100-fold lower (approximately 0.07% versus 6-12%; ref. 9). Yet, we observed a 7-fold higher mean LD10 for the five MGMT+ versus five MGMT lines (200 ± 152 versus 27 ± 6 μmol/L; P ≤ 0.035), as well as a significant correlation between MGMT activity and LD10 (P ≤ 0.006); for the latter calculation, we assumed a value of 0.25 fmol/106 cells, our lower limit of detection (35), for the MGMT lines. Four observations indicate that a factor(s) associated with expression of the MGMT gene, rather than the repair activity of the protein, is the major factor responsible for increased Me-lex resistance in the MGMT+ lines. First, O6-meG accounts for <0.1% of Me-lex-induced base adducts, whereas 3-meA accounts for >90% (9); this 1,000-fold difference would likely obscure any contribution of O6-meG to cell killing in a clonogenic assay. Second, one of the MGMT+ lines, SF767, was more sensitive than all other lines (Table 1), and a second, SF763, displayed no detectable resistance to low Me-lex doses (DT = 0; Table 1), suggesting that factors associated with MGMT expression, rather than or in addition to MGMT activity, may contribute to Me-lex resistance. Third, loss of MMR had no effect on Me-lex sensitivity as evidenced by unchanged LD10 for the MMR derivatives MR4 and Res256MSH6 relative to the parental MMR+ lines A1235 and Res256, respectively (Table 1). In contrast, loss of MMR had profound effects on sensitivity to other methylators: the LD10 for N-methyl-N′-nitro-N-nitrosoguanidine is 40-fold greater in MR4 than in A1235 (5.9 versus 0.15 μmol/L) and the LD10 for temozolomide is 10-fold greater in Res256MSH6 than in Res256 (200 versus 20 μmol/L).6 Fourth, ablation of MGMT with the substrate analogue inhibitor O6-benzylguanine had no effect on Me-lex cytotoxicity in SF767 (10.7 versus 10.4 μmol/L), SF763 (409 versus 398 μmol/L), and T98G (145 versus 144 μmol/L) cells. In contrast, the LD10 for temozolomide was reduced 2- to 20-fold in the three lines (22). We conclude that expression of the MGMT gene in our glioma lines is associated with additional phenomena that confer resistance to 3-meA.

Knowledge of the mechanisms responsible for resistance of human primary brain tumors to therapeutic alkylating agents is limited by our inability to assess the cytotoxicity of individual alkyl base adducts in DNA. The multiplicity of adducts induced by most alkylators (1, 2) and the numerous, often redundant repair and tolerance pathways that process DNA lesions (e.g., refs. 12, 36) pose major impediments to understanding. The adducts O6-meG and 3-meA are exceptions to this dilemma. Demonstration of the lethality of O6-meG was made possible by the exclusive role of MGMT in repair and specific ablation of MGMT activity with the substrate analogue inhibitor O6-benzylguanine (5). The lethality of O6-meG in human brain tumor cells is well established, and O6-benzylguanine and its derivatives are in clinical trials (6). Demonstration of the lethality of 3-meA was made possible by creation of knockout mice lacking AAG (17, 27), the enzyme that catalyzes initial removal of 3-meA in base excision repair of this lesion (13), and synthesis of the sequence-specific methylator Me-lex that induces 3-meA as the predominant adduct (8, 9). Although cell killing mediated by 3-meA has been documented in mouse embryonic stem cells, fibroblasts, and neurons (7, 37), the role of 3-meA in alkylator cytotoxicity in human brain tumor cells has yet to be established.

3-meA is cytotoxic in human glioma-derived cell lines. Here, we show that Me-lex is cytotoxic in 10 human glioma lines and that antisense suppression of AAG activity potentiates Me-lex killing. These results are in accord with the Me-lex hypersensitivity of embryonic stem cells derived from Aag knockout mice (17, 27). Suppression of AAG in glioma cells was accompanied by a decrease in the level of abasic sites; abasic sites are the product of AAG-mediated excision of 3-meA and are themselves cytotoxic. These results provide the first direct evidence that unrepaired 3-meA in DNA is a lethal lesion in human brain tumor cells. We also showed that antisense suppression of Ape1 increases the sensitivity of glioma cells to Me-lex; Ape1 is the major mammalian activity that incises abasic sites and is the second enzyme in base excision repair of 3-meA. Suppression of Ape1 was accompanied by an increase in abasic site abundance. Taken together, the foregoing experiments provide strong biochemical evidence that (a) 3-meA is cytotoxic in human glioma cells and that AAG promotes resistance to this adduct and (b) abasic sites resulting from 3-meA repair are cytotoxic and that Ape1 promotes resistance to these derivative lesions.

Resistance to Me-lex is multifactorial. Our data also indicate that as yet uncharacterized determinants in addition to AAG and Ape1 promote resistance to 3-meA in human glioma cells. This conclusion is consistent with the lack of correlation of Me-lex sensitivity with either AAG or Ap endo activity. The lack of association is exemplified by SF767, which has AAG and Ap endo activities comparable with that of nine other lines but is more sensitive to Me-lex as evidenced by 2- to 40-fold lower LD10 (Table 1). Our data show that one or more additional determinants of Me-lex resistance are associated with expression of the MGMT gene. We observed that, with the exception of SF767, MGMT-proficient lines are more resistant to Me-lex than MGMT-deficient lines (Table 1) and also that MGMT activity is correlated with survival. Yet, MGMT-mediated repair of O6-meG is not the mechanism underlying these observations or at least is not the primary mechanism. Two types of direct evidence support this conclusion. First, loss of MMR in two MGMT lines (A1235 and Res256) had no effect on Me-lex sensitivity (Table 1), whereas N-methyl-N′-nitro-N-nitrosoguanidine cytotoxicity was potentiated 40-fold for MR4 relative to the parental A1235 and temozolomide cytotoxicity was potentiated 10-fold greater for Res256MSH6 relative to the parental Res256.6 Because loss of MMR negates the lethality of O6-meG (5, 36), these findings indicate that the small proportion of O6-meG adducts (0.07%; ref. 9) caused by Me-lex was not responsible for detectable levels of killing. Hence, lack of repair by MGMT could not be responsible for the diminished survival of MGMT-deficient lines. Moreover, ablating MGMT with O6-benzylguanine did not change the Me-lex sensitivity of MGMT-proficient SF767, SF763, or T98G cells. We think it is probable that loss of MGMT expression, believed to be due to transcriptional silencing (6), is but one manifestation of a wider sensitivity to DNA alkylation damage mediated by epigenetic changes in gene expression. In accord with this hypothesis are reports that loss of MGMT expression in human primary brain tumors is accompanied by epigenetic silencing of other genes, including some that promote DNA repair (e.g., MLH1), damage response (e.g., p16INK4a and p14ARF), and cell survival (e.g., p53 and p73; refs. 3840). Epigenetic changes include not only decreases but also increases in levels of gene expression (41), some of which may affect Me-lex sensitivity as well.

Yet other likely determinants of Me-lex sensitivity have been identified. We have shown that suppression of the Werner syndrome protein WRN, a DNA helicase/exonuclease, increases Me-lex killing in SF767 and SNB19 (25). WRN promotes genomic stability, in part by mediating resumption of DNA replication at replication forks stalled by DNA damage (42), probably including 3-meA. In addition, the slow but measurable removal of 3-meA from Me-lex-treated Aag knockout mouse embryonic fibroblasts is completely ablated when the cells are also defective in nucleotide excision repair (43), implicating nucleotide excision repair in Me-lex resistance.

A hypersensitive cell line yields possible clues to mechanisms of Me-lex cytotoxicity. The extreme sensitivity of SF767 to Me-lex may afford clues to the mechanisms of Me-lex lethality. SF767 becomes more sensitive than other glioma lines to clinical and laboratory methylators when MGMT is ablated with O6-benzylguanine (21, 22), consistent with deficiency in resistance to 3-meA. As ∼10% of alkyl base adducts induced by temozolomide and procarbazine are 3-meA, elucidation of the sources of SF767 hypersensitivity could provide new insights into the mechanisms of glioma resistance to clinically used methylating agents. The source(s) of SF767 hypersensitivity to Me-lex does not include a defect in the initial steps of base excision repair because AAG and Ap endo levels are representative (Table 1), and suppression of either activity significantly increases sensitivity (Table 2). SF767 is not hypersensitive to UV light or 137Cs γ-rays,6 suggesting that it is not defective in nucleotide excision repair or in the major pathways of single-strand and double-strand break repair. In contrast, the LD10 for bleomycin is 3- to 12-fold lower in SF767 than in seven of eight glioma lines we have examined, including six used in this study.6 Approximately 40% of bleomycin-induced lesions are C4′-oxidized abasic sites, which, when opposite or closely opposed to a single-strand break, produce a double-strand lesion that cannot be resolved by base excision and strand break repair pathways (44). Such clustered lesions are strong blocks to DNA replication and promote double-strand break formation (45), raising the possibility that SF767 is deficient in a mechanism(s) that promotes recovery from stalled replication forks. The binding of Me-lex at A-T rich regions (4) may promote formation of clustered 3-meA adducts akin to the case of clustered bleomycin damage. These clustered lesions may tax mechanisms for recovery from stalled or collapsed forks, one or more of which mechanisms could be deficient in SF767 cells. As noted above, our evidence suggests that WRN is not a candidate (25).

Clinical implications. Our data suggest that Me-lex, either as a single agent or in combination with temozolomide, may have utility in the adjuvant therapy of malignant brain tumors. We have previously reported that 25% of adult malignant gliomas lack detectable MGMT activity (35, 46), suggesting that at least some of these tumors may be intrinsically sensitive to Me-lex due to factors associated with MGMT deficiency. Regression analysis of survival data for the lines used in this study showed no association between the LD10s for Me-lex and temozolomide (r = 0.26; P = 0.52; Table 1; ref. 22), a drug that is increasingly used in the chemotherapy of malignant brain tumors (20). Me-lex might serve as an alternative chemotherapeutic agent for tumors that are intrinsically resistant to temozolomide or that develop resistance during treatment. The lack of correlation between the cytotoxicity of the two methylators reflects in part the insensitivity of Me-lex cell killing to loss of MMR, a loss that greatly increases the temozolomide resistance of glioma cells by rendering them insensitive to O6-meG (36). A small study of adult glioblastomas has suggested that less favorable response to temozolomide is associated with an decreasing fraction of MMR protein-immunopositive cells (47). The clinical applicability of Me-lex awaits evaluation of toxicity and permeability of the blood brain barrier in animal models.

Our data also suggest suppression of either AAG or Ape1 as a complement to ablation of MGMT in overcoming temozolomide resistance in primary human brain tumors. We have previously shown that antisense suppression of the Ap endo activity of Ape1 increases temozolomide cytotoxicity in an MGMT adult glioma line (32) and in an MGMT+ pediatric medulloblastoma line (24). Our demonstration of the cytotoxicity of 3-meA and derivative abasic sites indicates that AAG and Ape1 are promising targets for antiresistance therapies and strengthens the rationale for developing pharmacologic inhibitors. The development of inhibitors for both AAG and Ape1 is ongoing in several laboratories (48, 49). Importantly, incomplete ablation of either AAG or Ap endo yields statistically significant reduction in 3-meA resistance in all the glioma lines we examined, including those insensitive to unrepaired O6-meG due to loss of MMR. The efficacy of partial suppression contrasts with the need to completely ablate MGMT activity for prolonged periods to increase methylator cytotoxicity in human primary brain tumor cell lines (e.g., ref. 24). These observations suggest that inhibiting repair of 3-meA or abasic sites may be a more readily achievable approach to suppression of glioma methylator resistance.

Grant support: American Cancer Society grant RSG 01 9101 CCE (M.S. Bobola) and NIH grants CA82622 and CA104593 (J.R. Silber), CA29088 (B. Gold), and T32 C09476 (S. Varadarajan).

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.

1
Beranek DT. Distribution of methyl and ethyl adducts following alkylation with monofunctional alkylating agents.
Mutat Res
1990
;
231
:
11
–30.
2
Newlands ES, Stevens MFG, Wedge SR, et al. Temozolomide: a review of its discovery, chemical properties, pre-clinical development, and clinical trials.
Cancer Treat Rev
1997
;
23
:
35
–61.
3
Lawley PD, Phillips DH. DNA adducts from chemotherapeutic agents.
Mutat Res
1996
;
355
:
13
–40.
4
Varadarajan S, Shah D, Dande P, et al. DNA damage and cytotoxicity induced by minor groove binding methyl sulfonate esters.
Biochemistry
2003
;
42
:
14318
–27.
5
Margison GP, Santibanez Koref MF, Povey AC. Mechanisms of carcinogenicity/chemotherapy by O6-methylguanine.
Mutagenesis
2002
;
17
:
483
–7.
6
Gerson SL. MGMT: its role in cancer aetiology and cancer therapeutics.
Nat Rev Cancer
2004
;
4
:
296
–307.
7
Fronza G, Gold B. The biological effects of N3-methyladenine.
J Cell Biochem
2004
;
91
:
250
–7.
8
Zhang Y, Chen FX, Mehta P, Gold B. Groove- and sequence-specific alkylation of DNA by sulfonate esters tethered to lexitropsins.
Biochemistry
1993
;
32
:
7954
–65.
9
Encell L, Shuker DEG, Foiles PG, Gold B. The in vitro methylation of DNA by a minor groove binding methyl sulfonate ester.
Chem Res Toxicol
1996
;
9
:
563
–7.
10
Engelward BP, Allan JM, Dreslin AJ, et al. A chemical and genetic approach together define the biological consequences of 3-methyladenine lesions in the mammalian genome.
J Biol Chem
1998
;
273
:
5412
–8.
11
Patel PH, Loeb LA. Getting a grip on how DNA polymerases function.
Nat Struct Biol
2001
;
8
:
656
–9.
12
Sedgwick B. Repairing DNA-methylation damage.
Nat Rev Mol Cell Biol
2004
;
5
:
148
–57.
13
Wyatt MD, Allan JM, Lau AY, et al. 3-Methyladenine DNA glycosylases: structure, function, and biological importance.
Bioessays
1999
;
21
:
668
–76.
14
Shah D, Gold B. Evidence in Escherichia coli that N3-methyladenine lesions and cytotoxicity induced by a minor groove binding methyl sulfonate ester can be modulated in vivo by netropsin.
Biochemistry
2003
;
42
:
12610
–6.
15
Monti P, Campomenosi P, Ciribilli Y, et al. Influences of base excision repair defects on the lethality and mutagenicity induced by Me-lex, a sequence-selective N3-adenine methylating agent.
J Biol Chem
2002
;
277
:
28663
–8.
16
Monti P, Iannone R, Campomenosi P, et al. Nucleotide excision repair defect influences lethality and mutagenicity induced by Me-lex, a sequence-selective N3-adenine methylating agent in the absence of base excision repair.
Biochemistry
2004
;
43
:
5592
–9.
17
Engelward BP, Weeda G, Wyatt MD, et al. Base excision repair deficient mice lacking the Aag alkyladenine DNA glycosylase.
Proc Natl Acad Sci U S A
1997
;
94
:
13087
–92.
18
Kelly JD, Inga A, Chen FX, et al. Relationship between DNA methylation and mutational patterns induced by a sequence selective minor groove methylating agent.
J Biol Chem
1999
;
274
:
18327
–34.
19
Macdonald DR, Kiebert G, Prados M, et al. Benefit of temozolomide compared to procarbazine in treatment of glioblastoma multiforme at first relapse: effect on neurological functioning, performance status, and health related quality of life.
Cancer Invest
2005
;
23
:
138
–44.
20
Stupp R, Hegi ME, van den Bent MJ, et al. Changing paradigms—an update on the multidisciplinary management of malignant glioma.
Oncologist
2006
;
11
:
165
–80.
21
Bobola MS, Blank A, Berger MS, Silber JR. Contribution of O6-methylguanine-DNA methyltransferase to monofunctional alkylating agent resistance in human brain tumor-derived cell lines.
Mol Carcinog
1995
;
13
:
70
–80.
22
Bobola MS, Tseng SH, Blank A, et al. Role of O6-methylguanine-DNA methyltransferase in resistance of human brain tumor cell lines to the clinically relevant methylating agents temozolomide and streptozotocin.
Clin Cancer Res
1996
;
2
:
735
–41.
23
Lari SU, Al-Khodairy F, Paterson MC. Substrate specificity and sequence preference of G:T mismatch repair: incision at G:T, O6-methylguanine:T, and G:U mispairs in DNA by human cell extracts.
Biochemistry
2002
;
41
:
9248
–55.
24
Bobola MS, Silber JR, Ellenbogen RG, et al. O6-methylguanine-DNA methyltransferase, O6-benzylguanine, and resistance to clinical alkylators in pediatric primary brain tumor cell lines.
Clin Cancer Res
2005
;
11
:
2747
–55.
25
Blank A, Bobola MS, Gold B, et al. The Werner syndrome protein confers resistance to the DNA lesions N3-methyladenine and O6-methylguanine: implications for WRN function.
DNA Repair
2004
;
3
:
629
–38.
26
Burton K. A study of the conditions and mechanism of diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid.
Biochem J
1956
;
62
:
315
–23.
27
Engelward BP, Dreslin AJ, Christensen J, et al. Repair-deficient 3-methyladenine DNA glycosylase homozygous mutant mouse cells have increased sensitivity to alkylation-induced chromosome damage and cell killing.
EMBO J
1996
;
15
:
945
–52.
28
Mandel HG, Straw A, Farmer PB, Martin J. Chromatographic detection of 7-methyladenine in urine of rats administered N-methylnitrosourea: a potential marker for monitoring exposure to methylating carcinogens.
Carcinogenesis
1989
;
10
:
757
–62.
29
Chen D, Herman V, Demple B. Two distinct human diesterases that act on 3′-fragments of deoxyribose in radical damaged DNA.
Nucleic Acids Res
1991
;
19
:
5907
–14.
30
Bobola MS, Blank A, Berger MS, et al. Apurinic/apyrimdinic endonuclease activity is elevated in human adult gliomas.
Clin Cancer Res
2001
;
7
:
3510
–8.
31
Tu GC, Cao Q, Zhou F, Isreal Y. Tetranucleotide GGGA motif in primary RNA transcripts.
J Biol Chem
1998
;
273
:
25125
–31.
32
Silber JR, Bobola MS, Blank A, et al. The apurinic/apyrimidinic endonuclease activity of Ape1/Ref-1 contributes to human glioma cell resistance to alkylating agents and is elevated by oxidative stress.
Clin Cancer Res
2002
;
8
:
3008
–38.
33
Nakamura J, Walker VE, Upton PB, et al. Highly sensitive apurinic/apyrimidinic site assay can detect spontaneous and chemically induced depurination under physiological conditions.
Cancer Res
1998
;
58
:
222
–5.
34
Evans AR, Limp-Foster M, Kelley MR. Going APE over ref-1.
Mutat Res
2000
;
461
:
83
–108.
35
Silber JR, Blank A, Bobola MS, et al. O6-methylguanine-DNA methyltransferase-deficient phenotype in human gliomas: frequency and time to tumor progression after alkylating agent-based chemotherapy.
Clin Cancer Res
1999
;
5
:
807
–14.
36
Stojic L, Brun R, Jiricny J. Mismatch repair and DNA damage signalling.
DNA Repair
2004
;
3
:
1091
–101.
37
Kisby GE, Lesselroth H, Olivas A, et al. Role of nucleotide- and base-excision repair in genotoxin-induced neuronal cell death.
DNA Repair
2004
;
3
:
617
–27.
38
Dong SM, Pang JC, Poon WS, et al. Concurrent hypermethylation of multiple genes is associated with grade of oligodendroglial tumors.
J Neuropathol Exp Neurol
2001
;
60
:
808
–16.
39
Alonso ME, Bello MJ, Gonzalez-Gomez P, et al. Aberrant promoter methylation of multiple genes in oligodendrogliomas and ependymomas.
Cancer Genet Cytogenet
2003
;
144
:
134
–42.
40
Gonzalez-Gomez P, Bello MJ, Arjona D, et al. Promoter hypermethylation of multiple genes in astrocytic gliomas.
Int J Oncol
2003
;
22
:
601
–8.
41
Esteller M. Aberrant DNA methylation as a cancer-inducing mechanism.
Annu Rev Pharmacol Toxicol
2005
;
45
:
629
–56.
42
Bachrati CZ, Hickson ID. RecQ helicases: suppressors of tumorigenesis and premature aging.
Biochem J
2003
;
374
:
577
–606.
43
Plosky B, Samson L, Engelward BP, et al. Base excision repair and nucleotide excision repair contribute to the removal of N-methylpurines from active genes.
DNA Repair
2002
;
1
:
683
–96.
44
Chen J, Stubbe J. Bleomycins: towards better therapeutics.
Nat Rev Cancer
2005
;
5
:
102
–12.
45
Weinfeld M, Rasouli-Nia A, Chaudhry MA, Britten RA. Response of base excision repair enzymes to complex DNA lesions.
Radiat Res
2001
;
156
:
584
–9.
46
Silber JR, Bobola MS, Ghatan S, et al. O6-methylguanine-DNA methyltransferase activity in adult gliomas: relation to patient and tumor characteristics.
Cancer Res
1998
;
58
:
1068
–73.
47
Friedman HS, McLendon RE, Kerby T, et al. DNA mismatch repair and O6-alkylguanine-DNA alkyltransferase analysis and response to Temodal in newly diagnosed malignant glioma.
J Clin Oncol
1998
;
12
:
3851
–7.
48
Madhusudan S, Hickson ID. DNA repair inhibition: a selective tumour targeting strategy.
Trends Mol Med
2005
;
11
:
503
–11.
49
Paik J, Duncan T, Lindahl T, Sedgwick B. Sensitization of human carcinoma cells to alkylating agents by small interfering RNA suppression of 3-alkyladenine-DNA glycosylase.
Cancer Res
2005
;
65
:
10472
–7.