Introduction

Top Summary Introduction Materials & Methods Results Discussion References

The MRS is a DNA repair pathway dedicated to the correction of biosynthetic errors occurring during DNA replication (for reviews, see Jiricny, 1996; Modrich, 1997). Cells with a defective MRS display high rates of spontaneous mutations (Jiricny, 1996; Modrich, 1997) and microsatellite instability (Jiricny, 1996; Modrich, 1997). Germline mutations in mismatch repair genes have recently been linked to hereditary nonpolyposis colon cancer (Jiricny, 1996; Modrich, 1997).

At least five proteins [i.e., hMSH2, hMSH3, hMSH6 (also called guanine/thymine binding protein), hMLH1, and hPMS2] are involved in the initial steps of mismatch repair in human cells (Jiricny, 1996; Modrich, 1997). The protein complex hMutS, consisting of hMSH2 and hMSH6, binds to base/base mismatches and to insertion/deletion loops of up to 14 nucleotides, whereas hMutS (a heterodimer of hMSH2 and hMSH3) binds to insertion/deletion loops but displays little affinity for base/base mismatches. The heterodimer of hMLH1 and PMS2, defined as hMutL, interacts with either hMutS or hMutS and thereby initiates the repair process (Modrich, 1997).

Although the MRS has evolved for the correction of replication errors, it is also implicated in the recognition of other types of DNA damage. Thus, tumor cell lines harboring mutations in the constituent proteins of hMutS and hMutL are resistant to the cytotoxic effects of MNNG and MNU (Kat et al., 1993; Koi et al., 1994; Carethers et al., 1996), 6TG (Aquilina et al., 1989; Hawn et al., 1995; Swann et al., 1996), and cisplatin, carboplatin, and doxorubicin (Anthoney et al., 1996; Drummond et al., 1996). These findings support the hypothesis that the MRS acts as a general sensor for genetic damage (Kat et al., 1993; Hawn et al., 1995) that, upon recognition of certain types of DNA modification, initiates a sequence of events resulting in cell death.

The cytotoxicity of MNU and MNNG has been attributed to the methylation of the O⁶-position of guanine in DNA. Cells with high levels of the DNA-repair enzyme OGAT, which specifically removes alkyl adducts from O⁶-G (Pegg, 1990), are more resistant to the cytotoxic effects of these alkylating agents than are OGAT-deficient cells (Pegg, 1990). Because a defect in the MRS rescues the OGAT phenotype, such that MRS-deficient, OGAT-deficient cells are generally tolerant to high levels of O⁶-G-methylating agents, cell death was proposed to result from the futile processing of O⁶-MeG/T and O⁶-MeG/C mismatches during DNA replication (Goldmacher et al., 1986; Karran and Bignami, 1992). Similarly, the cytotoxicity of 6TG has been attributed to the incorporation of the base analog into DNA and its successive S⁶-methylation by endogenous S-adenosylmethionine (Swann et al., 1996). During DNA replication, S⁶-methylthioguanine mispairs with thymine or cytosine, and the resulting mismatches activate the MRS.

We recently showed that treatment of human leukemic cell lines with the antitumor drug TMZ triggers apoptosis (Tentori et al., 1995). TMZ is a new antitumor triazene compound that is presently undergoing phase II clinical trials (Bleehen et al., 1995). Under physiological conditions, TMZ is rapidly metabolized to form the methylating species 5-(3-methyl-1-triazeno)imidazole-4-carboxamide (Stevens et al., 1987), which reacts with DNA bases to form methyl adducts, chiefly at N⁷-guanine, N³-adenine, and O⁶-G. Methylation of O⁶-G seems to be necessary for the triggering of apoptosis in TMZ-treated cells, because the process is prevented by high OGAT levels (Tentori et al., 1995). Moreover, we have shown that transfection of OGAT cDNA into OGAT-deficient cells confers resistance to the induction of apoptosis by TMZ (Tentori et al., 1997) and that pretreatment of OGAT-proficient cells with O⁶-benzylguanine, a potent OGAT inhibitor (Dolan et al., 1990), increases their susceptibility to TMZ-mediated apoptosis (Tentori et al., 1995, 1997). Methylation of O⁶-G might result in the triggering of apoptosis through the activation of the MRS. This hypothesis was substantiated by the recent observation that colon cancer lines harboring defects in the MRS are more resistant to cell growth inhibition by TMZ than are MRS-proficient lines (Liu et al., 1996).

As discussed above, there is ample experimental evidence linking the cytotoxic effects of DNA-modifying agents with the MRS. However, the molecular mechanisms implicated in this process are currently unknown. In this study, we show that TMZ induces apoptosis, cell cycle arrest at the G₂/M phase, and p53 accumulation in MRS-proficient cells and that these responses are absent in MRS-deficient cells.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Cell lines. The human B lymphoblastoid cell lines TK6 and MT1 were a generous gift of W. G. Thilly (Massachusetts Institute of Technology, Cambridge, MA). The TK6 line was originally isolated as thymidine kinase heterozygous by Skopek et al. (1978). The MT1 line was obtained from TK6 by treatment with IRC-191, followed by selection for MNNG resistance (Goldmacher et al., 1986). Both lines are OGAT-deficient and thus fail to remove alkyl adducts from O⁶-G (Goldmacher et al., 1986). TK6 cells are MRS-proficient (Kat et al., 1993), whereas the MT1 line is MRS-deficient (Kat et al., 1993), harboring different missense mutations in both alleles of the GTBP locus (Papadopoulos et al., 1995). The cells were cultured at 37° in a 5% CO₂ humidified atmosphere and were maintained in RPMI 1640 medium (Hyclone Europe, Cramlington, UK) supplemented with 10% heat-inactivated (56°, 30 min) fetal calf serum (Hyclone), 2 mM L-glutamine, and antibiotics (Flow Laboratories, Mc Lean, VA) (referred to as CM).

Drugs and reagents. TMZ was kindly provided by Schering-Plough Research Institute (Kenilworth, NJ). Because in aqueous solution the drug readily decomposes into 5-(3-methyl-1-triazeno) imidazole-4-carboxamide (Stevens et al., 1987), solutions were always prepared fresh by dissolving the drug in RPMI 1640 medium. The solutions were protected from light. Etoposide was purchased from Sigma (St. Louis, MO) and dissolved in dimethylsulfoxide. Aliquots of the stock solution were stored at 80°. Reagents for SDS-polyacrylamide gel electrophoresis were all purchased from Bio-Rad (Hercules, CA); other chemicals were obtained from Sigma.

Drug treatment and cell growth evaluation. Cells were suspended at 1 × 10⁵ cells/ml in CM or in CM containing the appropriate amount of TMZ or etoposide and were cultured in flasks (Falcon; Becton Dickinson Labware, Frankein Lakes, NJ) at 37° in a 5% CO₂ humidified atmosphere for 72 hr. Cell growth was evaluated, in terms of viable cell counts, every 24 hr. Cells were manually counted using a hemocytometer, and cell viability was determined by the trypan blue exclusion test. All determinations were made in quadruplicate.

Assessment of apoptosis by flow cytometric analysis. Cells from cultures were harvested by centrifugation, washed with PBS, and fixed with 70% ethanol at 20° for 18 hr. The centrifuged pellets were resuspended in 1 ml of hypotonic solution containing 50 µg/ml PI, 0.1% sodium citrate, 0.1% Triton X-100, and 10 µg/ml RNase. The cells were incubated in the dark, at room temperature, for 30 min. Data collection was gated using forward and side light scatter, to exclude cell debris and cell aggregates. The PI fluorescence was measured on a linear scale using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Apoptotic cells are represented by a broad hypodiploid peak, which is easily distinguished from the narrow peak of cells with diploid DNA content observed in the red fluorescence channel (Nicoletti et al., 1991). All data were recorded and analyzed using Lysis II software (Becton Dickinson).

Analysis of DNA fragmentation by PFGE. PFGE of high-molecular weight DNA fragments and preparation of plugs were performed according to the method of Walker et al. (1993). Briefly, cells (2 × 10⁶) were harvested from cultures, washed twice with ice-cold PBS, resuspended in 1.0 ml of a buffer containing 0.15 M NaCl, 2 mM KH₂PO₄/KOH, pH 6.4, 1 mM EGTA, and 5 mM MgCl₂, centrifuged at 10,000 rpm for 2 min, and washed twice with 0.1 ml of the same buffer. The pellets were then resuspended in 0.05 ml of the same buffer and transferred to an Eppendorf tube containing an equal volume of melted 1.5% low-melting point agarose and 0.4 mg/ml proteinase K. The mixture was then pipetted into a 1-ml syringe and refrigerated at 4° for 15 min. The resulting agarose plugs were placed in 1 ml of 10 mM NaCl, 10 mM Tris·HCl, pH 9.5, 25 mM EDTA, 1% N-lauroylsarcosine, supplemented with 0.1 mg/ml proteinase K, and were incubated at 37° for 3 hr. Plugs were rinsed in three changes of 10 mM Tris·HCl, pH 8.0, 1 mM EDTA, at 4° (10 min each), stored in 50 mM EDTA, pH 8.0, for up to 1 week, and then used for electrophoresis.

Assessment of apoptosis by electron microscopy. Cell suspensions were fixed in 2% glutaraldehyde, postfixed in 1% osmium tetroxide, dehydrated in graded alcohols, and embedded in Epon resin, and then semithin and ultrathin sections were cut with an ultramicrotome (Reichert Ultracut E; Leica, Wien, Austria). Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a transmission electron microscope (CM100; Philips, Eindhoven, The Netherlands).

Evaluation of 5-BrdU incorporation. Cells were harvested, washed with PBS, and resuspended in 70% ethanol at 20°. Fixed cells were washed with PBS, resuspended in 2 ml of 2 M HCl, and incubated at room temperature for 30 min. They were then centrifuged, resuspended in 2 ml of 0.1 M Na₂B₄O₇, pH 8.5 (to neutralize acidity), and washed with PBS containing 0.5% (v/v) Tween 20 and 1% (w/v) bovine serum albumin. The cell pellets were resuspended in 0.1 ml of PBS containing 0.4 mg of anti-5-BrdU (Boehringer Mannheim, Mannheim, Germany) and were maintained at room temperature in the dark for 30 min. They were then washed and stained with 0.1 ml of diluted (1/50) FITC-labeled (Fab')₂ rabbit anti-mouse IgG (H+L chains) (Dako, Glostrup, Denmark). Negative controls were represented by cell samples stained with FITC-labeled (Fab')₂ rabbit anti-mouse IgG only. After a 30-min incubation at room temperature, the cells were washed and resuspended in 1 ml of a solution containing 50 µg/ml PI, 0.1% sodium citrate, and 10 µg/ml RNase. They were then incubated in the dark, at 37°, for an additional 30 min. Cell samples were then analyzed for green (FITC, indicating 5-BrdU incorporation) and red (PI, indicating DNA content) fluorescence, using a FACScan flow cytometer. The results were displayed as representative two-dimensional frequency contour plots of green versus red fluorescence.

Immunoblotting. Cell pellets were resuspended in a hypotonic buffer (10 mM Tris·HCI, 1 mM EDTA, pH 7.4, with freshly added 1 mM phenylmethylsulfonyl fluoride), an aliquot was saved for protein concentration determination (using the Bio-Rad protein assay solution, with bovine serum albumin as the standard), and the remainder was immediately boiled in SDS sample buffer (50 mM Tris·HCl, pH 6.8, 100 mM dithiothreitol, 2% SDS, 0.1% bromphenol blue, 10% glycerol). Sixty micrograms of protein per sample were analyzed in 8% (for p53 protein) or 13.5% (for p21/waf-1) SDS-polyacrylamide gels. The proteins were then transferred to supported nitrocellulose membranes (Hybond-C; Amersham Life Science, Buckinghamshire, UK), using a Transphor TE 50X unit (Hoefer Scientific). Immunodetection was carried out using the Boehringer Mannheim chemiluminescence Western blotting kit, according to the manufacturer's instructions, except that the second antibody (horseradish peroxidase-linked) was purchased from Amersham Life Science. The anti-p53 (Ab-2; Calbiochem, Cambridge, MA) and anti-p21/waf-1 (Ab-1; Calbiochem) antibodies were used at 2.5 µg/ml. Filters were exposed to X-OMAT AR autoradiographic films (Kodak, Rochester, NY) for 5-45 sec, depending on the intensity of the signal.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Cell growth inhibition induced by TMZ or etoposide in TK6 and MT1 cell lines. To investigate whether cell sensitivity to TMZ was correlated with the presence of a functional MRS, the MRS-proficient TK6 and MRS-deficient MT1 cell lines, which are both unable to repair O⁶-MeG (Goldmacher et al., 1986), were incubated for 72 hr with graded concentrations of TMZ. Cell growth was evaluated every 24 hr. The results in Fig. 1, which refer to cell growth evaluations after 72 hr of drug exposure, show that TK6 proliferation was markedly impaired by TMZ, whereas MT1 cells were completely resistant to drug concentrations up to 50 µM. In the MT1 line, cell growth inhibition occurred at TMZ concentrations 70-fold higher than those required to obtain comparable levels of cell growth impairment in TK6 cells (data not shown). As a control, similar experiments were performed with the unrelated antitumor drug etoposide, which is known to exert its cytotoxic effects through the inhibition of topoisomerase II (Bender et al., 1990). In this case, no substantial differences in cell growth inhibition were observed between the two lines (Fig. 1).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Inhibition of cell growth induced by TMZ or etoposide in TK6 and MT1 cell lines. TK6 () and MT1 () cells (1 × 105 cells/ml) were incubated in medium alone or in medium containing the indicated concentrations of TMZ (A) or etoposide (B), and cell growth was evaluated after 72 hr. Each value represents the mean of two independent experiments, with error bars indicating standard errors. The figure shows that only the MRS-proficient line TK6 was affected by increasing concentrations of TMZ, whereas both lines were inhibited by etoposide.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Growth arrest in TK6 cells treated with TMZ or etoposide. TK6 cells (1 × 105 cells/ml) were incubated either in medium alone () or in medium containing 12.5 µM TMZ (), 25 µM TMZ (), or 250 nM etoposide () and were counted at 24-hr intervals for 72 hr. Each data point represents the mean of four independent experiments, with errors bars indicating standard errors. The control untreated cells doubled every 20-24 hr, whereas TK6 cells treated with etoposide failed to undergo cell division. Cells treated with TMZ appeared to undergo one doubling, and their numbers either stabilized (12.5 µM TMZ) or declined (25 µM TMZ) at later times.

Induction of apoptosis in TK6 and MT1 cells treated with TMZ or etoposide. The TK6 and MT1 cells were incubated with TMZ at 12.5 or 25 µM for 72 hr, and apoptosis was evaluated by flow cytometry after 6, 24, 48, and 72 hr of drug exposure. The cells were also treated with 250 nM etoposide, which was expected to induce apoptosis independently of the MRS status of the cells. The results illustrated in Table 1 show that TMZ induced apoptosis only in TK6 cells, whereas the effects of etoposide were comparable in the two lines. The percentage of apoptotic cells in TMZ-treated cultures increased substantially after 48 and 72 hr of drug exposure, whereas only a slight increase was observed after 24 hr. On the other hand, in etoposide-treated cultures a significant induction of apoptosis was evident by 24 hr after drug exposure.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Induction of apoptosis in TK6 and MT1 cells treated with TMZ or etoposide. TK6 and MT1 cells (1 × 105 cells/ml) were incubated in medium alone (A) or in medium containing 12.5 µM TMZ (B) or 250 nM etoposide (C). After 48 hr of culture, the cells were harvested and apoptosis was evaluated by flow cytometric analysis of DNA content. Data collection was gated using forward and side light scatter, to exclude cell debris and cell aggregates. The PI fluorescence was measured on a linear scale. The percentage of apoptotic cells exhibiting hypodiploid DNA content is also shown. The FACScan profiles show that, whereas apoptosis was induced by both TMZ and etoposide treatments in TK6 cells, only the latter drug seems to induce apoptosis in MT1 cells.

View larger version (55K): [in this window] [in a new window]   Fig. 4.   Apoptosis-associated DNA degradation in TK6 and MT1 cells treated with TMZ or etoposide. TK6 and MT1 cells were incubated (1 × 105 cells/ml) in medium alone or in medium containing the indicated concentrations of TMZ (A) or etoposide (B). High-molecular weight DNA was extracted from untreated or drug-treated cells, and apoptotic fragmentation was analyzed by PFGE. The sizes of the DNA fragments were estimated using molecular weight markers. The figure is a photograph of an agarose gel stained with ethidium bromide, and it shows that DNA of both cell lines was degraded after etoposide treatment, whereas treatment with TMZ resulted in the degradation of genomic DNA only in the TK6 cells.

View larger version (72K): [in this window] [in a new window]   Fig. 5.   Apoptotic TK6 cells after treatment with 12.5 µM TMZ for 48 hr. The cell surfaces appear completely devoid of microvilli (A and C), and several small vacuoles are distributed at the periphery of relatively well preserved cytoplasm (A). Typical chromatin condensation and margination are present in each cell (asterisks). A single micronucleus is visible in one cell (B), and several micronuclei, each surrounded by a double membrane, are evident in another (A). Electron-dense nuclear fragments free of membrane (arrowhead) are present in the cytoplasm of a cell in a later stage of apoptosis (C). Bars, 1 µm.

Cell cycle perturbations induced in TK6 and MT1 cells by TMZ or etoposide. The TK6 and MT1 cells were incubated with TMZ (12.5 µM) or etoposide (250 nM) for 72 hr, and the percentage of cells in each phase of the cell cycle was evaluated by bivariate analysis of 5-BrdU incorporation and DNA content after 24, 48, and 72 hr of drug exposure. 5-BrdU (10 µM) was added to the cultures 1 hr before harvesting. The results illustrated in Table 2 show that, in TK6 cells treated with TMZ, only a slight increase in the G₂/M fraction occurred during the first 24 hr of drug exposure, whereas after 48 hr about one third of the cells appeared to be arrested at the G₂/M phase of the cell cycle. Moreover, a significant fraction (30%) of the cells showed hypodiploid DNA contents at this point and were therefore considered to be undergoing apoptosis. A decrease in the G₂/M fraction associated with a further increase in the sub-G₁ cell fraction was observed at 72 hr (Table 2). On the other hand, the MT1 line exposed to TMZ failed to show any alteration of the cell cycle (Table 2).

p53 and p21/waf-1 accumulation in TK6 and MT1 cells after treatment with TMZ or etoposide. It is well established that p53 plays a crucial role in the regulation of the cell cycle and in the induction of apoptosis in response to DNA-damaging agents (for reviews, see Leonard et al., 1995; Harris, 1996). We therefore investigated whether there was an alteration in p53 protein expression in TMZ-treated cells and whether a correlation existed between p53 induction and MRS activity. In addition, we evaluated the level of the cyclin-dependent kinase inhibitor p21/waf-1 (Xiong et al., 1993), which is known to increase as a result of p53-mediated transcriptional activation (El-Deiry et al., 1994). To this end, TK6 and MT1 cells were incubated with TMZ (12.5 µM) or etoposide (250 nM) for 72 hr and the p53 and p21/waf-1 levels, in control and drug-treated cells, were evaluated by Western blot analysis after 6, 24, 48, and 72 hr of drug exposure. The results presented in Fig. 6 show that TMZ treatment of TK6 cells induced a progressive increase in the amount of p53, which was evident, although at low levels, after 6 hr of drug treatment. The induction of p53 was accompanied by a parallel increase in the amount of p21/waf-1 protein. Neither p53 nor p21/waf-1 protein levels increased upon TMZ treatment in the MT1 line, whereas exposure to etoposide resulted in increases in p53 and p21/waf-1 protein levels in both TK6 and MT1 cells.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   p53 and p21/waf-1 accumulation in TK6 and MT1 cells after treatment with TMZ or etoposide. TK6 and MT1 cells were incubated (1 × 105 cells/ml) in medium alone () or in medium containing 12.5 µM TMZ (A) or 250 nM etoposide (B) (+). The amounts of p53 and p21/waf-1 proteins were evaluated at the indicated times by immunoblotting. The figure shows that induction of p53 and p21/waf-1 protein expression in the MRS-proficient line TK6 occurred after treatment with either TMZ or etoposide. In contrast, only the latter drug induced these proteins in the MRS-deficient line MT1.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

DNA mismatch repair plays a critical role in the maintenance of genomic stability, as demonstrated by the observations that mutations in mismatch repair genes bring about a strong mutator phenotype and microsatellite instability and that the inheritance of MRS gene mutations is linked to hereditary colon cancer (for review, see Jiricny, 1996). Defects in the MRS are also associated with tumor cell resistance to the cytotoxic activities of MNNG, MNU, 6TG, cisplatin, doxorubicin, and TMZ, suggesting that cell death induced by these agents occurs through an active process that requires the participation of the MRS, although the molecular events involved in the process have not yet been elucidated.

We recently showed that TMZ exerts its cytotoxic effects mainly through the induction of apoptosis and that the persistence of O⁶-MeG in DNA is necessary for the triggering of the programmed cell death process (Tentori et al., 1995, 1997). In this article, we demonstrate that cell susceptibility to TMZ-induced apoptosis is correlated with the activity of the MRS.

The MRS-proficient cell line TK6 and its MRS-deficient subline MT1, both lacking OGAT and therefore being incapable of removing methyl groups from O⁶-G, were, indeed, differentially susceptible to TMZ, in terms of cell growth inhibition (Fig. 1) and apoptosis induction (Table 1 and Figs. 3 and 4). The growth of TK6 cells was readily inhibited by TMZ; the MT1 line was markedly more resistant to this agent, and apoptosis could not be induced even with TMZ concentrations up to 800 µM (Table 1, Fig. 4, and data not shown). It could be argued that the resistance of MT1 cells to TMZ is the result of a defect in one or more pathways that induce apoptosis after DNA damage. However, the fact that the two lines showed comparable susceptibility to cell growth inhibition (Fig. 1) and apoptosis induction (Table 1 and Figs. 3 and 4) in response to treatment with etoposide, a topoisomerase II inhibitor, eliminates this possibility and implies that the apoptotic response in the TMZ-treated TK6 cells is dependent on a functional MRS. The correlation between induction of apoptosis by TMZ and the activity of the MRS is further strengthened by the results presented in Table 2, which show that, upon treatment with TMZ, the MRS-proficient TK6 cells but not the MRS-deficient MT1 cells were arrested preferentially at the G₂/M phase of the cell cycle and underwent apoptosis, whereas the two lines exhibited comparable levels of G₂/M accumulation and apoptosis after treatment with etoposide.

Time course analysis of cell growth inhibition, apoptosis induction, and cell cycle perturbation in drug-treated cells showed that the effects of TMZ on TK6 cells became particularly evident after 48 hr of drug exposure (Fig. 2 and Tables 1 and 2). Our findings agree with the results of previous studies, performed with unsynchronized TK6 cells (Goldmacher et al., 1986) or other cell lines (Plant and Roberts, 1971; Black et al., 1989), which suggested that, after methylation of O⁶-G by MNNG or MNU, the cells proceed through a single cell division and are arrested at the G₂/M phase of the second cell cycle (Karran and Bignami, 1992). Although these results apparently differ from the findings of Carethers et al. (1996), who showed that several MRS-proficient cell lines treated with MNNG were arrested at the G₂/M phase of the first cell cycle, it is conceivable that the cells used in the latter study were arrested at the first G₂/M phase because of the level of DNA damage induced by drug concentrations significantly higher than those in the other series of experiments. The hypothesis that cell cycle arrest after treatment with TMZ occurs mainly at the G₂/M phase of the second cell cycle seems to be supported also by the different kinetics of cell growth inhibition, G₂/M arrest, and apoptosis induction observed in TK6 cells exposed to etoposide. In those cells, G₂/M arrest and apoptosis were evident after 24 hr of drug exposure (Tables 1 and 2), and cell numbers did not increase (Fig. 2). Moreover, if TK6 and MT1 lines were exposed to concentrations of etoposide that allowed cells to replicate during the first 24 hr of drug treatment (i.e., concentrations ranging between 15.62 and 62.5 nM), no substantial induction of apoptosis was observed (data not shown).

p53 is a critical participant in the signal transduction pathway that mediates apoptosis in response to DNA damage (Leonard et al., 1995; Harris, 1996). Induction of apoptosis by p53 may occur through up-regulation of Bax (Miyashita et al., 1994) and/or Fas/Apo-1 protein expression (Owen-Schaub et al., 1995), through activation of caspase 3 (Fuchs et al., 1997), or through down-modulation of Bcl-2 (Miyashita et al., 1994).

The results illustrated in Fig. 6 show that, in the MRS-proficient TK6 cell line, p53 was induced in response to TMZ and this increase was accompanied by up-regulation of p21/waf-1 protein levels. In contrast, neither p53 nor p21/waf-1 levels were increased in the MRS-deficient MT1 cells after TMZ treatment, although both TK6 and MT1 cells were capable of inducing p53 and p21/waf-1 expression in response to etoposide treatment (Fig. 6) or X-irradiation (data not shown). Therefore, a functional MRS seemed to be required for up-regulation of p53 in cells treated with TMZ, at least at the drug concentrations used, which are in the range of those achievable clinically (Newlands et al., 1992).

As mentioned above, the MRS seems to be involved in triggering the cytotoxic response to cisplatin. Indeed, several ovarian tumor cell lines isolated in vitro by single-step selection for resistance to the drug were found to be defective in the expression of hMLH1 (Drummond et al., 1996). Cisplatin resistance in these MRS-deficient lines has been associated with a reduced ability to activate a p53-dependent response, as indicated by absent or reduced up-regulation of p21/waf-1 mRNA, and to undergo apoptosis upon treatment with the drug (Anthoney et al., 1996). Although these results resemble our data in TK6 and MT1 cells, it must pointed out that the cisplatin-resistant sublines showed impaired p53 function (with no, or markedly reduced, G₁ arrest and transcription of the p21/waf-1 gene) also in response to ionizing radiation, whose cytotoxic activity does not depend on a functional MRS. Moreover, the cells were cross-resistant to multiple cytotoxic agents (Hamaguchi et al., 1993), suggesting a defect in the engagement of apoptosis after any DNA damage. Therefore, the reduced ability of the cisplatin-resistant lines to undergo apoptosis upon treatment with the drug is attributable not only to a deficiency in mismatch repair but also to a loss of p53 function, which was selected for during the generation of the resistant lines. Our results, showing that p53 and p21/waf-1 are induced in MRS-proficient and MRS-deficient cells upon treatment with etoposide (Fig. 6) or X-irradiation (data not shown) but not with TMZ (Fig. 6), represent the first experimental evidence that MRS-deficient cells are incapable of generating the signal leading to p53 induction in response to an O⁶-G-methylating agent, whereas they retain the ability to induce p53 after DNA damage not involving methylation of O⁶-G.

We previously showed that TMZ-mediated apoptosis can occur also in p53-null cells (Tentori et al., 1995, 1997). However, expression of wild-type p53 in p53-null HL-60 cells increased their susceptibility to TMZ-mediated apoptosis (Tentori L, Lacal PM, Benincasa E, Franco D, Faraoni I, Bonmassar E, Graziani G, manuscript submitted for publication). It is therefore possible to speculate that in MRS-proficient lines TMZ might activate both p53-dependent and p53-independent apoptosis and the two pathways might interact with each other, as described in other models (Peled et al., 1996). In cells expressing wild-type p53, the triggering of apoptosis might occur at lower drug concentrations or with faster kinetics. For instance, it has already been shown in a murine model that the p53 status of cells affects the rate of the onset, but not the overall extent, of apoptosis induced by doxorubicin (Han et al., 1997).

Further studies are required to establish whether apoptosis in TK6 cells occurs through a p53-dependent or p53-independent pathway. In fact, p53 has not been sequenced in this line and, although the protein shows an ability to trans-activate p21/waf-1, we cannot assume that the apoptotic function of p53 is intact in TK6 cells. Indeed, mutant forms of p53 that are capable of trans-activating p21/waf-1 but show impaired apoptotic function have already been described (Ludwig et al., 1996). Moreover, in TMZ-treated TK6 cells no changes in the levels of Bax, Fas/Apo-1, or Bcl-2 proteins have been observed (data not shown).

In conclusion, we demonstrated the existence of a link between a functional MRS and the trigger of apoptosis in cells exposed to clinically achievable concentrations of TMZ. We also showed that p53 induction in response to the methylation of O⁶-G is associated with a functional MRS. Studies designed to clarify these mechanisms, in particular to establish to what extent p53 and downstream effectors are involved in cell cycle arrest and apoptosis induction in TMZ-treated cells, are currently in progress.