• BCL-XL;
  • glioblastoma;
  • O6-methylguanine DNA methyltransferase;
  • p53;
  • temozolomide


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Temozolomide (TMZ) is a methylating agent which prolongs survival when administered during and after radiotherapy in the first-line treatment of glioblastoma and which also has significant activity in recurrent disease. O6-methylguanine DNA methyltransferase (MGMT) is a DNA repair enzyme attributed a role in cancer cell resistance to O6-alkylating agent-based chemotherapy. Using a panel of 12 human glioma cell lines, we here defined the sensitivity to TMZ in acute cytotoxicity and clonogenic survival assays in relation to MGMT, mismatch repair and p53 status and its modulation by dexamethasone, irradiation and BCL-XL. We found that the levels of MGMT expression were a major predictor of TMZ sensitivity in human glioma cells. MGMT activity and clonogenic survival after TMZ exposure are highly correlated (p < 0.0001, r2 = 0.92). In contrast, clonogenic survival after TMZ exposure does not correlate with the expression levels of the mismatch repair proteins mutS homologue 2, mutS homologue 6 or post-meiotic segregation increased 2. The MGMT inhibitor O6-benzylguanine sensitizes MGMT-positive glioma cells to TMZ whereas MGMT gene transfer into MGMT-negative cells confers protection. The antiapoptotic BCL-XL protein attenuates TMZ cytotoxicity in MGMT-negative LNT-229 but not in MGMT-positive LN-18 cells. Neither ionizing radiation (4 Gy) nor clinically relevant concentrations of dexamethasone modulate MGMT activity or TMZ sensitivity. Abrogation of p53 wild-type function strongly attenuates TMZ cytotoxicity. Conversely, p53 mimetic agents designed to stabilize the wild-type conformation of p53 sensitize glioma cells for TMZ cytotoxicity. Collectively, these results suggest that the determination of MGMT expression and p53 status will help to identify glioma patients who will or will not respond to TMZ.

Abbreviations used

O6-methylguanine DNA methyltransferase


mutS homologue




poly(ADP-ribose) polymerase


post-meiotic segregation increased 2



Temozolomide (TMZ) is a cytotoxic alkylating agent which has shown activity in recurrent anaplastic glioma and glioblastoma (Yung et al. 1999, 2000). Moreover, the recent European Organisation for Research and Treatment of Cancer/National Cancer Institute of Canada trial on concomitant and adjuvant TMZ in addition to radiotherapy as the first-line treatment of glioblastoma produced an increase in median survival from 12.1 to 14.6 months and an increase in the 2-year survival rate from 10 to 26% compared with radiotherapy alone (Stupp et al. 2005).

Temozolomide is rapidly and completely absorbed after oral administration and undergoes spontaneous hydrolysis at physiological pH to its active metabolite 5-(3-methyltriazeno)-imidazole-4-carboxamide. 5-(3-methyltriazeno)-imidazole-4-carboxamide causes DNA damage by methylation of the O6 position of guanine. However, DNA adducts different from O6-methylguanine might also be involved in the cytotoxicity induced by methylating agents and poly(ADP-ribose) polymerase (PARP) inhibitors may also sensitize tumour cells to TMZ-induced cell death (Tentori et al. 1998, 2001). Further, DNA lesions induced by alkylating agents activate the p53-controlled DNA damage response pathway (Hickman and Samson 1999).

The extent of methylation at the O6 position of guanine in DNA correlates well with the therapeutic activity as well as the toxicity of TMZ. The methyl group on the O6 position of guanine can be removed by the suicide DNA repair enzyme O6-methylguanine DNA methyltransferase (MGMT) (Pegg 1990). Thus, MGMT interferes with the cytotoxic effects induced by TMZ and other O6-alkylating agents and the MGMT levels in tumour cells correlate negatively with the efficacy of TMZ treatment, in particular when alterations of other DNA repair systems are absent (Baer et al. 1993; Gerson 2002; Kaina and Christmann 2002; Pagani et al. 2003; Pepponi et al. 2003; Barvaux et al. 2004a; Roos et al. 2004). Thus, deficiencies of the post-replication mismatch repair system render tumour cells resistant to methylating agents independent of MGMT levels (Gerson 2002; Pagani et al. 2003; Pepponi et al. 2003; Barvaux et al. 2004a).

O6-methylguanine DNA methyltransferase expression is retained in most gliomas (Preuss et al. 1995, 1996; Silber et al. 1998). Loss of MGMT protein as a consequence of MGMT gene promoter methylation predicts a good response to alkylating agent-based chemotherapy (Esteller et al. 2000). Accordingly, an inhibitor of MGMT, O6-benzylguanine (O6-BG), restores TMZ-induced cytotoxicity in glioma cells in vitro and in vivo (Wedge and Newlands 1996; Kokkinakis et al. 2001; Friedman et al. 2002; Ma et al. 2002; Kanzawa et al. 2003) and the therapeutic application of O6-BG in combination with different alkylating agents has been tested in clinical studies (Friedman et al. 2000; Quinn et al. 2002).

In addition to MGMT-mediated DNA repair, the DNA damage-sensitive p53-controlled cell cycle arrest and death pathway is likely to determine tumour cell responses to alkylating agents such as TMZ. The suppression of MGMT reporter gene activity by wild-type p53 (Harris et al. 1996; Srivenugopal et al. 2001) predicts a higher efficacy of TMZ in p53 wild-type than p53 mutant tumours. Given the emerging role of TMZ in the medical treatment of gliomas (Tentori et al. 1998; Stupp et al. 2005) and the probable dependence of its activity on MGMT (Hegi et al. 2004), we here characterize the molecular pathways mediating TMZ cytotoxicity in human malignant glioma cell lines with a focus on MGMT, mismatch repair and p53.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References


Temozolomide was kindly provided by Schering Plough (Kenilworth, NJ, USA). MGMT antibody was from BioTrend GmbH (Cologne, Germany), Bcl-XL, actin and p53 (Bp53-12) antibodies from Santa Cruz Biotechnology (Santa Cruz, CA, USA), mutS homologue (MSH)2 antibody from Calbiochem (San Diego, CA, USA), MSH6 antibody from Transduction Laboratories (Lexington, KY, USA) and post-meiotic segregation increased 2 (PMS2) antibody from Pharmingen (San Diego, CA, USA). The enhanced chemoluminescence system was from Amersham (Braunschweig, Germany) and CP-31398 was provided by Pfizer (Groton, CT, USA) (Foster et al. 1999). O6-BG was synthesized in the laboratory of B.K. (Kaina et al. 2004). Dexamethasone was purchased from Sigma (St Louis, MO, USA) and 3-aminobenzamide was purchased from Sigma (Deisenhofen, Germany). The PARP inhibitor NU1025 was kindly provided by N. Curtin (Newcastle upon Tyne, UK).

Cell culture

The human malignant glioma cell lines LN-18, U138MG, U87MG, LN-428, D247MG, T98G, LN-319, LNT-229, A172, U251MG, U373MG and LN-308 were kindly provided by Dr N. de Tribolet (Lausanne, Switzerland). Glioma cells stably expressing p53V135A were generated using the p53V135A hygro vector kindly provided by M.F. Clarke (Pittsburgh, PA, USA) (Naumann et al. 1998). The p53 siRNA cells were obtained by transfection using a pSUPERpuro p53 expression vector (Wischhusen et al. 2003). Bcl-XL-expressing cell lines were obtained by transfection using pSFFV-BCL-XL (Glaser et al. 2001). MGMT-expressing LNT-229 cells were obtained by transfection using the MGMT expression vector pSV-MGMT which carries a neomycin resistance gene (Kaina et al. 1991). These cell lines were maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 2 mm glutamine and penicillin (100 IU/mL)/streptomycine (100 µg/mL).

Viability assays

Acute growth inhibition/cytotoxicity assays involved the exposure of glioma cells seeded at 5000 cells/well in 96-well plates to increasing concentrations of TMZ (10–4000 µm) for 72 h. Clonogenic survival assays were performed by seeding 500 cells in six-well plates and exposing them to TMZ (10–1000 µm) for 24 h, followed by further observation for 7–14 days. Cell density or colonies were assessed using crystal violet staining. Colonies of more than 50 cells were counted. In some clonogenic survival assays, the cells were also pre-treated (24 h) and then cotreated with dexamethasone (100 nm) or irradiated at 0, 0.5, 1, 2, 4 or 8 Gy (GammaCell 1000 Elite; Nordion, Ontario, Canada).


Total RNA was extracted using an RNA purification system (RNeasy, Qiagen, Hilden, Germany). Five µg of total RNA was used for the RT reaction using oligo (dt)20. PCR primers and conditions were as follows: MGMT forward, AGAGTCGTTCACCAGACAGG (nucleotides 313–332) and MGMT reverse, GCCATTCCTTCACGGCCAG (nucleotides 524–542), 35 cycles, Tanneal 55°C. The PCR products were separated on agarose gels and stained with ethidium bromide.


To determine the protein levels of MGMT, BCL-XL and p53, soluble protein lysates of subconfluent glioma cells were obtained and sodium dodecyl sulphate–polyacrylamide gel electrophoresis with electroblotting was performed as described by Weller et al. (1994). Enhanced chemoluminescence was used for detection. Nuclear protein extracts were prepared to assess the level of mismatch repair proteins. The cell pellets were suspended in fractionation buffer A (HEPES-KOH, 10 mm; EDTA, 0.1 mm; EGTA, 1 mm; sucrose, 250 mm; Na3VO4, 1 µm; phenylmethylsulphonyl fluoride, 0.5 mm; dithiothreitol, 10 mm, pH 7.4). The cells were lysed by freeze/thaw/vortexing cycles. The lysates were centrifuged at 10 000 r.p.m. (7826 × g) for 10 min and the supernatant containing the cytoplasmic proteins was removed. The pellets containing nuclei, organelles and membranes were resuspended in fractionation buffer B (Tris-HCl, 20 mm; EDTA, 1 mm; β-mercaptoethanol, 1 mm; 5% glycerine; Na3VO4, 1 µm; phenylmethylsulphonyl fluoride, 0.5 mm; dithiothreitol, 10 mm, pH 8.5). This suspension was homogenized by sonication and centrifuged at 10 000 r.p.m. (7826 × g) for 10 min. The supernatant contained the nuclear proteins and the pellet contained the membrane fragments. The protein concentration was determined by the Bradford method and the nuclear proteins were separated using a 10% sodium dodecyl sulphate–polyacrylamide gel electrophoresis gel. Densitometry analysis was performed to determine relative protein expression.

Preparation of cell extracts and O6-methylguanine DNA methyltransferase activity assay

Cell extracts were prepared as described by Preuss et al. (1995). In brief, the cells were harvested and homogenized by sonication in buffer containing: Tris-HCl, 20 mm, pH 8.5, EDTA, 1 mm; β-mercaptoethanol, 1 mm; 5% glycerol and the protease inhibitor phenylmethylsulphonyl fluoride, 0.1 mm. The extract was centrifuged at 10 000 r.p.m. (7826 × g) (10 min) at 4°C in order to remove debris and the supernatant was snap-frozen in aliquots using liquid nitrogen and stored at −80°C until use. HeLa S3 cells expressing MGMT (588 ± 86 fmol/mg protein) and HeLa MR cells deficient in MGMT served as positive and negative controls. MGMT activity in cell extracts was determined as reported by Preuss et al. (1996). The method is based on the radioactive transfer of a tritium-labelled methyl group from the O6-position of guanine in the DNA to the MGMT protein in the cell extract. The protein concentration was determined as described by Preuss et al. (1996). For each assay, at least 100 µg of protein was used. The cell extract together with [3H-methyl]-nitrosourea-labelled calf thymus DNA containing O6-methylguanine (80 000 cpm/sample) was incubated in buffer containing HEPES-KOH, 70 mm, pH 7.8, dithiothreitol, 1 mm and EDTA, 5 mm for 90 min. Radioactivity was determined in a liquid scintillation counter. Data are expressed as fmol radioactivity transferred from 3H-labelled DNA to protein/mg of protein within the sample. A modulation of MGMT activity by γ-ray treatment (4 Gy) was determined in exponentially growing cells that were harvested 24 h after irradiation. To analyse the effect of glucocorticoids on MGMT activity, the cells were treated with dexamethasone (100 nm) 2 days after seeding and harvested 28 h later.

Statistical analysis

The data are representative of experiments performed at least three times with similar results. Viability assays were tested for significance by t-test. A computer-based program, GraphPad Prism, was used to determine correlations between EC50 data, MGMT activity and MSH2, MSH6 and PMS2 protein levels.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Glioma cell sensitivity to temozolomide in vitro

The cytotoxic effects and inhibition of colony formation mediated by TMZ were examined in a panel of 12 human malignant glioma cell lines. The EC50 values for the induction of cytotoxic cell death ranged between 87 µm (LNT-229) and 1290 µm (LN-308). Even high concentrations of TMZ resulted in the survival of up to 40% of the cells in some cell lines, indicating cytostatic rather than cytotoxic actions in these assays (Fig. 1a). The EC50 values for the prevention of clonogenic growth were much lower than the values obtained in the acute growth inhibition assays and ranged between 7 µm (U87MG) and 511 µm (LN-18). T98G had a similar EC50 value (502 µm) to LN-18 in the clonogenic assay and these two cell lines had significantly higher EC50 values than all other cell lines (Fig. 1b). There was no correlation between the EC50 values in the acute growth inhibition and clonogenic survival assays (r = –0.11, p = 0.74).


Figure 1. Temozolomide (TMZ)-mediated cytotoxicity and clonogenic cell death in human malignant glioma cell lines; the role of O6-methylguanine DNA methyltransferase (MGMT). The glioma cell lines were examined for EC50 values in acute growth inhibition (a) and clonogenic survival (b) assays (mean and SEM, n = 3), MGMT mRNA expression by RT-PCR (c) and MGMT protein levels by immunoblot using actin as a reference (d) and MGMT activity (e).  (f) Correlation of MGMT activity and EC50m) for TMZ in clonogenic survival assays.

Download figure to PowerPoint

O6-methylguanine DNA methyltransferase expression and activity: relation to sensitivity to temozolomide

Three cell lines, LN-18, D247MG and T98G, showed MGMT mRNA expression by RT-PCR (Fig. 1c). Only LN-18 and T98G also showed a strong immunoreactive band at the predicted size of 25 kDa on immunoblots (Fig. 1d). The same two cell lines also exhibited high levels of MGMT activity whereas the other 10 cell lines exhibited no or very low MGMT activity (Fig. 1e). These data correlated well with the EC50 data for clonogenic survival which showed the highest resistance to TMZ in LN-18 and T98G cells (Fig. 1b). In fact, MGMT activity and resistance to TMZ in clonogenic survival assays expressed as EC50 values were highly correlated (Fig. 1f). The role of MGMT in mediating TMZ resistance in T98G cells was confirmed by the coexposure to the MGMT inhibitor O6-BG which shifted the EC50 in the clonogenic survival assay from 555 to 150 µm. No such effect was seen in the MGMT-negative cell line U138MG (Fig. 2a). O6-BG had no effect on survival in these assays when administered alone and decreased MGMT activity at 24 h to 0 fmol/mg (data not shown). Conversely, ectopic expression of an MGMT transgene in LNT-229 cells resulted in a distinct protection from TMZ cytotoxicity (Fig. 2b), indicating that MGMT is a major predictor of response to TMZ in glioma cells.


Figure 2. Modulation of temozolomide (TMZ)-mediated clonogenic cell death by O6-methylguanine DNA methyltransferase (MGMT). (a) T98G (squares, dashed lines) or U138MG (circles, solid lines) cells were pre-incubated for 2 h with medium alone (open symbols) or O6-benzylguanine (O6-BG) (50 µm) (filled symbols) and then treated for 24 h with TMZ in the absence or presence of O6-BG in a clonogenic survival assay. (b) LNT-229 cells transfected with control vector (□) or with a plasmid encoding MGMT (•) were examined in clonogenic survival assays (n = 3, SEM < 10%). Transgene expression was verified by immunoblot (insert).

Download figure to PowerPoint

Collateral sensitivity to temozolomide and carmustine: probable link to O6-methylguanine DNA methyltransferase

The EC50 values for TMZ (Figs 1a and b) and the MGMT activity data were also analysed in view of previous data on the sensitivity of the same cell lines to irradiation and various other chemotherapeutic agents (Weller et al. 1998; Streffer et al. 2002). Interestingly, the cell lines with high MGMT activity were particularly resistant to carmustine in clonogenic cell death assays whereas no correlation with their sensitivity to irradiation, camptothecine, β-lapachone, doxorubicine, teniposide, cytarabine or vincristine became apparent. Like TMZ, carmustine is an alkylating agent but, unlike TMZ, its spontaneous non-enzymatic degradation leads to the formation of 2-chloroethyl carbonium ions which alkylate guanine, cytidine and adenine, resulting in inter- or intrastrand cross-linking of the DNA.

Mismatch repair protein expression: no correlation to temozolomide sensitivity

The expression levels of the MutSα proteins MSH2 and MSH6 were determined because of the importance of this heterodimer for the binding of the O6-methylguanine/thymine mismatch (Ceccotti et al. 1996; Duckett et al. 1996) (Figs 3a and b). The relative protein expression levels of MSH2 and MSH6 in the 12 cell lines correlated well (p < 0.0001, r2 = 0.82) (Fig. 3d), in line with the observation that binding of MSH6 to MSH2 stabilizes MSH6 in the MutSα complex (Chang et al. 2000). The protein levels of PMS2 were also determined (Fig. 3c). PMS2 is another mismatch repair protein that may serve as a sensor in DNA repair mechanisms, being critical for the decision between cell cycle arrest and apoptosis. PMS2 levels did not correlate with MSH2 levels (Fig. 3d). All three proteins were expressed by all 12 cell lines. Importantly, there was no correlation between the expression levels of the three mismatch repair proteins examined here and TMZ sensitivity (EC50) in acute cytotoxicity or clonogenic survival assays.


Figure 3. Mismatch repair protein expression in human glioma cell lines. mutS homologue (MSH)2 (a), MSH6 (b) or post-meiotic segregation increased 2 (PMS2) (c) protein levels were assessed by immunoblot, normalized to ERK2 protein levels and then expressed as relative protein levels with U87MG set to 1. (d) Correlation analysis between MSH2 and MSH6 or PMS2.

Download figure to PowerPoint

No modulation of temozolomide activity in glioma cells by poly(ADP-ribose) polymerase inhibition

As PARP has been attributed a role in inducing resistance to TMZ in MGMT-negative mismatch repair-deficient cells (Curtin et al. 2004), we also studied the effect of the PARP inhibitors 3-aminobenzamide (1 mm) and NU1025 (200 µm) on TMZ sensitivity in the glioma cell lines. 3-aminobenzamide did not significantly modulate clonogenic survival in response to TMZ in any of the 12 cell lines. Similarly, NU1025 had no effect in MGMT-positive LN-18 cells, MGMT-negative TMZ-sensitive LNT-229 cells or MGMT-negative TMZ-resistant LN-319 cells (data not shown), indicating that PARP is not an important factor for TMZ resistance in glioma cell lines.

Modulation of temozolomide activity by BCL-2 family proteins

A comparison of the data on TMZ sensitivity in acute growth inhibition and clonogenic survival assays (Figs 1a and b) with the endogenous expression levels of the BCL-2 family proteins BCL-2, MCL-1, BCL-XL, BCL-XS, BAX and BAK (Weller et al. 1998) revealed no apparent link. To directly address a modulation of TMZ cytotoxicity by the prototype antiapoptotic BCL-2 family protein BCL-XL, LN-18 and LNT-229 cells were transfected with the human BCL-XL expression plasmid pSFFV-BCL-XL(Fig. 4a). BCL-XL gene transfer had no significant effect on the MGMT levels in these cells (data not shown). Clonogenic survival assays revealed a significant increase in survival in BCL-XL-transfected LNT-229 cells which translated into an EC50 shift from 9.5 to 39.4 µm but had no such effect in LN-18 cells (Fig. 4b). Moreover, there was no modulation of TMZ cytotoxicity by BCL-XL in acute growth inhibition assays in either cell line although these cells were strongly protected from CD95L-induced apoptosis which was used as a positive control (data not shown). The protective effect of BCL-XL in LNT-229 cells was confirmed in an LNT-229 cell line expressing BCL-XL from a doxycyclin-sensitive promoter, indicating that the protection mediated by BCL-XL was not an artefact of long-term culturing of the BCL-XL-transfected cells examined in Fig. 4(b) (data not shown).


Figure 4. Modulation of temozolomide (TMZ) activity in glioma cells by BCL-XL. (a) BCL-XL expression was assessed by immunoblot. (b) LN-18 (circles, dashed lines) or LNT-229 (squares, solid lines) cells transfected with neo (open symbols) or BCL-XL (filled symbols) plasmids were assessed for TMZ sensitivity in clonogenic survival assays.

Download figure to PowerPoint

No modulation of temozolomide cytotoxicity by irradiation or dexamethasone

The possible inducibility of MGMT by irradiation (Chan et al. 1992; Rafferty et al. 1996; Grombacher et al. 1998) and the apparent benefit of adding TMZ to radiotherapy in the first-line treatment of glioblastoma (Kaina et al. 1991) prompted us to examine a modulation of MGMT activity and sensitivity to TMZ by single fractions of irradiation in vitro. However, MGMT activity was unaltered in all 12 glioma cell lines at 24 h after irradiation at 4 Gy (Table 1). Moreover, clonogenic survival assays showed additive effects of irradiation and TMZ in T98G and LNT-229 cells but no specific interactions (Table 2). Further, given the reported induction of MGMT by corticosteroids (Grombacher et al. 1996; Biswas et al. 1999), the protection from drug-induced glioma cell death in vitro by corticosteroids (Weller et al. 1997) and the widespread use of these drugs in human glioma patients, we performed similar experiments in the absence or presence of dexamethasone. Again, dexamethasone at clinically achieved concentrations of up to 100 nm for 28 h did not modulate MGMT activity in any of the 12 glioma cell lines and dexamethasone at 100 nm did not modulate the effects of TMZ in T98G and LNT-229 cells in acute growth inhibition or clonogenic cell death assays (data not shown).

Table 1.  O6-methylguanine DNA methyltransferase (MGMT) activity in glioma cell lines; no modulation by irradiation a
 MGMT activity (fmol/mg protein)
0 Gy4 Gy
  • a

    MGMT activity was determined in 12 glioma cell lines in control cells and 24 h after irradiation (4 Gy). Activity < 15 fmol/mg was determined as not significant.

Table 2.  Temozolomide (TMZ) sensitivity in glioma cell lines; no synergy with irradiation a
TMZ (500 µm)T98GTMZ (50 µm)LNT-229
  • a

    Cytotoxicity of TMZ was assessed in clonogenic survival assays. The cells were irradiated and treated with TMZ 24 h later for 24 h. Data are expressed as mean percentages and SEM (n = 3) of clonogenic survival relative to untreated control cells.

0 Gy100 ± 846 ± 90 Gy100 ± 357 ± 13
2 Gy65 ± 1041 ± 52 Gy72 ± 340 ± 6
6 Gy25 ± 1318 ± 36 Gy26 ± 815 ± 3

Modulation of temozolomide activity by p53

The p53 status of the glioma cell lines examined here has been determined previously (Wischhusen et al. 2003). U87MG, D247MG, A172 and LNT-229 cells are wild-type genetically and exhibit p53 reporter activity whereas the other eight cell lines do not. Thus, both MGMT-positive cell lines, T98G and LN-18, are mutant for p53. However, the mean EC50 values for TMZ in acute growth inhibition or clonogenic survival assays did not differ between these two groups of cell lines (p = 0.19). p53 function is negatively modulated by the levels of the murine double minute-2 protein. All glioma cell lines express murine double minute-2 protein, with U138MG, U87MG and LN428 expressing higher levels than the other nine cell lines (Weller et al. 1998). Even when considering p53 status and murine double minute-2 expression levels together, no apparent link to TMZ sensitivity became apparent. We then investigated the consequences of altering the p53 status in MGMT-negative glioma cell lines. The abrogation of wild-type p53 function by RNA interference protected LNT-229 cells and shifted the EC50 for the inhibition of clonogenic survival from 5.6 to 47.2 µm(Fig. 5b). A similar effect was seen when p53 function was abrogated by the dominant-negative p53 mutant p53V135A in LNT-229 (Fig. 5c) or U87MG siRNA (Fig. 5d) cells. These changes in TMZ sensitivity were not related to an increase in MGMT expression (data not shown). Conversely, the p53 rescue agent CP-31398 which restores p53 function in p53 mutant LN-18 cells (Wischhusen et al. 2003), sensitized LN-18 cells to TMZ (Fig. 5e) without modifying MGMT expression (data not shown), confirming a role of the p53 status in TMZ-induced cytotoxicity in glioma cells.


Figure 5. Modulation of temozolomide (TMZ) activity by p53. (a) p53 protein levels were assessed by immunoblot. LNT-229 puro or p53 siRNA (b), LNT-229 hygro or p53V135A (c) and U87MG puro or p53 siRNA (d) cells were assessed for TMZ sensitivity in clonogenic survival assays (mean and SEM, n = 3). (e) LN-18 cells were pre-treated with CP-31398 (4 µm) or vehicle for 6 h, treated with TMZ 24 h later for 24 h and monitored for clonogenic survival.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The emerging role of TMZ in the treatment of malignant glioma necessitates a better understanding of the molecular mechanisms underlying the therapeutic effects of this agent. Ideally, patients likely to benefit from TMZ chemotherapy would all receive this treatment whereas a priori non-responders would receive alternative treatments early in the disease course. Hypermethylation of the MGMT promoter region, which results in loss of expression, is strongly associated with prolonged survival in malignant glioma patients treated with TMZ (Hegi et al. 2004). The association of MGMT methylation status and progression-free and overall survival has been confirmed in the recent European Organisation for Research and Treatment of Cancer/National Cancer Institute of Canada trial (Hegi et al. 2005).

The present study employed a panel of human malignant glioma cell lines and transfection and gene-silencing strategies as a paradigm to answer some important questions regarding the use of TMZ in malignant glioma.

What is the role of O6-methylguanine DNA methyltransferase in determining resistance or sensitivity to temozolomide in vitro?

The median serum concentration of TMZ achieved in the treatment of human patients at a daily dose of 150 mg/m2 is 9.5 µg/mL corresponding to 50 µm (Hammond et al. 1999). Only one cell line, LNT-229, had an EC50 below 100 µm in the acute cytotoxicity assay (72 h continuous exposure) whereas nine of 12 cell lines had an EC50 below 100 µm in the clonogenic survival assay (Figs 1a and b). Accordingly, clinically relevant concentrations of TMZ have little acute cytotoxic activity but mainly act to inhibit proliferation and clonogenic survival in human malignant glioma cell lines. The two most resistant cell lines, LN-18 and T98G, were those with the highest MGMT protein levels and activity (Figs 1d and e) and both were strongly sensitized to TMZ by O6-BG, identifying MGMT as one important predictor of glioma cell sensitivity to TMZ. In this study of 12 glioma cell lines, there was overall strong correlation between clonogenic survival after TMZ exposure and MGMT activity (p < 0.0001). Furthermore, the resistance to TMZ and high MGMT activity reported here (Fig. 1) correlated well with previously reported resistance to carmustine (Weller et al. 1998), a nitrosourea compound previously considered the standard drug for the adjuvant treatment of malignant glioma (Glioma Meta-analysis Trialists Group 2002). A collateral sensitivity of diverse tumour cell lines, including glioblastoma cells to TMZ and lomustine or carmustine, has been shown previously (Baer et al. 1993; Gerson 2002; Pepponi et al. 2003). As the EC50 values in the clonogenicity assays (Fig. 1b) did not show strong variation among the MGMT-negative cell lines, it was unlikely that alterations in the mismatch repair system play a major role in the differential sensitivity of human malignant glioma cell lines to TMZ. Accordingly, the protein levels of MSH2, MSH6 and PMS2 (Fig. 3) did not correlate with clonogenic survival after TMZ exposure, confirming that mismatch repair is less important for TMZ sensitivity in glioma cells than MGMT, at least in vitro. That the activity of mismatch repair proteins is specifically regulated in vivo during the chemotherapy of gliomas and modulates clinical responses to TMZ in human patients cannot be excluded.

How significant are the clinically most relevant cotreatments, irradiation and dexamethasone, for the temozolomide sensitivity of malignant glioma cells?

Given the possible radiosensitization by TMZ mediating the beneficial effect on survival in the European Organisation for Research and Treatment of Cancer trial (Stupp et al. 2005) and the reports of MGMT induction by irradiation (Chan et al. 1992; Rafferty et al. 1996; Grombacher et al. 1998), we carefully assessed the effects of irradiation on MGMT activity but found none in cultured glioma cells (Table 1). Moreover, we observed only additive, but never synergistic, effects of irradiation and TMZ on clonogenic survival (Table 2). We had previously proposed that a synergy of TMZ with radiotherapy might rather derive from the inhibition by TMZ of the irradiation-induced increase in glioma cell motility (Wick et al. 2002). Further, unlike previous reports with other cell lines in which very high concentrations of glucocorticoids were applied (> 10 µm) (Grombacher et al. 1996; Biswas et al. 1999), dexamethasone at clinically achievable concentrations of 100 nm did not modulate MGMT activity and, unlike other chemotherapeutic drugs (Weller et al. 1997), did not decrease the cytotoxic effects of TMZ. An increase in MGMT mRNA levels in several glioma cell lines by 10 µm dexamethasone has been linked to resistance to nimustine (ACNU), although changes in MGMT activity were not ascertained (Ueda et al. 2004). Similarly, an induction of resistance of T98G glioma cells to TMZ by dexamethasone at 100–200 µm has been reported (Sur et al. 2005) but such concentrations are unlikely to be ever achieved in human glioma patients exposed to dexamethasone. The majority of studies suggest that dexamethasone levels in plasma do not exceed 200 nm unless excessive doses of dexamethasone are given (Brophy et al. 1983; Richter et al. 1983; Rohdewald et al. 1987).

What is the role of BCL-2 family proteins in temozolomide sensitivity of malignant glioma cells?

BCL-2 family proteins are key regulators of cell death induced by diverse stimuli. Down-regulation of BCL-2 with ensuing efflux of cytochrome c from mitochondria may mediate apoptosis triggered by O6-methylguanine in MGMT-deficient Chinese hamster fibroblasts or human lymphoblastoid cells (Ochs and Kaina 2000; Hickman and Samson 2004) and ovarian cancer cells (Barvaux et al. 2004b). Accordingly, we observed protection by BCL-XL from TMZ in the MGMT-negative cell line LNT-229 (Fig. 4). In contrast, the strong protection mediated by MGMT in LN-18 cells may have made it impossible to detect a moderate protection afforded by BCL-XL.

What is the role of p53 in temozolomide sensitivity of malignant glioma cells?

Mutations of the p53 gene are among the most common genetic alterations in human cancers and the mutations commonly result in a loss of transcriptional activity of p53. p53 mutations are also rather common (65%) in secondary glioblastomas developing by progression from grade II or III astrocytomas whereas only 10% of primary glioblastomas exhibit p53 mutations. The prediction, based on the suppression of MGMT expression by wild-type p53 (Harris et al. 1996; Srivenugopal et al. 2001), of higher sensitivity to TMZ in p53 wild-type than p53 mutant cell lines was not confirmed here. However, the cell lines retaining MGMT expression were both mutant for p53 and the higher incidence of MGMT expression in tumours in vivo (Preuss et al. 1995, 1996; Silber et al. 1998) than in cell lines in vitro (Fig. 1) may indicate that MGMT expression is lost during culturing, possibly precluding the detection of an overt link between p53 status and TMZ sensitivity.

A p53-mediated cell cycle arrest in response to TMZ may be necessary for TMZ-induced cytotoxicity (Hirose et al. 2001; Bocangel et al. 2002). We find that disruption of wild-type p53 function in two cell lines, LNT-229 and U87MG, by siRNA technology or ectopic expression of a dominant-negative mutant confers resistance to TMZ (Figs 5a–d). Moreover, the pharmacological rescue of mutant p53 function by CP-31398 sensitizes glioma cells to TMZ (Fig. 5e), strongly supporting the notion that wild-type p53 facilitates cytotoxic effects of TMZ. Of note, these effects of p53 were independent of alterations in MGMT expression or activity.

Accordingly, the combination of TMZ with novel agents such as CP-31398 (Foster et al. 1999; Wischhusen et al. 2003) or PRIMA-1 (Bykov et al. 2002), which aim at restoring p53 activity in mutant cell lines and may promote activation of the DNA damage response pathway in response to alkylating agents, appears to be a promising strategy to further expand the role of TMZ in the treatment of malignant glioma. Collectively, MGMT expression and p53 status may become valuable parameters to predict the response to TMZ in malignant glioma patients in vivo and to devise novel TMZ-based combination strategies using MGMT inhibitors as well as p53-agonistic drugs.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by an Award from the Jacqueline Seroussi Memorial Cancer Research Foundation to MW and grants from the German Research Council to BK (DFG KA724/12-1 and 13-1). The authors thank N. Curtin (Newcastle upon Tyne, UK) for providing NU1025.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Baer J. C., Freeman A. A., Newlands E. S., Watson A. J., Rafferty J. A. and Margison G. P. (1993) Depletion of O6-alkylguanine-DNA alkyltransferase correlates with potentiation of temozolomide and CCNU toxicity in human tumour cells. Br. J. Cancer 67, 12991302.
  • Barvaux V. A., Ranson M., Brown R., McElhinney R. S., McMurry T. B. and Margison G. P. (2004a) Dual repair modulation reverses temozolomide resistance in vitro. Mol. Cancer Ther. 3, 123127.
  • Barvaux V. A., Lorigan P., Ranson M., Gillum A. M., McElhinney R. S., McMurry T. B. and Margison G. P. (2004b) Sensitization of a human ovarian cancer cell line to temozolomide by simultaneous attenuation of the Bcl-2 antiapoptotic protein and DNA repair by O6-alkylguanine-DNA-alkyltransferase. Mol. Cancer Ther. 3, 12151220.
  • Biswas T., Ramana C. V., Srinivasan G., Boldogh I., Hazra T. K., Chen Z., Tano K., Thompson E. B. and Mitra S. (1999) Activation of human O6-methylguanine-DNA methyltransferase gene by glucocorticoid hormone. Oncogene 18, 525532.
  • Bocangel D. B., Finkelstein S., Schold S. C., Bhakat K. K., Mitra S. and Kokkinakis D. M. (2002) Multifaceted resistance of gliomas to temozolomide. Clin. Cancer Res. 8, 27252734.
  • Brophy T. R. O., McCafferty J., Tyrer J. H. and Eadie M. J. (1983) Bioavailability of oral dexamethasone during high dose steroid therapy in neurological patients. Eur. J. Clin. Pharmacol. 24, 103108.
  • Bykov V. J., Issaeva N., Shilov A., Hultcrantz M., Pugacheva E., Chumakov P., Bergman J., Wiman K. G. and Selivanova G. (2002) Restoration of the tumor suppressor function to mutant p53 by a low-molecular-weight compound. Nat. Med. 8, 282288.
  • Ceccotti S., Aquilina G., Macpherson P., Yamada M., Karran P. and Bignami M. (1996) Processing of O6-methylguanine by mismatch correction in human cell extracts. Curr. Biol. 6, 15281531.
  • Chan C. L., Wu Z., Eastman A. and Bresnick E. (1992) Irradiation-induced expression of O6-methylguanine-DNA methyltransferase in mammalian cells. Cancer Res. 52, 18041809.
  • Chang D. K., Ricciardiello L., Goel A., Chang C. L. and Boland C. R. (2000) Steady-state regulation of the human DNA mismatch repair system. J. Biol. Chem. 275, 18 42418 431.
  • Curtin N. J., Wang L. Z., Yiakouvaki A. et al. (2004) Novel poly(ADP-ribose) polymerase-1 inhibitor, AG14361, restores sensitivity to temozolomide in mismatch repair-deficient cells. Clin. Cancer Res. 10, 881889.
  • Duckett D. R., Drummond J. T., Murchie A. I., Reardon J. T., Sancar A., Lilley D. M. and Modrich P. (1996) Human MutSalpha recognizes damaged DNA base pairs containing O6-methylguanine, O4-methylthymine, or the cisplatin-d(GpG) adduct. Proc. Natl Acad. Sci. USA 93, 64436447.
  • Esteller M., Garcia-Foncillas J., Andion E., Goodman S. N., Hidalgo O. F., Vanaclocha V., Baylin S. B. and Herman J. G. (2000) Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N. Engl. J. Med. 343, 13501354.
  • Foster B. A., Coffey H. A., Morin M. J. and Rastinejad F. (1999) Pharmacological rescue of mutant p53 conformation and function. Science 286, 25072510.
  • Friedman H. S., Pluda J., Quinn J. A. et al. (2000) Phase I trial of carmustine plus O6-benzylguanine for patients with recurrent or progressive malignant glioma. J. Clin. Oncol. 18, 35223528.
  • Friedman H. S., Keir S., Pegg A. E., Houghton P. J., Colvin O. M., Moschel R. C., Bigner D. D. and Dolan M. E. (2002) O6-benzylguanine-mediated enhancement of chemotherapy. Mol. Cancer Ther. 1, 943948.
  • Gerson S. L. (2002) Clinical relevance of MGMT in the treatment of cancer. J. Clin. Oncol. 20, 23882399.
  • Glaser T., Wagenknecht B. and Weller M. (2001) Identification of p21 as a target of cycloheximide-mediated facilitation of CD95-mediated apoptosis in human malignant glioma cells. Oncogene 20, 47574767.
  • Glioma Meta-analysis Trialists Group (2002) Chemotherapy in adult high-grade glioma: a systematic review and meta-analysis of individual patient data from 12 randomised trials. Lancet 359, 10111018.
  • Grombacher T., Mitra S. and Kaina B. (1996) Induction of the alkyltransferase (MGMT) gene by DNA damaging agents and the glucocorticoid dexamethasone and comparison with the response of the base excision repair genes. Carcinogenesis 17, 23292336.
  • Grombacher T., Eichhorn U. and Kaina B. (1998) p53 is involved in regulation of the DNA repair gene O6-methylguanine-DNA methyltransferase (MGMT) by DNA damaging agents. Oncogene 17, 845851.
  • Hammond L., Eckardt J. R., Baker S. D. et al. (1999) Phase I and pharmacokinetic study of temozolomide on a daily-for-5-days schedule in patients with advanced solid malignancies. J. Clin. Oncol. 17, 26042613.
  • Harris L. C., Remack J. S., Houghton P. J. and Brent T. P. (1996) Wild-type p53 suppresses transcription of the human O6-methylguanine-DNA methyltransferase gene. Cancer Res. 56, 20292032.
  • Hegi M. E., Diserens A. C., Godard S. et al. (2004) Clinical trial substantiates the predictive value of O-6-methylguanine-DNA methyltransferase promoter methylation in glioblastoma patients treated with temozolomide. Clin. Cancer Res. 10, 18711874.
  • Hegi M. E., Diserens A. C., Gorlia T. et al. (2005) MGMT gene silencing and response to temozolomide in glioblastoma. N. Engl. J. Med. 352, 9971003.
  • Hickman M. J. and Samson L. D. (1999) Role of DNA mismatch repair and p53 in signalling induction of apoptosis by alkylating agents. Proc. Natl Acad. Sci. USA 96, 10 76410 769.
  • Hickman M. J. and Samson L. D. (2004) Apoptotic signaling in response to a single type of DNA lesion, O(6)-methylguanine. Mol. Cell 14, 105116.
  • Hirose Y., Berger M. S. and Pieper R. O. (2001) p53 effects both the duration of G2/M arrest and the fate of temozolomide-treated human glioblastoma cells. Cancer Res. 61, 19571963.
  • Kaina B. and Christmann M. (2002) DNA repair in resistance to alkylating anticancer drugs. Int. J. Clin. Pharm. Ther. 40, 354367.
  • Kaina B., Fritz G., Mitra S. and Coquerelle T. (1991) Transfection and expression of human O6-methylguanine-DNA methyltransferase (MGMT) cDNA in Chinese hamster cells: the role of MGMT in protection against the genotoxic effects of alkylating agents. Carcinogenesis 12, 18571867.
  • Kaina B., Mühlhausen U., Piee-Staffa A., Christmann M., Garcia Boy R., Rosch F. and Schirrmacher R. (2004) Inhibition of O6-methylguanine-DNA methyltransferase by glucose-conjugated inhibitors and effect on fotemustine and temozolomide-induced cell death. J. Pharmacol. Exp. Ther. 311, 585593.
  • Kanzawa T., Bedwell J., Kondo Y., Kondo S. and Germano I. M. (2003) Inhibition of DNA repair for sensitizing resistant glioma cells to temozolomide. J. Neurosurg. 99, 10471052.
  • Kokkinakis D. M., Bocangel D. B., Schold S. C., Moschel R. C. and Pegg A. E. (2001) Thresholds of O6-alkylguanine-DNA alkyltransferase which confer significant resistance of human glial tumor xenografts to treatment with 1,3-bis(2-chloroethyl)-1-nitrosourea or temozolomide. Clin. Cancer Res. 7, 421428.
  • Ma J., Murphy M., O'Dwyer P. J., Berman E., Reed K. and Gallo J. M. (2002) Biochemical changes associated with a multidrug-resistant phenotype of a human glioma cell line with temozolomide-acquired resistance. Biochem. Pharmacol. 63, 12191228.
  • Naumann U., Durka S. and Weller M. (1998) Dexamethasone-mediated protection from drug cytotoxicity: association with p21 WAF1/CIP1 protein accumulation? Oncogene 17, 15671575.
  • Ochs K. and Kaina B. (2000) Apoptosis induced by DNA damage O6-methylguanine is Bcl-2 and caspase-9/3 regulated and Fas/caspase-8 independent. Cancer Res. 60, 58155824.
  • Pagani E., Pepponi R., Fuggetta M. P. et al. (2003) DNA repair enzymes and cytotoxic effects of temozolomide: comparative studies between tumor cells and normal cells of the immune system. J. Chemother. 15, 173183.
  • Pegg A. E. (1990) Properties of mammalian O6-alkylguanine-DNA transferases. Mutat. Res. 233, 165175.
  • Pepponi R., Marra G., Fuggetta M. P., Falcinelli S., Pagani E., Bonmassar E., Jiricny J. and D'Atri S. (2003) The effect of O6-alkylguanine-DNA alkyltransferase and mismatch repair activities on the sensitivity of human melanoma cells to temozolomide, 1,3-bis(2-chloroethyl)1-nitrosourea, and cisplatin. J. Pharmacol. Exp. Ther. 304, 661668.
  • Preuss I., Eberhagen I., Haas S., Eibl R. H., Kaufmann M., Von Minckwitz G. and Kaina B. (1995) O6-methylguanine-DNA methyltransferase activity in breast and brain tumors. Int. J. Cancer 61, 321326.
  • Preuss I., Haas S., Eichhorn U. et al. (1996) Activity of the DNA repair protein O6-methylguanine-DNA methyltransferase in human tumor and corresponding normal tissue. Cancer Detect. Prev. 20, 130136.
  • Quinn J. A., Pluda J., Dolan M. E. et al. (2002) Phase II trial of carmustine plus O(6)-benzylguanine for patients with nitrosourea-resistant recurrent or progressive malignant glioma. J. Clin. Oncol. 20, 22772283.
  • Rafferty J. A., Clarke A. R., Sellappan D., Koref M. S., Frayling I. M. and Margison G. P. (1996) Induction of murine O6-alkylguanine-DNA-alkyltransferase in response to ionising radiation is p53 gene dose dependent. Oncogene 12, 693697.
  • Richter O., Ern B., Reinhardt D. and Becker B. (1983) Pharmacokinetics of dexamethasone in children. Ped. Pharmacol. 3, 329337.
  • Rohdewald P., Möllmann H., Barth J., Rehder J. and Derendorf H. (1987) Pharmacokinetics of dexamethasone and its phosphate ester. Biopharm. Drug Dispos. 8, 205212.
  • Roos W., Baumgartner M. and Kaina B. (2004) Apoptosis triggered by DNA damage O6-methylguanine in human lymphocytes requires DNA replication and is mediated by p53 and Fas/CD95/Apo-1. Oncogene 23, 359367.
  • Silber J. R., Bobola M. S., Ghatan S., Blank A., Kolstoe D. D. and Berger M. S. (1998) O6-methylguanine-DNA methyltransferase activity in adult gliomas: relation to patient and tumor characteristics. Cancer Res. 58, 10681073.
  • Srivenugopal K. S., Shou J., Mullapudi R., Lang F. F., Rao J. S. and Ali-Osman F. (2001) Enforced expression of wild-type p53 curtails the transcription of the O(6)-methylguanine-DNA methyltransferase gene in human tumor cells and enhances their sensitivity to alkylating agents. Clin. Cancer Res. 7, 13981409.
  • Streffer J. R., Rimner A., Rieger J., Naumann U., Rodemann H. P. and Weller M. (2002) Bcl-2 family protein expression modulates radiosensitivity in human glioma cells. J. Neuro-Oncol. 56, 4349.
  • Stupp R., Mason W. P., Van Den Bent M. J. et al. on behalf of the European Organisation for Research and Treatment of Cancer (EORTC) Brain Tumor and Radiotherapy Groups and National Cancer Institute of Canada Clinical Trials Group (NCIC CTG) (2005) Radiotherapy plus concomitant and adjuvant temozolomide for patients with newly diagnosed glioblastoma. N. Engl. J. Med. 352, 987996.
  • Sur P., Sribnick E. A., Patel S. J., Ray S. K. and Banik N. L. (2005) Dexamethasone decreases temozolomide-induced apoptosis in human glioblastoma T98G cells. Glia 50, 160167.
  • Tentori L., Lacal P. M., Benincasa E., Franco D., Faraoni I., Bonmassar E. and Graziani G. (1998) Role of wild-type p53 on the antineoplastic activity of temozolomide alone or combined with inhibitors of poly(ADP-ribose) polymerase. J. Pharmacol. Exp. Ther. 285, 884893.
  • Tentori L., Portarena I., Bonmassar E. and Graziani G. (2001) Combined effects of adenovirus-mediated wild-type p53 transduction, temozolomide and poly(ADP-ribose) polymerase inhibitor in mismatch repair deficient and non-proliferating tumor cells. Cell Death Differ. 8, 457469.
  • Ueda S., Mineta T., Nakahara Y., Okamoto H., Shiraishi T. and Tabuchi K. (2004) Induction of the DNA repair gene O6-methylguanine-DNA-methyltransferase by dexamethasone in glioblastomas. J. Neurosurg. 101, 659663.
  • Wedge S. R. and Newlands E. S. (1996) O6-benzylguanine enhances the sensitivity of a glioma xenograft with low O6-alkylguanine-DNA alkyltransferase activity to temozolomide and BCNU. Br. J. Cancer 73, 10491052.
  • Weller M., Frei K., Groscurth P., Krammer P. H., Yonekawa Y. and Fontana A. (1994) Anti-Fas/APO-1 antibody-mediated apoptosis of cultured human glioma cells. Induction and modulation of sensitivity by cytokines. J. Clin. Invest. 94, 954964.
  • Weller M., Schmidt C., Roth W. and Dichgans J. (1997) Chemotherapy of human malignant glioma: prevention of efficacy by dexamethasone? Neurology 48, 17041709.
  • Weller M., Rieger J., Grimmel C., Van Meir E. G., De Tribolet N., Krajewski S., Reed J. C., Von Deimling A. and Dichgans J. (1998) Predicting chemoresistance in human malignant glioma cells: the role of molecular genetic analyses. Int. J. Cancer 79, 640644.
  • Wick W., Wick A., Schulz J. B., Dichgans J., Rodemann H. P. and Weller M. (2002) Prevention of irradiation-induced glioma cell invasion by temozolomide involves caspase 3 activity and cleavage of focal adhesion kinase. Cancer Res. 62, 19151919.
  • Wischhusen J., Naumann U., Ohgaki H., Rastinejad F. and Weller M. (2003) CP-31398, a novel p53-stabilizing agent, induces p53-dependent and p53-independent cell death. Oncogene 22, 82338245.
  • Yung W. K., Prados M. D., Yaya-Tur R. et al. (1999) Multicenter phase II trial of temozolomide in patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. Temodal Brain Tumor Group. J. Clin. Oncol. 17, 27622771.
  • Yung W. K., Albright R. E., Olson J. et al. (2000) A phase II study of temozolomide vs. procarbazine in patients with glioblastoma multiforme at first relapse. Br. J. Cancer 83, 588593.