SEARCH

SEARCH BY CITATION

Keywords:

  • O6-methylguanine-DNA methyltransferase;
  • glioblastoma;
  • carmustine wafer;
  • methylation-specific polymerase chain reaction;
  • methylation;
  • immunohistochemistry

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND:

O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status was proposed as a prognostic biomarker for patients with glioblastoma. However, the prognostic impact of MGMT in patients with newly diagnosed glioblastoma who receive carmustine-releasing wafers (Gliadel) along with temozolomide (TMZ) is still unknown.

METHODS:

MGMT promoter methylation status and protein expression were analyzed in formalin-fixed, paraffin-embedded tumor specimens obtained from 111 French patients with newly diagnosed glioblastoma. Patients received the Gliadel wafers followed by radiotherapy plus concomitant and adjuvant TMZ chemotherapy while they were enrolled in a French multicenter prospective study.

RESULTS:

For the whole cohort, the median overall survival (OS) was 17.5 months, and the progression-free survival was 10.3 months. Patients with tumors that harbored MGMT methylation had a significantly longer OS compared with patients who had wild-type MGMT (21.7 months vs 15.1 months; P = .025). Similarly, patients who had low MGMT protein expression (≤15%) had a significantly improved OS compared with patients who had high MGMT expression (27.0 months vs 15.1 months; P = .021). The extent of resection was the strongest clinical predictor of outcome. In multivariate Cox models that were adjusted for sex, performance status, and extent of surgery, both MGMT methylation and protein expression were identified as independent prognosticators, and the finding was validated internally using a bootstrap resampling technique. Discrepancies were identified between protein expression and MGMT methylation status, thus suggesting that the 2 assays probably assess different biologic features.

CONCLUSIONS:

MGMT promoter methylation status and low MGMT expression both were identified as positive prognosticators in patients with newly diagnosed glioblastoma who underwent surgical resection and received Gliadel wafer implants followed by adjuvant radiotherapy and concomitant oral TMZ chemotherapy (the Stupp protocol). Cancer 2012. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

O6-methylguanine-DNA methyltransferase (MGMT) is a key enzyme in the base-excision pathway of DNA repair (BER). MGMT removes mutagenic and cytotoxic adducts from O6-guanine in DNA, the preferred point of attack for alkylating drugs, such as carmustine (BCNU) or temozolomide (TMZ), which are used in the treatment of glioblastoma. Hypermethylation of the CpG islands located in the region of MGMT is primarily responsible for the loss of MGMT function in several tumor types.1 In glioblastomas, the reported frequency of MGMT inactivation ranged from 45% to 68%.2-4 Thus, an analysis of epigenetic silencing may be a valid prognostic marker in patients with high-grade glioma.2, 5 Indeed, it was demonstrated that MGMT-hypermethylated tumors were more sensitive to the killing effects of alkylating drugs that induced 06-guanine DNA adducts, because tumor cells with low MGMT expression were unable to repair such DNA lesions and, thus, were prone to apoptosis.2 Recognizing MGMT status as a prognostic or predictive biomarker in brain tumors raises the issue of using standardized assays for the routine evaluation of MGMT status. At least 3 methods are available to assess MGMT status: 1) determining MGMT enzyme activity, 2) evaluating MGMT protein expression by immunohistochemistry (IHC), and 3) establishing MGMT promoter methylation status based on methylation-specific polymerase chain reaction (MSP) or pyrosequencing. MGMT activity can be evaluated only by using snap-frozen tumor samples, which may limit routine use, whereas MGMT protein expression and MSP may be performed using paraffin-embedded tumor tissues.6 Because contradictory results have been reported across laboratories based on these methods, agreement on the best and most reliable technique for evaluating MGMT status still needs to be achieved.7, 8

In January 2005, biodegradable BCNU wafers for implantation in the resection cavity (Gliadel implant; Eisai, Inc., Woodcliff Lake, NJ; Archimedes Pharma, Reading, United Kingdom) were approved in France as adjuvant treatment in association with surgery and radiotherapy (RT) for patients with newly diagnosed glioblastoma. In the same year, adjuvant radiotherapy and concomitant oral TMZ chemotherapy (European Organization for Research and Treatment of Cancer [EORTC] protocol 26981; also known as the “Stupp protocol”) were considered standard treatment for patients with glioblastoma.9 Consequently, BCNU wafers were involved in the combined use of both therapeutic strategies for patients with newly diagnosed glioblastoma. The wafers slowly released BCNU over a period of 2 to 3 weeks. Thus, the rationale underlying such local therapy was to fill the gap between surgery and TMZ treatment to maintain alkylating treatment pressure on residual tumor cells. Because MGMT loses its enzymatic activity after receiving the alkyl group from alkylated DNA, BCNU implants are likely to induce the functional depletion of MGMT tumors, thereby enhancing the efficacy of TMZ. Conversely, the risk of adverse MGMT gene induction or selection of resistant tumor cells after alkylating agent (ie, BCNU) exposure has been suggested.10, 11

Given the aforementioned considerations, the primary objective of our current study was to assess the prognostic significance of MGMT status in patients with newly diagnosed glioblastoma who received BCNU wafers followed by the Stupp protocol, and the secondary goal was to compare the utility of MSP and IHC, 2 laboratory tests currently used to determine MGMT status. For this purpose, both MGMT promoter methylation and MGMT protein expression were analyzed in a series of 111 consecutive French patients with newly diagnosed glioblastoma who received Gliadel wafers followed by RT plus concurrent and adjuvant TMZ chemotherapy. Patients were treated through a network of 11 neurosurgical university departments in France.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Study Design and Patient Inclusion Criteria

Tumor samples were collected from 111 patients who were included in 2 observational studies that analyzed the daily practice of French neuro-oncologists in relation to Gliadel implantation. A patient flow chart is provided in Figure 1. Clinical details from 1 of those observational studies were previously published.12 In both studies, patients were followed until death, and clinical and radiologic evaluations were conducted in a standard manner. The extent of resection (with total resection defined as the disappearance of contrast enhancement, subtotal resection defined as the disappearance of ≥90% of contrast enhancement, and partial resection defined as the disappearance of <90% of contrast enhancement) was assessed using postoperative magnetic resonance imaging (MRI) or computed tomodensitometry (CT) scans with 2-dimensional measurements.

thumbnail image

Figure 1. This is a flow diagram of the study patients. RT indicates radiotherapy.

Download figure to PowerPoint

MGMT status was evaluated by a steering committee according to a detailed protocol (available on request). The inclusion criteria for the current study were as follows: 1) newly diagnosed, glioblastoma; 2) the use of Gliadel implants followed by adjuvant treatment according to the Stupp protocol; and 3) the availability of routine formalin-fixed, paraffin-embedded (FFPE) tumor tissues. Overall, 11 centers with at least 5 recruited patients provided specimens. All tumor specimens were reviewed by 2 independent pathologists (E.L.-Z. and F.C.) according to the histopathologic classification system adopted by the World Health Organization in 2007. Immunostaining and MSP were performed at Caen University Hospital. All patients provided informed consent regarding the collection of tumor specimens and their molecular evaluation, as required by French laws. Tissue collection was registered in accordance with French regulatory requirements (collection number DC-200866588), and tumor specimens were analyzed in a blinded manner with respect to patient identity. The study was approved by the Institutional Ethics Committee of Caen University Hospital (Caen, France).

Immunohistochemistry for O6-Methylguanine-DNA Methyltransferase

In short, 3-μm paraffin sections were cut from each block. Immunostaining was performed using heat-induced epitope retrieval, pH 9.0, a labeled method (EnVision+ Kit; Dako SA, Trappes, France), and automate immunostainer (Dako SA) according to the manufacturer's protocol. A mouse primary antibody against MGMT (clone MT3.1; Chemicon International, Temecula, Calif) was used at 1:200 dilution. Negative controls consisted of omitting the primary antibody and replacing it with an irrelevant antibody of similar isotype. Endothelial staining was used as an internal positive control.3, 13

A pathologist (E.L.-Z.) who was blind to the patients' clinical and MGMT methylation data independently evaluated MGMT staining using a light microscope at ×400 magnification. Specimens without valid internal positive controls were excluded from the analysis. For each specimen, 5 to 10 images of representative fields were then acquired at ×400 magnification. Between 360 and 1790 tumor cells were counted in specimen, and the percentage of positive tumor nuclei was calculated. Endothelial and inflammatory cells were excluded from the cell counts.

DNA Extraction and Methylation-Specific Polymerase Chain Reaction Analysis

For all tumor samples, a hematoxylin-stained section was evaluated by a neuropathologist (E.L.-Z.), who labeled the tumor area that contained the smallest amount of necrosis and inflammation (<40%). DNA from FFPE tissue was extracted from that area using the QIAmp DNA FFPE Tissue Kit (Qiagen SA, Courtaboeuf, France) according to the manufacturer's recommendation. MSP was performed using a 2-step approach. Bisulfite modification of genomic DNA was undertaken by means of the Epitect Kit (Qiagen SA) according the manufacturer's recommendation. PCR amplification was carried out as described by Esteller et al.1 PCR products were loaded onto 5% agarose gels, stained with GelRed (Interchim, Montlucon, France), and observed under ultraviolet illumination.

Statistical Analysis

To identify any selection bias when grouping patients from 2 clinical studies, baseline characteristics of the 2 patient groups were compared using chi-square tests, Fisher exact tests, Student t tests, and Mann-Whitney tests. Patients who remained alive were censored at the date of their last visit. Survival rates and curves were estimated using the Kaplan-Meier method, and comparisons of prognostic subgroups were conducted using the log-rank test. The relation between MGMT promoter methylation and MGMT expression was assessed using the Mann-Whitney test. MGMT expression was classified as low (reactivity <15%) or high (reactivity ≥15%), because this cutoff was the median value of reactivity in our glioblastoma series. To account for the effects of prognostic factors on overall survival (OS) and progression-free survival (PFS), multiple regression analyses were performed using a Cox proportional hazards model, estimating the adjusted hazard ratio (HR) with 95% confidence intervals (CI) for each variable relative to the risk of death or disease progression. For both OS and PFS, in multiple regression analyses, 1 model was computed using MSP, and another was computed using IHC, and the 2 models were adjusted for potential prognostic factors, notably, age (≥50 years), sex, Karnofsky performance status (KPS) (≤70), and extent of surgery (<90%). The significance level for selecting variables in multivariate analyses was set at .20, and the bilateral significance level was fixed at .05 for each test. A bootstrap resampling analysis on 300 samples was performed; all steps of the model-development process were replicated for each bootstrapped sample,14 and optimism-corrected concordance indexes (C-indexes) were calculated.14

A parallel analysis was conducted to combine both methods (MSP and IHC) into a single double-check test in which patients were classified as “methylated-low expression,” “unmethylated-high expression,” or “discrepancy.” OS was estimated with the Kaplan-Meier method and was compared using a multivariate Cox model. The R software package (version 2.12.1; R Foundation for Statistical Computing, Vienna, Austria) was used for statistical analyses and to create the figures.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patient Characteristics

Overall, 11 centers were able to provide 1 tumor block for a total of 111 patients enrolled in 2 observational studies that analyzed the daily practices of French neuro-oncologists in relation to using Gliadel implantation. The first study was conducted between January 2005 and March 2006 and recruited a total of 163 patients, including 83 patients (51%) with newly diagnosed malignant glioma and 80 patients (49%) with recurrent malignant glioma. That study coincided with the introduction of the Stupp protocol as the treatment standard for newly diagnosed glioblastoma.9 Therefore, only 43 of 83 patients with newly diagnosed malignant glioma in that study received concomitant radiochemotherapy with TMZ. Among those 43 patients, only 31 patients were eligible for our current study. In the second study, which was conducted between April 2006 and January 2009, the Stupp protocol was offered systematically to all 91 enrolled patients. Because of tumor tissue availability, 80 of the 91 recruited patients were eligible for our current MGMT study. Tumor specimens and patient flow are illustrated in Figure 1.

The final patient population consisted of 111 patients, including 38 women and 73 men, who had a median follow-up of 13.6 months (range, 0-37.6 months). Patient characteristics are summarized in Table 1. The median age was 58 years (range, 33-77 years), and the mean ± standard deviation KPS score was 80.2 ± 13.5. At the final analysis, 56 of 111 patients (50.5%) had died. Patients from the 2 studies were comparable in terms of clinical and molecular factors (Table 1).

Table 1. Baseline Patient Characteristics
 No. of Patients (%) 
CharacteristicFirst Observational Study, n = 31Second Observational Study, n = 80Total Cohort, n = 111P
  • Abbreviations: IHC, immunohistochemistry; KPS, Karnofsky performance status; MSP, methylation-specific polymerase chain reaction; SD, standard deviation.

  • a

    Five KPS score values were missing in the second study.

  • b

    One MSP methylation value was missing in the first study.

  • c

    Four IHC expression values were missing in the first study, and 1 was missing in the second study.

Sex   .54
 Men19 (61.3)54 (67.5)73 (65.8) 
 Women12 (38.7)26 (32.5)38 (34.2) 
Age    
 Median [range (years)57 [34-77]58 [33-76]58 [33-77].39
KPS score    
 Mean±SDa76.8±17.681.6±11.380.2±13.5.17
Extent of resection   .87
 Total: 100%16 (51.6)45 (57)61 (55.5) 
 Subtotal: ≥90%10 (32.3)23 (29.1)33 (30) 
 Partial: <90%5 (16.1)11 (13.9)16 (14.5) 
Follow-up: Median [range], mo    
 All patients13.6 [2.6-27.3]13.4 [0-37.6]13.6 [0-37.6].69
 Surviving patients14.2 [4.7-19.9]15.3 [0-37.6]15.2 [0-37.6].27
MSP    
 Methylationb20 (66.7)43 (53.8)63 (57.3).22
IHC    
 Low expressionc9 (33.3)42 (53.2)51 (48.1).08

O6-Methylguanine-DNA Methyltransferase Analysis

MGMT promoter methylation was assessed in 110 of 111 tumor specimens (99%) using a standard MSP assay and was detected in 63 tumors (57.3%), whereas MGMT protein expression was quantified successfully in 106 specimens (95.5%). In the absence of an internal control, 5 specimens were considered nonassessable for IHC. Initially, the percentage of stained nuclei as a continuous variable was analyzed. The methylated group exhibited statistically significant, lower percentages of MGMT-stained nuclei (median, 9%; interquartile range, 5%-18%) compared with the unmethylated group (median, 50.5%; interquartile range, 15.75%-90%; Mann-Whitney test; P < .0001) (Fig. 2). Quantitative IHC allowed for classifying tumors into 2 groups with the cutoff set at the median percentage of MGMT-stained nuclei, as previously proposed.13, 15 Because the median percentage of tumor cells that exhibited MGMT protein expression when using IHC was 15% (range, 0%-99%), 51 of 106 samples (48.11%) were classified as tumors with “low MGMT” protein expression (Fig. 3a), and 55 samples (51.89%) were classified as tumors with “high MGMT” protein expression (Fig. 3b).

thumbnail image

Figure 2. The distribution (%) of O6-methylguanine-DNA methyltransferase (MGMT) protein expression is illustrated according to MGMT promoter methylation status. Box plots represent medians, quantiles, and extreme points. Gray dots represent all MGMT protein expression (Mann-Whitney test comparing those 2 distributions: P < .0001).

Download figure to PowerPoint

thumbnail image

Figure 3. Representative results of O6-methylguanine-DNA methyltransferase (MGMT) immunoreactivity in glioblastoma are illustrated for (a) low MGMT protein expression and (b) high MGMT protein expression. Endothelial staining (arrows) was used as an internal positive control (original magnification, ×40).

Download figure to PowerPoint

Overall, concordant results between MGMT promoter methylation and MGMT protein expression were observed in 73 of 105 tumor specimens (69.52%). However, 21 tumors with high MGMT protein expression had MGMT promoter methylation, whereas 11 tumors with low MGMT protein expression lacked MGMT promoter methylation, as detected by our MSP assay. The potential association between discordant MGMT profiling and preanalytic variables, such as fixative or histologic variables (ie, cytoplasmic background), was examined next, although no statistical correlation was observed to account for the aforementioned discrepancies.

Progression-Free Survival

During follow-up, 87 of 111 patients were diagnosed with presumed local tumor relapse. The median PFS was 10.3 months (95% CI, 8.7-12.0 months) for the whole patient series, and the PFS rate was 76.6% at 6 months, 40.7% at 12 months, 15.6% at 18 months, and 7.8% at 24 months. In univariate analysis, the extent of resection was the only clinical factor that influenced PFS: The median PFS for patients who underwent partial resection was 4.4 months versus 10.8 months for those who underwent macroscopically total resection (P < .001). A nonsignificant trend favoring patients with tumors that had unmethylated MGMT promoter was observed (PFS: 9.7 months vs 10.7 months for patients without; P = .155) (Fig. 4a), whereas PFS was significantly longer for patients with low MGMT protein expression (9.6 months vs 11.5 months for patients with high protein MGMT expression; P = .049) (Fig. 4b). Other factors, such as age, sex, and KPS score, did not significantly influence PFS in univariate analysis.

thumbnail image

Figure 4. Kaplan-Meier survival curves illustrate the progression free survival of patients with glioblastoma according to (a) O6-methylguanine-DNA methyltransferase (MGMT) methylation status and (b) MGMT protein expression. Kaplan-Meier survival curves illustrate the overall survival of patients with glioblastoma according to (c) MGMT methylation status and (d) MGMT protein expression.

Download figure to PowerPoint

In the multivariate Cox proportional hazards model, after adjusting for sex, KPS score, and extent of resection, patients who had methylated MGMT promoter status had significantly longer PFS compared with those who had unmethylated MGMT promoter status (HR, 0.61; 95% CI, 0.39-0.97; P = .036) (Table 2). When MGMT expression was evaluated with IHC, multivariate analysis did not reveal any significant difference in PFS between patients who had low and high MGMT protein expression (Table 3).

Table 2. Multivariate Analysis of Factors Associated With survival, N = 104
 Overall SurvivalProgression-Free Survival
VariablePHR95% CIPHR95% CI
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; KPS, Karnofsky performance status; MGMT, O6-methylguanine-DNA methyltransferase; NS, not significant.

MGMT promoter methylation.0080.460.26-0.82.0360.610.39-0.97
MenNS  .0451.651.01-2.69
Incomplete resection (<90%).0023.241.52-6.91<.00014.312.34-7.94
KPS score ≤70%NS  NS  
Table 3. Multivariate Analysis of Factors Associated With Survival, N = 100
 Overall SurvivalProgression-Free Survival
VariablePHR95% CIPHR95% CI
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; KPS, Karnofsky performance status; MGMT, O6-methylguanine-DNA methyltransferase; NS, not significant.

Low MGMT expression (<15%).0100.450.24-0.82NS  
Men.0422.011.02-3.95NS  
Incomplete resection (<90%).0192.511.16-5.43.00023.141.73-5.7
KPS score ≤70%.0072.261.26-4.07NS  

Overall Survival

The median OS was 17.5 months (95% CI, 15.1-22.6 months) for the overall study population, with OS rates of 71.7% (95% CI, 63%-81%) at 12 months, 34.2% (95% CI, 24%-50%) at 24 months, and 18.1% (95% CI, 9%-36%) at 36 months. In univariate analysis, the factors that influenced OS were MGMT promoter methylation status (21.7 months vs 15.1 months; P = .025) (Fig. 4c), MGMT protein expression (15.1 months vs 27.0 months; P = .021) (Fig. 4d), and extent of resection (20.2 months for total/subtotal resection [≥90%], with 1-year, 2-year, and 3-year OS rates of 74.6%, 37.7%, and 19.9%, respectively, vs 14.2 months for partial [<90%] resection, with 1-year and 2-year OS rates of 51.9% and 17.3%, respectively; P = .016). Other factors, such as age, sex, and KPS score, did not significantly influence OS in univariate analysis.

In a multivariate Cox model that was adjusted for sex, KPS score, and extent of resection, patients who had methylated MGMT promoter status experienced significantly longer survival compared with those who had unmethylated MGMT promoter status (HR for death, 0.46; 95% CI, 0.26-0.82; P = .008) (Table 2). Similarly, patients who had low MGMT protein expression according to IHC analysis had significantly longer survival compared with those who had high MGMT protein expression (HR, 0.45; 95% CI, 0.24-0.82; P = .010) (Table 3). The C-indexes from multivariate analysis were 0.74 for MSP and 0.70 for IHC. The prognostic value of MGMT status for OS was validated further by the resampling bootstrap procedure based on 300 replications. MGMT methylation status and low MGMT expression by IHC were associated with significantly longer OS in 81.67% and 84.00% of samples, respectively, according to univariate analysis, and in 79.67% and 66.33% of samples, respectively, in multivariate analyses, confirming the strength of MGMT methylation status and MGMT expression by IHC for the prognostic evaluation of OS. In the same analyses, the extent of resection was selected as a prognostic factor in 80%, 77.67%, and 62% of samples according to univariate analysis, multivariate analysis with MSP, and multivariate analysis with MGMT IHC, respectively. The optimism-corrected C-indexes were 0.73 and 0.69 for MSP and IHC, respectively.

When combining both methods (MSP and IHC) into a single double-check test, the methylated-low expression group exhibited very good survival (median OS, 27 months) compared with patients who had a discrepancy between the 2 methods (median OS, 17.5 months), whereas the unmethylated-high expression group had the worst survival (median OS, 14.7 months; overall P = .0013) (Table 4). The 2 different discrepancy subgroups had the same OS (log-rank test; P = .94). Considering only the concordant results derived from this double-check test, MGMT methylation combined with MGMT protein expression (HR, 2.7; 95% CI, 1.3-5.3) was associated with better OS results compared with MSP (HR, 1.8; 95% CI, 1.1-3.2) or MGMT protein expression alone (HR, 1.9; 95% CI, 1.1-3.3). The same result was observed in multivariate analyses, as indicated in Tables 2, 3, and 4.

Table 4. Multivariate Analysis of Factors Associated With Overall Survival, N = 99
 Overall Survival
VariablePHR95% CI
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; KPS, Karnofsky performance status; MGMT, O6-methylguanine-DNA methyltransferase; NA, not applicable; Ref, reference group.

MGMT methylation and expression.0013  
 Unmethylated and high expression (Ref) 1.0NA
 Discrepancy 0.40.21-0.86
 Methylated and low expression 0.30.12-0.55
Men.0152.41.18-4.72
Incomplete resection (<90%).0043.21.58-6.34
KPS score ≤70%.0152.11.15-3.80

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

In this study, our objective was to further define the prognostic significance of MGMT in a series of patients with newly diagnosed glioblastoma who underwent surgical resection and received Gliadel BCNU implants followed by the Stupp protocol according to the princeps publications.2, 9 To investigate MGMT expression and MGMT promoter methylation status simultaneously, paraffin sections from the same block were used for DNA extraction and immunostaining. Patient baseline characteristics were similar to those reported in previous studies and, thus, were likely to reflect daily practice in neuro-oncology.

Age at diagnosis is a classic prognostic factor in glioblastoma series, although such a link was not observed in our series.9 An explanation may be the homogeneity of our purely surgical series, because all patients, including the elderly (aged >65 years), underwent debulking surgery, and 85.5% achieved total or subtotal resection (≥90%). In our study, the extent of surgery was recorded by postoperative imaging. A central review of immediate postoperative imaging was not performed. Despite these limitations, total/subtotal extent of resection (≥90%) was identified as the main prognostic factor for PFS and OS. Our findings are in agreement with previous studies, which demonstrated that patients who underwent complete tumor resection survived longer than those who underwent partial resection.9, 12, 16

Several retrospective studies have suggested that BCNU wafers within multimodal treatment strategies may increase survival with manageable adverse events,12, 17 whereas a recent study by Noel et al did not reveal any difference.16 Our results indicated improved OS for the overall study population (17.5 months) compared with the results reported by Stupp et al.9 Whether our results were better because of the use of BCNU wafers is speculative, since our study was observational and was not randomized. When considering the extent of resection, our results were comparable to those reported from another study by Stupp et al.18 Because no efficacy data are currently available in the setting of a randomized phase 3 trial compared with the Stupp strategy alone, the advantage of BCNU wafers as part of multimodal treatment strategies is undefined and, thus, is still debated.12, 16, 17

It has been demonstrated that MGMT methylation status is a potent prognostic factor for patients who receive either RT or TMZ concomitant with and adjuvant to RT.2 Whether this epigenetic feature also is a consistent biomarker for patients who receive BCNU wafers remains to be confirmed. In addition, it has been suggested that TMZ-induced DNA damage leads to the activation of MGMT transcription in at least some tumors with hypomethylated MGMT promoter status, with the up-regulation of MGMT expression possibly contributing to TMZ resistance.11 To our best knowledge, our study, which was focused on MGMT status, involved the largest multicenter series of glioblastoma patients to date who received treatment with the Gliadel BCNU wafers followed by the Stupp protocol. MSP proved to be a sensitive and robust method for assessing MGMT promoter methylation in almost all FFPE tumor samples (99%), as previously demonstrated.6 Optimal results with MSP were obtained with cryopreserved tumor specimens, although collecting and preserving frozen tumor specimens is costly and is not always available in routine practice. In contrast, tissue-preservation protocols that are used for histopathologic diagnosis, workup, and tissue archiving involved formalin-fixation and paraffin-embedding. In fact, pathologic laboratories are now considering more appropriate protocols for tumor tissue preservation, such as early fixation in buffered formalin, so that DNA does not deteriorate. In our study, MGMT promoter methylation status was analyzed by MSP using forward and reverse primers that were matched, respectively, with the +31 to +59 and +95 to +123 regions downstream from the transcription initiation site. These CpG hotspot regions, as reported recently, were correlated significantly with messenger RNA transcription and protein expression in terms of survival.4, 19 We observed that the rate of methylation was 57.3%, in line with previous reports (range, 44%-68%).2, 4, 6 Because our results identified MGMT promoter methylation status as an independent and favorable prognostic marker for OS, we speculate that BCNU implants do not induce TMZ resistance, because all of our patients underwent this local procedure.

An additional issue that was addressed in our study was whether IHC may be a reliable method for clinically evaluating MGMT expression. In our study, MGMT expression was readily quantified in 105 of 111 specimens (95.5%), and only 4.5% specimens were nonassessable because of the lack of positive internal controls. Because the 100% of same specimens were available and assessable for the IHC analysis of another nuclear marker, Ki67/Mib-1 (E3 ubiquitin-protein ligase), we suggest that the MGMT epitope may be highly sensitive to the fixation procedure, which may result in a loss of signal. This explanation also may account for specimens without MGMT promoter methylation status that had negative MGMT protein expression. Our results indicated that, together with MGMT promoter methylation, low protein expression (<15%) was associated independently with longer OS and PFS, suggesting the potential usefulness of this approach in the daily evaluation of clinical samples, particularly when fresh tumor tissue is unavailable. Although different cutoff values were used in previous studies, ranging from 5% to 80%,7 it is noteworthy that our value of 15% previously corresponded to the best cutoff value for identifying patients with significantly better survival rates in a series of 75 patients with primary glioblastoma.13 Obviously, our study highlights that, practically speaking, the relation between methylation status and protein expression is not unconditional. Our concordance rate of 69.52% was in the upper end of the previously reported range of 50% to 72%.7, 20 To our surprise, 21 tumors with high MGMT protein expression contained a methylated MGMT promoter. Thus, as suggested by others, we cannot exclude the possibility that such a discrepancy results from false-positive IHC data because of the difficulty in discriminating between MGMT tumor and nontumor nuclear staining (ie, endothelial and inflammatory cells) within glioblastomas. To avoid confusion, Sciuscio et al performed double immunostaining with nontumor cell markers.8 However, this procedure is costly and difficult to perform in a routine pathologic laboratory. Conversely, the codetection of methylated and unmethylated alleles in most of glioblastoma specimens revealed the existence of contaminating normal tissue.8 Another explanation for the inconsistent concordance rate in our study may be related to tumor heterogeneity, a particular feature of glioblastomas, which often are composed of various clonal subsets of cells within a single tumor.8 We also identified 11 tumors with low MGMT protein expression without MGMT promoter methylation. In this context, several authors have described other silencing gene mechanisms, including MGMT gene deletion and histone acetylation of the MGMT promoter.8 Finally, it should be noted that p53 gene status also may influence the level of MGMT transcription.21 To some extent, our results indicated that MSP and IHC are complementary assays that provide a more reliable prognostic factor. The combined use of both IHC and MSP could be used as a first rapid screening test to discriminate between samples with a clear MGMT status from gray-zone samples, which should be explored using more sophisticated and costly assays. We believe that, as proposed in other types of cancer, a composite test should be validated in an independent, prospective cohort. Clearly, there is still much debate around the subject, although we hope that studies currently in progress will help to decipher the epigenetic regulation of MGMT expression.

In conclusion, the current study revealed that both MGMT promoter methylation status and low protein expression were independent prognostic markers for longer OS in patients with newly diagnosed glioblastoma who received the Gliadel BCNU implants followed by concomitant radiochemotherapy and adjuvant chemotherapy. Therefore, our data suggest that BCNU implants did not induce TMZ resistance, because all patients underwent this local procedure. Despite the significant relation observed between IHC and MSP, we also obtained discordant results, notably concerning MGMT promoter methylated tumors with high MGMT protein expression as well as unmethylated tumors with low protein expression. Our study results confirm that MGMT status is a strong and independent prognostic factor for survival in patients with glioblastoma, suggesting the potential usefulness of this approach in the daily evaluation of formalin-fixation and paraffin-embedding samples, even when fresh tumor tissues are unavailable.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

We thank the Neuro-Oncology Club of the French Neurosurgery Society as well as the neurosurgeons, radiotherapists, oncologists, pathologists, and study coordinators who participated to the study. We give special thanks to S. Michalak (Angers); G. A. Czorny, J. Godard, H. Katranji, and G. Viennet (Besancon); T. Riem, A. Rougier, P. Monteil, I. Catry-Thomas, E. Gimbert, P. Monteil, D. Liguoro, H. Loiseau, and D. Pedenon-Peyrichout (Bordeaux); D. Hieu and I. Quintin-Roue (Brest); A. Borha, J. M. Derlon, R. Gadan, J. Geffrelot, S. Khouri, and V. Leon (Caen); N. Jovenin and N. Heil (Reims); D. Hoffmann, B. Pasquier, and C. Salon (Grenoble); L. Bauchet and V. Rigau (Montpellier); P. Paquis and F. Vandenbos (Nice); C. Miquel and P. Page, (Paris; St. Anne); and L. Capelle and K. Mokhtari (Paris, La Pitie Salpetriere).

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

This study was funded by grants from Caen University Hospital, Archimedes Pharma Laboratory, the Centre National de la Recherche Scientifique (CNRS), the French Ministry for Higher Education and Research, and the University of Caen Basse-Normandie.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES
  • 1
    Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG. Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res. 1999; 59: 793-797.
  • 2
    Hegi ME, Diserens AC, Gorlia T, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med. 2005; 352: 997-1003.
  • 3
    Brell M, Tortosa A, Verger E, et al. Prognostic significance of O6-methylguanine-DNA methyltransferase determined by promoter hypermethylation and immunohistochemical expression in anaplastic gliomas. Clin Cancer Res. 2005; 11: 5167-5174.
  • 4
    Everhard S, Tost J, El Abdalaoui H, et al. Identification of regions correlating MGMT promoter methylation and gene expression in glioblastomas. Neuro Oncol. 2009; 11: 348-356.
  • 5
    van den Bent MJ, Dubbink HJ, Sanson M, et al. MGMT promoter methylation is prognostic but not predictive for outcome to adjuvant PCV chemotherapy in anaplastic oligodendroglial tumors: a report from EORTC Brain Tumor Group Study 26951. J Clin Oncol. 2009; 27: 5881-5886.
  • 6
    Cankovic M, Mikkelsen T, Rosenblum ML, Zarbo RJ. A simplified laboratory validated assay for MGMT promoter hypermethylation analysis of glioma specimens from formalin-fixed paraffin-embedded tissue. Lab Invest. 2007; 87: 392-397.
  • 7
    Brell M, Ibanez J, Tortosa A. O6-Methylguanine-DNA methyltransferase protein expression by immunohistochemistry in brain and non-brain systemic tumours: systematic review and meta-analysis of correlation with methylation-specific polymerase chain reaction [serial online]. BMC Cancer. 2011; 11: 35.
  • 8
    Sciuscio D, Diserens AC, van Dommelen K, et al. Extent and patterns of MGMT promoter methylation in glioblastoma- and respective glioblastoma-derived spheres. Clin Cancer Res. 2011; 17: 255-266.
  • 9
    Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005; 352: 987-996.
  • 10
    Metellus P, Coulibaly B, Nanni I, et al. Prognostic impact of O6-methylguanine-DNA methyltransferase silencing in patients with recurrent glioblastoma multiforme who undergo surgery and carmustine wafer implantation: a prospective patient cohort. Cancer. 2009; 115: 4783-4794.
  • 11
    Kitange GJ, Carlson BL, Schroeder MA, et al. Induction of MGMT expression is associated with temozolomide resistance in glioblastoma xenografts. Neuro Oncol. 2009; 11: 281-291.
  • 12
    Menei P, Metellus P, Parot-Schinkel E, et al. Biodegradable carmustine wafers (Gliadel) alone or in combination with chemoradiotherapy: the French experience. Ann Surg Oncol. 2010; 17: 1740-1746.
  • 13
    Capper D, Mittelbronn M, Meyermann R, Schittenhelm J. Pitfalls in the assessment of MGMT expression and in its correlation with survival in diffuse astrocytomas: proposal of a feasible immunohistochemical approach. Acta Neuropathol. 2008; 115: 249-259.
  • 14
    Heng DY, Xie W, Regan MM, et al. Prognostic factors for overall survival in patients with metastatic renal cell carcinoma treated with vascular endothelial growth factor-targeted agents: results from a large, multicenter study. J Clin Oncol. 2009; 27: 5794-5799.
  • 15
    Karayan-Tapon L, Quillien V, Guilhot J, et al. Prognostic value of O6-methylguanine-DNA methyltransferase status in glioblastoma patients, assessed by 5 different methods. J Neurooncol. 2010; 97: 311-322.
  • 16
    Noel G, Schott R, Froelich S, et al. Retrospective comparison of chemoradiotherapy followed by adjuvant chemotherapy, with or without prior Gliadel implantation (carmustine) after initial surgery in patients with newly diagnosed high-grade gliomas [published online ahead of print February 5, 2011]. Int J Radiat Oncol Biol Phys. 2011.
  • 17
    McGirt MJ, Than KD, Weingart JD, et al. Gliadel (BCNU) wafer plus concomitant temozolomide therapy after primary resection of glioblastoma multiforme. J Neurosurg. 2009; 110: 583-588.
  • 18
    Stupp R, Hegi ME, Mason WP, et al. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol. 2009; 10: 459-466.
  • 19
    Shah N, Lin B, Sibenaller Z, et al. Comprehensive analysis of MGMT promoter methylation: correlation with MGMT expression and clinical response in GBM [serial online]. PLoS One. 2011; 6: e16146.
  • 20
    Preusser M, Charles Janzer R, Felsberg J, et al. Anti-O6-methylguanine-methyltransferase (MGMT) immunohistochemistry in glioblastoma multiforme: observer variability and lack of association with patient survival impede its use as clinical biomarker. Brain Pathol. 2008; 18: 520-532.
  • 21
    Shamsara J, Sharif S, Afsharnezhad S, et al. Association between MGMT promoter hypermethylation and p53 mutation in glioblastoma. Cancer Invest. 2009; 27: 825-829.