Status of O 6 -methylguanine-DNA methyltransferase [MGMT] gene promoter methylation among patients with glioblastomas from India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.103190
Source of Support: None, Conflict of Interest: None
Background: O 6 -methylguanine DNA methyltransferase [MGMT] gene promoter methylation has emerged as a promising marker in determining resistance to temozolomide, used in the treatment of patients with glioblastomas. Aim: To determine the frequency of MGMT promoter methylation among patients with glioblastomas using methylation-specific polymerase chain reaction (MSP) and compare it to the results obtained by bisulfite sequencing of a subset of samples. Materials and Methods: DNA obtained from the frozen tissue of 27 samples of glioblastomas and three other gliomas, were analyzed for MGMT promoter methylation using a nested MSP assay. Sixteen samples were also subjected to bisulfite sequencing to determine the methylation status of 27 CpG sites within the sequenced region of the MGMT promoter. Data with respect to radiation, chemotherapy and survival outcome was also collected. Results: MGMT promoter methylation was seen in 67% of the cases included in the study using frozen tissues by MSP analysis, while 62% were methylated among glioblastomas alone. There was a 100% concordance between the results obtained by MSP analysis and bisulfite sequencing. Clinical outcome was known among 67% of cases and methylation was higher among those patients who had no recurrence, though it was not statistically significant [P=0.44]. Conclusion: The frequency of methylation seen in this study concurs with that reported earlier from the country. MSP was easy to perform and interpret. However, the utility of this testing system in a routine diagnostic setting is still being debated.
Keywords: Bisulfite sequencing, glioblastoma multiforme, O 6 -methylguanine DNA methyltransferase, methylation specific PCR
Glioblastomas are among the commonest primary brain tumors in adults, where the prognosis has remained extremely poor despite multimodal treatment including surgery, radiotherapy and chemotherapy. , However, the introduction of an oral alkylating agent, temozolomide [TMZ], has helped to prolong median survival among glioblastoma patients by a few months. ,, The current standard of care has now evolved to include surgery, radiotherapy with concomitant TMZ chemotherapy followed by adjuvant TMZ chemotherapy. 
TMZ damages DNA by introducing multiple alkyl groups and the mainstay of resistance to the drug is through the DNA-repair protein O 6 -alkylating DNA alkyltransferase [AGT], encoded by the gene O 6 -Methylguanine-DNA-Methyltransferase [MGMT].  The MGMT gene activity is in turn controlled by a promoter, the methylation of which, has been shown to be associated with increased sensitivity to TMZ and with better survival among patients treated with radiotherapy and concomitant TMZ chemotherapy. ,,, Consequently, MGMT promoter methylation has become an important predictive marker of TMZ resistance. However, clear consensus has been found to be lacking with respect to the optimal method for detection of methylation and with regard to the region of the promoter most predictive of response to therapy. Promoter methylation frequency reported from various parts of the world for glioblastomas appears to be heterogeneous ranging from 16.8% to as high as 70%. ,, Studies from India have reported promoter methylation to range between 50-60% in the pediatric and adult populations. , However, all reports from India have been from a single center. We therefore, attempted to determine the frequency of methylation among patients with glioblastomas from our center using methylation specific-PCR [MSP] and correlated this information with clinical outcome when follow-up was available. Further, a subset of samples were also bisulfite-sequenced to determine the methylation status of 27 CpG sites within the MGMT promoter region and the results compared to that obtained by MSP.
Tumor and patient data
Twenty-seven cases of primary glioblastomas and one case each of anaplastic oligodendroglioma, anaplastic astrocytoma and oligoastrocytoma that were histopathologically confirmed  were included in the study after obtaining patient consent. These cases were diagnosed during the period 2009-2010. These tumors were collected directly from the operation theatre during surgery, snap frozen in liquid nitrogen and stored at -80°C. Formalin-fixed paraffin-embedded tissues [FFPE] available for 24 cases, were also included in the study. Data with respect to radiation therapy and chemotherapy were retrospectively obtained from hospital records. The treatment protocol for glioblastoma was as follows: Patients received concurrent TMZ at a dose of 75 mg per square meter of body surface area daily during standard fractionated radiotherapy (RT) for seven days a week for six weeks.  Following RT and concurrent TMZ, TMZ was given once every 28 days at a dose of 150 mg per square meter of body surface area for the first cycle, followed by 200 mg per square meter of body surface area from the second to sixth cycle. Patients treated prior to the use of TMZ in the institution were given lomustine along with radiotherapy.
Survival outcome was determined either when the patient presented for follow-up or was telephonically ascertained. Patients were considered to have a tumor recurrence only if (a) gross total resection of the initial tumor was documented in the postoperative images or (b) if there was an increase in size of the tumor in those with a partial excision or biopsy.
DNA extraction and bisulfite conversion
DNA was extracted from 25 mg of fresh tissue or from four 10μm paraffin sections using the QIAamp DNA mini kit [QIAGEN, Gmbh, Germany]. DNA was quantitated using the Nanodrop [NanoDrop technologies, USA]. One microgram of DNA was used for bisulfite conversion [EpiTect® bisulfite conversion kit, QIAGEN, Gmbh, Germany]. Briefly, DNA was added to a bisulfite mix and the bisulfite conversion was carried out in a thermal cycler as per manufacturer's instructions. The product was cleaned in spin columns using wash buffers and finally desulfonated with the desulfonation buffer before elution.
Methylation-specific PCR [MSP] was performed as a two-step nested PCR as described previously  using the primers of Esteller et al.  One μl of bisulfite-converted DNA was amplified using 20 picomoles of primers (Sigma Aldrich, India) 1.25 m MgCl 2 (Applied Biosystems, USA) and 1.5 units of recombinant taq polymerase (Amplitaq gold, Applied Biosystems, USA). The thermal cycling conditions used were as follows: 95°C for 10 min, and 40 cycles of 95°C for 45 sec, 52°C for 50 sec, 72°C for 60 sec with a final extension of 72°C for 10 min. Each step included methylated (LN 229, ATCC CRL-2611) and unmethylated (LN 18, ATCC CRL-2610) controls along with non-template control. One μl of the amplified first-round product was used for second round of amplification. The following thermal cycling conditions were followed: 95°C for 10 min, and 20 cycles of 95°C for 45 sec, 65°C for 25 sec, 72°C for 30 sec with a final extension of 72°C for 10 min. The amplified products were run on a 2% agarose gel with an expected size of 81 bp for methylated product and 93 bp for an unmethylated product.
Sodium bisulfite sequencing
Primers described by Mikaska et al.,  targeting a 266 bp region of the MGMT gene were used for sequencing 16 of the 30 samples included. The thermal cycling conditions used were as follows: 95°C for 10 min, and 40 cycles of 95°C for 45 sec, 52°C for 50 sec, 72°C for 60 sec with a final extension of 72°C for 10 min. The PCR products were purified and sequenced using the BigDye Terminator v3.1 cycle sequencing kit [Applied Biosystems, Foster city, CA, USA]. The criteria described by Grasbon-Frodl  were used for interpretation of sequences.
Descriptive statistics were used to summarize all study variables. Continuous variables were described using means and standard deviations, if normally distributed. Medians with ranges were used to describe non-normally distributed variables. Categorical variables were expressed as frequencies with percentages. Association between methylation status and other categorical variables (e.g. recurrence status) were assessed using Chi-square tests. All analyses were done using STATA 10.0 (StataCorp, College Station, Texas, US).
MGMT promoter methylation was determined for 27 primary glioblastoma samples along with three samples of other glioma subtype by MSP. These samples were from patients whose age ranged from 7-67 years [mean 42.5± 13.6] with a male to female ratio of 1:1. All glioblastomas were classified as World Health Organization (WHO) Grade IV tumors and were histologically typical. Twenty-six per cent [n=7] of glioblastomas in the study also had an oligodendroglial component (GBMOs).
Overall, MGMT promoter methylation was seen in 67% [n=20] of the cases included using frozen tissue, while 62% [n=17] were methylated among glioblastomas alone [Figure 1]. [Table 1] provides details of the tumor type along with the methylation status. Further, when the methylation status was compared between glioblastomas and GBMOs, 60% (12/20) of Grade IV glioblastomas were methylated while the percentage of methylation was slightly higher (71%) among the GBMOs (5/7). FFPE sections were available in 24 of the 30 cases and 79% [n=19] of the samples yielded amplifiable DNA. A comparison between the methylation pattern obtained from frozen tissues and FFPE showed good concordance [93%] except in two cases where the methylated samples appeared unmethylated using FFPE sections.
Sixteen of the 30 cases included in the study were sequenced after bisulfite conversion [Figure 2] and there was a 100% concordance between the results obtained by MSP analysis and bisulfite sequencing. Samples with 50% or more 'complete methylation' on sequencing were considered 'methylated'. Four samples showed 'weak methylation' by sequencing and were considered 'unmethylated' as suggested by the Grasbon-Frodl criteria.  MSP also classified these samples 'unmethylated'. Sequencing also accurately classified both methylated cases that appeared unmethylated using FFPE sections. The range of methylation varied from 83% in methylated cases to 5% in unmethylated cases.
Clinical outcome was known among 67% of cases [n=20] with the median follow-up period of 11.5 months [range 6-24 months]. The 20 cases included ten dead, eight with no recurrence and two cases that had recurrence. While the number of cases with and without methylation was fairly even among those dead and among those with recurrence, the number of cases of methylation was higher among those with no recurrence [n=6 of 8], though this did not reach statistical significance [P=0.44]. A similar pattern of higher methylation [n=5] was also seen among those cases that were histologically grouped as GBMOs, though the small numbers among these cases did not make this a statistically significant finding. Interestingly, none of these five cases of GBMOs with methylation presented with a recurrence. However, there was no follow-up available for the two unmethylated cases with the oligodendroglial component. All cases with no recurrence of the tumor, with the exception of one case, received radiotherapy with concomitant TMZ followed by adjuvant TMZ.
MGMT is a DNA repair protein, encoded by the MGMT gene, that has the ability to remove cytotoxic adducts from O 6 -guanine in the DNA.  Epigenetic silencing of the MGMT gene promoter has been shown to be associated with a loss of MGMT protein and consequently, diminished DNA-repair activity in the cell. , Loss of MGMT protein is now known to provide a therapeutic advantage to glioblastoma patients. Those with methylated tumors receiving both radiotherapy and TMZ chemotherapy [concomitant and adjuvant] are known to have longer overall survival.  Hence determining the methylation pattern among patients with glioblastomas on a routine basis has become essential since it could help both as a predictive and prognostic marker. 
The frequency of methylation [62%] observed among glioblatomas in this study was slightly higher than the 56.8% reported previously among glioblastomas in India. , However, such minor differences in frequency of methylation might not be significant considering the heterogeneity of MGMT promoter methylation data worldwide , which has ranged from as little as 17% of the cases characterized to as high as 70%.  This heterogeneity has [been an interesting point for debate, especially since it has been noted that promoter methylation might not be related to the ethnicity of the population under study. 
While the differing frequencies of methylation in different studies has gained acceptance, there is no consensus on a testing system that can be reliably used across laboratories worldwide. ,, Assays for methylation ranging from qualitative, semi-quantitative and quantitative testing, have been reviewed in detail elsewhere.  Despite the availability of several testing systems, MSP testing and analysis based on immunohistochemistry [IHC] are used frequently in laboratories, though both are known to have limitations. A majority of studies that have investigated the utility of MGMT IHC have shown that IHC does not correlate with patient survival and might not be the appropriate test to be used as a predictive or prognostic marker in glioblastomas. ,, Similarly, the MSP method used in this study also has limitations since it detects methylation only in a limited number (~5) of CpG sites and provides only a qualitative result. Despite these limitations the MSP system was used in the study since it is easy to use and also considering its financial feasibility in our resource-limited setting.
We obtained good-quality DNA from all frozen tissues, while only 79% of FFPE tissues provided amplifiable DNA. Hegi et al., in a multi-centric study reported that working with paraffin sections provided variable results with a median success rate of 75%.  Further, investigators have also suggested that results from frozen tissues may be more reliable than from paraffin sections. , In our dataset two FFPE samples were found to be unmethylated though the sections were manually macrodissected ensuring that the area chosen had adequate tumor cellularity with none to very little of non-neoplastic tissue to confound the results. The corresponding frozen tissues of the same two cases consistently provided a methylated result. We considered the issue resolved after the results of bisulfite sequencing concurred with that of frozen tissue. It would be pertinent to point out that small amounts of contaminating non-neoplastic tissues could become a source of contamination and could provide false negative results. ,, These issues emphasize the need to carefully choose samples for detection of methylation status, therefore avoiding potential confounders. Studies have described regions within the tumor being variably methylated, even within the individual clones of the same tissue sample and that the tumor could be a mosaic of variable degrees of methylation. ,, Though speculative, mosaicism could have contributed to our findings here.
A subset of samples [n=16] in the study was bisulfite-sequenced and the results compared to MSP. Though bisulfite sequencing is tedious and more time-consuming, it was included since sequencing is currently considered the standard for determination of methylation patterns ,,,, and is more quantitative than MSP.  There was complete concordance between the two testing systems and sequencing also helped to resolve the discordant results of MSP between frozen and FFPE tissues. There were also four samples that showed 'weak methylation' by sequencing that were classified as 'unmethylated' and were also 'unmethylated' by MSP. Though this does not amount to discordance between the two systems, it highlights the fact that bisulfite sequencing might detect methylated CpGs that MSP might not, since it covers a larger number of CpG sites (n=27) in the MGMT promoter region than investigated by MSP (n=5).  However, it is also important to note that these 27 CpG characterized by sequencing, account for only a fraction of the 97 CpG sites present in the MGMT promoter  and regions upstream and downstream of this targeted region might also be predictive and prognostic. While a consensus does not exist on the exact region or the best testing system for determining the methylation status, several newer testing systems are being investigated that might help to overcome the limitations of these above-mentioned testing systems. 
The survival outcome was known in 67% of cases with a median follow-up of 11.5 months. Methylated and unmethylated tumor numbers were fairly similar among the 10 patients who died.Treatment details were available for only four; all four had received RT and concomitant and adjuvant TMZ. In contrast the pattern of methylation was different among the cases with no recurrence, where 75% [6/8] samples were methylated and five of these methylated cases received RT and concomitant and adjuvant TMZ. While this group could have had a better outcome both due to promoter methylation and because of the appropriate chemotherapy, it is hard to ascertain this fact due to the small number of cases in the group and limited follow-up available. However, it is now fairly well established that tumor methylation correlates with progression-free survival [PFS] ,,,, and even the exact region of promoter methylation correlating with improved PFS has been identified. 
Another interesting feature of the five methylated cases with no recurrence of the tumor is that they were also histologically distinct glioblastomas which had an oligodendroglial component [GBMO]. Though there are a few reports on the prognosis of GBMO vs glioblastomas, the reports have not provided uniform results. While some investigators ,, have found GBMOs to be associated with longer overall survival [OS], others have found no difference.  A very recent report that included extensive molecular characterization of both glioblastomas and GBMOs, however, shows that GBMOs are associated with better median survival.  Whether molecular mechanisms that result in this histological phenotype have a synergy with the mechanisms that contribute to methylation of the promoter thereby leading to a better outcome, is perhaps a feature worth investigating in a larger set of GBMOs in future.
This study highlights the methylation pattern of glioblastoma in India comparing the frequency of methylation, to the few reports from the country. The study also determines the validity of results obtained by MSP by comparing it to a subset that were characterized by bisulfite sequencing. The study however has some limitations both in terms of the number of samples analyzed and in terms of the follow-up available. However, taken with these limitations the study still provides useful information on promoter methylation pattern among glioblastomas from the southern part of India, along with emphasizing the utility of MSP in determining MGMT promoter methylation.
We would like to acknowledge the help of Ms. Janet Paul, in storing and cataloguing all fresh tissue samples. We would also like to thank Dr. Monica Hegi for providing controls and the help lent during standardization of the assay.
[Figure 1], [Figure 2]