Artikel
Individual CpG methylation status of the DMR-2 island of the MGMT promotor determined by Sanger sequencing improves predictive stratification of glioblastoma patients receiving temozolomide
Der individuelle CpG-Methylierungsstatus im DMR-2 Abschnitt des MGMT-Promotors mithilfe der Sanger-Sequenzierung verbessert die prediktive Stratifizierung von Glioblastom-Patienten unter Chemotherapie mit Temozolomid
Suche in Medline nach
Autoren
Veröffentlicht: | 26. Juni 2020 |
---|
Gliederung
Text
Objective: Determination of MGMT promotor methylation usually relies on the evaluation of a limited number of CpG sites of the DMR-2 island. The predicitve role of the remaining usually not analyzed CpG sites remains unknown. We here analyzed the individual methylation patterns of the DMR-2 island and their respective predictive impact in a large and homogeneously treated cohort of glioblastoma multiforme (GBM) patients undergoing radiotherapy plus concomitant and adjuvant temozolomide. The study protocol was approved by the institutional ethical board.
Methods: The individual methylation pattern of CpG-sites 74-98 was assessed by Sanger sequencing technique (SSeq) in 222 consecutive patients with histologically proven GBM assigned for radio-chemotherapy. In parallel, conventional methylation-specific PCR (MSP) was used for MGMT promoter methylation classification. Prognostic models were obtained from multivariate proportional hazards estimations.
Results: According to MSP, patients with methylated (MSP-positive) and unmethylated (MSP-negative) tumors experienced a median overall survival (OS) of 21.3mos and 12.5mos, respectively (p<0.001). Individual methylation patterns as determined by SSeq were heterogeneous: Among MSP positve tumors, methylation level was highest in CpG sites 79-94 & 98 (median methylation rate: 82%), intermediate in CpG sites 74-78 (median methylation rate: 60%), and lowest in CpG sites 95-97 (median methylation rate: 33%) indicating a site-dependent methylation propagation. Correlation analyses additionally demonstrated a neighborhood-dependent methylation propagation: the methylation status of a given CpG site usually matched that of its neighborhood. Multivariate prognostic/predictive modelling indicates the cumulative number of methylated CpG sites to be linearly associated with OS (p<0.001). Stratification of MSP-positive tumors indicates that median OS ranged from as low as 17.5 months (<18 methylated CpG sites) to as high as 27 months (≥18 methylated CpG sites).
Conclusion: Site- and neighborhood-dependent methylation propagation contribute to heterogeneous methylation patterns. Extensive analyses of CpG sites in the DMR-2 island improves predictive stratification of GBM patients, which can be used for more precise treatment concepts. SSeq overcomes limitations of conventional techniques such as MSP and can be implemented in daily clinical practice.