High specificity of quantitative methylation-specific PCR analysis for MGMT promoter hypermethylation detection in gliomas

Abstract

Normal brain tissue from 28 individuals and 50 glioma samples were analyzed by real-time Quantitative Methylation-Specific PCR (QMSP). Data from this analysis were compared with results obtained on the same samples by MSP. QMSP analysis demonstrated a statistically significant difference in both methylation level (P = .000009 Mann Whitney Test) and frequencies (P = .0000007, Z-test) in tumour samples as compared with normal brain tissues. Although QMSP and MSP showed similar sensitivity, the specificity of QMSP analysis was significantly higher (93%; CI95%: 84%-100%) as compared with MSP (64%; 95%CI: 46%-82%). Our results suggest that QMSP analysis may represent a powerful tool to identify glioma patients that will benefit from alkylating agents chemotherapy.

Figure 1

Box plot for the MGMT/ACTB ratios determined by Quantitative Methylatiom Specific PCR in normal brain tissues and tumour samples. NBT: Normal brain tissues (n = 28); FFPE: Formalin fixed paraffin-embedded samples (n = 28); SnF: Snap-frozen specimens (n = 22); Tum: Total tumour samples (n = 50). The boxes mark the interquartile range, (interval between the 25th and 75th percentile). The lines inside the boxes denote median values. The whiskers represent the interval between the 10th and 90th percentiles. *indicates the extreme cases with more than three boxes length upper or down from the interquartile range.

Figure 2

ROC curve analysis of Quantitative Methylation-specific PCR results. (a) ROC curve for QMSP assay was designed on the basis of MGMT/ACTB ratios in determined in normal brain tissues (n = 28) and total tumour samples (n = 50). The area under the curve (AUC) is 0.74 (P = .000001). (b) Comparison of ROC curves were designed on the basis of MGMT/ACTB ratios determined in normal brain tissues (n = 28) and FFPE (n = 28) or snap-frozen specimens (n = 22). The area under the curve (AUC) is 0.70 (P = .001) for FFPE samples and 0.79 (P = .0000001) for snap-frozen specimens. (c) Scatter plot in logarithm scale of MGMT/ACTB ratios in normal brain tissue (NBT), tumour FFPE samples, and snap-frozen (SnF) specimens. *indicates samples with an MGMT/ACTB ratio of 0 which cannot be plotted on a logarithm scale (n = 21 for NBT, n = 13 for FFPE and n = 6 for SnF). The horizontal line indicates the optimal cutoff value of 0.35.

Figure 3

Comparison of QMSP analysis MSP . Representative results of Methylation-Specific PCR analysis and comparison with Quantitative Methylation-Specific PCR. As control of the bisulphite conversion the upper panels show PCR reactions specific for the ACTB gene promoter region not containing CpG; the middle panels show PCR reactions specific for the methylated sequence of the MGMT promoter; lower panels show results of QMSP. MW: Molecular weight marker. M, positive control (CpGenome Universal Methylated DNA); UM: Negative control (Universal Unmethylated DNA); N: Normal brain tissue; G: Glioma sample. Normal brain tissues N4, N7, N26 and snap-frozen tumour sample G43 show faint bands in the gel electrophoresis but MGMT/ACTB ratios below the cutoff value of 0.35 by QMSP, snap-frozen sample G45 show no methylated band by MSP but an MGMT/ACTB ratio above the cutoff value of 0.35 by QPCR.