[Hypermethylation as a potential prognostic factor and a clue to a better understanding of the molecular pathogenesis of medulloblastoma--results of a genomewide methylation scan].

BACKGROUND

The molecular mechanisms controlling initiation and progression of medulloblastomas are largely unclear. Changes in DNA methylation of promoter regions have been shown to disturb the expression of growth regulatory genes.

PATIENTS AND METHODS

We evaluated DNA methylation patterns in 17 medulloblastomas, 5 stPNETs and 5 medulloblastoma cell lines using Restriction Landmark Genomic Scanning (RLGS), a method displaying up to 2.000 potential gene loci in a single gene. To test whether previously characterized tumor suppressor genes are affected by hypermethylation we performed MS-PCR for p15INK4B, p16INK4A, VHL, TP53 and E-cadherin.

RESULTS

The analysis of RLGS profiles from tumors revealed an abundance of hypermethylation in primary tumors and cell lines. Extrapolated to the human genome with its approximately 36,000 genes a total of 420 loci become hypermethylated in the tumor genomes. The previously characterized medulloblastoma breakpoint cluster in 17p11.2 appears to be a hotspot for aberrant methylation. Cox regression analysis of survival data identified seven CpG islands for which hypermethylation is suggestive of a poor prognosis. MS-PCR analysis of known genes demonstrated hypermethylation of p16INK4A in a limited number of tumors. The pattern of DNA hypermethylation was similar in medulloblastomas and stPNETs. However, some CpG islands were shown to be specific for a tumor type, while others were shared targets.

CONCLUSIONS

Hypermethylation is a common abnormality in primary medulloblastomas and supratentorial PNETs. Several hundreds of CpG islands are potential targets for methylation in medulloblastomas including the breakpoint cluster in 17p11.2. The methylation status of certain gene sequences appears to be associated with the clinical outcome. Promoter hypermethylation has an outstanding potential as a marker for the identification of novel tumor suppressors as well as diagnostic and therapeutic targets in medulloblastomas.