[HTML][HTML] Methylation-associated down-regulation of RASSF1A and up-regulation of RASSF1Cin pancreatic endocrine tumors

G Malpeli, E Amato, M Dandrea, C Fumagalli… - BMC cancer, 2011 - Springer
G Malpeli, E Amato, M Dandrea, C Fumagalli, V Debattisti, L Boninsegna, G Pelosi
BMC cancer, 2011Springer
Background RASSF1A gene silencing by DNA methylation has been suggested as a major
event in pancreatic endocrine tumor (PET) but RASSF1A expression has never been
studied. The RASSF1 locus contains two CpG islands (A and C) and generates seven
transcripts (RASSF1A-RASSF1G) by differential promoter usage and alternative splicing.
Methods We studied 20 primary PETs, their matched normal pancreas and three PET cell
lines for the (i) methylation status of the RASSF1 CpG islands using methylation-specific …
Background
RASSF1A gene silencing by DNA methylation has been suggested as a major event in pancreatic endocrine tumor (PET) but RASSF1A expression has never been studied. The RASSF1 locus contains two CpG islands (A and C) and generates seven transcripts (RASSF1A-RASSF1G) by differential promoter usage and alternative splicing.
Methods
We studied 20 primary PETs, their matched normal pancreas and three PET cell lines for the (i) methylation status of the RASSF1 CpG islands using methylation-specific PCR and pyrosequencing and (ii) expression of RASSF1 isoforms by quantitative RT-PCR in 13 cases. CpG island A methylation was evaluated by methylation-specific PCR (MSP) and by quantitative methylation-specific PCR (qMSP); pyrosequencing was applied to quantify the methylation of 51 CpGs also encompassing those explored by MSP and qMSP approaches.
Results
MSP detected methylation in 16/20 (80%) PETs and 13/20 (65%) normal pancreas. At qMSP, 11/20 PETs (55%) and 9/20 (45%) normals were methylated in at least 20% of RASSF1A alleles.
Pyrosequencing showed variable distribution and levels of methylation within and among samples, with PETs having average methylation higher than normals in 15/20 (75%) cases (P = 0.01). The evaluation of mRNA expression of RASSF1 variants showed that: i) RASSF1A was always expressed in PET and normal tissues, but it was, on average, expressed 6.8 times less in PET (P = 0.003); ii) RASSF1A methylation inversely correlated with its expression; iii) RASSF1 isoforms were rarely found, except for RASSF1B that was always expressed and RASSF1C whose expression was 11.4 times higher in PET than in normal tissue (P = 0.001). A correlation between RASSF1A expression and gene methylation was found in two of the three PET cell lines, which also showed a significant increase in RASSF1A expression upon demethylating treatment.
Conclusions
RASSF1A gene methylation in PET is higher than normal pancreas in no more than 75% of cases and as such it cannot be considered a marker for this neoplasm. RASSF1A is always expressed in PET and normal pancreas and its levels are inversely correlated with gene methylation. Isoform RASSF1C is overexpressed in PET and the recent demonstration of its involvement in the regulation of the Wnt pathway points to a potential pathogenetic role in tumor development.
Springer