TGFBR2 mutation is correlated with CpG island methylator phenotype in microsatellite instability-high colorectal cancer.

TitleTGFBR2 mutation is correlated with CpG island methylator phenotype in microsatellite instability-high colorectal cancer.
Publication TypeJournal Article
Year of Publication2007
AuthorsOgino S, Kawasaki T, Ogawa A, Kirkner GJ, Loda M, Fuchs CS
JournalHum Pathol
Volume38
Issue4
Pagination614-20
Date Published2007 Apr
ISSN0046-8177
KeywordsCohort Studies, Colorectal Neoplasms, Core Binding Factor Alpha 3 Subunit, CpG Islands, DNA Methylation, Female, Humans, Male, Microsatellite Instability, Mutation, Phenotype, Polymerase Chain Reaction, Protein-Serine-Threonine Kinases, Proto-Oncogene Proteins, Proto-Oncogene Proteins B-raf, Proto-Oncogene Proteins p21(ras), ras Proteins, Receptor, Transforming Growth Factor-beta Type II, Receptors, Transforming Growth Factor beta
Abstract

The transforming growth factor-beta receptor type 2 gene (TGFBR2) is mutated in most microsatellite instability-high (MSI-H) colorectal cancers. Promoter methylation of RUNX3 (runt-related transcription factor 3; encoding a transcription factor downstream of the TGF-beta pathway) is observed in colorectal cancer with CpG island methylator phenotype (CIMP), which is characterized by extensive promoter methylation and is associated with MSI-H and BRAF mutations. However, no study to date has examined interrelationship between TGFBR2 mutation, RUNX3 methylation, and CIMP in colorectal cancer. Using 144 MSI-H colorectal cancers derived from 2 large prospective cohort studies, we analyzed a mononucleotide repeat of TGFBR2 and quantified DNA methylation (by MethyLight technology) in 8 CIMP-specific promoters (RUNX3, CACNA1G [calcium channel, voltage-dependent, T type alpha-1G subunit], CDKN2A [p16], CRABP1 [cellular retinoic acid binding protein 1], IGF2 [insulin-like growth factor 2], MLH1, NEUROG1 [neurogenin 1], and SOCS1 [suppressor of cytokine signaling 1]). Among the 144 MSI-H tumors, the presence of TGFBR2 mutation (overall 72% frequency) was correlated positively with CIMP-high (with >/=6/8 methylated promoters; P < .0001), RUNX3 methylation (P = .0004), BRAF mutation (P = .0006), and right colon (P = .05); inversely with KRAS mutation (P = .006); but not significantly with sex, tumor differentiation, and p53 status (assessed by immunohistochemistry). After stratification by sex, location, tumor differentiation, RUNX3 status, KRAS/BRAF status, or p53 status, CIMP-high was persistently correlated with TGFBR2 mutation. In contrast, RUNX3, KRAS, or BRAF status was no longer correlated with TGFBR2 mutation after stratification by CIMP status. In conclusion, TGFBR2 mutation is associated with CIMP-high and indirectly with RUNX3 methylation. Our findings emphasize the importance of analyzing global epigenomic status (for which CIMP status is a surrogate marker) when correlating a single epigenetic event (eg, RUNX3 methylation) with any other molecular or clinicopathologic variables.

DOI10.1016/j.humpath.2006.10.005
Alternate JournalHum Pathol
PubMed ID17270239
Grant ListP01 CA55075-13 / CA / NCI NIH HHS / United States
P01 CA87969-03 / CA / NCI NIH HHS / United States
Related Faculty: 
Massimo Loda, M.D.

Pathology & Laboratory Medicine 1300 York Avenue New York, NY 10065 Phone: (212) 746-6464
Surgical Pathology: (212) 746-2700