Defining CIMP in epigenetics - a walk in the park?

Defining CIMP for correct classification of epigenetically altered tumors can be a challenging task, yet correct classification is very important for understanding tumor biology, identifying molecular classes that can explain differences in outcome or response to therapy, and for comparison between results from different groups.

We have recently added to the debate by comparing a set of commonly agreed panels to assess CIMP. The discrepancy between some of the markers and panels - and, depending on the number of genes and the number of probes in each gene included - was considerable and (maybe) even disturbing!

read the full text paper published in Plos One, if you would like to read beyond the abstract:

PLoS One. 2014 Jan 21;9(1):e86657. doi: 10.1371/journal.pone.0086657. eCollection 2014.

Comparison of CpG Island Methylator Phenotype (CIMP) Frequency in Colon Cancer Using Different Probe- and Gene-Specific Scoring Alternatives on Recommended Multi-Gene Panels.

Berg M1, Hagland HR2, Søreide K3.


Abstract

BACKGROUND:

In colorectal cancer a distinct subgroup of tumours demonstrate the CpG island methylator phenotype (CIMP). However, a consensus of how to score CIMP is not reached, and variation in definition may influence the reported CIMP prevalence in tumours. Thus, we sought to compare currently suggested definitions and cut-offs for methylation markers and how they influence CIMP classification in colon cancer.

METHODS:

Methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA), with subsequent fragment analysis, was used to investigate methylation of tumour samples. In total, 31 CpG sites, located in 8 different genes (RUNX3, MLH1, NEUROG1, CDKN2A, IGF2, CRABP1, SOCS1 and CACNA1G) were investigated in 64 distinct colon cancers and 2 colon cancer cell lines. The Ogino gene panel includes all 8 genes, in addition to the Weisenberger panel of which only 5 of the 8 genes included were investigated. In total, 18 alternative combinations of scoring of CIMP positivity on probe-, gene-, and panel-level were analysed and compared.

RESULTS:

For 47 samples (71%), the CIMP status was constant and independent of criteria used for scoring; 34 samples were constantly scored as CIMP negative, and 13 (20%) consistently scored as CIMP positive. Only four of 31 probes (13%) investigated showed no difference in the numbers of positive samples using the different cut-offs. Within the panels a trend was observed that increasing the gene-level stringency resulted in a larger difference in CIMP positive samples than increasing the probe-level stringency. A significant difference between positive samples using 'the most stringent' as compared to 'the least stringent' criteria (20% vs 46%, respectively; p<0.005) was demonstrated.

CONCLUSIONS:

A statistical significant variation in the frequency of CIMP depending on the cut-offs and genes included in a panel was found, with twice as many positives samples by least compared to most stringent definition used.