Gastroenterology
Volume 129, Issue 3 , Pages 1121-1124, September 2005

CIMP, at Last

  • Jean-Pierre J. Issa

      Affiliations

    • University of Texas, M. D. Anderson Cancer Center, Houston, Texas
    • Corresponding Author InformationAddress requests for reprints to: Jean-Pierre Issa, MD, Department of Leukemia, M. D. Anderson Cancer Center, Unit 428, 1515 Holcombe, Houston, Texas 77030; fax: (713) 794-4297.
    • ,
  • Lanlan Shen

      Affiliations

    • Sapporo Medical University, Sapporo, Japan
    • ,
  • Minoru Toyota

      Affiliations

    • Sapporo Medical University, Sapporo, Japan

Article Outline

 

In cancer biology, the 1980s and 1990s were the decades of the genetic revolution. A series of landmark observations connected structural DNA changes with the development of both hereditary and sporadic cancers, capping years of speculation on the molecular basis of malignant transformation.1 The late 1990s also saw the blossoming of a different concept–an epigenetic etiology for cancer which, at first, appeared as an anathema to the genetic basis of cancer.2 This hypothesis proposed that some of the neoplastic pathology was related to permanent changes in gene expression (“epigenetic”), which were not due to structural damage in DNA (mutations), but to changes in chromatin structure associated with decorations of and around DNA, such as DNA methylation and histone posttranslational modifications.3 A spectacular confirmation of the concept came when it was demonstrated that several tumor-suppressor genes were altered in cancer by aberrant promoter methylation rather than DNA structural changes,4, 5 and a mutually exclusive relationship between mutations and methylation strongly suggested that these two distinct molecular processes provided equivalent selective advantages to affected cells.6 The epigenetic basis of cancer was initially met with much resistance in the genetics field, and there continue to be dissenting opinions arguing that epigenetic changes are epiphenomena that simply accompany (rather than cause) cancer development.7

One of the strongest arguments in favor of a genetic etiology of cancer comes from the discovery of multiple genetic instability syndromes in neoplastic cells,8 including the now-famous microsatellite instability (MSI) phenomenon in colorectal cancer (and others), which is caused by mismatch repair deficiency.9 In MSI cancers, mutations of specific sequences accumulate at much higher rates than normal, providing a sound mechanism for genetic diversity leading to cancer via Darwinian selective processes. A similar phenomenon involving epigenetic changes was described in 1999, whereby some colorectal cancers were found to accumulate high rates of aberrant promoter methylation events, including several tumor-suppressor genes.10 This phenomenon, termed descriptively CpG island methylator phenotype (CIMP) appeared to provide a similar level of evidence for a mechanistic contribution of methylation events to phenotypic diversity, leading to epigenetic-based Darwinian selective pressures. In a rather dramatic twist, CIMP itself was implicated in genetic instability via the methylation-based silencing of the mismatch repair gene MLH1.10 Not surprisingly, this concept proved quite controversial.11 In the current issue of Gastroenterology, an article by Samowitz et al12 reports on a relatively large, population-based study of DNA methylation in colorectal cancer that confirms most of the previously published aspects of CIMP.

The controversy over CIMP will almost certainly one day be considered a tempest in a teapot. Nevertheless, it is instructive to review its history. After the first report, CIMP was rapidly demonstrated in multiple other malignancies,13, 14, 15, 16, 17, 18, 19 and several groups confirmed the original findings using similar markers and technology.20, 21 Other groups, however, did not.22, 23 Recently, a publication concluded that all methylation events in colorectal cancer were related to aging rather than neoplasia,24 and directly reignited the issue of methylation as an epiphenomenon rather than a cause of cancer. Some investigators took the issue quite at heart; a paper that reported on CIMP in brain tumors25 received the following comments by one of the reviewers: “The CpG island methylator phenotype (CIMP) does not exist, as it has been widely demonstrated by many groups. Thus, delete all the speculations and references about this subject” (as quoted in the author’s response to reviewers that is available online, an illuminating demonstration of the usefulness of an open review system). This vehemence was surprising given that, in parallel, a body of data from several groups conclusively linked CIMP to distinct clinical and epidemiological features including older age, female gender, proximal location in the colon, high rates of BRAF and KRAS mutations, low rates of p53 mutations, specific precursor lesions (the serrated pathway), specific histology (mucinous poorly differentiated), familial occurrence, and a distinct clinical course.11, 20, 26, 27, 28 Was it possible that so many researchers were reporting on consistent artifacts? The article by Samowitz et al12 convincingly demonstrates that many of these previously reported features can in fact be linked to aberrant methylation of multiple genes, almost exactly as initially reported.

Why the discrepant results? The first CIMP demonstration made it clear that the majority (70%–80%) of aberrant DNA methylation events in colorectal cancer were age-related.10 The phenotype only became clear when those were filtered out. Subsequent studies also showed that CIMP affects methylation frequency and density,15 thus necessitating quantitative methylation measurements for optimal detection. It was clearly argued that an overly sensitive technique would overestimate methylation and blur the distinctions between the groups.29 A common feature of all the reports that could not confirm the presence of CIMP was the use of unselected genes and/or a nonquantitative, very sensitive methylation detection method. Remarkably, all the papers that used the originally described genes and methods confirmed CIMP. It is difficult to explain how scientists could unequivocally reject the concept without trying to replicate it using the same initial methodology. Why the heated debate? What are the stakes in the case of CIMP? From a cancer biology standpoint, they are not trivial. As discussed previously, methylation instability in some cancers but not others invalidates the argument that methylation is an epiphenomenon of cancer development, and argues strongly that methylation contributed to cancer pathophysiology by creating epigenetic diversity. More fundamentally, the data suggest that aberrant methylation (or at least CIMP) has a cause that could eventually be identified. On the minus side, this concept also predicts that methylation events (alone) may not provide the ideal universal cancer marker they were once thought to be because CIMP target genes will not be useful to screen for all colorectal cancers (many false-negatives are predicted), and non-CIMP target genes will likely yield a high rate of false-positives because they are also methylated in normal appearing mucosa of older individuals without tumors.

As the debate winds down toward an acceptance of the concept, it is useful to review the current state of knowledge on CIMP, its characteristics, definition, and causes. In the article by Samowitz et al, CIMP was most clearly associated with the MLH1 methylated tumors that also exhibited microsatellite instability. These cancers have well-recognized distinct epidemiology, histology, genetics, and clinical course and the current data confirm most of these observations. However, the strong association of these characteristics with methylation of multiple genes20, 30 (beyond MLH1) raises the possibility that some of the distinctive features of these cases are related to CIMP rather than MSI. Indeed, in comparing inherited MSI cancers (where MSI is typically caused by germline mutations in MSH2 or MLH1) to sporadic MSI cancers (where MSI is typically caused by CIMP), shared features including specific gene mutations, infiltration by intra-epithelial lymphocytes, and a lower propensity to metastasize and result in patient death are all likely related to MSI. On the other hand, distinct characteristics, such as a high rate of BRAF mutations in the sporadic MSI cancers, are most likely attributable to CIMP effects.31

In addition to CIMP/MSI cancers, Samowitz et al12 note a distinct but less well-defined group of cancers that are CIMP but lack MSI. These cancers also tend to be proximally located, occur in older patients, and have poorly differentiated/mucinous histology, all features that are shared with CIMP/MSI cases, suggesting that they relate to methylation rather than mutations. Interestingly, this subgroup of CIMP has less intense methylation and is characterized by mutations in KRAS rather than BRAF. It is intriguing to consider then that there are at least 2 distinct CIMP subsets, CIMP1–MSI associated, CIMP2- non-MSI associated, which may also have distinct causes. A pressing issue in this regard is to arrive at validated molecular markers for CIMP. This is a complex issue because of the plethora of methylation methods in use, some of which sometimes yield conflicting results, and because research into “optimal” CIMP markers had to await resolution of the conflict over its very existence.11 The article by Samowitz et al12 uses a practical definition of CIMP that is adequate for confirmation and current studies. However, it is likely that a better definition will be arrived at using studies of a larger group of genes in well-defined patient populations.

The most pressing question, the real issue where a debate should be ongoing, is the cause of CIMP. Six years after its initial description, CIMP causes remain mysterious. There are 2 dominant possibilities–genetic and environmental.11 The genetic cause postulates that CIMP results from a specific molecular event activating methylases or inactivating methylation-protection factors. There is no direct evidence for this as yet, and no convincing mutations in DNA methyltransferases have been described to date. Indeed, most studies do not find associations between methylation in cancer and DNA methyltransferase expression. However, one recent study showed a higher rate of positive family history in CIMP-positive cases,32 and another group described families where many cancers had mutations in BRAF and methylation of MINT31, two reliable CIMP markers in the colon.33 Thus, it remains likely that genetic lesions could underlie CIMP, and it is hoped that family studies will help identify the culprit(s).

There are equally tantalizing data for an environmental exposure related CIMP etiology. In the colon, chronic inflammation such as is found in ulcerative colitis has been associated with higher degrees of DNA methylation in both normal34 and neoplastic tissues,35 and some of the genes affected are known CIMP targets. In the liver, CIMP has been associated with chronic inflammation of either viral or alcoholic etiology.12 In gastric cancers, there is a remarkable association between CIMP and the presence of the Epstein–Barr virus genome in the tumors.36 The link between methylation and inflammation states is also present in tumors where CIMP has not yet formally been studied, including Barrett’s esophagus22 and smoking-related lung cancers,37 2 situations where hypermethylation (presumably reflecting at least in part the presence of CIMP) is very common. This conserved pattern of association across multiple organs and tumor types (which remains to be mechanistically explained) raises the possibility that CIMP is a molecular signature and one of the mediators of inflammation-related neoplasia. How can one reconcile these multiple observations on an environmental etiology of CIMP with the genetic data discussed above? A possible hypothesis is that the gene(s) involved in genetic predisposition to CIMP are not directly linked to the methylation machinery, but to a propensity to chronic inflammation (or to an exaggerated response to inflammation).

With the publication of the paper by Samowitz et al,12 it is hoped that the issue of the existence of CIMP will be finally put to rest. Research should now focus on the fascinating clinical characteristics of this phenotype, on arriving at a definition (genes and methods of analysis) that could be used by multiple laboratories, on determining whether any of the mouse models commonly used to study colon tumors is relevant to human patients with CIMP and, importantly, on deciphering the causes of this phenotype. The advent of therapies based on targeting epigenetic pathways38 also raises intriguing questions regarding the treatment of patients with CIMP-positive colorectal cancers.

Back to Article Outline

References 

  1. Vogelstein B , Kinzler KW . Cancer genes and the pathways they control . Nat Med . 2004;10:789–799
  2. Baylin SB , Herman JG , Graff JR , Vertino PM , Issa JPJ . Alterations in DNA methylation–A fundamental aspect of neoplasia . Adv Cancer Res . 1998;72:141–196
  3. Jones PA , Baylin SB . The fundamental role of epigenetic events in cancer . Nat Rev Genet . 2002;3:415–428
  4. Sakai T , Toguchida J , Ohtani N , Yandell DW , Rapaport JM , Dryja TP . Allele-specific hypermethylation of the retinoblastoma tumor-suppressor gene . Am J Hum Genet . 1991;48:880–888
  5. Herman JG, Latif F, Weng Y, Lerman MI, Zbar B, Liu S, et al. Silencing of the VHL tumor-suppressor gene by DNA methylation in renal carcinoma . Proc Natl Acad Sci U S A . 1994;91:9700–9704
  6. Herman JG , Baylin SB . Gene silencing in cancer in association with promoter hypermethylation . N Engl J Med . 2003;349:2042–2054
  7. Bestor TH . Unanswered questions about the role of promoter methylation in carcinogenesis . Ann N Y Acad Sci . 2003;983:22–27
  8. Loeb LA . A mutator phenotype in cancer . Cancer Res . 2001;61:3230–3239
  9. Kinzler KW , Vogelstein B . Lessons form hereditary colorectal cancer . Cell . 1996;87:159–170
  10. Toyota M , Ahuja N , Ohe-Toyota M , Herman JG , Baylin SB , Issa JPJ . CpG island methylator phenotype in colorectal cancer . Proc Natl Acad Sci U S A . 1999;96:8681–8686
  11. Issa JP . CpG island methylator phenotype in cancer . Nat Rev Cancer . 2004;4:988–993
  12. Samowitz WS, Albertsen H, Herrick J, Levin TR, Sweeney C, Murtaugh MA, et al. Evaluation of a large, population-based sample supports a CpG island methylator phenotype in colon cancer . Gastroenterology . 2005;129:837–845
  13. Shen L, Ahuja N, Shen Y, Habib NA, Toyota M, Rashid A, et al. DNA methylation and environmental exposures in human hepatocellular carcinoma . J Natl Cancer Inst . 2002;94:755–761
  14. Garcia-Manero G, Daniel J, Smith TL, Kornblau SM, Lee MS, Kantarjian HM, et al. DNA methylation of multiple promoter-associated CpG islands in adult acute lymphocytic leukemia . Clin Cancer Res . 2002;8:2217–2224
  15. Toyota M , Kopecky KJ , Toyota MO , Jair KW , Willman CL , Issa JP . Methylation profiling in acute myeloid leukemia . Blood . 2001;97:2823–2829
  16. Ueki T, Toyota M, Sohn T, Yeo CJ, Issa JP, Hruban RH, et al. Hypermethylation of multiple genes in pancreatic adenocarcinoma . Cancer Res . 2000;60:1835–1839
  17. Toyota M, Ahuja N, Suzuki H, Itoh F, Ohe-Toyota M, Imai K, et al. Aberrant methylation in gastric cancer associated with the CpG island methylator phenotype . Cancer Res . 1999;59:5438–5442
  18. Strathdee G, Appleton K, Illand M, Millan DW, Sargent J, Paul J, et al. Primary ovarian carcinomas display multiple methylator phenotypes involving known tumor suppressor genes . Am J Pathol . 2001;158:1121–1127
  19. Abe M, Ohira M, Kaneda A, Yagi Y, Yamamoto S, Kitano Y, et al. CpG island methylator phenotype is a strong determinant of poor prognosis in neuroblastomas . Cancer Res . 2005;65:828–834
  20. Whitehall VL, Wynter CV, Walsh MD, Simms LA, Purdie D, Pandeya N, et al. Morphological and molecular heterogeneity within nonmicrosatellite instability-high colorectal cancer . Cancer Res . 2002;62:6011–6014
  21. van Rijnsoever M , Grieu F , Elsaleh H , Joseph D , Iacopetta B . Characterisation of colorectal cancers showing hypermethylation at multiple CpG islands . Gut . 2002;51:797–802
  22. Eads CA, Lord RV, Wickramasinghe K, Long TI, Kurumboor SK, Bernstein L, et al. Epigenetic patterns in the progression of esophageal adenocarcinoma . Cancer Res . 2001;61:3410–3418
  23. Esteller M, Tortola S, Toyota M, Capella G, Peinado MA, Baylin SB, et al. Hypermethylation-associated inactivation of p14(ARF) is independent of p16(INK4a) methylation and p53 mutational status . Cancer Res . 2000;60:129–133
  24. Yamashita K , Dai T , Dai Y , Yamamoto F , Perucho M . Genetics supersedes epigenetics in colon cancer phenotype . Cancer Cell . 2003;4:121–131
  25. Yu J, Zhang H, Gu J, Lin S, Li J, Lu W, et al. Methylation profiles of thirty four promoter-CpG islands and concordant methylation behaviours of sixteen genes that may contribute to carcinogenesis of astrocytoma . BMC Cancer . 2004;4:65
  26. Toyota M , Ohe-Toyota M , Ahuja N , Issa JP . Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype . Proc Natl Acad Sci U S A . 2000;97:710–715
  27. Jass JR . Pathogenesis of colorectal cancer . Surg Clin North Am . 2002;82:891–904
  28. van Rijnsoever M , Elsaleh H , Joseph D , McCaul K , Iacopetta B . CpG island methylator phenotype is an independent predictor of survival benefit from 5-fluorouracil in stage III colorectal cancer . Clin Cancer Res . 2003;9:2898–2903
  29. Shen L , Kondo Y , Hamilton SR , Rashid A , Issa JP . p14 methylation in human colon cancer is associated with microsatellite instability and wild-type p53 . Gastroenterology . 2003;124:626–633
  30. Chirieac LR , Shen L , Catalano PJ , Issa JP , Hamilton SR . Phenotype of microsatellite-stable colorectal carcinomas with CpG island methylation . Am J Surg Pathol . 2005;29:429–436
  31. Jass JR . HNPCC and sporadic MSI-H colorectal cancer (a review of the morphological similarities and differences) . Fam Cancer . 2004;3:93–100
  32. Frazier ML, Xi L, Zong J, Viscofsky N, Rashid A, Wu EF, et al. Association of the CpG island methylator phenotype with family history of cancer in patients with colorectal cancer . Cancer Res . 2003;63:4805–4808
  33. Young J, Barker MA, Simms LA, Walsh MD, Biden KG, Buchanan D, et al. Evidence for BRAF mutation and variable levels of microsatellite instability in a syndrome of familial colorectal cancer . Clin Gastroenterol Hepatol . 2005;3:254–263
  34. Issa JP , Ahuja N , Toyota M , Bronner MP , Brentnall TA . Accelerated age-related CpG island methylation in ulcerative colitis . Cancer Res . 2001;61:3573–3577
  35. Sato F, Harpaz N, Shibata D, Xu Y, Yin J, Mori Y, et al. Hypermethylation of the p14(ARF) gene in ulcerative colitis-associated colorectal carcinogenesis . Cancer Res . 2002;62:1148–1151
  36. Kang GH, Lee S, Kim WH, Lee HW, Kim JC, Rhyu MG, et al. Epstein-Barr virus-positive gastric carcinoma demonstrates frequent aberrant methylation of multiple genes and constitutes CpG island methylator phenotype-positive gastric carcinoma . Am J Pathol . 2002;160:787–794
  37. Toyooka S, Maruyama R, Toyooka KO, McLerran D, Feng Z, Fukuyama Y, et al. Smoke exposure, histologic type and geography-related differences in the methylation profiles of non-small cell lung cancer . Int J Cancer . 2003;103:153–160
  38. Egger G , Liang G , Aparicio A , Jones PA . Epigenetics in human disease and prospects for epigenetic therapy . Nature . 2004;429:457–463

PII: S0016-5085(05)01548-9

doi:10.1053/j.gastro.2005.07.040

Refers to article:

  • Evaluation of a Large, Population-Based Sample Supports a CpG Island Methylator Phenotype in Colon Cancer

    Wade S. Samowitz, Hans Albertsen, Jennifer Herrick, Theodore R. Levin, Carol Sweeney, Maureen A. Murtaugh, Roger K. Wolff, Martha L. Slattery
    Gastroenterology September 2005 (Vol. 129, Issue 3, Pages 837-845)

Gastroenterology
Volume 129, Issue 3 , Pages 1121-1124, September 2005

Article Tools