A systematic profile of DNA methylation in human cancer cell lines

Maria F. Paz, Mario F. Fraga, Sonia Avila, Mingzhou Guo, Marina Pollan, James G. Herman, Manel Esteller

Research output: Contribution to journalArticlepeer-review


Human cancer cell lines are commonly used in basic cancer research to understand the behavior of primary tumors. Aberrations in the DNA methylation patterns are nowadays recognized as a hallmark of the cancer cell. However, no comprehensive study defines the DNA methylation environment present in the established cancer cell lines used in everyday laboratory-based research. To address this matter, we have analyzed 70 widely used human cancer cell lines of 12 different tumor types for CpG island promoter hypermethylation of 15 tumor suppressor genes, global 5-methylcytosine genomic content, chemical response to the demethylating agent 5-aza-2'-deoxycytidine, and their genetic haplotype for methyl-group metabolism genes. Several conclusions arise from our study: (a) a specific profile of CpG island hypermethylation exists for each tumor type, allowing its classification within hierarchical clusters according to the originating tissue; (b) cancer cell lines generally have higher levels of CpG island hypermethylation than primary tumors, because of the contribution of particular CpG islands and tumor types; and (c) there are no major differences between cell lines in their 5-methylcytosine DNA content, efficacy of 5-aza-2'-deoxycytidine treatment, and distribution of allelotypes of methyl-group metabolism genes. Our data provide a basis for a better use of human cancer cell lines in basic and translational research with respect to their DNA methylation environment.

Original languageEnglish (US)
Pages (from-to)1114-1121
Number of pages8
JournalCancer Research
Issue number5
StatePublished - Mar 1 2003

ASJC Scopus subject areas

  • Cancer Research
  • Oncology


Dive into the research topics of 'A systematic profile of DNA methylation in human cancer cell lines'. Together they form a unique fingerprint.

Cite this