Identification of chemical modifications in DNA may provide better diagnostic markers of disease, according to research published in the March 16 issue of Proceedings of the National Academy of Sciences.

City of Hope investigators have used a molecular approach developed at the institution to identify pieces of DNA aberrantly modified in lung cancer cells and then searched the human genome for the corresponding genes.

What they found could facilitate earlier cancer detection strategies.

The DNA that flanks genes controls whether genes are switched off or on, researchers explained. Some repressed genes can be detected biochemically because their flanking DNA is modified through a process called methylation. Methylated DNA recruits repressors that keep a gene inactive and often is found in regions of DNA with abundant cytosine (C) and guanine (G) residues — areas known as CpG islands.

Gerd Pfeifer, Ph.D., professor and chair of the Division of Biology, and colleagues screened microarray chips representing large portions of the human genome and found hypermethylated genes most likely silenced in lung cancer cells. The study was done in collaboration with Arthur Riggs, Ph.D., professor and director of Beckman Research Institute, and Kemp Kernstine, M.D., Ph.D., professor of surgery and director of the Lung Cancer and Thoracic Oncology Program. 

Pfeifer noted that the genes identified could be useful to catch lung cancer early, when patients have the best chance for a cure. “There is a population of high risk heavy smokers who could be screened both by advanced imaging techniques and molecular testing to provide the highest level of sensitivity and specificity for detecting malignant early stage tumors,” Pfeifer said.

Currently, only 16 percent of the more than 170,000 cases of lung cancer diagnosed each year are detected at an early, operable stage. In addition to enabling early diagnosis, molecular markers could be evaluated through noninvasive tests, possibly by analyzing sputum samples. 

In a 2006 Cancer Research paper, Pfeifer and postdoctoral fellow Tibor Rauch, Ph.D., devised a method called MIRA (short for methylated CpG island recover assay) to capture small pieces of methylated DNA from lung cancer cells. They used the method to show that several members of the much-publicized HOX (short for homeobox) gene family were hypermethylated in lung cancer cells.

In this new study, they cast an even wider net and used MIRA-captured DNA fragments to screen microarrays containing DNA representing 27,900 CpG islands associated with human genes. Human HOX genes are clustered on four different chromosomes, and the investigators observed aberrantly methylated CpG islands in genes of all four HOX clusters.
 
The flanking regions, or promoters, of two genes called HOXA7 and HOXA9 constituted significant methylation “hot spots.”

“We identified the HOXA9 gene’s promoter region as a very early target of DNA methylation in lung cancer,” Rauch said. “In this way, HOXA9 could be a very promising marker candidate for diagnostic kits.”

Pfeifer agrees that aberrant methylation is most useful as a diagnostic tool. “I speculate that this is not a cause of cancer,” he said. “Instead, I see methylation of homeobox genes as a cancer marker.”

Initially characterized in the fruit fly Drosophila, HOX genes are renowned embryonic regulators governing position along the body axis in organisms from insects to humans. Embryonic activity of HOX genes determines whether a fly will sprout legs from its head or whether a budding limb will become an arm or leg. Why HOX genes are apparently silenced in cancer is intriguing, Pfeifer said, but their silencing “supports the hypothesis that cancer arises from stem or progenitor cells.”

Members of the Pathology Core Laboratory provided valuable assistance to the researchers in preparing tissue samples. Also participating in the study were Zunde Wang, Ph.D., and Xueyan Zhong of the Division of Biology, Xiwei Wu, Ph.D., of the Division of Information Sciences, Sean Lau, M.D., of the Division of Pathology, and Xinmin Zhang, Ph.D., of Nimblegen Systems Inc.

Grants from the National Institutes of Health funded the study.