A team of scientists at Cold Spring Harbor Laboratory and Stanford University has identified a tumor-suppressor function for Chd5, a member of the chromatin remodeling protein (CHD) family that maps to the 1p36 region of human chromosome 1. Their results are published February 9 in Cell.

Mutations in 1p36 are coincident with many diseases, including neural cancers, melanoma, hematopoietic cancers, and epithelial cancers of the thyroid, cervix, colon, and breast. These mutations are usually deletions, sparking a widespread search for a tumor suppressor as the missing link.

Finding this link was the ultimate goal, says lead author Dr. Alea A. Mills, of Cold Spring Harbor Laboratory. "There was a risk," she says. "If we had hit a haploinsufficient gene, for example, we wouldn't have been able to study cancer in these models, but may have discovered other interesting genes in the process. It takes a big commitment to see this sort of chromosome engineering project through to the end."

Using embryonic stem cells, Dr. Mills' team created mouse models with a region corresponding to human 1p36 that was deleted or duplicated. This allowed them to study what happens when the region was lost (as in human cancers) or when an extra copy was added.

They found that cells with an extra copy of the region proliferated poorly and underwent senescence, but deletion of the same region caused immortality. In vivo, mice with duplications had excessive apoptosis, while those with deletions were normal but cancer prone.

Once they pinpointed an important tumor suppressive region that corresponded to 1p36, Dr. Mills' team tested which of the candidate genes in the area was the tumor suppressor by seeing which one was able to correct the poor proliferation of cells with an extra copy of the region when knocked down by RNAi. Only one, Chd5, scored positive. Importantly, when Chd5 was depleted in normal cells, the result was the same as when the whole region was deleted. Thus, Chd5 was the tumor suppressor gene in the region.

Chd5 is related to a family of proteins known as chromatin remodelers, proteins that help unravel DNA so that genes can be read by a cell's transcription machinery.

"The role of Chd5 as a chromatin remodeler is based purely on its homology to other members of the Chd family - there has been no functional data before this," says Dr. Mills. "How this affects tumorigenesis is not well understood, but a better knowledge in this arena is likely to offer novel ways to design more effective anticancer therapies."

The methods behind this research are an important contribution and will have significant implications for mouse modeling and cancer, says Dr. Stephen Chanock, who leads the Genomic Variation Section of NCI's Pediatric Oncology Branch within the Center for Cancer Research and the Core Genotyping Facility within the Division of Cancer Genetics and Epidemiology.

"Not only have they identified a functionally important novel tumor suppressor mapping to human 1p36, but they have provided a useful road map of where and how to go after other similar genes; moreover, they've used a technically clever and very interesting way to accomplish this," Dr. Chanock says.

For Dr. Javed Khan, who leads the Oncogenomics Section of NCI's Pediatric Oncology Branch, an important aspect of this research is the finding that heterozygosity of a single gene, Chd5, predisposed cells to malignancy and that the wild-type locus was retained in immortalized cells, as well as in spontaneous tumors. This finding implies that the "two-hit hypothesis" - which states that if one copy of a gene is defective, then the other copy must be deleted or silenced for a cancer to develop - is not necessary for tumorigenesis in certain cancers. "This study provides convincing evidence that haploinsufficiency of a single gene can lead to certain cancers and this represents a paradigm shift," says Dr. Khan.

By Brittany Moya del Pino