Science Editor

You might not expect to find a lab full of mice practicing glass-ceiling discrimination between the sexes. Yet knocking out a gene in a cohort of rodents caused their male contingent to acquire epithelial tumors while the females tend to develop lymphomas and sarcomas.

This gender-based parsing of the malignancies remains a mystery to cancer biologist Stephen Baylin and his team at the Johns Hopkins University in Baltimore.

Baylin is senior author of a paper in the February 2003 issue of Nature Genetics titled: "Heterozygous disruption of Hic1 predisposes mice to a gender-dependent spectrum of malignant tumors."

Screening tumor samples a few years ago for genes with too many methyl groups attached, one of Baylin's graduate students proposed the acronym "hypermethylated-in-cancer 1" - that is, "Hic1." "Methyl groups," Baylin explained, "are among the many types of additions to the DNA sequence that help control gene expression.

"Most genes linked to cancer," he told BioWorld Today, "were first identified from mutations in families, and subsequently some of the same genes were found to be hypermethylated - but Hic1 was identified solely because of extra methylation in cancer cells. Now in this article," he added, "we've shown directly that loss of the gene's function leads to cancers in mice."

Since excess methylation turns the gene off, the team tried to measure its impact on cancer by completely knocking out both copies of Hic1 in their mice, but the homozygous embryos died in utero before birth. So the researchers switched to heterozygous mice, carrying only one of the gene's two copies. "Loss of a gene, already implicated in human cancers, also leads to age- and gender-linked cancers in mice," Baylin pointed out. "Our findings in Nature Genetics validate the gene's dysfunction as a possible early step in development of cancer in people."

A First' For Epigenetic Alterations

"The finding also provides a powerful message to other scientists studying genes linked to cancer," he observed, "through epigenetic' changes such as abnormal methylation. Genes whose cancer-causing abilities are traced to mutations in their DNA sequence have long been studied in knockout mice. However," Baylin observed, "this is the first evidence, to our knowledge, that KO mice can also reveal the cancerous effects of epigenetic alterations in genes."

"Epigenesis," Baylin explained, "is regulation of gene expression activity without a change in genetic structure. Epigenesis emerges from heritable cell changes, not by a sequence of DNA, but by methylation or hypermethylation. What's more, its heritability can be reversed by certain anticancer drugs such as 5' azo-cytidine, which is a demethylation agent."

As their population of KO heterozygous mice matured, Baylin recounted, "for a long time, nothing happened. Hic1 plus/minus mice (heterozygotes) developed normally throughout their first year of life, but thereafter evolved two distinct phenotypes over 100 weeks of observation. The heterozygotes showed a marked age-dependent disposition to malignancies after 70 weeks. In fact, not until the mice were more than 70 weeks of age - pretty old for a mouse - did excessive cancers begin to appear. By 90 weeks of age, we detected 12 malignant tumors in the 73 heterozygotes. That is, 16.4 percent of our experimental rodents had developed tumors, while none of the 28 normal, wild-type littermates had. By 100 weeks, these mice had gone on to develop 25 malignant tumors, three times the rate of malignancies as wild-type ones - 34.2 percent vs. 14.3 percent of normal animals.

"Twenty-two percent of male heterozygotes developed carcinomas," Baylin continued, "whereas their wild-type littermates acquired no carcinomas. By contrast, heterozygous and wild-type females developed lymphomas and sarcomas at a similar rate. Thirty-three percent of female heterozygotes acquired these lymphomas and sarcomas, whereas their wild-type littermates had only one lymphoma."

Sex Scores Resemble Cancers In Humans

"Overall, 44 percent of the malignant tumors were epithelial cancers in Hic1 heterozygous mice. This percentage is high when compared with the usual preponderance of lymphomas and mesenchymal tumors observed in the targeted single-gene disruptions of other tumor suppressors, or in aged wild-type mice. It resembles more closely the distribution of cancer types seen in adult humans.

"Each cancer lacked any sign of Hic1 expression. Furthermore, loss of the one existing copy's function was due solely to hypermethylation of one of the gene's two promoters,' which are strips of DNA that allow the gene to be read. Which promoter was silenced," he noted, "generally depended on the cancer type, and hence gender. While Hic1 hypermethylation in mice mirrors the range of human cancers better than the individual loss of most tumor-suppressor genes, it is not known whether the gender effects seen in the mice are present in human cancer. We have ongoing experiments," he said, "to unravel the gene's mechanism."

"Our study suggests a chromatin region is harbored and hypermethylation reversed by resorting to a KO impact. We conclude that Hic1 is a candidate tumor-suppressor gene for which loss of function in both mouse and human cancers is associated only with epigenetic modifications."