By David N. Leff

Science Editor

Upon emerging from its mother’s womb, a newborn mammal’s first instinct is to make a beeline to the nearest mammary nipple, and start to nurse.

Impressed by the nipple’s soft, smooth surface, the ancient Greek wordsmiths named its membrane “epithelium” – after “thelia,” Greek for nipple. They also observed that it covered the tender borderline area of the lips, which did the suckling.

In fact, the vast, seamless stretches of epithelial cells pave all free surfaces of the human body, mainly skin and mucous membrane. When an embryo sets out to mass-produce these cells, their exuberant output is held within limits by a cytokine called transforming growth factor-beta (TGFb).

“All mammals probably produce TGF-beta,” observed molecular biologist Adam Glick, a principal investigator in the National Cancer Institute’s Laboratory of Cellular Carcinogenesis and Tumor Promotion in Bethesda, Md. “Most cells in the body,” he continued, “make one of several forms, and respond to it. TGF-beta is a pluripotent and very powerful growth factor that regulates proliferation, differentiation and extracellular matrix production. And it’s quite important in human cancer.

“Many cancer cells have some kind of defect in their signaling pathway, which blocks responses to TGF-beta,” Glick explained, “so they’re no longer responsive to growth inhibition. TGF-beta is considered primarily a tumor suppressor gene, so knocking out its signaling pathways has a positive effect on cell growth. Most epithelial cancers – colon, breast, head and neck, skin – have that TGF-beta defect in individual tumor types. But I think the defect is fairly prevalent in those tumors. It would be specific genes in the signaling pathway that are inactivated.”

Glick is senior author of a paper in the current Proceedings of the National Academy of Sciences (PNAS), dated July 31, 2001. Its title: “Conditional epidermal expression of TGFb1 blocks neonatal lethality but causes a reversible hyperplasia and alopecia.”

Transforming Growth Factor’s Many Hats

“Our long-range goal,” Glick told BioWorld Today, “is trying to understand how TGF-beta can have multiple effects in cancer. The mouse skin is a model tissue for multistage carcinogenesis, and it parallels a number of human epithelial lining tissues, such as the gastrointestinal tract. What’s curious about TGF-beta’s action in cancer,” he went on, “is that it has both a tumor suppressive effect – that is, it can block outgrowth of benign tumors – and it also seems to enhance the malignant phenotype. So if you have overexpression of this growth factor in a benign tumor, it actually makes it normal. There’s compelling data from retrospective studies in human patients suggesting that people whose tumors overexpress TGF-beta have a worse prognosis than those who do not.

“We wanted to devise a system,” he recalled, “where we could turn TGF-beta expression on and off at different stages of carcinogenesis. That is, looking at the biological readout in both a benign tumor and a malignant one, seeing what happens to those tumors. Do the benign ones grow? Do they regress? Do they undergo conversion to a malignant tumor? Also, using microarray analysis, to get at different genes that are turned on in these different tumor types when TGF-beta is overexpressed.

“What we needed,” he went on, “was a conditional expression system in mice where we could suppress expression of TGF-beta until specific tumors grew on their skin. This initial PNAS publication describes the mice, and the interesting phenotype that we found in their normal skin.”

Glick explained the difference between a conditional and a constitutive system: “Conventional transgenic technology links the target gene of interest to a constitutive promoter. For instance, in the skin, you get very high levels of the protein expressed by your target genes. But expression is always on,” Glick added. “There’s nothing regulating the promoter. That’s constitutive targeting.

“In the conditional system,” he went on, “there’s an extra layer of complexity, which can turn the promoter on and off at will by addition or removal of a broad-spectrum antibiotic, doxycycline, administered either orally or topically.

“The way the system worked,” he recounted, “is we made two transgenic mouse lines. One contained the TGF-beta target gene, the other the regulator of that gene – itself regulated by doxycycline – which prevents activation of expression. When we mated these animals together, we looked for the double-transgenic progeny that had both transgenes. If we set up breeding in which there was no doxycycline antibiotic present, we never obtained any double-transgenic animals.

“We reasoned that because we were getting such high levels of TGF-beta expression during embryogenesis that it was lethal. So those double-transgenic animals were never born. Now if we added to the embryonic diet a subthreshold level of doxycycline, we knew we had partially lowered expression of TGF-beta. Then we did obtain double-transgenic mice that went through complete development. But most of those animals were born dead, and they had a skin phenotype associated with inhibition of cell proliferation.

“When we completely suppressed expression of TGF-beta by feeding higher levels of doxycycline, normal double-transgenic animals were born. Those overcame the lethality entirely, because they completely suppressed expression of the target TGF-beta gene.”

TGF-Beta Genes As Therapeutic Targets?

“At weaning,” Glick related, we removed the antibiotic from their diet, which allowed the conditional expression of the system to work. And once we did that, in three weeks they had lost their hair – which is the major finding in the paper. And when we added doxycycline back to the feed, to suppress TGF-beta expression, in two weeks the hair had grown back. So we had a reversible hair-loss phenotype.

“Then we were able to allow those mice to grow to adulthood before we turned on expression of the gene. At that point we saw that the mice had lost their hair, but instead of showing a lack of proliferation in the epidermis, which is what one might expect because we were overexpressing a growth inhibitor, we saw elevated proliferation, and thickening of the skin.

“What we learn from these mice,” he concluded, “is that the TGF-beta genes activated in the malignant tumors that contribute to the malignant phenotype could themselves be potential targets for therapeutic agents, which would be useful in a clinical setting.”