By David N. Leff

The first animal model for a human disease is said to be the armadillo. This low-key, armor-plated mammal, Dasypus novemcinctus, apparently can be infected with Mycobacterium leprae, then serve as a surrogate for studying leprosy in man.

Until gene cloning came on the scene in the early 1980s, the only source of mammalian models was the occasional spontaneous mutant, such as mice prone to obesity and diabetes.

Now, of course, genetically engineered transgenic animals that mimic human maladies are virtually off-the-shelf, or made-to-order, items.

As such, though, they have their problems.

"Because of the deleterious effects of constitutive expression of transgenes, which frequently result in prenatal or postnatal death," wrote molecular endocrinologist Bert O'Malley, at the Baylor College of Medicine, "only a limited number of disease models have been established in transgenic mice."

Among their shortcomings, he told BioWorld Today, is the static state of the inserted transgene, which expresses its target protein non-stop. "Certain things," he pointed out, "you don't want on all the time, because they may have a toxic effect over time. If you could have the gene silent during early development, and then be able to switch it on in the adult, you could then examine the biologic ramifications of whatever gene you're expressing."

Thanks to a stroke of serendipity, O'Malley has constructed just such a transgenic mouse, with an organ-targeting transgene that toggles between on and off expression.

His report of this invention, titled "Ligand-inducible and liver-specific target gene expression in transgenic mice," appears in the March issue of Nature Biotechnology.

O'Malley chairs the department of cell biology at Baylor. His research lab there focuses on steroid hormone receptors.

"Looking for mutations to the progesterone receptor," he recalled, "we found one that had an aberrant response to RU486."

This synthetic compound, a.k.a. mifepristone, is known to the public as a chemical alternative to surgical abortion. RU486 acts to interrupt early-stage pregnancy by blocking the receptor for progesterone, the steroid hormone that maintains gestation.

"Normally, RU486 will keep a progesterone receptor off," O'Malley explained. "In the presence of the hormone, the levels of RU486 will compete for the receptor, and shut down the natural hormone."

He continued: "This mutant receptor we came upon by accident, got rid of a repressor domain in the molecule, without which an antagonist could turn the receptor on * backward to RU486's normal effect as an anti-progestin.

"This gave us the thought that if we could modify that somehow, it might be a ligand-activating gene switch, which you would activate only when you took something exogenously into the body, such as oral or injectable RU486."

He and his colleagues modified that mutant steroid receptor so it would no longer respond to any natural hormone, and inserted a DNA-binding domain that would couple up with a target transgene.

They then proceeded to generate two strains of transgenic mouse, one that expressed the modified switch only in its liver; the other equipped with a target gene that had the binding site for the RU486-activated switch. In this demonstrator model, they used human growth hormone (hGH) as the gene of interest, "because it was something we could conveniently measure, by showing growth in the animals."

Once they mated these two murine versions, their bigenic progeny were fully equipped to activate or silence the target gene by a flip of the RU486 switch.

As their paper reported, ". . . after a single intraperitoneal injection of mifepristone, strong induction [5,800- to 33,000-fold] of hGH expression was detected in the serum." This effect decayed with time, but was reactivated by a new dose of RU486.

Principle Proven: hGH Switched On, Mice Grew

Bigenic animals treated with mifepristone every other day for 10 days showed the effect of the growth hormone * expressed in their liver * by increasing their body weight 30 percent to 38 percent; it reached 50 percent to 60 percent growth 17 days after the first injection.

This proof-of-principle experiment, O'Malley observed, "demonstrated that for any tissue for which you have put in a tissue-specific promoter, you can now use the switch, made in that animal, combined with a target gene of your choice. We could pop out hGH and put any other DNA sequence in, to make any other protein.

"Just add a little RU486 to the animals' drinking water, and operate the switch."

One feature of the transgene was its "insulation" in the genome by flanking the DNA construct with the insulator sequences isolated from the 5' end of chicken beta-globin locus. "Around a gene that you're putting into a transgenic animal," O'Malley explained, "they help insulate the gene from surrounding regulatory influences. Where this insert goes into the genome is by chance. If it goes in next to a strong promoter, it will be turned on by that promoter instead of by the RU486 switch."

Baylor has an issued U.S. patent on O'Malley's toggleable switch, which it has licensed to GeneMedicine Inc., of The Woodlands, Texas. It's No. 5,364,791, issued Nov. 15, 1994, headed, "Progesterone receptor having C-terminal hormone-binding domain truncations."

"GeneMedicine is a gene therapy company," O'Malley concluded. That's what they'll use it for." *