By David N. Leff

A teaser that tantalized research oncologists a few years ago goes like this: In some tumors they examined, only half of the cells had mutated tumor-suppressor genes * a plausible reason for their malignancy. But the other half showed no such mutations, yet were just as cancerous.

It turned out that instead of being mutated, some of those genes had been methylated * with equally tumorigenic effect.

"What they found," said cancer researcher Eric von Hofe, "was methylation substituting for mutation."

von Hofe, associate director of new targets at Hybridon Inc., of Worcester, Mass., explained: "Of the four DNA bases in the human genome * cytosine, guanine, adenine and thymine, one in particular, cytosine, can become methylated by one enzyme in particular, DNA methyltransferase.

"Basically," he told BioWorld Today, "if you methylate the cytosine, it's like creating a lock to keep the gene from being turned on. And tumor suppressor genes, such as p16, for example, are rich in cytosine DNA bases, which are targets for methylation."

So the hunt is on, at several academic and commercial laboratories, for a way to demethylate tumors, by finding a means of inhibiting DNA methyltransferase. One such agent is an antisense compound, under development at Hybridon and its research partner, biochemist Moshe Szyf, at McGill University, Montreal.

"Moshe is a clear leader in the methyltransferase arena," Hybridon's von Hofe said, "and we're the antisense people."

Both are senior co-authors of a paper in the latest Proceedings of the National Academy of Sciences (PNAS), dated Jan. 21, 1997. Its title: "Inhibition of tumorigenesis by a cytosine-DNA methyltransferase antisense oligonucleotide."

The experiments reported in their PNAS article tested the hypothesis that the methyltransferase enzyme is a candidate target for anticancer therapy, using their antisense inhibitor.

Inhibitor Slowed Tumor Growth 'Significantly'

A mouse strain known to have harbored a spontaneous tumor of the adrenal gland's cortex, namely Y1, served as in vivo model. To begin with, Szyf and his colleagues at McGill implanted one million Y1 cancer cells subcutaneously into the flanks of 40 such mice.

Three days later, the animals received five milligrams per kilogram of the 20-base-pair test antisense compound, or one of two control preparations. The first dummy agent consisted of scrambled antisense DNA, the second of a reversed sequence.

"We took those 20 bases," von Hofe recalled, "and jumbled them around until they were no longer recognizable. Our reverse control consisted of flipping the sequence direction between its terminals, while preserving the base-pair complementarity."

Thirty days later, all of the mice had developed growing tumors, but those on the true antisense recipients were significantly smaller.

Since submitting these preliminary in vivo results to PNAS last June, von Hofe told BioWorld Today, "we now have in vitro inhibition of the human methyltransferase, as distinct from the murine enzyme." He went on: "The real push scientifically now is on identifying appropriate preclinical in vivo models for testing that human antisense compound, as well as looking at other direct inhibitors of methyltransferase."

So far, the group is developing nude mouse models to look at the effects of antisense inhibition on human lung, colon and bladder tumor cells. "We have some preliminary data showing antisense activity against human lung cancer," von Hofe observed, "which is furthest along."

While emphasizing that many of the mechanisms on which this work is based have yet to be validated, von Hofe proposed a hypothetical scenario for treating pulmonary carcinoma in the clinic:

Setting Stage For Clinical Use

"The most obvious scenario right now," he said, "would be to find a cancer where there's a methylated tumor suppressor gene, and for which there's good evidence that the loss of its activity is very important in the growth of the tumor.

"Demethylating that tumor suppressor gene," he pointed out, "would restore the cancer to a more normal growth pattern.

"So we would for instance take a lung cancer biopsy, and look for methylation of the p16 gene, because in lung cancer there's a very good case for p16 being involved. If it's found to be methylated, that patient would then get therapy with an antisense or direct methyltransferase inhibitor, and one would expect that tumor to regress."

He made the added point that "The tumor suppressor gene in this case, p16, is involved in growth, but may also be implicated in a type of cell death. So if that were to happen, you're tipping the balance away from cells that are dividing, and toward tumor regression."

Tumor suppressor cells are very prone to be methylated," McGill's Moshe Szyf told BioWorld Today. "But another sensitive site, on which we have some data, may be DNA's origin of replication." This is the place on a genome where the DNA forks into two single strands, as the first step in forming two double helices.

Should that site become methylated, Szyf continued, "that enables the cell to grow without control, and also to shut down functions very important to the cell."

As an example, he cited the same type of adrenocorticoid cancer tested in his mice. "Once these cells becomes cancerous, they stop making their normal steroid hormones. They will shut down genes that control growth, and may send the wrong cues to origins of replication. The combination is very brutal."

Szyf pointed out another aspect of methylation: "Viruses," he said, "also turn it on, probably to shut down genes that are expressed on the surface of the target cell, the ones that provide cues for the body's immune surveillance system.

"One of the most terrible viruses," he continued, "T-antigen SV40, does that. Epstein-Barr virus does that. HIV probably does that. It's a whole area," he concluded, "that we haven't explored yet * but there are indications."

Szyf is a scientific founder of MethylGene Inc., in Montreal, and serves on its scientific board. The company is a joint venture between Hybridon and Canadian investors.

"Its mission," von Hofe said, "is to develop methylation inhibitors as therapeutics for the clinic." *