By David N. Leff

Science Editor

Mice don't get AIDS. As a rule, the HIV retrovirus doesn't attack them. Instead, a different retrovirus - the murine leukemia virus - may infect these lab animals.

"Humans who inherit the AIDS virus from their mothers get AIDS," observed mouse cancer geneticist Neal Copeland. "Just as the HIV retrovirus is transmitted in human infants from mothers to offspring, so the murine retrovirus can be transmitted through infection in the uterus."

Copeland is a scientist in the Frederick (Md.) Research and Development Center of the National Cancer Institute. "The leukemia-prone mice that we use," he pointed out, "are recombinant inbred strains, and they spontaneously develop leukemia. We don't do anything to them to get the disease; it's natural susceptibility.

"The way it works," he went on, "is that these mouse viruses like to infect lymphoid cells. As they are retroviruses, part of their life cycle is to make a DNA copy of themselves, and that provirus inserts itself into the mouse's genome, where it makes RNA. This gets translated, and produces more virus, which buds in the cell and infects a neighboring cell. That's its life cycle.

"When that retrovirus integrates into the genome," Copeland continued, "it does so at random. So every lymphoid cell that's infected spawns a provirus integrating some place in the genome. The way this works is that during the life span of one of our leukemia-prone mice, billions of lymphoid cells are being infected by the virus, and billions of proviruses are integrating at random in the genome.

"And what can happen, by accident, is that the provirus, when it integrates, can land near a proto-oncogene, and maybe affect its expression. That is, how to convert it into a cancer gene. Or else it could land in a tumor suppressor gene, and inactivate it, making that cell become a cancer cell. So every cell in that tumor that grows out would have this provirus linked to the gene that it mutated by insertional mutagenesis to cause the disease."

Copeland is senior author of an article in the November 1999 issue of Nature Genetics titled, "Leukemia disease genes: large-scale cloning and pathway predictions."

"We were asking if we could develop better ways to use this insertional mutagenesis to identify cancer-causing genes," he told BioWorld Today. "Over the last 20 years, we and several other labs throughout the world have been using this retroviral process to identify cancer genes. It's a slow, arduous process. The classical, genetics-based way that we did it was to clone out these viral integration sites into the lambda phage [bacterial virus], which is slow and laborious."

Adding Genomics To Genetics

Copeland ticked off "a whole litany of things that we went through: We had to identify unique sequence-flanking probes to see if the loci of the original viral inserts were rearranged in multiple tumors. This long harangue has been productive, to be sure. We've identified a number of genes, and many of them have turned out to be human cancer genes. So it's a very nice method - but slow.

"Our whole purpose, as we report in this paper," he went on, "was to see if we could devise a higher-throughput strategy - based on genomics as well as genetics - for using these insertional mutagens to identify cancer genes."

So he and his co-authors modified a procedure they call "inverse polymerase chain reaction," or IPCR. "This made it much more rapid to clone the genes out," Copeland recalled. "Then we used a sequence approach to identify the genes in the vicinity of these viral integration sites. In that, we made use of the new tools provided by the NIH Human and Mouse Genome Projects - in particular, all the new databases that have all the cancer genes in them. So then when we did BLAST searches - comparing the sequences of everything in all the databases against our sequence that we get from these flanking regions - we often got hits, because most of the genes were now in the databases.

"If we had done that 10 years ago," Copeland observed, "when there were just 100 genes in the database, we wouldn't have hit much. But the fact that there are thousands of genes now in the databases, we hit genes very often by this BLAST approach. And in the process, we've screened a lot of our mouse leukemias, and identified a whole host of new candidate cancer genes. So we developed a better technology, while identifying a lot of potential new cancer genes."

Copeland continued, "We wanted to see if in that flanking DNA sequence there are any coding regions for known genes, or for that matter expressed sequence tags - ESTs. And when we did that we got about 90 genes out of the 400 or so that we cloned and characterized. Some of those are already known to be cancer genes, or are really good candidates.

"Through this IPCR method," he went on, "we identified all the cancer genes that we had cloned out by the more laborious method of cloning into lambda phage. We've been doing that for 19 years, during which time we'd cloned out only a handful of genes. In this IPCR approach we got them all, plus all these new ones. So the system clearly works."

From Gene Knowledge To Drug Power

"Clearly, there are many more mouse leukemia genes to be identified, and many of the genes that we're cloning out on the mouse are also involved in human leukemia," he said. "So we think that through this IPCR approach we're going to be able to clone out a lot of new human leukemia genes that weren't easy to clone in humans before."

Copeland surmises that "eventually IPCR will lead to new cancer therapies, in solid tumors as well as leukemias. What the field wants to know," he pointed out, "is the identity of every cancer-causing gene in the human genome. Once you know those, they provide targets for developing new anticancer drugs. Knowing the genes is the first step. Then the second step is figuring out what pathways those genes function in, and how they induce cancer. And once you know that, you can then specifically target those key genes for drug design."