Taking advantage of the latest developments in lentiviral expression systems, Massachusetts Institute of Technology researcher Luk Van Parijs has been using RNA interference to screen hundreds of genes within transgenic animals in order to learn about the genetics of cancer. So far, this approach has been quite successful — data from the work has already been submitted for publication — and now Van Parijs and colleagues have begun to investigate whether they can get similar results for diabetes.
Using what he calls a “forward genetics strategy”, Van Parijs used lentiviruses engineered to express the oncogene c-Myc and short-hairpin RNAs targeting particular genes. “We’ve made a collection of 700 of these vectors that target about 200 genes total, and we basically use those to make transgenic animals” by putting them into hematopoietic stem cells that are injected into mice, he told RNAi News.
“We know what the oncogene does by itself: It will make the mice get cancer at a certain rate and speed,” Van Parijs said. “We [were] just looking for short-hairpins that change incidence of cancer — fast, slow, more, or less.” Once a change in cancer progression was observed, the researchers looked to see if the shRNA was targeting a novel or interesting gene.
“It’s forward in the sense that we’re just looking at phenotype — the only thing we care about is that we change how these animals get diseases, and we’re less biased about what genes should be doing it,” he said.
Van Parijs said that in the tumor study, the results of which are currently under review by Cancer Cell, the shRNAs used were primarily against genes that are regulated by growth factors. “The way an oncogene works is that it co-opts these pathways that usually need a growth factor — basically, meaning that you don’t need a growth factor anymore, which is why you get a cancer,” he explained. “Our hypothesis was: The reason c-Myc is such a potent oncogene is because it does this really well, but we don’t know all the different genes it can control.”
According to Van Parijs, the screen confirmed that a number of genes already suspected as having ties to cancer, such as p53 and Akt, are indeed linked to the disease. “But some of the genes that we put in there — there’s one of the members of the NF-kappaB family, there’s another gene that is basically controlling Ras signaling in ways we don’t understand — are genes that nobody has implicated [in cancer] before,” he said.
“That was with the first initial set and now we’ve got libraries covering many hundreds of genes, so our current experiments are focused on getting those wrapped up and that’s where we’re going to start some really cool and unexpected stuff,” Van Parijs added.
Given the success of the tumor work, Van Parijs said that he and his lab are now trying to conduct similar experiments in type I diabetes.
“What has happened in diabetes is that people have done a lot of human population or mouse population genetics,” he said. “So they’ve got these genetic intervals … each [of which] contain about 10 to 15 genes. We’re setting up to systematically test most of those — that’ll be a grand total of maybe 200 [genes].”
In the diabetes experiments, Van Parijs said that he and his colleagues are injecting the lentiviruses into the embryos of non-obese diabetic (NOD) mice, resulting in the global expression of shRNAs against a particular gene.
“Diabetes occurs spontaneously anywhere after 10 weeks of age in the mouse [model], so what we do is make a cohort of [transgenic] mice … and you just follow those [and] monitor whether they get diabetes,” and how quickly, he said. “There is no way you could do this using traditional genetic approaches — it would just be murder.”
Van Parijs said that his lab is currently capable of screening in a cancer or diabetes model a couple of hundred genes in three to four months. “As you can imagine, we’ve had a lot of interest from industry, and I think a number of companies are establishing platforms based on these ideas.”
Already, Van Parijs’ work has caught the attention of Swiss drug maker Serono. “We’ve set up a strategic collaboration with Serono,” he said. “It’s a large-scale cancer screen to cover many hundreds of genes, including those that they’re interested in as far as druggable targets. It’s a fairly substantial research agreement — many millions of dollars.”
Van Parijs sees his work leading in two directions: “From our perspective, we want to get a fundamental understanding of a pretty complicated biological process,” he said. “From the company’s perspective, we’ve got a hundred genes that could potentially be targets for drugs — which of those are likely to be the most effective or relevant in a particular disease setting?”
As for the possibility of RNAi-based therapeutics: Van Parijs expressed doubts as to their viability, especially given the difficulty of delivery, in a November interview with RNAi News (see RNAi News, 11/21/2003). However, recent advances have led him to consider changing his stance.
“It still hasn’t happened, but … results from people like Judy Lieberman [from Harvard’s CBR Institute for Biomedical Research] suggest that [researchers] are … delivering to certain tissues fairly effectively,” he said. “It seems like you can get some success, but whether it will be enough to get therapeutic value [is unclear].
“But, I’d be happy to have to retreat on my words on that one,” Van Parijs added.
— DM