By David N. Leff

Adeno-associated virus (AAV) is a ubiquitous microbe that has infected 80 percent of the human race.

When a vaccinologist or a gene therapist needs to employ a specific virus to confer immunity or deliver genes, he or she knocks out its virulence factors to make it non-infectious.

AAV is one virus that comes into the world avirulent. It's a human parvovirus, with a single-stranded, 4.7-kilobase DNA genome.

Although it has a lifelong lease on human chromosome 19, this virus has never been known to cause its human hosts the slightest ill or ailment. Its viral DNA integrates in the chromosome, settles down in a state of latency, and waits.

What it's waiting for is the arrival of a pathogenic virus, from the adenovirus that causes the common cold to the herpesvirus of cold sores. That's when AAV comes to life. These intruding disease-bearing viruses help AAV to switch from its latent stage to lytic mode. That's why AAV's other name is Dependavirus, because it depends on helper viruses to rescue it from the chromosome's embrace.

Gene therapists depend on mouse retroviruses (RVs) to deliver their genes of interest to cells in need of missing proteins.

"All the early approaches to gene therapy relied on RVs," microbiologist John Monahan told BioWorld Today, "simply because they were available as vectors for delivering DNA. They worked fine. Over 1,000 patients to date have been treated with that approach." Monahan is president and CEO of Avigen Inc., a gene therapy company based in Alameda, Calif.

"But retroviral vectors," he continued, "always had two fundamental problems: They could only transfect dividing cells, and the genes they carried got expressed at random. Every single cell in the ex vivo preparation has DNA at a different location. No two cells are alike."

Retroviral Vectors' Hangups

While RVs' inability to infect non-dividing cells limited their effectiveness, he pointed out, their random expression posed a potentially more serious problem. "Once you start randomly shotgunning DNA into any human chromosome," he said, "you have a finite probability of inactivating a tumor suppressor gene, or activating a nearby oncogene. The last thing you want is an approach with a tumorigenic background."

Monahan made the point that "if you really want to put a piece of DNA somewhere, and you don't care where, as long as it won't cause any side effect, that's where adeno-associated virus comes in."

The fact that AAV's sole abode in the human genome is chromosome 19 commended it to AAV's scientists as a promising take-off point for constructing a safe and predictable gene delivery system. It's based not on the conventional tack of removing the virus' virulence genes, but, in effect, harvesting only two key viral proteins and leaving the rest of AAV behind.

"The virus produces a protein called Rep," explained Avigen's vice president of research and development, Gary Kurtzman. "When you introduce the gene for Rep into the target cell, together with a plasmid DNA sequence," he told BioWorld Today, "it targets a second molecule for integration into chromosome 19 at a specific site."

Kurtzman continued: "That second plasmid can contain any gene of interest. The only other thing on it is a small DNA sequence, about 145 nucleotides long, which also comes from the virus, from a site near its tip. This inverted terminal region binds to the Rep protein, which drags it into chromosome 19.

"The reason Rep binds to a sequence on chromosome 19," he pointed out, "is because it's very analagous to a sequence in the wild-type AAV, which in nature integrates itself with that chromosome. It's evolved this mechanism," Kurtzman said, "and we're just taking it out of the virus and trying to build from scratch a gene delivery system that has this property."

The Journal of Virology for October 1997 carries an article by five Avigen researchers and a Japanese colleague titled: "The adeno-associated virus Rep proteins target DNA sequences to a unique locus in the human genome." Kurtzman is a co-author.

"What we have discovered for the first time," Monahan observed, "is a way wherein we can put a gene into that chromosome 19, piggybacking on what the wild-type virus normally does.

"To our surprise," he went on, "as a sort of bonus, we found that it doesn't matter how much DNA you're adding. We've put in as much as 35,000 nucleotides, bigger than most viruses. It allows for inserting half a dozen to a dozen gene sequences. We'd like to start delivering multiple genes for treating polygenic conditions."

Future Clinical Outlooks

Closer to realization, though still a ways off, are a number of therapeutic scenarios.

"Most human genetic diseases," Monahan observed, "can be treated by putting in a second, correct, copy of the dysfunctional gene. For example, recombinant Factor IX is already on the market for self-injection by hemophiliacs. Expressing the correct protein in the patient's bloodstream can repair the genetic defect."

Kurtzman pictured human insulin as a remote but reasonable target. "One would need a large piece of DNA, larger than what is currently available in gene therapy, with all the regulatory regions in the gene, for an orally available small molecule." He foresaw that such a gene fix would provide long-term control, rather than the daily injections that many diabetics now require.

On this score, Monahan suggested "the bulk of gene therapy applications in the future will not be correcting point mutations in genetic diseases, but producing protein continuously for extended periods of time. Have the cells secrete it, rather than injections or skin patches, or other modes of administration."

Sickle-cell anemia is on Avigen's agenda. "One would, in theory," Monahan said, "have to put the corrective gene into a single bone-marrow stem cell that would propagate. That's a little bit toward science fiction," he concluded, "but it's the direction for the next generation of gene therapy." *