By David N. Leff

From cave wall to clay tablet to papyrus to paper, writing has been one of the prime skills that entitled Homo to add sapiens to his or her name.

And all through recorded history, the main way to correct written mistakes was to cross them out and write over them.

Much the same is a novel form of gene therapy developed by molecular geneticist David Russell. The usual method for correcting genomic errors has been simply to insert genes carrying the proper sequence. His approach is akin to crossing out, or whiting out, the mutant form and removing it from the chromosome.

Russell, a professor of medicine at the University of Washington, in Seattle, has recruited adeno-associated viruses (AAVs) as gene-delivery vectors. He reports his method for removing defective sequences while inserting normal ones, (so far in vitro), in the April 1998 issue of Nature Genetics. The article's title: "Human gene targeting by viral vectors."

"My basic strategy," Russell told BioWorld Today, "is trying to fix mutations already present in the chromosomes of human cells, using a viral vector that contains a fragment of normal human DNA. This has homologous sequences identical to the chromosomal sequence, except for the corrected bases.

"Once we have created this viral vector," he added, "we just infect the cells and look to see if the gene has been fixed. We add the AAV vector stock to normal human fibroblasts, for example, growing in a dish. The viral particle binds to the cells, and delivers the DNA right to the nucleus."

A novel aspect of this approach, Russell pointed out, "is that we used a particular kind of virus, the AAV. It has potential advantages that we thought might just make it much more effective than any other method.

"First, you can deliver the DNA really efficiently to the nucleus in multiple copies — say, 10,000 particles. This is really hard to do with any other method, except probably microinjection.

"AAV's other advantage," Russell continued, "at least potential —we haven't formally proven it — is that it's single-stranded. So we reasoned that this single strand would be searching for homologous sequences to pair with. That is the first step in doing the gene correction; it's got to pair with the matching site on the cell's chromosome. Then we'll have the same DNA sequence. It will form a nice, stable double helix, where one strand comes from the vector and one from the chromosome.

"I'd like to draw the distinction," he emphasized, "of gene correction vs. gene addition. We're doing gene correction. Typically, almost every other use of a viral vector is for gene addition. Correction overcomes the problems of addition, such as: the gene isn't regulated properly or it might be silenced. All because it's integrated in the wrong place."

And the strongest drawback of additive gene therapy, Russell pointed out, "is that it doesn't eliminate the bad gene, which — depending on the disease — may or may not be necessary to cure it."

The One-Percent Solution

In one experiment, the Seattle geneticist infected normal human fibroblast cells with AAV vectors containing a mutant exon of HPRT — the gene that encodes hypoxanthine phosphoribosyltransferase. "This happens to be the gene that causes Lesch-Nyhan syndrome and gout," Russell observed. "We chose it just for convenience, because one can very easily detect cells in which the gene has been modified simply by growing them in the presence of a drug.

"In addition to targeting exactly the right part of the cell," he went on, "the repaired copy and the mutated copy of the genetic material were exchanged at a frequency of nearly one percent of targeted cells. This level of exchange," he pointed out, "is perhaps 10,000 to 100,000 times better than has been achieved in normal human cells."

While sufficient to prove efficiency, Russell allowed, "there are very few diseases you can cure with a one-percent correction. You probably don't need to fix every single cell to begin to get improvement. I think if you start to get closer to 10 percent, you'd have a whole range of diseases. Hemophilia," he suggested, "would be a good one. Perhaps even cystic fibrosis or sickle-cell anemia. And a lot of metabolic disorders, such as the lipid storage diseases."

Russell is now moving from in vitro to in vivo experiments. "We're trying to do this in animals," he vouchsafed. "We need a good animal model, affording an easy assay of whether correction has occurred. They will enable us to try to deliver the vectors by intravenous or intramuscular injection, which is more of a model of human disease than what we'd done until now."

How would such parenteral administration find the vector's cellular target?

"We don't know," Russell said, "except that it seems to go to the liver. At least, as far as people can tell from intravenous injections into mice, most of the vector ends up in the liver — but I doubt if that's the whole story. And targeting via intramuscular injection depends on which muscle you inject."

Russell concluded: "The wide host range of AAV, and its potential for transducing large numbers of cells by a simple viral infection procedure, could allow for gene targeting in cell types that are resistant to DNA delivery by other methods." *