News FocusGene Therapy

Putting the Fingers On Gene Repair

See allHide authors and affiliations

Science  23 Dec 2005:
Vol. 310, Issue 5756, pp. 1894-1896
DOI: 10.1126/science.310.5756.1894

The struggling field of gene therapy could regain its momentum if proteins called zinc finger nucleases live up to their promise of efficiently and safely repairing mutations

Imagine one of your car's headlights winks out, but instead of simply replacing the bulb, you attach a third headlight. That's typically how genetic engineering works today. When molecular biologists want to boost a plant's drought resistance, for example, or repair the cells of a patient with an inherited disease, they paste a new gene into a random spot on a chromosome and hope it does the job. Nobody has yet figured out a good way to directly repair a cell's defective genes.

Now a technology is emerging that could enable scientists to much more readily repair or alter a cell's existing genes. The key is an engineered protein called a zinc finger nuclease that latches onto a specific gene and snips its DNA. The cell then heals the broken strand using copies of a replacement gene that researchers also supply—in the case of gene therapy, the copies would lack the disease-causing mutation in the original.

In the past 3 years, researchers have shown that zinc finger nucleases can successfully modify existing genes in fruit flies and plants. They've even fixed, in a lab dish at least, human cells bearing a mutation that causes a deadly inherited immune disease. Although no disease gene has yet been repaired in a mammal, much less in a person, researchers are hopeful that the work will lead to clinical applications. The ability to routinely edit genes in lab-grown cells, animals, and plants would also be a boon for basic scientists exploring gene function. “This would be a phenomenal research tool. It could change the way we do science,” says molecular biologist Matthew Porteus of the University of Texas Southwestern Medical Center in Dallas.

Glowing success.

Human cells began to shine after zinc finger nucleases repaired a gene for green fluorescent protein.

CREDIT: NATURE 435, 646 (2005)

There remain several obstacles to this vision for zinc finger nucleases. For one, they must be tailored for each target gene, which only a few labs can do at the moment. Several of these groups are gearing up to make the customized proteins more widely available, hoping to get around the fact that one company owns the broadest collection of zinc finger nucleases as well as sweeping patents on their use. Some biologists are concerned that this firm will stifle the field's progress.

Finally, like other gene-therapy strategies, the use of zinc finger nucleases poses serious safety questions. “It's a very exciting approach,” says Richard Mulligan of Harvard University. But he cautions, “As an old gene-therapy person, my view is this runs the risk of moving too quickly, leaving out too many biological details, and suffering the same fate as the gene-therapy field overall.”

Product placement

Gene therapists and other biologists would like to be able to modify a cell's existing genes because simply inserting a new gene into a cell's genome poses problems. First, that new gene may not function in the same way as the one it's meant to replace. The introduced gene usually lands in a random location, far from the promoters and other noncoding regions that control the natural gene. That often means the cell makes too much or too little of the added gene's protein product.

Moreover, the random nature of the gene insertion has led to serious side effects. In the first clear success in gene therapy, scientists in the past 6 years have apparently cured nearly two dozen children with severe combined immunodeficiency disease (SCID) by stitching a corrective gene into patients' blood cells. But three patients with X-linked SCID developed leukemia, seemingly because the retrovirus carrying the corrective gene inserted its package of DNA near an oncogene. In theory, says Mulligan, repairing the endogenous gene that causes X-SCID should be much safer.

Scientists have tried to exploit one of the cell's natural repair mechanisms to edit genes, but with limited success. When a chromosome is damaged, cellular enzymes can restore it using a corresponding strand of DNA as a template, usually from the cell's other copy of the chromosome—a process called homologous recombination. Scientists can piggyback on this natural repair system by tricking the cell into performing homologous recombination using added DNA as the template instead. Although this strategy works well enough in yeast and is routinely used to make “knockout,” or transgenic, mice, the rate of repair—one in a million cells—is too low to be useful in other species. Another gene-repair technique, chimeraplasty, has not proven to be easily reproduced, if it works at all (Science, 13 December 2002, p. 2116).

More recently, researchers seeking a way to make gene repair via homologous recombination work better turned to zinc fingers, discovered by Aaron Klug's group at the Laboratory of Molecular Biology in Cambridge, U.K., in 1986. Molecular structures containing about 30 amino acids and held together by a zinc ion, they're key components of many proteins involved in transcription, the process by which a gene's information is converted from DNA into RNA. Indeed, zinc fingers determine where so-called transcription factors bind. Each finger nestles into the DNA helix at a specific set of three bases (such as GCG), allowing a transcription factor to turn on a specific gene. Klug's lab and others next showed that they can mix and match different zinc fingers to latch onto specific sequences of DNA—there are 64 possible three-base combinations.

Researchers then began exploiting zinc fingers to ferry molecules to a unique position along a chromosome—for example, fusing them to proteins that turn genes on or off so that such proteins would regulate a specific gene. And that inspired the idea of zinc finger nucleases as a way to spur homologous recombination. The strategy is to attach zinc fingers to enzymes called endonucleases that make double strand breaks in DNA. When these enzymes are added to a cell, the usual rate of homologous recombination—1 in a million cells—rises to at least 1 in 1000. In 1996, Srinivasan Chandrasegaran's group at Johns Hopkins University in Baltimore, Maryland, reported that by attaching three different zinc fingers to these DNA-snipping enzymes, they could cut a piece of a free-floating DNA at a precise location. (The researchers add two nucleases that first land on each side of the point they wish to cut and then combine to snip the DNA.) With Dana Carroll's group at the University of Utah, Salt Lake City, they later showed that when new DNA was inserted into frog eggs and cleaved by a zinc finger nuclease, the cells then fixed the break.

The next step was to see whether zinc finger nucleases could alter specific genes in a cell's chromosomes. In 2002, Carroll's group showed in fruit fly larvae that the nucleases could mutate a gene that controls the insect's color. Some of the resulting flies had patches of yellow where they would normally be dark.

That work didn't attempt to replace the cleaved portion of the color gene, but Carroll's team reported doing that in 2003 in Science (2 May 2003, p. 764). In addition to the zinc finger nucleases, they added copies of a different version of the color gene into the fly larvae and showed that the larvae incorporated that variant via homologous recombination. In the same issue, Porteus and David Baltimore of the California Institute of Technology in Pasadena reported a similar success. They showed for the first time in human cells that zinc finger nucleases could be used to repair a mutation in the gene, albeit a nonhuman reporter gene inserted into the cells.

The first proof of principle that zinc finger nucleases can correct a human disease gene came this spring. In the April online edition of Nature, Porteus and scientists at Sangamo BioSciences Inc. in Richmond, California, showed that such nucleases could make a one-base change in a functional copy of IL2Rγ, the gene that causes X-SCID, in human cells. The zinc finger nucleases worked with relatively high efficiency—18% in primary blood cells and 5% in T cells, the cells that would need to be targeted in X-SCID patients.

Fancy finger work.

Three zinc fingers (ribbon structures) attached to a nuclease (purple oval) can be used to latch onto either side of a mutated gene and snip it. The cell then fixes the break with supplied DNA.


Sangamo also intends to use zinc finger nucleases to correct mutations in other blood diseases, such as hemophilia. The general strategy is to isolate bone marrow or other blood-forming stem cells from a patient, correct the mutation in the cells in lab dishes, and put the stem cells back.

And in a twist on repairing disease genes, Sangamo is also testing whether zinc finger nucleases can treat HIV patients by disabling the gene for a protein, called CCR5, that the HIV virus uses to enter cells. In 2006, Sangamo and collaborators hope to begin clinical trials in which a person's HIV-susceptible immune cells would be replaced with bone marrow cells that have had their CCR5 genes knocked out.

Delivery problems

Whereas Sangamo may be optimistic that zinc finger nucleases will soon enter the clinic, others say the technology needs to be mature. “It's certainly not going to be a slam dunk. There's a huge number of things standing in the way,” says Michael Blaese of the Institute for Inherited Disease Research in Newton, Pennsylvania.

One of the first hurdles is to get enough zinc finger nucleases into the right kinds of cells. Scientists don't just add the proteins to cells. Instead, DNA encoding an engineered nuclease is coaxed into cells with a jolt of electricity. But the immature T cells that researchers would like to target in X-SCID patients are too fragile for such electroshock. The company is now working on ferrying DNA encoding the zinc finger nucleases into cells using disabled lentiviruses. “It's perfect for our technology,” says Sangamo Vice President of Research Philip Gregory. The team hopes to show they can this way repair the IL2Rγ gene in cells extracted from an X-SCID patient by next summer.

Yet safety issues remain. Zinc finger nucleases can create double strand breaks at DNA sequences other than the target gene, which in theory could lead to cancer. Sangamo says it has greatly reduced that risk by using very specific nucleases, ones with an extra zinc finger. They're also using nucleases that cannot pair up in wrong combinations. “It's a very neat solution,” says Carroll, who collaborates with Sangamo. With this technique, the company sees only a minimal increase over the background rate of double strand breaks, says Gregory. Mulligan, who has tested many zinc finger nucleases for off-target effects, is skeptical that this solves all the problems, however.

Despite such safety issues, the flurry of successful experiments with zinc finger nucleases has created a demand for the proteins among many other groups. “People are lining up,” says Daniel Voytas, a plant biologist at Iowa State University in Ames. Plant scientists, for example, are keenly interested because some critics of engineered foods containing foreign genes may be more willing to accept crops made by tweaking an existing gene. Voytas's group reported in a November issue of The Plant Journal the first demonstration of gene modification using zinc finger nucleases in plants.

Still, the problem, says Porteus, is that the work is so challenging that you have to be an experienced zinc finger biologist to craft nucleases that work well. The simplest way to create a nuclease that targets enough DNA sequence to hit a specific gene is a “modular design” approach, favored by Carlos Barbas of The Scripps Research Institute in San Diego, California, that yokes together three different zinc fingers and a nuclease to home in on a nine-nucleotide sequence of DNA. But there is debate about whether these nucleases will work in all cases. If a nuclease isn't specific enough, many cells die from off-target breaks, and efficiency is low; only a tiny fraction of cells receive the desired change. Others optimize their nucleases—that is, they try out many design variations to identify the best one. Harvard's J. Keith Joung, for example, generates libraries of zinc finger combinations that vary slightly and then tests them in cells.

Sangamo, which has published some of the zinc fingers it uses to make nucleases and maintains a huge proprietary library of other zinc fingers generated by a company Klug founded, also optimizes the proteins but declines to publicly reveal exactly how. Selecting a zinc finger nuclease “is the beginning of the design problem, not the end,” Gregory says. The company's Nature paper shows the benefits of the tweaking, he says—the nuclease used became five times more efficient after optimization. “That process is not easy,” Gregory says, defending the company's decision to keep its technology confidential. Klug, whose institution licensed his work to Sangamo and could receive royalties, agrees, adding that the company is still refining zinc fingers. “I wouldn't release the technology until it's fully developed,” he says.

Sangamo CEO Ed Lanphier also points out that the company collaborates with dozens of academic labs and has no objections to independent efforts to develop the technology. “If they want to go out and work hard in this area, that's great,” he says.

But some say Sangamo is hindering progress. For instance, last year, Voytas attempted to license one of Sangamo's zinc finger nuclease patents in order to launch his own plant biotech firm, but the two sides failed to reach an agreement. Barbas, too, contends that Sangamo is “inhibit[ing] the technology from proliferating.” He says that the company recently called Scripps to question a Web site he created where biologists can type in a gene's DNA sequence and learn how to create zinc fingers to target it ( (Sangamo's Lanphier said he could not comment on the matter.)

Spreading the word.

Scripp's Carlos Barbas launched a Web site to share zinc-finger technology with the academic community.


Other researchers agree that without more involvement from academic researchers, the technology will never mature. Chandrasegaran at Johns Hopkins is seeking funding for a multi-institution collaboration that would design zinc finger nucleases to target 30 disease genes. Meanwhile, Joung, Voytas, Porteus, and Andrea Cristani of Imperial College London have formed a consortium to publicly share zinc finger nuclease technology. The group has posted a Web site (, and it expects, by next year, to disburse at a nominal cost materials that will allow others to make and test the nucleases.

The consortium will also seek to answer questions such as whether using more zinc fingers—six on each nuclease rather than the usual three or four, for example—improves specificity. “My interest is not to circumvent Sangamo's patents. I just want to make the technology available, easy to use, efficient, and robust,” Joung says. If that happens, gene therapists trying to repair disease genes may have finally found the tool they've sought.

Navigate This Article