Volume 3, Issue 19 March 2006
Engineering Evolution
David Schaffer comes from a medical family, with parents who are biomedical researchers and a sibling who is a doctor. Yet when he decided to design new techniques to tackle some of humankind's toughest diseases, he bucked the family traditions and did it his way—as a chemical engineer at UC Berkeley.
Schaffer is working to improve the delivery of gene therapies that have the potential to fight neurodegenerative diseases like Parkinson's, Lou Gehrig's disease (Amyotropic Lateral Sclerosis or ALS), and even cancer. "Chemical engineering gives us a molecular approach to problems of mass transfer and complex reaction networks," says Schaffer. "These are problems that chemical engineers are trained to solve."
Chemical Engineering Associate Professor David Schaffer has had success in developing new gene therapy tools.
Schaffer, Associate Professor of Chemical Engineering and a member of the Helen Wills Neuroscience Institute, has reported on his success in developing new gene therapy tools in the current issue of Nature Biotechnology. His coauthors include Berkeley chemical engineering graduate students Narendra Maheshri and James Koerber, and Brian Kaspar of Ohio State's Columbus Children's Research Institute.
Gene therapy involves inserting genes into cells to program them to function in ways that are critical for health—whether it is producing a chemical like dopamine that is required for normal brain function but is depleted in people with Parkinson's disease, or producing specialized antibodies to kill cancer cells.
Schaffer and his colleagues are helping overcome one impediment to gene therapy—the human immune response. The immune system is an amazingly complex and effective system for protecting our bodies against bacteria, viruses, and the illnesses they cause. Yet in some cases, the immune response can get in the way of treating disease.
A gene therapy vector is some sort of microscopic device, often a modified virus, that has been engineered to deliver genes into living cells. Schaffer has devised a technique to circumvent the immune response to potential gene therapy vectors like the adeno-associated virus (AAV), a common, though innocuous, resident of the body. Using a technique based on directed evolution, Schaffer has created new versions of AAV that are good candidates for gene therapy vectors. Better yet, he has done so by forcing the virus itself to do the work.
A graphical representation of the adeno-associated virus (AAV). Schaffer's work has led to the creation of new versions of AAV that could one day assist in gene therapy.
"It is almost impossible for a researcher like me to rationally create gene therapy vectors than can slip past the immune defenses," says Schaffer. "The beauty of directed evolution is that we can use viral evolution to generate 'designer' gene delivery vectors without having to rationally design them ourselves."
AAV consists of two genes enclosed within a ball, or capsid, of proteins. The capsid proteins are what antibodies recognize, and Schaffer's goal was to alter the capsid proteins to allow them to bypass the immune response.
Schaffer's team first mimicked evolution by creating random mutations in AAV. By introducing small variations in the genes through an error-prone polymerase chain reaction (PCR) coupled with a test tube recombination technique, the researchers created new varieties of AAV.
After reassembling the mutant viruses inside their capsids, the researchers forced the AAV strains to run the gauntlet of rabbit blood serum rich in antibodies to AAV. Only the mutant viruses good at evading the antibodies survived the serum.
After passing the viruses three times through increasingly more potent serum, the researchers isolated the survivors and subjected them to another round of PCR that introduced more mutations. Then they forced these second-generation viruses to run the antibody gauntlet again.
After two generations of directed evolution, one strain of virus was 96 times more effective as a gene therapy vector than the wild AAV in cell culture, and two evolved strains survived injection into mice harboring nearly 1,000 times the level of antibodies normally required to neutralize the wild virus.
By sequencing the survivor strains, Schaffer and colleagues discovered that the capsid proteins of the survivors differed from those of the original strain by only seven amino acid building blocks, two of which were responsible for most of the altered interaction with antibodies.
"Starting from scratch, just trying to rationally decide which two amino acid changes to make on the virus, we could have never identified these two," Schaffer said. "Instead, we used the algorithm nature created called evolution to solve the problem for us. This virus is kind of a gift from nature–very safe and efficient–but nature never evolved it to be a human therapeutic. So we had to re-evolve it for that purpose."