A revolutionary new way to make monoclonal antibodies (MAbs) was born during a walk on the beach in La Jolla, Calif. During that stroll in 1984, Steve Benkovic, a professor of organic chemistry from Penn State University, and Richard Lerner, director of the Research Institute of Scripps Clinic in La Jolla, vented their frustrations - frustrations that arose from screening hybridoma-derived MAbs for catalytic (enzyme) activity.
Monoclonal antibodies are of vital importance in diagnostics because they provide a targeted way of recognizing a molecule. Benkovic and Lerner's problem was that because enzymes catalyze very specific types of reactions, the amino acid sequence of a particular catalytic antibody also had to be very specific. Hybridomas just weren't providing enough different antibodies to choose from.
Says Benkovic, "It had become obvious to both of us, in giving talks, that we couldn't answer two questions. First, how many antibodies in a response show catalytic activity? We would screen 100 MAbs at a time, and sometimes find a few with catalytic activity. It was too small a statistical number."
The second stumbling block was that a mammal might not normally manufacture a catalytic antibody, and therefore a way had to be found to tap the immune system's potential to do so. "We knew we were throwing away lots of information," says Benkovic. "The immune response could make millions of potential antibodies, and we were looking at not even a fraction of that," he adds.
The investigators' seaside conversation clarified what was missing in their research: the need to generate monoclonal antibodies much faster. At the time, Benkovic was spending two to three months each year at Scripps with Lerner, working on catalytic antibodies. It proved a productive pairing.
"We are from different backgrounds, but the two of us can cross over into our respective disciplines quite easily. We were so successful together that we decided to pool our efforts from then on, and the two labs would work together," recalls Benkovic.
Benkovic was trained as an organic chemist, and Lerner as an immunologist. Their papers on catalytic antibodies obtained via hybridomas poured out in the mid- and late 1980s, as the team geared up to pioneer a better approach. Among their many publications were articles in Science (244:437, 1989), Proceedings of the National Academy of Sciences (PNAS) (85:5355, 1988), and Journal of the American Chemical Society (JACS) (110:4835, 1988). Other research done by the team was cited in Science (237: 1041, 1987 and 241:1188, 1988), PNAS (83:6736), and JACS (110: 2282, 1988).
Also hot on the trail of genetically engineered antibodies is Greg Winter, of the Medical Research Council, Cambridge, England, who created libraries based on the variable heavy chains. Chemist Peter Schultz of the University of California, Berkeley, is the second major force in catalytic antibody research. He is so excited by the team's approach that he recently came down to Lerner's lab in La Jolla to learn how to use it.
Genetic engineering provided a key step to the researchers' combined approach. It could generate the vast numbers of antibodies needed to hunt down those with catalytic activity in less time, and could produce pure human antibodies, circumventing the immune rejection problems associated with rodent-based antibodies. "To really make the system work," says Lerner, "we realized it's important to assay at the level of the bacteriophage plaque."
They planned to use recombinant DNA techniques to endow phage with the portions of the antibody molecule that forms the all-important antigen binding site (heavy and light variable chains), and then entice the antibodies to mix and match the sequences. This is precisely what happens to the immunoglobulin genes in the immune system's B cells when they are stimulated to churn out antibodies to specific antigens.
And so was born the idea of "combinatorial libraries," collections of E. coli bearing bacteriophage lambda that harbor heavy and light chain genes - the information to manufacture the antigen binding portions of antibodies.
But building combinatorial libraries required a team with broader expertise than that of the Lerner-Benkovic partnership. Enter Stratagene, a La Jolla-based biotech company down the road from Scripps. "Richard Lerner called and asked if we knew a way to make antibody libraries in E. coli," recalls Joe Sorge, Stratagene's chief executive officer. "Together we came up with the methodology, and in those first six months, there were perhaps half a dozen scientists from Scripps and half a dozen from Strategene working on it. We learned how to make the libraries, which entailed using the polymerase chain reaction to amplify the immunoglobulin genes and different tricks to express the antibodies," adds Sorge, who contributed his phage expertise and, later, business acumen to the fledgling technology.
The rather sophisticated genetic engineering fell to Scripps' William Huse, a molecular biologist, who, along with Lakshmi Sastry, a chemist, built libraries for both heavy and light antibody chains and combined them. Huse makes light of what must have been very exacting work. "I figured if we put the proper orientation of restriction sites in the [phage lambda] vector, it would simply combine light and heavy libraries, and generate hundreds of millions of combinations. That's more or less what the immune system does in generating different B cells. Actually, combining the heavy and light libraries was the simple part," says Huse.
It worked. The two libraries could indeed be mixed and cut in such a way that only those bearing heavy and light chain genes survived. With the problem of rapidly generating numerous and diverse MAbs solved, the next step was to find a way to select antibodies that would bind to a specific antigen.
This was the job of Dennis Burton and Angray Kang, both on sabbatical leave from the Krebs Institute at the University in Sheffield, England. They perfected the phage assay system, based on a nitrocellulose filter binding technique. The method involved a labeled version of an antigen with which a mouse that donated the original genetic information had been immunized. The antigen, p-nitrophenyl phosphon-amidate antigen 1 (NPN), was known to successfully induce MAb formation. A paper in the Dec. 8, 1989, issue of Science (W.D. Huce et al., 246:1275-81) presented the details of the elegant, multifaceted work.
A compelling idea based on existing tools that solves a pressing problem, combined with a team whose members' scientific talents mesh and complement each other beautifully, leads to the creation of a very marketable technology. But the collaboration between the nonprofit Scripps team members and those at Stratagene was apparently short-lived, a brief coming together to conceive, gestate, and give birth to the technology.
Since their landmark collaboration, the academic researchers and the scientists from the entrepreneurial firm have taken completely different directions in their pursuit of the combinatorial library approach to generating MAbs.
The divergence in ideologies was apparent in December, when the publication of the seminal Science paper dovetailed with Stratagene's announcement that it was spinning off a subsidiary, Stratacyte, to deal exclusively with the genetic engineering of MAbs.
The business plan is three-pronged, says Sorge: to license the technology nonexclusively for a "reasonable" price, to offer contract research services, and, in its own research and development efforts, to discover human MAbs with possible therapeutic value. Stratacyte started with 10 scientists from the parent company, but the subsidiary is actively recruiting and hopes to at least double that number by June, according to the firm's CEO.
The Scripps faction, by contrast, seems more concerned with digging in and using the technology to fish out rare antibodies than with business details. But there is no acrimony involved at all, Lerner maintains, just different approaches to both the work's availability and applications. Lerner, Huse, and Benkovic each said he would like to just give the lambda vectors to any researcher who wants them. "It's a very powerful technique, and it's user-friendly. People can just walk in and do it," says Lerner.
Yet even the Scripps team is diverging in their interests. "Benkovic and I will be pursuing catalytic antibodies, full stop, looking for cofactors," says Lerner. While both scientists are interested in the stereospecificity of antibodies, Lerner wants to concentrate on catalysis of hydrolytic reactions, and Benkovic on nonhydrolytic reactions. As for the other team members, says Lerner, "Bill Huse and Dennis Burton are working on human antibodies; that's their thing. A lot of people will be going to human antibodies."
Although the Scripps researchers appreciate the value of being able to generate MAbs using recombinant DNA technology, they seem a little surprised at the enthusiasm in the biotech community for their combinatorial libraries.
And although their behavior sometimes resembles that of little kids given a vast new set of blocks, the researchers haven't lost site of the science or the thrill of discovery. "It's like a runaway train. But what's important is to keep doing good experiments," Lerner concludes. And to keep taking those fruitful walks on the beach.
Ricki Lewis teaches biology at the State University of New York, Albany.