In the March 14, 2006, issue of the Proceedings of the National Academy of Sciences, researchers from Monrovia, Calif.-based biotech company Xencor Inc. reported on engineering antibodies so they bind more strongly to immune system effector cells.
The discovery "starts to let you think about targeting different kinds of cancer cells," Bassil Dahiyat, president and CEO of Xencor, told BioWorld Today. Currently, there are eight monoclonal antibodies for cancer treatment on the market, out of a total of 17 approved monoclonals. But for every approved antibody, there are several that have failed clinical trials for lack of efficacy.
Most optimization efforts to date have focused on improving an antibody’s binding to its target protein, for example, on a tumor cell. But an antibody can be conceptualized as a hinge between the tumor cell it marks for destruction and the immune system cells that do the actual dirty work of killing it. Xencor has focused on improving the binding of the latter type of cell, through the constant region of the antibody, also known as its Fc region.
There are three families of Fc receptors, and the richest of them is the Fc-gamma receptor family. Humans have at least five Fc-gamma receptors.
"All of them are related structurally, and all bind in essentially the same place," Dahiyat said. Fc receptors can mark antibody-bound cells for either cytotoxicity - the killing of antibody-bound cells by natural killer cells - or phagocytosis, where macrophages lunch on antibody-bound cells. Besides fetal Fc receptors, the third family, the C1q receptors, activates the complement cascade, which also marks antibody-bound cells for phagocytosis.
In the PNAS paper, the researchers used a combination of directed mutations to the Fc region and high-throughput screening to identify mutations that were likely to confer tighter antibody binding to immune system cells. Several mutants with changes in one to three amino acids of the Fc region showed tighter binding of antibodies, including trastuzumab (Herceptin, Genentech Inc.) rituximab (Rituxan, Genentech/Biogen Idec Inc.) and cetuximab (Erbitux, ImClone Systems Inc./Merck & Co. Inc.) to both natural killer cells and macrophages. That in turn increased both cytotoxicity and phagocytosis; in the strongest effects the Xencor team observed, when using triple mutants, cytotoxicity was increased more than 100 times.
The researchers tested whether the engineered antibodies were able to improve antibody binding to cells with low levels of antigen. For some antibodies, "even if the biological activity is there, it is not enough," Dahiyat said. For example, while the her2/neu receptor is expressed only on breast cancer cells, the drug currently is useful only to women who overexpress the receptor moderately to strongly; low levels of expression are not sufficient for the antibody to be effective. The Xencor team found that Fc-engineered Herceptin was able to mediate cytotoxicity at lower levels of her2 expression than the antibody as it is currently in clinical use.
The Xencor team also tested the ability of engineered Rituxan to deplete B cells in a monkey model of non-Hodgkin’s lymphoma.
Dahiyat said the in vivo study was designed to address the question, "If you use these souped-up antibodies, aren’t you going to start killing all these bystander cells?"
The answer, at least in this case, was no; Fc-engineered Rituxan was up to 50 times as effective as regular Rituxan in depleting B cells, and those B cells were targeted without a concomitant destruction of several T-cell types. There was an initial reduction in natural killer cells, at roughly the same level as occurs with regular Rituxan.
Dahiyat noted that the same mutations in the Fc region affect the binding of a wide range of antibodies, making the enhanced Fc regions recyclable. This is one advantage of tweaking the Fc site over the antibody-antigen interaction site, which must of course be tailored to each antigen. It also has made for a nice business model for Xencor, which grants nonexclusive licenses (four to date) to its engineered Fc regions. "We license the same Fcs" to different customers, Dahiyat said. "We don’t have to redesign the Fc over and over again,"
At the same time, it’s not quite a one-size-fits-all model; for example, different Fc engineering strategies can tune activity toward Fc-gamma receptors and away from the complement system, or vice versa.
Neither Xencor itself nor its licensees have any Fc-engineered antibodies in the clinic to date, but Xencor plans to enter the clinic with an antibody of its own in 2006. (See BioWorld Today, Oct. 20, 2005.)