Main

It started about 20 years ago, with the idea that antibodies would make good therapeutics; by targeting a protein on a diseased cell they should swiftly and specifically bring about the destruction of this cell. A compelling idea indeed, but according to Willem Stemmer—an antibody engineer—not a hugely successful one because of problems in antibody production and stability. Stemmer and his colleagues from the biotech company Avidia have now taken an approach—recently described in Nature Biotechnology—that deviates from classical antibody engineering. Instead of using large antibody scaffolds with one high-affinity domain for the target, they linked small protein domains, each with relatively low target affinity, resulting in a cumulative effect of strong binding to the target.

Their starting point were A domains, repetitive stretches of 35 amino acids, mainly found on the extracellular portions of human receptor proteins, which bind to different epitopes on the same target. Stemmer explains nature's ingenuity in this approach: “Combinatorial methods for creating binding proteins are the most efficient. Each domain by itself has a small affinity for the target but in combination you get strong binding through an avidity effect.” His group developed phage display libraries that started with the human repertoire of A domains and created a highly diverse pool of monomers by synthetic recombination. They then screened the monomers against a target protein; once candidates were found, they added another monomer and screened the new library of dimers against the target. After iteration, Stemmer obtained a trimer with very high binding affinity for its target protein, the cytokine interleukin 6 (Fig. 1). They appropriately named these molecules avimers, for avidity multimer, and showed that their anti–interleukin 6 avimer binds with high affinity and inhibits the proliferation of leukemia cells by stimulation with interleukin 6.

Figure 1
figure 1

Katie Ris

Avimers are made up of multiple small protein domains, each with many disulfide bridges and unique binding regions for the target protein.

Full size image

Stemmer argues that the combinatorial nature of avimers gives them unique possibilities as therapeutics, for example, in cancer therapy. “You can treat a disease at multiple different levels,” he says. “You can have domains that inhibit proteins necessary for extravasation and domains that induce killing of the tumor cell.” Although the cure for cancer with avimer therapy is not yet around the corner, their application to the clinic is not far off, and Stemmer is confident that clinical trials will begin next year on the use of their interleukin-6 avimer to treat autoimmune diseases.

One requirement for the clinical application of a molecule is low immunogenicity, a requirement the avimers fulfill, partly owing to their small size and partly because of a large number of disulfides. Work with other high-disulfide density proteins has shown that such proteins yield low immune responses, most likely because the presence of disulfide bonds makes the protein hard to cleave by intracellular proteases in antigen presenting cells. This cleavage is a prerequisite to the presentation of the peptides on the surface of these cells, which in turn triggers an immune response.

Another requirement for clinical application is the ability to manufacture the drug. Stemmer presents data showing the efficient production of avimers in bacteria, which should greatly reduce the cost of manufacturing relative to that for antibodies. In addition, the multiple disulfide bonds contribute to a high-temperature and stress stability, which may lead to longer drug shelf lives, and simplify the shipping and storage processes.

With their high target affinity, specificity and low immunogenicity, avimers appear to be well positioned for their first appearance in the clinic.