Impact of Affinity on the Activity of Immunomodulatory Antibodies
Mark Cragg, PhD, Chair in Experimental Cancer Biology, University of SouthHampton
Affinity Tuning Could Make Immune-Stimulatory Antibody Therapies More Effective
Developers working on antibodies designed to stimulate the immune system should consider engineering them to bind their targets less strongly, according to a new study.
For some antibodies, high affinity — the ability to tightly bind a target — is a useful characteristic.
But for others that are designed to boost the immune response, the opposite is true, according to Mark Cragg, PhD, Chair in Experimental Cancer Biology, University of Southampton in the UK.
Cragg made the claim at Antibody Engineering and Therapeutics Europe in June, citing the findings of a research paper published in Nature earlier this year.
“So with antibodies like rituximab or Herceptin or an antibody that you want to block like a checkpoint blockade receptor, you want something to bind very tightly and not come off,” he said.
“What wasn’t clear is about immune-stimulatory antibodies. Do we need or want a high level of affinity? And so that’s really the question that we asked.”
CD40 Antibodies
Cragg and his team used antibodies that target the CD40 receptor that is present on the surface of antigen presenting cells as the model primarily because they had previously made structural models.
“Because we had the crystal structure, we were able to guide where we should make mutations to then lower affinity. Chris [Christian M. Orr, a researcher at the UK synchrotron science facility, Diamond Light Source] basically made some predictions about where we could make those mutations,” Cragg said.
“We made a whole series of molecules with various mutations, and we got a very nice range of molecules that had different affinities. Interestingly, some of these were molecules we’d already got when we chimeric-ized the antibody. So we got a little shift in affinity when we changed some of those framework regions.”
Furthermore, “We went in with the crystal structure to change regions in the CDRs. So both of those properties could change the affinity. But overall, we basically had a series of about 12 different antibodies of different affinity,” Cragg said.
The team then tested these antibodies for activity using B-cell-based assays, and the key finding was that as affinity reduced, activity increased to a point.
“We’d clearly gone from the wild-type molecule to a much more agonistic molecule, which has lower affinity,” he said.
Cragg noted the findings remained consistent across other measures of activation, including CD23, CD86, and proliferation.
Therapeutic Potential
The team tested the theory in mouse cancer models, and the findings were also positive, Cragg said.
“We then went back to the in vivo assays [to ask] could we expand OT1s [engineered mouse immune cells used to measure immune response]. And essentially what we saw is that as we dropped affinity across the series, we got better OT1 expansion.
“We could do the same thing in the presence of the tumor and we could get better therapeutic efficacy with our lower affinity antibody,” he said.
Cragg added that “it does exactly the same thing in human cells — CD23, CD86, proliferation — you get the same bell-shaped curve. [In human] dendritic cells, we [get] exactly the same thing.”
The same approach was also tested on other antibodies, including the candidate therapy utomilumab, which is being developed to treat advanced cancers, he noted.
“We went through the same trick [process]. We didn’t have so many mutants — in this case, we only made three — but essentially had exactly the same property: dropping affinity, increased activity, and increased clustering.”
