Infinimmune sees the human body as a kind of prototyping platform for antibody medicines, one that holds considerable advantages over model organisms, display libraries, or computational algorithms. Complete human antibodies naturally reflect te human milieu, which includes antibody targets that cannot be replicated outside the human body, as well as selection processes that help reduce immunogenicity and enhance pharmacodynamic properties, half-lives, and tissue distributions. Consequently, antibody medicines derived from complete human antibodies have the potential to be tightly targeted, minimally immunogenic, and highly potent.
According to Infinimmune, the human body conducts the equivalent of 100 billion antibody clinical trials every day, testing each antibody for safety and efficacy in parallel. The human body, then, has an enormous pipeline of antibody medicine candidates that Infinimmune can tap into during discovery and development.
Myriad possibilities
Whole human antibodies “have potential in several key areas,” says Wyatt McDonnell, PhD, CEO, Infinimmune. “The first is as direct drug products that are manufactured at scale for subcutaneous or intravenous administration.” That is Infinimmune’s primary focus, though the antibodies Infinimmune develops can be formatted into multispecific antibodies, antibody-drug conjugates, and other antibody formats.
Less obvious applications include cell therapy and B-cell engineering. Cell therapies today are limited by high development costs and scalability, and as a result relatively few patients currently benefit. Whole human antibodies, McDonnell suggests, may be able to help cell therapies reach the right target, potentially enhancing efficacy and safety and lowering costs. Infinimmune is beginning to explore some of those options through partnerships.
Asset development
Last December, Infinimmune partnered with Grid Therapeutics to identify antibody drug candidates in the antibody repertoires of exceptional survivors, that is, patients who had durable responses to cancer therapies. The companies intend to discover more drugs like GT-103, Grid Therapeutics’ lead candidate for the treatment of non-small cell lung cancer. GT-103 is currently in Phase II trials. Together, Grid and Infinimmune hope to discover which facets of antibody biology protect exceptional survivors, and whether there are other promising autoantibodies waiting to be found.
“One of the really fun things about this type of partnership,” McDonnell says, “is that we get to ask, ‘What are the most interesting human samples you can think of?’ Then we can go and learn what actually is present in the antibody repertoire. There are very few places where you can do research on the antibody repertoire and drug development at the same time.”
Infinimmune’s most advanced asset is a preclinical candidate that targets autoimmunity for skin indications. Industry-wide, only a few commercial-stage molecules for this indication are antibodies or antibody derivatives, and all emerged from transgenic or standard mice. “Our asset hits the same target,” he notes, “but from a different angle of attack because it comes from the human immune system.”
Consequently, McDonnell suggests this asset may be particularly competitive in chronic disease indications while bringing new assets with regulatory exclusivity and improved safety. “Elite human antibodies are some of the least immunogenic antibodies to have reached the clinic,” he observes. Human-derived antibodies have been patented and commercialized before, and Infinimmune is evaluating opportunities related to up-coming patent cliffs.
Another asset (for immuno-oncology) is in the discovery phase. Additional assets in the inflammation and immunology space are in preclinical development.
Digging deeper into the genome
McDonnell and four former colleagues from 10x Genomics formed the company in the summer of 2022. “We fished through the government’s public Sequence Read Archive and identified human antibodies that seemed impossible to generate in transgenic systems,” he recalls. “There are types of human antibodies that you will never see made in a mouse. More important, we found an unbelievable amount of diversity in the constant region of human antibodies. [This region is] substantially more edited in vivo by B cells than previously thought.”
Scientists hadn’t noticed this diversity, he hypothesizes, because they usually use only a limited number of molecular approaches to screen “that distracting 1014 number of binding regions, while totally forgetting about the tail end of the antibody, which is what actually dictates the function of the antibody in vivo.”
Infinimmune has discovered nearly 1,500 unique protein-coding immunoglobulin G1 “constant” sequences. “This is far more variation than is currently understood to exist in the entire human population,” he points out. “This is a rich sequence space for the exploration of novel functions.”
Human VHH antibodies
“We are finding molecules people think don’t exist in the human antibody repertoire because they haven’t had the right tools,” McDonnell says. Infinimmune, however, has its Complete Human immunosequencing technology. It has been used to find heavy-chain-only antibodies, which are commonly associated with camelids and sharks.
The human immune system produces canonical Y-shaped antibodies comprised of two heavy chains and two light chains. Nonetheless, according to McDonnell, when you look deep enough, you can see that “every human produces heavy-chain-only antibodies natively in their immune system.”
Heavy-chain-only antibodies have a single variable domain (VHH) attached to one constant domain, making them attractive to drug developers. “The cost of goods is lower because manufacturers only need to knock-in one protein to the production cell line,”
McDonnell explains. “They are smaller proteins, as well, so they can be easier to stabilize and purify.” Using whole, human-derived heavy-chain-only antibodies could confer additional benefits in terms of safety and efficacy. “Only four VHH-like molecules have reached the market,” McDonnell notes. “So, it’s an emerging modality.”
After the antibodies are sequenced, Infinimmune’s Anthrobody drug discovery platform enables the company to identify related antibody lineages and predict the target specificity for many antigens in parallel. Results specify where the antibody came from, the number of cells that produce that antibody or one closely related to it, and the target specificity of the antibody. With that data, developers can screen drug products that are part of the immune response to a given target.
“By starting with products that human B cells have edited in vivo, you are starting with a protein that [nature] has tested for cell expression inside mammalian cells,” McDonnell says. “So, unless an antibody is inherently expressible by a human mammalian cell, we won’t see it in our platform. We get survivorship bias of the best kind. The biology is doing the work for us.”
To do more, faster
Infinimmune was founded by scientists who “saw opportunities to apply advanced technology directly to antibody drug discovery,” McDonnell says. One of these opportunities came from research McDonnell conducted at Vanderbilt University on the human adaptive immune repertoire.
They published a paper reporting that they had observed a new property of antibodies called light chain coherence (Jaffe et al. Nature 2022; 611: 352–357). This property, they suggested, could be explained by sequence space constraints for light chains. They added that “antibody designers would be wise to actively look for optimal light chains used widely in nature, rather than focusing on the heavy chain alone.”
By the time the paper appeared, Infinimmune had launched. Although industry funding was scarce, Infinimmune closed a $12 million seed round in three months.
Infinimmune looks forward to forming more partnerships, including partnerships with AI companies and pharma companies. Infinimmune is also focused on building and selecting its own lead candidates for end-to-end development and commercialization. “For antibody companies like ours,” McDonnell declares, “that’s table stakes.”
