Antibodies, in the most basic sense, are Y-shaped proteins that bind like a lock-and-key to a target. In the body, binding can trigger a cascade of actions that exterminate the invader. In the lab, antibodies can be anchors to fluorescent beacons for imaging, recognizing a molecule within a cell. But can antibodies do more than just bind?
Biolojic Design, based in Tel Aviv and Boston, aims to transform antibodies into intelligent medicinal solutions through computational design. By creating a pipeline of single and multi-specific antibodies that precisely target predetermined epitopes to carry out unique biological programs, the technology company believes it can offer therapies for various diseases. Biolojic employs its unique AI platform to design antibodies with novel therapeutic properties, focusing on cancer and autoimmune illnesses.
In April 2022, the company announced that an antibody designed using its AI technology entered a clinical trial in patients, becoming the first computationally designed antibody to be tested in humans. The trial will assess AU-007, a highly distinct monoclonal antibody that uses the body’s own interleukin-2 (IL-2) to attack solid cancer tumors. Aulos Bioscience, a Biolojic spinoff, founded with $40 million in Series A funding from ATP (Apple Tree Partners), is leading the clinical development of AU-007. Aulos is enrolling patients and performing clinical studies in Australia to release preliminary results later this year.
GEN Edge spoke with Yanay Ofran, PhD, founder and CEO of Biolojic Design and a former professor of computational biology at Bar-Ilan University, about why antibodies have so much potential as therapeutics.
GEN Edge: What spurred the founding of Biolojic Design?
Yanay Ofran: Biolojic Design was born in the basement of my laboratory at Bar-Ilan University near Tel Aviv, where we were developing computational ways to study biomolecular recognition—how proteins affect biology by specifically binding to a target. Proteins are the most sophisticated nanomachines that exist in nature.
However, the pharma world uses proteins, mostly antibodies, like small molecules that happen to be larger. They don’t consider these clever nanomachines that can be programmed to execute dynamic plans. This is crazy because antibodies are amazing tools! They can agonize and target. Antibodies can do one thing under one condition and do something else if conditions change. This is what proteins do in nature. Why not make them do the same things as drugs?
Our mission at Biolojic is to unlock the power that evolution gave proteins, including antibodies, to execute dynamic programs and use them in the clinic. This differentiates us from how the drug industry typically uses antibodies and AI technology in drug discovery. AI is often used to make discoveries faster and cheaper. We use AI not only to accelerate the process but also to design antibodies with novel capabilities that other technologies cannot discover.
With time and success, Biolojic grew out of the university. Initially, we approached pharma companies and told them to give us the most ambitious discovery projects that had failed using traditional methods. We offered to take a crack at these failed projects, and almost invariably, we were able to give the pharma partner the antibody they wanted. Now the dynamic inside pharma companies is such that once a project dies, it’s hard to resuscitate. We often gave them the antibody they were looking for two years earlier but were too late to get it into clinical development.
We realized two things; first, the best way to bring a revolutionary antibody into the clinic is to do it on our own. And secondly—this is more subtle but more profound—we should not limit ourselves to the creativity of pharma companies. If we believe antibodies have capabilities beyond the uses envisioned by the pharma industry, then we should design antibodies according to what we think they can do. Even though we had a quite powerful collaboration system that generated steady income years ago, we decided to focus on building our pipeline.
Fast forward a couple of years, we announced in April that the first ever computationally designed antibody—an antibody we designed—had entered human clinical trials. So, that’s a revolution that we pioneered. This antibody, AU007, is a conditional agonist/antagonist. Under some circumstances, for some cells, it is an antagonist, and it’s an agonist for other cells. This type of antibody was not considered possible. And this is one of the most conservative antibodies in our pipeline.
GEN Edge: Why has Biolojic Design classified these therapeutics as antibodies?
Ofran: Because in their sequence and structure they are indistinguishable from any other human antibody. Human genes—V(D)J genes—make antibodies. The mature antibodies in our immune system combine V(D)J sequences with somatic hypermutation. Maybe 5–10% of the sequence gets mutated during the development of the antibody against the antigen. We are keeping this set of rules. We are using human antibodies and introducing a handful of mutations. That’s it.
But again, it’s not that we are so clever. We are just using the fact that this system has evolved to make antibodies that are smart machines. We’re learning from billions of real examples how to extract functions from these sequences. That’s the Big Data and AI component. For example, we are learning what it takes to make an antibody bind to a chemotherapeutic in the periphery and what it takes to make an antibody release the chemotherapy in the tumor micro-environment in exchange for binding to a tumor-specific antigen. These are all things that everybody knows antibodies can do. We’re just putting them together into one developable antibody.
GEN Edge: What is Biolojic Design’s approach to drug development?
Ofran: First, we have shown that we can design antibodies with a conditional function. For example, an antibody that does one thing for one type of cell and another thing for another type of cell. We designed antibodies that bind to one conformation while avoiding another, which allows us to lock the target protein into a specific functional state while avoiding other states. This allows us to design agonistic antibodies. None of the approved antibodies in the clinic are agonists that activate a target by stabilizing an active conformation.
An important capability of our platform is to design precision binding—binding to what we want, where we want, making sure that we interfere in the biological process or initiate some biological intervention by controlling the target function. Most drugs today focus on disrupting protein function. We propose an approach that will allow antibodies to activate or initiate processes. Another capability of our platform is to design an antibody that binds to one target under one set of circumstances and to another target under another set of circumstances.
This brings me to what we call exchangers. It’s an antibody with a competitive binding site that can bind to either chemotherapy or a tumor-specific antigen. In the periphery, it acts as a mop to the chemotherapy, binding to the molecules and preventing them from doing any harm to healthy tissue. When the antibody gets to the tumor micro-environment, it finds the tumor-specific antigen that it likes, drops the chemotherapy, and binds to the tumor-specific antigen. Thus, increases the concentration of the chemotherapy around the tumor, while also attacking the tumor through the validated mechanism of the tumor-specific antigen. That’s one example.
GEN Edge: What is Biolojic Design targeting with these antibodies?
Ofran: Many companies look for new biology and novel targets to attack. That comes with a very high risk. After all, new biology often turns out to be way more complex than we initially thought. Most new targets fail. Instead of saying, “Let’s tackle new biology with old technology,” we’re saying, “Let’s tackle old biology with new technology.”
We focus on validated, well-understood targets that have delivered suboptimal benefits. Once we understand why, we devise a clever antibody to solve the problem. This is a nice business strategy with lower risk. We are working on validated biology—the targets, assays, animal models, and pathways have all been paved—but the differentiation is high because our antibodies are doing something nobody else does.
GEN Edge: What clinical indications is Biolojic Design pursuing?
Ofran: This technology is relevant for any disease area. We are focusing our internal pipeline on the immune system—inflammatory disease and immuno-oncology therapies that fight cancer. We think the best way to demonstrate the value of our capabilities is in control of the immune system. But there are many other indications—brain diseases, metabolic diseases, fibrosis, and other diseases—where our technology can revolutionize care.
For example, we are working with the scientists of Eli Lilly about how our technology of smart antibodies can change the care of diabetes. We came up with the idea of designing an antibody to do something very unusual, and Eli Lilly will develop it. So that’s our structure: our internal pipeline focuses on validated targets in the immune system, and anything involving a new target or a different disease area we will pursue with partners.
GEN Edge: What is Biolojic Design’s vision?
Ofran: We want to get to the point where drugs are smarter. When we think about smart technologies, like smart cars, phones, and homes, these technologies allow some apparatus to respond to environmental changes. It’s funny that we do it with phones, cars, and homes, but we don’t do it with drugs. Our vision is to develop smarter drugs. It will be great for patients because it could bring about a whole new generation of therapeutics that are safer and more efficacious.
We designed the first AI-designed antibody to be tested in humans, AU-007, in our lab in the south of Tel Aviv, and we spun it out into Aulos Bioscience—a company that we formed with Apple Tree Partners. The next antibody from our platform to enter the clinic will be wholly owned by Biolojic Design, and we believe it could revolutionize the treatment of inflammatory diseases, such as asthma. The four or five assets expected to follow could change the way we harness the immune system to attack cancer. Each of them will validate a new capability of the platform, allowing us to treat patients with the most severe disease more effectively.
