Public discussion about the role of biotech and pharma in the pandemic tends to focus on one of two areas.
Vaccines are the great hope, and occupy center stage. Gilead’s remdesivir has captured considerable attention, largely by being the first approved drug for COVID-19, despite offering a limited benefit for hospitalized patients.
Off in the periphery of the public mind is an entire category of pharmaceutical intervention that many scientists believe has decent odds to pay off this fall – likely several months before a vaccine would arrive in the best-case scenario.
We’re talking today about therapeutic neutralizing antibodies. Several companies – Regeneron Pharmaceuticals, Eli Lilly, VIR Biotechnology, and Amgen among the leaders – have been sprinting ahead with recombinant antibodies that have shown an ability to neutralize the SARS-CoV-2 virus and which can be made and scaled up rather quickly. These treatments should, in theory, be able to knock down active infections in COVID-19 patients, and potentially could provide temporary prophylaxis (at least a month or two) for high-risk populations in need of protection, such as healthcare workers.
These kinds of antibodies play to the strength of the biotech industry. They can be manufactured at large scale. They should be amenable to repeat dosing. Since these treatments aren’t meant to provide long-lasting immunity like a vaccine, they can also be tested in rigorous studies that can quickly provide important answers – like whether the antibody or antibody cocktail immediately wipes out the virus, and whether it is safe.
The public isn’t focused on this particular battlefront at the moment, but therapeutic antibodies could provide an essential “bridge to a vaccine” as some in public health and industry hope.
To discuss the state of therapeutic antibody development, Timmerman Report correspondent Kshithija (KJ) Mulam spoke June 24 with David Weinreich of Regeneron Pharmaceuticals. He’s the company’s senior vice president of global clinical development.
Timmerman Report: Can you tell me a little bit about Regeneron’s antibody cocktail for COVID-19 and how it’s supposed to prevent viral escape?
David Weinreich: So in order for the virus, any virus to get into the cell, it has to have a protein on the surface that interacts with a receptor or protein on the cell surface in the host, in the human. In this case, that allows the virus to get in and ultimately destroy the cell, cause inflammation, replicate, make more viruses and so forth. For coronavirus, the protein is called spike. If you were to block the spike protein from interacting with the receptor on the human cell, you could prevent the virus from causing infection. You essentially block the cycle of the virus getting into the cell, replicating, causing damage, cell dies, more virus gets out. You break that cycle.
That protein is made up of a lot of amino acids. Some of them are more critical than others. Some of them are at parts of the protein that are not relevant to interacting with the receptor on the human cell. Some are critically involved in interacting with the cell. You want to interfere at a critical interaction between the protein in the virus and the protein on the human cell.
Small changes in the structure or the amino acids of the protein may or may not inhibit the ability for the protein on the virus to interact with the protein on the cell. But if they’re in the right spot, they may block the antibody from binding to the spike protein anymore. It all depends on where they are.
When we constructed our cocktail, we found two spots that are non-competing. That means the two antibodies do not fight for the same binding site on spike. Both of them independently inhibit the ability for the spike protein to bind to the human cell. So they’re non-competing, neutralizing anti-spike antibodies.
Why is that important? If you have one antibody, wherever it binds to that spike protein, you may get a subtle alteration in the protein structure and mutation that prevents that antibody from binding to the spike protein and neutralizing it.But it doesn’t prevent the spike protein from working by binding to the receptor on the human cell and causing infection. That’s the worry.
So then the next question is, is it a real worry? We’ve done this, we’ve created antibodies against a number of viruses. This isn’t our first time creating a cocktail. We have experience with a Regeneron antibody that targets respiratory factors — the respiratory syncytial virus (RSV). This is a virus that affects essentially premature babies and causes a pretty debilitating lung infection, and in that case, we did not use a cocktail: it was one antibody. And what we learned from our Phase III study, which failed, was that in in the time in which we ran that trial, a minor variant of a minor strain of the RSV virus that had a point mutation right at the spot where [the Regeneron antibody candidate] bound. That prohibited the binding of our antibody, so it didn’t work. And in the end, it happened to be a very prevalent variant, when we were running our trials, and the entire study failed.
Is this going to happen within human treatments to treat coronavirus? And the short answer is, we don’t know. We personally don’t want to take the risk, because what we don’t want to have happen is escape variants being generated. This is kind of the same idea of how do you treat really resistant infections like tuberculosis? You don’t treat them with one drug. You treat them with a cocktail. HIV? You treat them with a cocktail. And we’re following that same pattern. And there’s a variety of scientific reasons that we think it’s a better way to go so long as the antibodies are non-competing, and both neutralizing. Our cocktail has been designed specifically to achieve that.
TR: What is the clinical trial going to look like, and what are the parameters of evaluation, or the expected endpoints going to look like in the future?
DW: We’re going to run multiple clinical trials. The two therapeutic intervention trials are ongoing. What I mean by that is these are individuals who already have contracted the coronavirus. We take a swab from their nose, and we put it into a PCR machine, and it turns out positive, so we know they have the disease. Those trials were initiated last week [week of June 15]. We’re still in startup, but they’re going incredibly well, and we’re interested in a number of different potential endpoints.
You have to evaluate safety. We know a lot about our antibodies. We have structured our antibodies in a very controlled fashion, that one Regeneron antibody and the next Regeneron antibody share a lot of similarities from cell lines, to what the structure looks like, and so forth, except for the business end of what it’s actually binding to. While we haven’t had a problem with an antibody in the past, you never know until you’ve actually tested it in humans. So, the first thing is we have to make sure that there’s no untoward safety issues.
The second is we want to know whether or not we alter the viral kinetics of an infection that a patient presents at the hospital or at the doctor’s office. They’re positive. We can essentially tell how much virus is floating around, kind of like a viral load, although it’s measured by how many times the polymerase chain reaction machine has to cycle before it turns positive. The fewer times it cycles, the more virus there was, so lower numbers are higher viral types we can test over time. So that’s the control arm.
And you’ll generate a curve of essentially viral load versus time. We would expect that giving a large dose of our monoclonal antibody will dramatically alter that viral kinetics curve. It’s the reason why vaccines don’t instantaneously work. The first 10 or 14 days, you really don’t have a lot of coverage. It isn’t until the body’s immune system finally builds up that you have an immune response. When you give a big intravenous dose of our antibodies, you essentially have coverage that circulates through the body. You know, minutes, two hours, we would expect there to be a difference in the viral kinetics curve. We’re exploring that.
Finally, you would expect that the patient’s symptoms would either get better or not get as worse. In one of our trials, we’re taking patients who are already hospitalized, who already need supplemental oxygen. And we’re evaluating whether or not the patients get off oxygen faster and whether they improve in a second study. Patients don’t have to have symptoms. They’re positive for the virus, but they don’t have symptoms. So here we’re measuring whether or not they develop symptoms, whether or not they have to rush to medical care, where they have to get admitted to the hospital. So the clinical parameters are slightly different, but we’re looking for better clinical outcomes.
In addition to that, we’re also looking at studies in prophylaxis. Here it’s people who are not positive with coronavirus. They’re PCR negative. The question is, because of their particular circumstance, they might be at a very high likelihood for developing coronavirus.The first trial we’re going to be kicking off is in a household context.
So imagine this. Your significant other comes down with cough and fever, goes to the doctor. They test you. Your significant other is positive for coronavirus. But you don’t have a lot of symptoms. You’re not going to go into the hospital. They’re going to send you home. Call the doctor if it gets worse and self-quarantine. But in your household, you have your significant other, and let’s say, a couple college-age kids, so your household has four or five other people potentially there or intermingling with you. They’re not positive, butwell, they’re at very high risk for developing coronavirus unless you happen to live in 20,000 square feet, and you can send the person to the pool house, which is not the case for the vast majority of people. In that case, what we’re asking is, can you take the household contacts who are not positive, but there is a household contact who is positive? Can you give it to the people who are negative and prevent them from converting to become positive? And in that case, the primary endpoint is whether or not they turn PCR positive by the virology test.
Now, we’re also evaluating whether or not if they turn positive, do they develop symptoms? But our goal is to show that you prevent the transmission from infected individual to a household contact, and that proves the point that you can block transmission generally. That study is going to start in a couple weeks. For a number of reasons, it had to be slightly delayed. There are a number of other variations of this prophylaxis theme that we’ve been talking about, and we’re still contemplating exactly how to do this, such as can you do this in health care workers or can you do it in nursing homes? Every study brings its own operational challenge. We’re trying to get this done as quickly as possible; we think the quickest way to get this answer is the household contact study, so we’re prioritizing that.
All of this is a bridge to the definitive thing. We hope that there’s a real vaccine – those trials are kicking off. Those are going to be very large trials. They may take some time. But if we can deliver a therapeutic that prevents the bad outcome, that’s a win that helps you reopen the world while we wait for the vaccine.
TR: You mentioned that the cocktail has a neutralizing approach, and I know that there are companies working on single neutralizing antibody approaches. So what exactly are the pros and cons of going with a single neutralizing antibody versus multiple neutralizing antibodies?
DW: The risk is escape. So the risk is that a variant strain of this virus that has a mutation that renders it immune to a single antibody then becomes the dominant strain that’s out there. And this has already happened. There’s a slightly different variant that is out there. I believe it’s in Scotland. We’ll double check that. Compared to what’s here in the United States, it’s a little bit different. This is classic selection pressure. If you have a variant out there that is immune to your therapeutic, it will grow to become more of a dominant strain. And then your therapeutic is worthless.
If you have a cocktail in order for that to happen, so long as both antibodies are not competing and both neutralizing, you would have to simultaneously have two mutations in two different places in the viral protein at the same time for that to happen. That is exponentially less likely. It’s never zero. But it is a far more rare event. It’s the reason why in our Ebola cocktail, Regeneron EV3 against Ebola virus, we actually have three antibodies. Not one. Not even two. Because that virus is obviously incredibly deadly. It kills very, very quickly, and it also can mutate. So in this particular case, we went with three. We actually contemplated whether or not to go with more than two antibodies for coronavirus. It turns out that the spike protein isn’t all that large. And we could find multiple pairs of antibodies that met our criteria. They had to neutralize. They had to prevent the spike protein from binding to be receptor on the human cell and not compete. But we just couldn’t squeeze a third antibody in there. It always wound up competing with one of the other antibodies. And then it doesn’t add anything to the cocktail.
TR: There are other related programs that Regeneron has, and one is an IL-6 inhibitor Kevzara for fighting cytokine storms. What’s the latest on that program?
DW: There’s sort of three components to this. The first was, is there something on the shelf that you could potentially use? The second is the passive vaccination with an antibody, and then the third is a real vaccine. So really Kevzara, our IL-6 receptor antagonist, fell into bucket one. This was an off-the-shelf. We already had it. It’s already approved for rheumatoid arthritis. We saw some encouraging data from a single arm study coming out of China for a competitor molecule that is very, very similar. And we decided that this was worth testing in a rigorous fashion. We’ve already announced that the independent data monitoring committee made us make some alterations to the Kevzara study for safety reasons. And we stopped treating patients with so-called severe disease, which means that they’re in the hospital, but they’re on relatively low flow oxygen. The study is still ongoing in patients who are on a ventilator, who were on a ventilator when it initiated. We will have data this month from that part of the trial. And we still plan on having data this month in June.
TR: How scalable do you think Regeneron’s COVID-19 efforts are to a larger platform nationally?
DW: I’ll admit I’m a little biased, but I think they’re incredibly important. We have a proven track record for helping neutralizing antibodies for viral infections. We didn’t decide to get into coronavirus just for the sake of ‘everybody’s talking about it.’ We had already been in antibodies against viral diseases long before anybody knew about coronavirus. So we had the platform, the capability, the know-how to do this. We’re taking all of our learnings on how to get these trials up and running rapidly and get them recruited. And we have the manufacturing capability to scale this quite extensively. Now, I will admit we can’t treat everybody in the U.S., let alone everybody in the world. If this works, we would certainly need manufacturing support from other companies to be able to produce enough antibody to treat the world or at least big chunks of the world. I will say, without going into any details, we are having those conversations.
TR: So what do you think are strong contenders, I guess, in the industry beyond Regeneron in terms of therapeutic antibody production?
DW: This isn’t your typical situation where multiple companies are competing to develop an antibody for a target for some disease, and you know where one is trying to win over the other. This is different. The world needs something to work. And if it isn’t ours, then I want it to be somebody else’s, because we need something as a bridge to the vaccines, and I’m rooting for them. We are rooting for the vaccine companies too. We’re in this because we really believe, given our track record of doing this several other times, for RSV, MERS, Ebola. We have a track record. I have very high hopes that this is going to be the next one that works. We think we have a very strong track record that ours should work. But if it doesn’t, somebody else has to be there to save the world, whether it’s Lilly or Amgen or any one of these other companies. The only caveat, the only thing that sits in the back of my head, I just hope we don’t have so many people who just want to be in this because it’s become a political hot potato that everybody wants to participate, that everybody’s running their own clinical trials, and then no one actually gets enough data to prove whether or not their cocktail works. I don’t think we’re there. But that is the one thing that is a theoretical problem.
TR: Where do you see gaps maybe in industry efforts to create therapies for COVID-19, either production-wise or ideologically?
DW: I don’t think we have many gaps. I think that there are enough companies trying variations of a theme. We’ve got a lot of diversity in approaches. If something works, I think as a world we are going to have a production issue. The antibody doses you need, the milligrams of an antibody that we think you have to give are pretty sizable. Regeneron has a lot of manufacturing capacity, but it’s not enough. It’s going to require collaboration for whatever works to provide enough doses globally.
TR: Do you think that therapeutic antibody efforts are a long-term viable solution to COVID-19 ?
DW: I do not. I think there are a short-term bridge and then a long-term safety net. What do I mean by that? We would be better off if we had a vaccine like the pneumococcal vaccine, or even a regular old influenza vaccine. If it were good enough, that would generally prevent infection. It’s more economical. It’s more scalable. However, until that happens, the antibodies can be developed faster than the vaccines. It’s a good bridge until those are available. But then even after the vaccine becomes available, despite getting a vaccine, an individual patient may still develop the disease. If they do develop the disease, they have the possibility of transmitting the disease. And not everybody may get vaccinated for one reason or another. I think in the long term, an antibody cocktail, if it works, can fill that gap of patients who still develop the disease despite having been vaccinated, those who haven’t been vaccinated and get the disease. Because you have to remember, a vaccine takes time. If you become positive, and you haven’t gotten the vaccine, it’s too late to give the vaccine. Then you would have to intervene with a therapeutic antibody. So I see a long-term role, but it’s a smaller role once the vaccines have demonstrated efficacy. It’s a much bigger role early. That’s perfectly fine with us. We recognize that this is a bridge to an active vaccine. And that’s why we’re investing a lot of time and energy to try and develop this cocktail.