Transcript of interview with Professor Barton Haynes* by Dr Lara Marks, 24 June 2021

(this transcript has been edited for clarity and brevity)

*Director, Duke Human Vaccine Institute, Professor of Medicine and Immunology Duke University

Lara

I really would appreciate you starting from scratch to say where you've come from in relation to HIV work and how the vaccine you're contemplating [for coronaviruses] works.

Barton

Let's talk about HIV for just a second. So HIV is an RNA virus, just like COVID is an RNA virus. [RNA] is [the] genetic material that it uses to replicate. HIV is one of the fastest evolving life forms on Earth, so there are billions of different strains of the virus. And so for us to make an antibody against one particular [HIV] strain doesn't help anybody. [What we need is for the body] to make antibodies against regions of the virus that are relatively conserved so that the antibodies will react with many, many, many different strains. Those antibodies are called 'broadly neutralising antibodies'. Dennis Burton at Scripps and Herman Katinger in Vienna isolated the first HIV broadly neutralising antibodies (Burton).

Lara

Right.

Barton

Our strategy for the HIV vaccine is to make a vaccine that will induce antibodies against the conserved regions on the envelope, which is the target of broadly neutralising antibodies. The HIV envelope is the equivalent of the spike protein for COVID-19. For a pancoronavirus vaccine our strategy is to make antibodies against conserved SARS-CoV-2 regions.

HIV has evolved in primates for over 30,000 years and jumped from the great apes into humans only within the last 100 years; it has evolved evasion of the primate immune system by looking like ourselves, so that our bodies don't readily recognise these broad conserved spots as foreign. And when we do recognise them as foreign, the sites are so unusual and they're covered in sugars so that the space for the antibodies to get into is very narrow. So only a certain kind of antibody can get in there that has a certain shape. And that shape frequently looks like a type of antibodies that can attack our bodies and therefore is disfavored by our immune system (Williams et al).

Since 1985 I have been working to develop a successful HIV vaccine. It took 20 years to figure out what the problem was. So, since about 2005, what we've been doing… is to study the pathways of development of broadly neutralising antibodies. More recently, we have been designing pieces of the HIV envelope that will stimulate each antibody development step along the way, and guide B cells that make broadly neutralising antibodies to acquire the changes, called mutations, that broadly neutralising antibodies need to be successful in blocking HIV. The way that antibodies get stronger is they acquire mutations that make them fit tighter and tighter to the pathogen. We design a series of sequential immunogens to lead the antibody part of the immune system where we want it to go (Saunders et al, 2019).

There's never been a vaccine that has required this level of immune control. It's very difficult. But once we learn how to do it and learn what the rules are, then not only will we be able to make an HIV vaccine, but also the universal flu vaccine, and we hope other difficult-to-make vaccines will be approachable with this strategy.

Lara

Is the flu virus as fast evolving in terms of mutations as HIV. As I understand it, HIV is much bigger in terms of the mutations?

Barton

Of the three viruses, HIV, influenza and SARS-CoV-2, HIV is the fastest in acquiring mutations. Influenza is evolving new mutations but not as fast as HIV, and SARS-CoV-2 is slower but is certainly also evolving mutant viruses. And SARS-CoV-2 is picking up steam. The more people that don't get vaccinated, the more time it has to mutate, and in doing so can cause more people to become sick.

Lara

Do you think the mutations are gathering speed over time?

Barton

Yes, it appears so. You look at the timing of what happened during the first part of the epidemic and then what happened starting last September, the rate of SARS-CoV-2 mutations began to increase and that's when the Alpha variant, the Beta variant, and other variants of concern [were spotted].

Lara

And did you expect that?

Barton

Yes, SARS-CoV-2 is an RNA virus, and RNA viruses are prone to mutate.

Lara

Right

Barton

There are two strategies for making a beta pancoronavirus vaccine: one, induction of broad neutralising antibodies, and two, put pieces of all the types of viruses you want to neutralise in one vaccine-called a multivalent approach. We are trying to use the first strategy to induce broad neutralising antibodies with our approach, and we happened upon a design that brought out this broad neutralising antibody response.

Now the second strategy for making a pan-coronavirus vaccine is to use the spike for the Beta variant and the Delta variant and mix them together for variant specific responses. Or you could make a messenger RNA for spikes of all these different variants. The problem is there is always going to be another variant as new variants can evolve.

Lara

And would you be able to code for all those variants in one mRNA vaccine?

Barton

No, for an mRNA vaccine you would just mix them all together.

Lara

Okay.

Barton

[Now] regarding the vaccine that Kevin Saunders and I and our team just published in Nature, we originally made it to be a boost for existing COVID vaccines, but it turned out that it preferentially induced antibodies against a broad neutralising site on SARS CoV-2 (Saunders et al, 2021).

Lara

So originally you were looking for a booster for SARS-CoV-2, but it turned out to elicit this real broad neutralising antibody?

Barton

Correct. That's the essence of the vaccine. [We achieved this because from the start we focused] on the concept of broad neutralising antibodies that will hit lots of variants. That's inherent in our [HIV] strategy. So for HIV, as I said, if you take an HIV variant and induce a neutralising antibody that's specific for that one variant that's not going to help prevent HIV, because there are millions of other variants out there. We've got to have HIV broad neutralising antibodies to neutralise HIV variants. In contrast, for SARS-CoV-2 there are not that many clinically relevant circulating variants of concern. But there are a myriad of animal coronaviruses that may jump to humans in the future. So one strategy for a pancoronavirus vaccine is to induce broad neutralising antibodies.

Lara

Okay.

Barton

The pancoronavirus vaccine that Kevin Saunders at Duke and I reported in the Nature paper is a protein vaccine. Kevin took a scaffold protein called a ferritin molecule and put 24 copies of the spike receptor binding domain (RBD) on it. Ferritin is a bacterial protein that self-assembles. The RBD is what COVID neutralising antibodies target. In our HIV work, we used a similar strategy to induce HIV broad neutralising antibodies. We just shifted the same strategy over to COVID vaccine work. And instead of putting the HIV envelope on the scaffold, or different pieces of the envelope, we put the SARS-CoV-2 RBD on it.

Lara

Right.

Barton

Kevin and I found in the HIV vaccine work that nanoparticles that have multiple arrayed HIV envelope proteins on them give higher titres of neutralising antibodies than just one envelope by itself. We decided to try the same strategy to make a potent boost for existing COVID vaccines. We reasoned, back in March 2020 when this work began, that if the current COVID-19 vaccine candidates vaccines did not induce sufficient titres of neutralising antibodies or were not durable, then we wanted to have a boost that would raise neutralising titres. In monkeys, what we were getting were very high titres to the HIV envelope, so we decided to try the same nanoparticle strategy with CoV-2.

Data emerged last winter that the SARS-CoV-2 variants of concern like the Alpha, Beta and now the Delta virus can be neutralised by antibodies induced by the current vaccines but at reduced potency. In the case of the Beta Cov-2 variant, it is six to eight fold less sensitive to current COVID-19 vaccines. But if you start with a high enough titre to the variant it is still sufficient to neutralise the variant, and hopefully protects against severe disease as well. So we thought, let's work to develop something that will induce very high neutralising antibody titres.

Lara

Do you mean titres of antibodies within the serum that you're measuring?

Barton

Yes. Serum.

Lara

So the nanoparticle you're developing, were you aiming to get those very high titres?

Barton

Right. What we showed in the Nature paper was when we immunised monkeys with the nanoparticles plus a strong adjuvant, and compared it with animals immunised with a vaccine that was designed to be similar to the existing mRNA vaccines, the mRNA vaccine induced titres in the monkeys of approximately 7000. But with the protein nanoparticle we had titres averaging 40,000 and one of the monkeys had a titre of 162,000. We know monkeys can respond to vaccines better than humans, but it usually is just one order of magnitude better. So if we found anything approaching these titres in humans it would be an advance.

Next we tested our vaccine against bat viruses, as well as against the SARS-CoV original virus from 2003. Our RBD-nanoparticle vaccine-induced antibodies [that] neutralised all of the animal coronaviruses tested and CoV-2 variants of concern as well. The antibodies induced by this vaccine hit a broad spectrum of what's called group 2B beta coronaviruses. Group 2B beta coronaviruses are SARS CoV-2, SARS CoV, the original SARS virus in 2003, and bat viruses that [also] use ACE2 as the receptor. The beta coronavirus group 2C are Middle Eastern Respiratory Syndrome (MERS)-like viruses; their RBD is only approximately 40% similar to SARS-CoV-2, whereas the original SARS virus RBD and SARS CoV-2 are about 80% similar.

What we highlighted in the Nature paper was that our vaccine with the SARS CoV-2 RBD was inducing a broad neutralising antibody specificity that neutralises quite well both the SARS-CoV-2 variants of concern and other group 2B human and animal beta coronaviruses. In contrast, when we immunised with one spike encoded for by a messenger RNA, we did not see as high serum levels of this broad neutralising antibody. Therefore, the protein RBD nanoparticle vaccine, appeared to preferentially induce a broad neutralising antibody response to group 2B viruses.

Lara

Right. I was also speaking to Andrew Ward about the 2P [2 prolene] and what he found on a spike protein and how that's used within the mRNA vaccines. Really what you're doing is you're developing vaccines for that conserved part of the virus. But what you're doing is slightly different isn't it?

Barton

What Barney Graham and Jason McClellan at the NIH found is that when they tried to make the MERS spike protein, it was very difficult to express. Putting the two prolene stabilisation mutations allowed it to express better (Pallesen et al). And now Jason McClellan has what's called HexaPro, which is six prolenes (Goldsmith et al). This stabilises the spike protein even better. Actually, the two prolene set of mutations is good for MERS spike expression, but less so for SARS-CoV2. But it's good enough so that the vaccines that use it work well and we're all grateful for that.

Lara

Right. But that's only step one. The problem is that then if these start to mutate you need a more powerful approach, which is what you're talking about here.

Barton

That's right. The two prolenes are down in the middle part of the spike protein. What we're talking about is just using the RBD that's right at the top of the SARS-CoV-2 virus and just using that alone, not the full spike on the surface of this ferritin nanoparticle.

Lara

Okay. Now, how did you work out that was the bit to go for?

Barton

Well, that was known, because that's the RBD part of the SARS-CoV-2 spike that binds to the host cell receptor called the ACE2 molecule. So we compared our RBD nanoparticle with other vaccine forms in the Nature paper, and we found that those that had the multimeric RBD had the highest neutralising titres.

Lara

And when you say you tested them, how do you go about that? If I was coming in, totally new to the laboratory, and I was watching your process, how would you show me how you make that?

Barton

Kevin Saunders at Duke made a gene that encodes for a piece of this bacterial protein called ferritin. Ferritin comes as 24 pieces that self-assembles [into] a ball. It looks like a soccer ball. And on that ball, each of those 24 pieces have a place where you can put a linker, several amino acids, to which an enzyme can bind and couple something to it. Then one takes another gene and we make the SARS-CoV-2 RBD. And on to that is a linker that will join to the linker on the ferritin. Then an enzyme called sortase is added that will join those two ends together. The result is a sphere with 24 pieces of the CoV-2 RBD on it.

Lara

And how long does it take to assemble that in time?

Barton

The lovely part of it is that once you've made the gene and you learn to express the soccer ball, and you've made the gene and enzyme sortase, then it takes only a few weeks to make the RBD gene and express the protein and hook it onto the ferritin sphere.

We did this for the HIV envelope and then another bug, the SARS-CoV-2, came along. We [already] had the soccer ball, the ferritin sphere, and the enzyme. Kevin Saunders just made the SARS-CoV-2 RBD nanoparticle, and we immediately immunised monkeys to see how it worked. This started last fall and we got quick results because we had all of our monkeys ready to go. And by February, we were looking at these data and said, 'Oh, my goodness'.

Lara

You must have been stunned

Barton

Kevin and I were very surprised, and we felt very lucky because we weren't trying to make something that neutralised a whole group of coronaviruses. We were just trying to make a boost for SARS CoV-2 when we saw the broad neutralisation by the serum. It took us another three months to figure out why it worked, and that was because it was inducing cross-reactive neutralising antibodies.

We knew what antibody we were looking for because we had spent the last year isolating over 2000 antibodies from either CoV or CoV-2 previously infected or convalescent individuals. We have a huge group of antibodies that we've isolated from these folks [which] we screened for those that cross-reacted. Because the first thing we had to do to make a pancoronavirus vaccine was to find out if that can even be done. We knew there were broad neutralising antibodies for HIV, and the first question we asked in our pancoronavirus work was 'Are there such things as broad neutralising antibodies for beta coronaviruses?'

By screening many coronavirus antibodies from several coronavirus antibody-positive individuals we found a cross-reactive antibody called DH1047 that had all the qualities that we wanted. And we were setting about to use it as a template to design a vaccine that would induce cross-reactive coronavirus antibodies when we found that our RBD-nanoparticle induced cross-reactive antibodies. We used the DH1047 antibody to tell that this type of broad neutralising antibody was in the vaccine serum, using serum competition assays.

Lara

Was there a particular patient you found that had this neutralising antibody? Or is it just that you've got a whole series of samples that you were testing?

Barton

This was a blood sample provided to us from a collaborator that the collaborator had studied, and this person had the original SARS CoV infection many years ago.

Lara

Okay, but did they also have SARS-CoV-2 as well?

Barton

No. We were looking for cross-reactive coronavirus antibodies. And that was the beauty of studying someone with SARS CoV from the past. We figured if we got an antibody that bound to both SARS-CoV and SARS CoV-2 RBDs, then by definition it would be a cross-reactive antibody.

Lara

Amazing. So basically, what you're going for is to understand the host defence system and then work backwards to then work out the domain.

Barton

That is the whole secret we believe will be critical to developing the AIDS vaccine. And that is why so much of what's been done for the AIDS vaccine has helped the COVID vaccine effort. Because the AIDS vaccine is the first vaccine ever in the history of vaccinology where the success of making that vaccine depended upon understanding at a very deep level the host virus interactions.

Do you know the history of the Merck Vaccine Institute and the leader of that Institute and all the vaccines that came out of there?

Lara

Do you mean Maurice Hilleman?

Barton

Yes. Maurice was a dear friend and passed away in 2005. He was on many of our committees during the very early stages of the HIV vaccine development in the late 1980s and 1990s. When I was a young person, I was a member of many NIH committee meetings on HIV vaccine development and Maurice would get so frustrated ... he always challenged the field to work quickly.

HIV had been discovered by Françoise Barré-Sinoussi and Luc Montagnier in 1983. I was on one of those original papers that Bob Gallo showed HIV was the cause of AIDS in 1984. At that time, for the vaccine we all thought that we're just going to do just like the hepatitis B vaccine, that is to make the HIV envelope and mix it with an adjuvant, put it in monkeys, and induce protective neutralising antibodies. That would be the vaccine, and we'd be done in two years.

And then the field realised that all the viruses that we were working with were the wrong viruses. They were all lab-adapted viruses, and they all had open envelopes which the natural viruses did not have. The envelopes of all natural viruses were closed and the envelope regions available on clinically relevant viruses are the conserved, broad neutralising epitopes. It took another 20 years to figure out why the body didn't want to make broad neutralising antibodies, which we did with the Center for HIV/AIDS Vaccine Immunology (CHAVI) programme and the Collaboration for HIV/AIDS Vaccine Immunology funded by the Gates Foundation from 2005 to 2012. So we started working on the host-HIV biology in about 2000, and by 2013, we'd figured out the many reasons why broad neutralising antibodies are rarely made (Mascola, Haynes).

Lara

So what would Hilleman say now?

Barton

He'd still be saying that we are not working fast enough, but I think he would be pleased. He went along with us at the beginning of the journey of the difficult neutralising antibody biology and was a great supporter of all that we were doing. Maurice is one of my major inspirations and role models. He made more successful vaccines than anyone else. That have saved so many lives.

Lara

I studied the history of the hepatitis B vaccine and actually that's one thing I want to go back to working on is the breakthrough with a subunit of the virus that was linked to Baruch Blumberg’s work.

But here, you've gone a step backwards to look at the immune system and look at how that reacts with the pathogen.

Barton

Yes, the fundamental issue is how do you make a vaccine to induce an immune response that is disfavored by immune tolerance mechanisms, the frequency of the precursors are few for the kinds of antibodies that need to be made. And even when the antibodies are made, they require very unusual changes in the antibody. They don't occur very often. The only way to capture those changes is to have the right envelope immunogens present, such that when the antibodies develop the right mutations the envelope can stimulate those B cells making the right antibodies. If the right envelope is not present in the vaccine, then the B cells making the right antibodies will die.

Lara

What is so different between SARS-CoV-2 and HIV vaccines?

Barton

This is a very important question which I have written about (Haynes). There are two major differences in SARS CoV-2 and HIV vaccines. First, for HIV, the body does not want to make broad neutralising antibodies to the HIV envelope, whereas with SARS-CoV-2, the body is quite happy to make broadly neutralising antibodies to the coronavirus vaccine. And the diversity of coronavirus is limited, such that even though our pancoronavirus vaccine also makes a SARS-CoV-2-specific response, it also induces antibodies that neutralise other group 2B coronaviruses.

The second reason the HIV vaccine is so difficult is that HIV is an integrating RNA retrovirus. HIV inserts its genes into our own genetic material, and once it gets there, it is invisible to the immune system. It does so within about the first three days of transmission. So the host does not have time to make a secondary immune response. Therefore once in the genetic material of the host the immune system cannot clear it.

In contrast, all vaccines that have ever been successfully made, including the SARS CoV2 vaccines, have been made against viruses that are non-integrating viruses. Most people clear the virus for SARS-CoV-2 and for influenza and other pathogens with the vaccines that we have. In contrast, for an HIV vaccine to be successful, we have to prevent infection 100%. If we prevent 99% of the viruses from infecting a person, but 1% of HIV virions infect the host, then the vaccine will fail and irreversible infection occurs. Therefore, an HIV vaccine has to be 100% effective for preventing transmission, a very high bar.

That's the big difference. Those two things: the body doesn't want to make the protective antibody responses and the virus is an integrating-virus, whereas the SARS-CoV-2 virus does make protective antibody responses and it's a non-integrating virus.

Lara

That is fascinating. How long do you think it will take for this new pancoronavirus group 2B vaccine to start being tested in clinical trials in humans?

Barton

Well, it depends on when we get the money to make it. We’ve got three grants submitted right now.

Lara

And if you have that go ahead, how long will it take?

Barton

Probably a year. It's a protein. We've got a manufacturing facility in the Duke Vaccine Institute which can do Fill/Finish, the process that involves putting vaccine doses into vials to prepare them for shipment. So we can make our own. The limiting factors now are raw materials because they're all going for existing COVID vaccines.

Lara

Would you then partner with a company to develop it further?

Barton

Oh, absolutely. Yes. We'll partner whenever we can and we're talking to folks now about that. But if we can get it into a phase 1 trial that will speed everything up.

Lara

And what do you think the reception has been to this work generally? I mean, how are people looking at what you've done?

Barton

It's been good. You can go online and see the White House Press Conference on May 13th on COVID. It was discussed there by Tony Fauci (White House)

References

Burton, D (20 Nov 2019) 'Broadly neutralising HIV antibodies: from the beginning', Nature Microbiology. Back

Goldmsith, CL, et al (18 Sept 2020) 'Structure-based design of prefusion-stabilized SARS-CoV-2 spikes', Science, 369/6510, 1501-05. Back

Haynes, B (16 July 2021) 'SARS-CoV-2 and HIV-1 — a tale of two vaccines', Nature Reviews Immunology. Back

Mascola, JR , Haynes BF (July 2013)'HIV-1 neutralizing antibodies: understanding nature's pathways', Immunology Review, 254/1, 225-44, doi: 10.1111/imr.12075.Back

Pallesen, J, et al (29 Aug 2017) 'Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen', PNAS, 114/35, E7348-357. Back

Saunders, KO et al (6 Dec 2019) 'Targeted selection of HIV-specific antibody mutations by engineering B cell maturation', Science, 366/6470, eaay7199 DOI: 10.1126/science.aay7199. Back

Saunders, KO et al (10 May 2021) 'Neutralizing antibody vaccine for pandemic and pre-emergent coronaviruses', Nature, 594, 553-59. Back

The White House (13 May 2021) Daily Briefing.Back

Williams, WB et al (27 May 2021) 'Fab-dimerized glycan-reactive antibodies are a structural category of natural antibodies', Cell, 184/11, 2955-72, E25. Back