3.5 Daria Hazuda: Discovery and Development of Integrase Inhibitors

Daria Hazuda: [00:00:00] First, let me say it really is a tremendous honor to be here at this meeting with so many luminaries in the field. I feel tremendously awed and privileged. I'm going to start my talk with this slide. I know it was shown earlier by Tony Fauci. I thought it was worth showing again for several reasons. First, as somebody who's actually been doing this for a living for over 20 years, and I just can't believe it [00:00:30] when I say that, I still find this just incredibly remarkable. Because if you think about it, it was a little more than a decade from the discovery of HIV as the cause of AIDS that we not only had one, but we actually had three drugs to put together in a cocktail.

If you think about it from a drug discovery perspective, given the fact that most things that we do actually fail and never see the light of day, and then it generally takes 10 to 15 years [00:01:00] to develop any one new molecular entity, the fact that we had three new molecular entities that we could put together to form this cocktail within a decade of actually identifying HIV, to me is still a remarkable achievement for the field. The second reason—Let me go back. 

The second reason I like to show this slide is I remember sitting in the audience in Vancouver (11th International AIDS Conference) when [00:01:30] these data were first shown publicly. I had just started working on integrase maybe a few years, it had made a little bit of progress, but clearly, we were far from where we needed to be. And I just remember being in that audience when this data was shown, and seeing the response to this data, and people actually even thinking that, well, perhaps we can cure HIV. Sitting there and thinking to myself, "Geez, maybe I ought to [00:02:00] find something else to do for a living."

It was a discussion or a comment that was made in one of the sessions by Paul Volberding, who's here today, that really convinced me that maybe there still was a lot that we needed to do. What Paul had said was: really think about what we're asking our patients to do. We've demonstrated what's possible. I'm paraphrasing, [00:02:30] obviously. Now let's go back and actually do it right. That's been very foundational for me and thinking about anything that I do in drug development for the next 20 years. It actually convinced me that there was still a lot that we needed to do. 

Of course, over the next decade, that's exactly what people did, and many companies focused on developing next-generation agents, [00:03:00] which addressed the liabilities that we had with the early compounds addressing toxicity, tolerability—and as you'll hear in the next presentation, I'm sure, convenience. If you think about it, during the course of the next 10 years, we had over two dozen new agents to treat HIV infection, but they still resided in the same three oral drug classes that we knew about in 1996, 1997, when that [00:03:30]data was first presented. From a historical perspective, which is what we're here to do today, I think the year 2007 was really a pretty remarkable year, because for the first time in more than a decade we now had two new classes of oral agents to treat HIV infection: the CCR5 antagonists (maraviroc), which I relied on the brilliant work of Ed Berger that you'll hear about [00:04:00] tomorrow, and the integrase inhibitors (raltegravir). (1) As you all know integrase inhibitors work by blocking a critical step in the viral life cycle, the insertion of the fully reverse-transcribed viral DNA into the DNA of the host.

Now, it had been known for decades, of course, largely from the work of many people that are in this room, that in fact, integrase was essential to the replication of all retroviruses. It's actually [00:04:30] fundamental to retroviral replication. The question that I frequently get asked is, why did the discovery of integrase inhibitors take so long?

There are many reasons for this, but I've listed—from a drug development perspective, I've listed two relatively important ones here. First, there was no relevant structural information. I think it's worth pointing out that, in the context of HIV drug discovery, HIV protease inhibitors were really [00:05:00] among the first true examples of structure-based drug design, and all the next generation agents certainly relied on structure-based drug design. In the case of integrase, we didn't have a relevant structure information. We had structural information on each of the independent domains when we started the program, but there was no structure of the entire protein, and importantly, as you'll hear in the talk, we didn't have structure of the intasome, which is the integrase DNA complex. [00:05:30]

This is what we knew. We knew that there were three domains and the business part of the molecule was generally considered to be the catalytic core domain and within the catalytic core domain, there were two acidic residues that are coordinated by two magnesium ions and the work of Bob Craigie, Alan Engelman, Rick Bushman, and many others had really elegantly shown that these residues and this act of slight confirmation was absolutely critical [00:06:00] per function. (2)

In fact, this sort of motif is not only conserved in all retroviral integrases, but it's conserved in transposases and actually many magnesium-dependent phospho-transferase enzymes. We had structures of each of the domains, we had no idea how they all fit together to make an intact enzyme. In fact, if you even looked at just the catalytic core domain, you couldn't do even if you just took the catalytic core domain, it was impossible to do structure-based [00:06:30] drug design off of the catalytic core domain itself, despite the fact that that is really the business part of the molecule.

I like to show this is a paper that was published as late as 2007 by people who actually should know a bit about drug discovery—people at Pfizer, and people at Amgen—who, based on the structure actually deemed integrase completely undruggable. (3)

The second reason, I think, was that there was no relevant high throughput [00:07:00] biochemical assays. Again, the work of many folks, including people that are sitting in this room like Anne Skalka and Duane Grandgenett and the folks at the NCI had really—we knew a lot about the mechanism of integration, we knew there were three steps. We knew that the enzyme had to bind to very specific sequences at the ends of the viral DNA (LTRs) . The context of that complex, the pre-integration complex [00:07:30] it cleaved the dinucleotide to form a 3’-OH which was a nucleophile for the strand joining event. 

These events, this process, these events, this whole process was almost impossible or very, very difficult to recapitulate in any high throughput way in biochemical assays. Pretty much all of the initial drug discovery efforts focused on the first two steps of the reaction, assembly and processing, because these were very easy to reproduce. [00:08:00] In fact, the work of Rick Bushman and Alan [Engelmann] had actually suggested that the same active site performed both of these processes, and if you were interested in active site inhibitors, why not just focus on steps that were easy to reproduce in a test tube.

Many groups, academic groups, many pharmaceutical companies had used these kinds of assays to screen for molecules that would inhibit integration. A lot of compounds were discovered, the patent literature, the literature is filled with examples [00:08:30] and they work really brilliantly in biochemical assays, but none of them had antiviral activity. At this point, the field was, I think, what I would call incredibly frustrated. And, in fact, there were many people who actually argued that it would be impossible to find integrase inhibitors because if you really think about what you're trying to do, you're actually trying to inhibit an irreversible reaction. 

I remember even having an argument with John Coffin at a drug resistance meeting, who basically said this didn't make any [00:09:00] sense because it's impossible to inhibit an irreversible reaction with a reversible inhibitor because any potential transfer escape, you'd actually get a productive infection event.

We persevered despite the skepticism. I had a chance conversation at this point with one of the most accomplished enzymologists [00:09:30] at Merck, trying to get some advice on what we were doing wrong and how to fix it. He said, well, the answer is very simple. He said, "You need to develop a robust enzymatic assay." He said, "But the problem is integrase isn't an enzyme, so I really don't know how to do that.” So that wasn't very helpful, but it was very insightful, Because it was that idea that you really have to think about this as an enzyme. And we went back and we thought about it.

When you think about it as an enzymatic reaction [00:10:00]. in the context of viral replication, you realize, in fact, that the most rate-limiting step in the context of viral replication is the last step, strand joining or strand transfer. This, in principle, should be the most sensitive step. The step is most sensitive to inhibition. The problem was that we didn't have an assay that could measure or screen for inhibitors of strand joining. 

We went back to the drawing board and the basic science [00:10:30] and made an observation, in fact, that we could actually make stable complexes between the recombinant enzyme and viral DNA substrate that mimicked short oligonucleotide substrates that mimic the viral DNA end.

It was this observation that was, I think, fundamentally important, because using that observation then allowed us to take these complexes, and mobilize them on these or microtiter plates and actually use these preformed complexes [00:11:00] to specifically look for compounds that would block joining of an artificial substrate representing the cellular DNA in an in vitro biochemical assay. 

So we screened the Merck chemical collection and we found a series of hits. One of the things that was really so interesting about these hits is they were structurally very different from all the hits that have been described previously which didn't work. And they all were very structurally similar. They all had this [00:11:30] diketo-acid, diketocarboxylate motif. It turned out that these hits came from a previous program that I worked on called the “Flu Endonuclease Program.” When the chemists looked at these compounds, they were very dismissive and they said, "All they're doing is chelating metals. They have a metal coordinating group. They're just chelating metals." 

But because they had come from the Flu Endonuclease Program that I had worked on in collaboration with a virologist at Merck, Joanne Thomasini-Lynch [00:12:00]. She was one of the most neurotic people I'd ever met. She had actually kept every single compound that was ever made for the Flu Endonuclease Program in her personal freezer. We went back to her personal freezer, took out all of the of those inhibitors, and screen them for inhibition of strand joining. We were able to show that the structure-activity relationship for integrates was entirely different from flu endonuclease. This actually convinced the [00:12:30] chemists that there was something actually interesting here and it was just not nonspecific chelation.

I give a lot of credit, just as Raymond [Schinazi] did, to people who I think are among the many unsung heroes in this field, Joe Vacca and Steve Young, because at that point, they really believed that this was a true medicinal chemistry problem, that these compounds were truly working in a very specific way. It just [00:13:00] became a chemistry issue to actually fix them and make them drugs. Joe and Steve have made huge contributions to the field of HIV because they were involved both in, not just in the development of raltegravir, but also the development of efavirenz, Sustiva, as well as Crixivan (indinavir).

Some of the compounds not only had very interesting properties from a structure-activity [00:13:30]relationship perspective but they were also really unique biochemically and virologically. (4) Just as we had set out to do, we actually had found very, very highly potent inhibitors of the strand joining reaction. But much to my surprise and many others, they actually had no effect on 3’-processing. 

Remember, the work of many people have suggested the same active site actually performs both functions. And I remember, several years before we actually [00:14:00] even publish these data, sitting at Cold Spring Harbor meeting, talking to Bob Craigie and telling him, in fact, that we had molecules that did this and they didn't do that. He was scratching his head and he said that just doesn't make any sense to him. Nonetheless, that's how they worked in the biochemical assays. In fact, we could also show that they had antiviral activity, so that was, obviously, a major achievement, considering the past history of integrase inhibitors in terms of the [00:14:30] lack of antiviral activity. Importantly, we could actually show that the antiviral activity was, in fact, due to this very selective inhibition of strand joining. So, if you looked at what was happening to the viral DNA in the context of concentrations of these compounds that completely abolished HIV replication, you could show that the ends of the viral DNA were still processed but there was no integration. 

How does this work? How is it actually possible to inhibit an irreversible reaction with a reversible compound? [00:15:00] These compounds are actually functionally reversible. Well, it turns out that biology, inevitably is far more complicated than biochemistry. What we had learned from folks like Mark Muesing (1953–2017) and Mario Stevenson is that if the viral DNA doesn't integrate, there are cellular processes that recognize the viral DNA as something foreign and either degrade it or recombine it to form this one into 1 & 2 LTR circular products. What integrase strand [00:15:30] transfer inhibitors actually do is they stall the integration process long enough that these competing cellular processes actually prevail. It turns out that inhibiting integration is a very effective means of inhibiting viral replication because it results in functionally irreversible inhibition of integration in the context of HIV infection. All integrase inhibitors, in fact, integrase transfer inhibitors result in functional inhibition of HIV [00:16:00] replication because you get these recombined products or you get degradation of the viral DNA. 

This was all great. We had leads, we had a very stringent proof that they were working as expected, but we didn't have a structure and so we felt it was very important to really understand as much about the biochemical and molecular mechanism as possible. These two papers that we published really formed the fundamental basis for understanding how all [00:16:30] integrase strand transfer inhibitors work. (5, 6)

First, with respect to biochemical mechanism: we demonstrated that these compounds did not bind to the enzyme free in solution, but required an integrase, a very specific integrates DNA complex and then they inhibited strand joining by competing with the host or cellular target DNA. Then with respect to molecular mechanism, a lot of SAR allowed us to develop a model, which suggested that the diketocarboxylate motifs which were common to all of our compounds were interacting [00:17:00] with the two metals. But that wasn't sufficient. You needed to have this hydrophobic group, this piece of grease which had to be very selective for integrase. And this led us to the hypothesis that the R group was interacting with a hydrophobic pocket at the enzyme active site. 

And it was on the basis of this structural motif, that the structural model that the chemists then rationally design alternatives to the diketo acid motif, the naphthyridine carboxamides. [00:17:30] This allowed us to demonstrate proof of efficacy and a SHIV (SIV) infection model. (7, 8) Also, it was the first compound that we took into clinical development.

We were really excited because they had robust anti-viral activity but unfortunately, in fact, this happened during a CROI (Conference on Retroviruses and Opportunistic Infections), I distinctly remember Joe Vacca and I, at CROI meeting, getting a call from our toxicology group saying that there would be some terrible, terrible toxicology issues in the dog studies and we had to stop the clinical study with our first [00:18:00] compound 810. It was a real tour de force to figure out the mechanism of toxicity and it turned out to be a) dog-specific and b) something that had to do with naphthyridine carboxamides structure.

At this point, now we were stuck. We couldn't progress the diketo acids because they weren't pharmacologically stable, the naphthyridines were toxic, so we needed to identify new leads. It was this fundamental understanding of the mechanism and how they work that made us look at compounds [00:18:30] that had been identified for another magnesium-dependent phosphotransferasem hepatitis C polymerase. These very ugly compounds shown on the left were identified as leads for Hep C polymerase, another magnesium-dependent phosphotransferase. They had been progressed to these dihydroxypyrimidines by the chemists. Unfortunately, they're not very potent inhibitors of HCV.

But we thought, "Well, looking at these compounds are probably working by the same mechanism, being magnesium binders." This [00:19:00] allowed the chemists to make hybrid molecules between the diketo acids and the dihydroxypyrimidines. To make a very long story short, this ended up with a compound that was first called 612, then 518, raltegravir, which was discovered in February 2002 and proved in the US as the first integrase inhibitor approved for the treatment of HIV infection in 2007.

You see the metal-binding group, the hydrophobic group that we actually hypothesized based on our structural model. [00:19:30] We're really excited, when, about three years after approval, this crystal structure came out which was the first retroviral intasome structure from Peter Cherepanov. (9, 10) We could actually say not only how the viral DNA interacts with the enzyme but also how raltegravir interacts in the enzyme active site, confirming very nicely our structural hypothesis that the interaction with the two metal ions in the active site, but [00:20:00] really explaining for the first time why you need the intasome, because the halogenated benzyl group actually makes a very important stacking interaction with the viral DNA. This is the first HIV inhibitor that interacts with two critical components of the virus. It interacts both with the viral DNA ends, as well as the active site of the enzyme. 

Now we have three integrase inhibitors that are approved for HIV infection, [00:20:30] many compounds that have been described in the patent literature, and they retain the same structural features identified in the original diketo acids, and they have exactly the same mechanism of action. (11) I'm happy to say integrase inhibitors are now commonly used as first-line therapy on the treatment of HIV infection. 

One of the things that's remarkable about integrase inhibitors is they really exemplify the concept that improved tolerability actually in drugs actually translates to higher efficacy. [00:21:00] Going back to the original comments that were made by Paul Volberding at that CROI meeting in Vancouver, this is the ARDENT study that was done by the ACTG (2009–2013) with Isentress (raltegravir), but this has also been shown for dolutegravir (DTG).

Then the last thing I just want to emphasize is there's still a lot of work to do. Non-adherence still remains an issue in the treatment of all chronic diseases. HIV is actually better than most surprisingly, but it's still [00:21:30] a major issue. Integrase inhibitors cabotegravir is leading the way in terms of the paradigm shift of thinking about long-acting injectable regimens for the treatment of HIV: it shows what's possible. This compound that Raymond alluded to MK-8591 is an incredibly potent nucleoside with a very long terminal half-life in HIV infected patients, doses as low as a milligram, [00:22:00] actually gives sustained inhibition of viral replication out to a week. (12, 13, 14, 15)

This incredible potency allows us to actually think completely differently now about how we can actually deliver HIV medicines for both treatment and prevention because you can develop parental formulations such as implants, which can actually give you a sustained release of active concentrations of the compound out to a year. I think there's been many innovations [00:22:30] in HIV. I hope that the future will enable more innovation such as extended duration dosing. Thank you.

[applause]

Sandra Lehrman (Moderator): Thank you, Daria. Questions, we'll go to John and then the back of the room.

John Coffin: Thank you. Very nice talk, Daria and a very [00:23:00] really remarkably important contribution and innovative contribution to the field. I would like to modify slightly something you said. I was indeed highly skeptical about integrase inhibitors as good targets for the reasons you expressed, but I view it slightly differently. I would like to think, and maybe it's just my own ego, perhaps, but I'd like to think that you forged ahead because of my skepticism, not despite it.

Daria: [00:23:30] Absolutely. That's the reason I'm also working on a cure, by the way.

Sandra: Got a question back there?

Mark Harrington: I just wanted to ask a doubleheader question. You're a role model to me. I have tremendous admiration for the stubbornness and persistence with which you went through this whole effort. 

Daria: Talk to my husband.

Mark: But I'm going back to what Françoise [Barré-Sinoussi] said before and thinking about industry, how did you [00:24:00] obtain the support within a company that was going through several leaderships. I know Emilio [Emini] was there at the start to help you out and he's another visionary for whom I have tremendous admiration, but how did you do it? How can we broaden that experience so that other women scientists and scientists and companies can have such success?

Daria: Well, I've always been very lucky at Merck, to even when things aren't going well. As long as [00:24:30]I had other things that were going well, I had the freedom to actually pursue things that were much, much higher risk. I think that's a challenge in industry these days, but I think it's still very important. You need to balance doing things that are potentially high risk and very transformative with other things that perhaps are where the bar is perhaps a bit lower.

Sandra: Question mark over there. [00:25:00]

Carol A. Carter (SUNY Stony Brook): Just comments. I can't help but say this. As somebody in other people in the room also who've collaborated over the years with pharmaceutical companies, we have to say that both Emilio and Daria who are both alums of Stony Brook, we really feel that these guys are special. Because even when other companies had reasons for backing out, [00:25:30] I think it was the spirit and the persistence and the dream goal of these two individuals that really kept things pushing at Merck. I don't know whether you guys would take that credit, but I really think you should.

[applause]

[00:25:53] [END OF AUDIO]

Citations

Index 

• 1.5 John Coffin — The Origin of Molecular Retrovirology 
• 2.1 Paul Volberding — The First Patients 
• 2.3 Mark Harrington — The Importance of Activism to the US Response 
• 2.5 Françoise Barré-Sinoussi — Discovery of HIV 
• 3.4 Raymond Schinazi — Discovery and Development of Novel NRTIs 
• 5.0 Anna Marie Skalka — Introduction, Session 5 
• 5.4 Edward Berger — Discovery of HIV Co-receptors 
• 6.5 Emilio Emini — Issues in HIV Vaccine Development: Will the Future be any Easier than the Past? 
• 11th International AIDS Conference, Vancouver, Canada, July 1996
• ACTG Study 5257 "ARDENT" (2009–2013) 
• adherence, patient compliance
• Amgen (Applied Molecular Genetics) 
• basic vs. applied research 
• Bushman, Frederic D. "Rick" 
• cabotegravir 
• Carter, Carol A. 
• CCR5 receptor antagonist 
• chemistry, chemists 
• Cold Spring Harbor Laboratory (CSHL) 
• Conference on Retroviruses and Opportunistic Infections (CROI)
• Craigie, Robert A. 
• cure vs. remission of HIV/AIDS
• drug discovery and development 
• DTG (dolutegravir) 
• EFV (efavirenz) 
• Engelman, Alan N. 
• freezer 
• Grandgenett, Duane P. 
• iconoclasm in science 
• implants, drug delivery implants, implantable antiretroviral therapy 
• in vitro vs. in vivo
• indinavir (IDV, Crixivan) 
• integrase 
• integrase inhibitors 
• LTR (long terminal repeat) 
• maraviroc (Selzentry, Celsentri) 
• Merck & Co., Inc. (Merck Sharp & Dohme) 
• microplate 
• models (model systems, model organisms, modeling) 
• Muesing, Mark A. (1953–2017) 
• National Cancer Institute (NCI) 
• observation 
• Pfizer 
• prevention of HIV/AIDS
• protease inhibitors 
• raltegravir (RAL, Isentress) 
• scientific revolution 
• sensitivity and specificity; false positive, false negative; biological specificity
• Session 3: Antiretroviral Therapy 
• simian immunodeficiency virus (SIV) 
• Stevenson, Mario 
• structural biology
• structure-based drug design 
• transposase 
• Vacca, Joseph P. 
• virology 
• women in science 
• Young, Steven D.