This was actually a follow-up on an observation we first reported in Nature in January 1995, showing that HIV replication in an infected person is highly dynamic, which means that the virus is constantly being produced from infected cells and is constantly being cleared by the body as well, and infected cells are dying rapidly only to be replaced by newly infected cells. So there is a constant turnover of both virus and infected cells.

It’s a bit more dynamic than others. What is different is the perspective it gave us about HIV. Prior to that paper, people had the notion that HIV was probably a pretty quiescent disease because it took so long for a person to get sick. The average incubation period is about ten years—that’s what we call the clinical latency period. And that led people to incorrectly think, "Well, the virus isn’t doing so much because it takes so long to do it." In fact, we showed that the virus is constantly destroying T-cells and constantly replicating, and the body is actually engaged in a constant struggle to clear the virus and knock off the infected cells.

“...the AIDS epidemic is the worst plague in history, having affected 70 million people already.”

A lot of things converged in the 1994-95 time period. One was our ability to measure the virus more accurately. The so-called "viral load" test emerged in the early 1990s, and by 1994 this was already well-validated and put into clinical use. Another practical development in that period was the development of drugs to treat people with HIV, and the drugs became a tool to probe this whole issue and to reveal the dynamic nature of HIV replication.

Let me give you an example of what I’m talking about. Let’s say you’re looking at somebody who is running on a treadmill, but you can’t see his feet and you can’t see the machine. You can only see the head bobbing up and down mostly at the same position. That’s sort of what it was like when we looked at HIV in the blood. Over a prolonged period of time, we’d see a steady concentration, well-maintained for month and years. Now if you ask, how fast is this guy running and how fast is the treadmill moving, all you can say is that they’re both obviously moving but in opposite directions. If I can’t see either one, though, I can’t tell if he’s doing a slow walk or a rapid run. If I had a mechanism to abruptly shut the treadmill off, then the guy’s head suddenly moves forward, and if I’m able to capture that information, I can figure out how quickly he was running.

If we come back to HIV, what we had been doing is measuring the concentration of the virus in the blood of an infected person and seeing this steady state, but we didn’t know how rapidly the virus was reproduced and how rapidly it was being cleared. We knew those two parameters were approximately equal, but they could be equally high or equally low. Once we had drugs to block the viral production part, then we could look at the acute decrease in virus concentration and that would tell us about these two parameters. That’s analogous to that snapshot of the head moving forward after we shut down the treadmill, and that’s effectively what we published in that 1995 Nature paper. We described this whole concept, and it became our most-cited paper.

We only knew from that first study that the dynamics were very fast, but we didn’t have individual numbers for virus particles and infected cells. So we engaged Alan Perelson, who was the first author and a mathematician, to write all this out, do a simulation, and then we could go back and do the patient experiments to confirm it. And that’s what this 1996 paper covered.

The observation itself changed our perspective of what HIV is doing. If you go into the discussion section of that ‘96 paper, you’ll see we really went into the implications for treatment. So now we had within our hands the minimum estimate of replication rate. We knew how fast HIV was dividing, we could then calculate the error rates, and we realized that HIV is mutating so quickly that if we treat this virus one drug at a time, HIV is always going to escape from the drug because of its rapid mutation rate.

We also did the calculations suggesting that if we combine drugs and force the virus to mutate in multiple positions throughout its genome, we can come up with a probability for the virus to successfully do that, and it becomes increasingly difficult the more drugs we add to the regimen. This was the theoretical foundation for combination drug therapy. Although we actually had realized that in 1995, even before we had all the specific numbers.

So in parallel with doing the follow-up study for this 1996 paper, we had already initiated combination therapies. By the middle of 1996, these treatment results were already emerging from this combination of drug and anti-retroviral therapy. The reason this is such a highly cited paper is the fact that it’s linked to the treatment effort.

No, that was in a 1997 Nature paper: also Perelson et al. ("Decay characteristics of HIV-1-infected compartments during combination therapy," Nature 387[6229]: 188-91, 8 May 1997). I presented those results first at an international conference in the summer of 1996.

I’m only a co-author on that one. That was work from Bob Siliciano’s laboratory. And that paper is one of three from 1997 reporting the same phenomena. They’re prominent because they were also touching on big issues. Our paper impacted on treatment strategy. These pointed out that the reservoir precludes the eradication of HIV-1.

We continued to work on the dynamics of HIV replication until three or four years ago. We became more and more detailed in defining, for example, the half-life of the virus particles and the infected cells. We drilled down quite deeply into that subject. And then gradually, over the last six or seven years we took to concentrating on vaccine development, because that’s a higher priority in addressing the epidemic. So for the past few years I’ve moved away from dynamics entirely.

It wasn’t an easy shift, especially considering that vaccine development for HIV has been so difficult. But the shift was not all of a sudden. We did it gradually over quite a few years.

Yes, and we have been pursuing it. We made two candidate vaccines that are already in human trials, and we just got a pretty big grant from the Gates Foundation to do the next generation of HIV vaccine. And, of course, the proposed idea in the new grant is pretty novel.

We are trying to direct a viral antigen to the dendritic cell, which is the principal antigen-presenting cell in the body. So we’re constructing platform technologies that will take the HIV protein to that cell population at the same time that they’re activating that cell population. Ideally, to generate an optimal immune response, it’s best to get the antigen to the dendritic cell and turn it on at the same time. So our platform technologies are all based on that principle. That’s the new program we’re starting.

I think the biggest problems are still those presented by the virus itself. It seems that, structurally speaking, HIV shields its crucial proteins on the viral surface quite well. It has a sugar coat, and it also has sequences that change rapidly. This variable sequencing and the sugar coating help protect the virus from antibodies, so it’s very hard to get antibodies to neutralize the virus.

That’s why we’re hoping that by taking the antigen to the dendritic cells, those will then present the HIV protein to B cells and that will lead to a stronger antibody response with higher affinity. And that might partially overcome this shield that the virus has. Nobody can ever tell you with confidence that any approach is going to work with HIV, but in theory this should lead to a better antibody response. We have no way of knowing whether we’ll get there or not until we do it.

One is a DNA vaccine, the other is a viral vector vaccine, using vaccinia as the backbone, the smallpox vaccine. There we’re trying to induce T-cell mediated immune response.

The point is that it takes so long to get any answer in vaccine development, especially when you need to do human trials. You have to spend a few years in the lab, then a year or a year and a half working with animals. If you clear all the hurdles there, then you have to manufacture a product suitable for human testing, which raises all sorts of manufacturing and regulatory issues with the FDA. So it’s five years, at least, before you can test a concept out in human beings.

If you have a new idea today, it’s unlikely to be fully tested for another five years down the line. So you can’t simply wait for an idea to go through that whole life cycle before starting another one. That’s why most of us doing vaccine development take multiple approaches, staggered in time.

I think it’s valid to talk about DNA vaccines, then and now. DNA vaccines first looked pretty good when they were checked out in small animals. Then the experience over the past decade suggested that once you move into monkeys or humans, DNA vaccines don’t look so good. We don’t fully understand why that is. Although we know that how we deliver the DNA is very important. If you just give it as an intramuscular injection, then the immune response is not sufficient in humans.

Recent studies have shown that DNA is actually not very good at inducing an immune response in itself, but it’s pretty good as a priming agent. That is, if you come and follow it up with another vaccine, in what we call a prime-boost approach, then the DNA actually does quite a bit of a good. So DNA is being resurrected in this prime-boost strategy.

DNA can also give pretty good results if you deliver it differently. Biotech companies are now pursuing all sorts of injection devices that will do a better job of delivering DNA into the muscle cells. When you do that, you can get a much better immune response. So the pessimism a few years ago wasn’t unwarranted, but now DNA is being resurrected, using these novel delivery devices and in this prime-boost approach.

The lay public or a scientific audience?

I think the message to the general public is that the AIDS epidemic is the worst plague in history, having affected 70 million people already. Twenty-five million have died; 45 million are living with a disease that, if untreated, would be lethal. For me it has been a real privilege and a fascinating ride to be involved in this research, to be able to make a contribution to the basic understanding and the treatment of this disease. That’s a real honor, and now we’ve shifted to take on an even tougher challenge, and that’s vaccine development. No one knows whether we will succeed in that, but it’s probably the most noble goal within AIDS research today.

To a scientific audience, I’d say that it’s really amazing what we have learned about this virus. It is so small and yet so clever in ways that we could never have appreciated in the past. It’s so dynamic; it makes so many mistakes, and yet these turn out to be a survival-strategy for the virus: through rapid mutation, it’s able to generate many structural features that combat the immune system and the host rather well. Ninety-nine percent of the time, the virus beats the host. You come to appreciate what a little organism with only a 10,000 base-pair genome can accomplish.