How Far Can You Knock Down HIV?
An Interview with David D. Ho, MD
- The primary infection experiment
- Not just "math and modeling"
- Thinking ahead with anti-HIV regimens
- "Not a time to pull back" on quest for new drugs
- The eradication wager
When attempting to write a pithy preamble to an interview with David Ho, one is tempted to consult the latest edition of The Toastmaster's Handbook for snappy alternatives to the much-abused "Our speaker tonight needs no introduction." Even before the December 30, 1996 issue of Time, telling this Journal's readers something new about the director of the Aaron Diamond AIDS Research Center would have been difficult.
But perhaps there is a way to illumine further his contribution to HIV research, and that is to announce here that the Journal has selected Ho as the first recipient of what is certain to become a highly coveted prize: Author of the Most-Stolen Slide of the Year award. In fact, if the Journal had begun bestowing this trophy a few years back, Ho would now have at least three platinum AMOSSYs festooning his office. The first would commemorate the celebrated image from his 1989 New England Journal article1 demonstrating astoundingly high titers of HIV in plasma and PBMCs of seropositive people with and without AIDS-defining symptoms. The second would show plummeting levels of plasma RNA in people whose steady state of viral replication had just been perturbed by a protease inhibitor,2 evidence that led to the current view of viral dynamics. The third would outline the "compartments" in which HIV is sequestered,3 and from which Ho and others hope it can be pried by a few years of aggressive antiretroviral therapy.
The cloning of these figures by countless speakers has elevated them to near-iconographic status. More importantly, they trace the uninterrupted stream of clinical research through which Ho and his colleagues have helped transform the treatment of HIV infection, and they testify to the frequent superiority of visual to verbal suasion when trying to change minds. But the Journal allowed Ho to rely solely on words when interviewing him at the April Keystone Conference on AIDS Pathogenesis. (A report on this conference will appear in the July issue of the Journal.)
The primary infection experiment
Journal: Have there been any breakthroughs in viremia in your study of treatment for primary HIV infection?
Ho: We have 20 patients who remain on the study, taking either indinavir or ritonavir with AZT plus 3TC. Everyone is still aviremic. Most of them have now had tissue biopsies, tissue cultures, tissue in situ hybridization and PCR, and these assays have shown no evidence of ongoing HIV replication. The patient treated for the longest time has been on therapy for about 21 months. Because these patients began treatment so early, their immune responses are good. Their CD4 counts are high, if not completely normal, and the CD4/CD8 ratios are generally above 1. So everyone is doing well and we're very pleased with the results. We're getting to the point where some of the patients who have been treated longer are coming around to have a second set of biopsies. We've been taking gut biopsies that contain lymph follicles and are planning to biopsy lymph nodes and -- if patients have them -- tonsils. We're also going to do spinal taps for cerebrospinal fluid. All of these samples will be assessed to make sure there's really no evidence of active virus replication. And, at that juncture, we would think about stopping medication in some of these individuals.
Journal: When you get to that point, let's assume that some people agree to stop...
Ho: We already know that several patients would like to stop to see what would happen.
Journal: You've said you want to continue treatment for two and a half to three years before you stop. When you do stop, how long would it take for you to be convinced that the therapy had worked and had eradicated the virus?
Ho: We're in uncharted territory here. We really don't know, but we have the feeling from previous observations, when therapy was stopped earlier, that when HIV returns it returns quickly. We plan to monitor the patients very, very frequently and carefully, and I think we will feel pretty good if we don't see any virus in the first month or two. Obviously, there could be a late appearance, but knowing how quickly HIV doubles, we really should see detectable levels of virus within days or weeks -- unless there's something we don't understand. In that case there could be occurrences months or years later. Obviously, we will continue to follow these patients very carefully even if HIV can't be detected in the first few months after treatment stops.
Journal: Let's say that it doesn't work and virus isn't eradicated in these studies. Based on what's known now, are there still good reasons to start treating people with primary infection, even if you can't clear the virus?
Ho: The only patient data to justify intervention during primary infection is the study by Perrin and his colleagues in Geneva.4 They showed that, over the short term, AZT treatment of primary infection has some clinical benefit compared with placebo. Beyond that, I don't think there's any real information to guide us. I think we have to rely on our current understanding of what HIV does and what the drugs do in other settings: since there's now growing evidence that multidrug combination therapy is better than monotherapy or dual therapy,5-7 and since we know we get the best viral response when the patient is treated earlier, one could extend that logic to the primary setting. But that's only a theoretical argument rather than something that's supported by clinical data at this point.
Journal: You've been careful to say that you think of these studies as proof-of-concept studies rather than as an attempt to establish a standard of care. But what exactly does that mean? There are some clinicians who already have adopted this strategy as the standard of care for primary HIV infection. Do you think they shouldn't until your results and those of others are more mature?
Ho: It would be incorrect for me to say that clinicians should or should not do certain things if they see a patient with primary infection, simply because there's not much data. So I am not going to come out and propose a certain therapy and speak against other approaches. I know what I would do clinically, but you have to remember that what we're doing in these patients is an experiment. Our initial question was not whether HIV could be eradicated, it was how far down we could knock HIV with the most potent therapy at the time. Then, when the first results looked promising, the question got extended to whether HIV can be eradicated.
But what we're doing is an experiment, and when you're doing an experiment to prove a principle, you want to set it up in a more or less idealized manner. So you begin by saying, "Obviously it's easier to treat earlier than later." Then logically you move "early" to "as early a possible," when the viral population is as homogeneous as possible and there are the fewest viral variants to deal with. As a result the chance of success goes up by some factor -- we don't know how much.
That's how this experiment came about, but people sometimes forget that and think we should be doing other things. They have different questions, but they want us to answer them. For example, some people think that we ought to have a control group. But that's not the question we asked. Our experiment is not designed to show improved prognosis with an early treatment group compared with an untreated group. Our question was how far down we could knock this virus as early as possible. So if some people want to ask different questions, they should ask those questions in different trials.
But some clinicians, as you point out, have begun to treat people with primary infection, simply because the early results of our study and others are so dramatic. And it may turn out to be the correct strategy, but I think time will have to tell.
Not just "math and modeling"
Journal: What are the most troubling misconceptions about your work on viral dynamics?
Ho: I think the viral dynamics model3 is accepted by most workers in this field. Everyone may not understand the full details of the different viral turnover rates in each of the compartments -- productively infected CD4 lymphocytes, latently infected T cells, macrophages, and so on -- but I think those details are gradually becoming better understood. If people don't understand the details, they do know that there's very rapid turnover and that HIV infection is characterized by this persistently active process.
But there are technical details on which there are general misconceptions. People think we study only viral particles, but by studying viral particles we're really studying the cells that make them. That's obviously more important. Another common misconception is that all of our work on viral dynamics is just theory. They say, "Oh, that's math and modeling." Those who think that way generally have not read the work carefully or listened to the presentations carefully. They think we're just pulling things out of the air. In reality what we do is go and gather patient information, and we try to be very quantitative and precise with the analysis of patient data. It's this analysis that relies on mathematical approaches -- but it's analysis of hard clinical data.
Journal: You mentioned in your presentation the other night that you don't disagree with the finding of Frank Miedema's group that telomeres* on the chromosomes of CD4 cells in people with HIV infection are not unusually shortened,8 but you think their interpretation of that finding is wrong. Their interpretation is that CD4 T lymphocytes are not turning over as rapidly as you and your colleagues calculate. How do you square their results with your calculations of CD4 -cell turnover?
Ho: I think the whole field would have to be turned upside down if they turned out to be right. There's just too much staring them in the face. You take the data that CD4 telomeres are not shortened, and you say, "Let's see what the explanations are." One of them is that CD4 cells are not turning over rapidly, and that's the interpretation they adopted. But there are two other explanations as well. One is that telomerase activity is enhanced in people with HIV infection, and that that's what keeps the telomeres close to their normal length. Miedema's group argues that they have done assays to measure telomerase activity and didn't find that it is enhanced. But I would argue that their assays are not sensitive enough to measure telomerase activity in cells from these patients. You want to be very sensitive and very accurate when the range of activity is very small, because a difference of 0.1 percent and 0.5 percent can be substantial. But the assay system they used can't determine such small differences reliably. You have to use modified assays, which are available in the field. So the issue of telomerase activity is unresolved.
A more important issue is that when you assay the telomere length of a mixed cell population, you have to know something about population dynamics. If I give you an extreme example, this becomes very evident. Let's say we have 100 cells, and 50 are turning over rapidly. If these 50 that are turning over rapidly are preferentially eliminated by HIV, which infects activated cells, they're gone and you're left with the 50 that are not turning over. You measure the telomere lengths of these 50 remaining cells and they're normal. So, because the cells that are turning over are gone, you conclude that there's no turnover. Nothing could be further from the truth. This is why we haven't published our results, because we don't think you can analyze it in such a simplistic manner.
In fact, people who have done modeling and careful thinking about cell dynamics immediately realize that the cell you're looking for is the cell that's disappearing from the sample. Modelers like Alan Perelson [who works with the Aaron Diamond group] and Rob de Boer [who works with Miedema's group and other Dutch investigators] have gone through these results and have shown that the data are consistent with a high turnover and a preferential elimination of the cells that are turning over. HIV goes for activated cells that are turning over. That aside, if you just think about how many cells are expressing the activation antigen -- the cell cycling antigen -- it's pretty clear that their conclusion cannot be right. Plus, I think the experiments we are doing now, labeling T cells in monkeys exposed to virus, will show that their conclusion is not right.
Journal: Another question raised about your work is the significance of proviral HIV DNA. You say that the proviral DNA that can be detected in latently infected T cells months after fully suppressive therapy begins represents "archival dead sequences." But Bob Siliciano at Johns Hopkins reported at this meeting and elsewhere10 that the tiny amount of integrated HIV DNA he finds in resting cells is replication competent.
Ho: There's no inconsistency between his findings and our analysis of viral turnover. In our analysis we also looked for a pool of latently infected cells with an infectious genome, and we know we have a pretty good idea of its turnover rate. Bob is looking at untreated individuals, and that small population of latently infected calls -- the ones with integrated DNA capable of making replication-competent virus -- is turning over. Our studies indicate that, after six months or so of therapy, all of those latently infected cells should be gone because their average half-life is about a week. So the remaining viral DNA that can be detected after months of therapy is not the replication-competent form he's speaking of. It would be the other [turning over] form that he's talking about.
Whether integrated or not, they're defective. We know that because, let's say nine months into therapy, the patient still has 1000 copies of proviral DNA per million cells. That's a fair amount of viral DNA. Yet, no matter what you do with those cells, you cannot activate them into releasing HIV. Obviously you could come up with fancy explanations of why these cells containing viral DNA cannot be activated to produce new virions, but the straightforward explanation is that the DNA is defective. And that explanation is consistent with a large literature showing that most viral DNA is defective.
Thinking ahead with anti-HIV regimens
Journal: I want to ask you just a couple of questions about the majority of people with HIV infection -- those who test positive years after being infected. Right now they have a difficult choice to make because lots of factors argue for starting maximally suppressive therapy. On the other hand, given the drugs available right now, drug-naive people with chronic infection have to be concerned that if they start AZT, 3TC, and indinavir, for example, and if that combination fails, they've probably blown their chance to get maximal benefit from another protease inhibitor combination. What can you say to those patients when they're trying to make those decisions?
Ho: I would say, "Let's find out where you are in the course of HIV infection." I think if viremia levels are high, let's arbitrarily pick a number of 5000 HIV RNA copies or higher -- then it's best to go ahead and treat. Of course, every physician has a different comfort level when it comes to an RNA threshold, but I think most physicians would recommend starting treatment now if they think the viral load is too high. But if the viral load is lowish, then I think there's a more difficult decision to be made between observing carefully and not treating, or going ahead and starting therapy.
When you do start therapy, obviously you should not just think about immediate therapy, with the combination you mentioned, for example. Clinicians should think about one regimen and a second regimen. Construct a regimen that is good and potent, then think, "If that should fail, what would be the next option?" Fortunately, given the drugs that are available -- two NNRTIs, five nucleoside RT inhibitors, four licensed protease inhibitors, with more of each in the pipeline -- I think one could come up with two, possibly three combinations that are reasonable to go with. And I generally recommend thinking ahead that way -- if the patient is drug naive. Unfortunately, the problem is that so many patients are drug experienced.
Journal: A difficult moral issue has arisen about starting antiretroviral therapy. Some clinical investigators are saying, "I won't start a combination including a protease inhibitor in a patient I don't think will be compliant, because resistance will emerge and that's bad for the patient and bad for public health." At the same time some clinical investigators are saying, "If you don't start with maximal therapy, don't start at all, because the virus will become resistant more quickly to suboptimal therapy." Both arguments seem to make sense. Is there any way out of that dilemma?
Ho: I appreciate the problem. I'm not sure I have any real answers to that tough question. You're getting into an area where the discussion needs involvement by ethicists and people who are doing behavioral studies. I think this is beyond my level of expertise. I face the same problems and participate in these discussions, but I think this is an area that requires input from experts other than clinical investigators.
"Not a time to pull back" on quest for new drugs
Journal: You were the person who showed that the first attempt to block the primary receptor of HIV -- with soluble CD4 -- was destined to fail.11 Do you think there's any better chance of finding agents that will effectively block secondary receptors?**
Ho: I think they will find small agents that will block secondary receptors, but I think this overall strategy is a lot tougher than many people appreciate. Many of the experts here at this conference know that once you find a small chemical blocker, you still might not have an effective therapy because there are so many alternative pathways by which the virus can enter the cell. So I think this strategy won't be easy. If blocking the principal receptor did not work so well, why would blocking a secondary receptor work any better, especially when there are multiple secondary receptors? And if you block CCR5, for example, those HIV variants that favor, say, CXCR4, might well have a selective advantage. In that case, are you doing the patient a favor by blocking CCR5? There's a real debate, but clearly we won't know for sure until such an experiment is done. Perhaps one could do some of these experiments in monkeys. So I view that as a very, very tough problem. I have no doubt that the pharmaceutical companies, once they set up their screens for inhibitors of CCR5 and gp120 interaction, will find "hits" and that "hits" will be developed. But ultimately how effective they will be in patients, I don't know.
Journal: Do you think it's going to be necessary to develop other classes of drugs besides nucleosides, NNRTIs, and protease inhibitors? Let's say that drug developers succeed in making really tolerable protease inhibitors that are even stronger than the ones licensed now, and that they can be combined with a new generation of better nucleosides and good nonnucleosides. Are clinicians really going to need integrase inhibitors and chemokine receptor blockers?
Ho: Without getting into social, political, and economic issues, overall my answer would be that -- although we've made tremendous therapeutic advances against HIV -- we haven't gotten to the point of curing patients. So this is not a time to pull back on the development of new drugs. I think if we're beating up on the virus a little bit, now is the time to really go after it and come up with a few more agents, hoping that these few more, when added on top of what we have, could do the trick. That's how I view it.
The other thing, as you know, is that these meds are not without side effects. We need newer agents. Hopefully, we can come up with equally potent or more potent agents with fewer side effects. That's how the antibiotic field evolved. Who knows? Maybe some integrase inhibitor, for example, could be a small chemical that's very easy to make, and one could make buckets of that stuff for very little compared with the protease inhibitors. And if those agents are as active or more active, then you could see some of the newer agents replacing the old ones, with tremendous benefit to the patient and to society if the cost factor is a substantial one.
So I think there are multiple reasons to push ahead. There are people from the pharmaceutical industry who say, "Well, we've come so far with the new protease inhibitors and the RT inhibitors doing such a great job, we should just back off." My reply is the one I've just given you. That's quite the wrong attitude.
The eradication wager
Journal: Your colleague John Moore has argued strongly that certain basic questions about HIV vaccines have to be answered before investigators can proceed to large-scale efficacy trials. What do you see those questions as, and how close do you think vaccine research is to answering them?
Ho: Let me just clarify John's point a little bit. You summarized it well by saying he believes we should not proceed with large efficacy trials at this point. I would agree with that. But that doesn't preclude clinical research going on right now. I think we need to have basic research, and we need to have clinical testing of new strategies to get information in patients. We can't just do everything in the laboratory and figure things out that way. So it will have to be an integrated process between the clinic and the laboratory, back and forth.
But I agree that we are not ready for large-scale trials involving thousands of patients. I really do think we need to understand more about what protects against HIV infection. These mysterious correlates of protection are still not well understood. We have a vaccine that works in monkeys, the live attenuated SIV [simian immunodefiency virus] vaccine, but we don't know how it protects. Is it a CTL [cytotoxic T lymphocyte] response? Is it an antibody response? Is it something else? We need to understand that because that will guide us. I think most people will agree that, while the live attenuated virus strategy is very efficacious in the monkey model, it's not a strategy that will be adopted for humans any time soon -- anywhere. But we can use that system to understand what's going on, and hopefully once we understand what the mediators of protection are, then we can think of other safer strategies to induce them.
Of course there are many other questions. Once you've decided on an approach you also have to decide how to present that antigen to the immune system to elicit the effect you're looking for. So there are many issues, and many of these things are not well worked out.
Journal: My last question has to do with how far HIV research has come and how far you think it has to go. Let me put it to you this way: There are about a million people in the United States who have HIV infection. How many of them do you think will die of AIDS?
Ho: Without giving you specific numbers, I think there's a good chance that this population of one million infected people will get substantial benefit from the current therapies, so that their life span could be lengthened substantially by the developments of the past few years and by the developments that will be coming in the next five years. If the eradication of HIV is possible, I think there's a chance that, some time down the line, a portion of the people infected today could come to benefit from this development. You have to put your money down somewhere on the situation. And I think that's a reasonable wager to make.
Mark Mascolini writes about HIV (firstname.lastname@example.org).
*Telomeres, repeated sequences on the tips of chromosomes, become shortened as cells divide. As a result, the telomere length of a cell population can be used as one measure of how "aged" that population is -- in other words, how often it has "turned over." For example, both Miedema's group and researchers at the University of California, Los Angeles,9 determined that the telomeres of CD8 T lymphocytes in people with HIV infection are about as short as those of healthy 100-year-olds and concluded that this is proof of the constant activation of CD8 cells by HIV. Miedema's interpretation of his CD4 telomere data is more controversial.
1. Ho DD, Moudgil T, Alam M. Quantitation of human immunodeficiency virus type 1 in the blood of infected persons. N Engl J Med 1989;321:1621-1625.
2. Ho DD, Neumann AU, Perelson AS, et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 1995;373:123-126.
3. Perelson AS, Neumann AU, Markowitz M, et al. HIV-1 dynamics in vivo: virion clearance rate, infected cell lifespan, and viral generation time. Science 1996;271:1582-1586.
4. Kinloch-de Loes S, Hirschel B, Hoen B, et al. Zidovudine versus placebo in primary HIV infection: a randomized study. N Engl J Med. 1995;333:408-413.
5. Gulick RM, Mellors J, Havlir D, et al. Potent and sustained antiretroviral activity of indinavir (IDV), zidovudine (ZDV) and lamivudine (3TC). Presented at the XI International Conference on AIDS; July 7-12, 1996; Vancouver. Abstract Th.B.931.
6. Hirsch M, for the Protocol 039 (Indinavir) Study Group. Indinavir (IDV) in combination with zidovudine (ZDV) and lamivudine (3TC) in ZDV-experienced patients with CD4 counts <50 cells/mm3. Presented at the 4th Conference on Retroviruses and Opportunistic Infections; January 22-26, 1997; Washington, DC. Abstract LB7.
7. Conway B, Montaner JSG, Cooper D, et al. Randomized, double-blind one-year study of the immunologic and virologic effects of nevirapine, didanosine and zidovudine combinations among antiretroviral naive, AIDS-free patients with CD4 200-600. Presented at the Third International Congress on Drug Therapy in HIV Infection; November 3-7, 1996; Birmingham, UK. Abstract OP7.1.
8. Wolthers KC, Wisman GBA, Otto SA, et al. T cell telomere length in HIV-1 infection: no evidence for increased CD4+ cell turnover. Science 1996;274:1543-1547.
9. Effros RB, Allsopp R, Chiu C-P, et al. Shortened telomeres in the expanded CD38- CD8+ cell subset in HIV disease implicates replicative senescence in HIV pathogenesis. AIDS 1996;10:F17-F22.
10. Chun T-W, Carruth L, Finzi D, et al. Quantitative analysis of latent tissue reservoirs and total body viral load in HIV-1 infection. Nature 1997;387:183-187.
11. Daar ES, Li LX, Moudgil T, Ho DD. High concentrations of recombinant soluble CD4 are required to neutralize primary human immunodeficiency virus type 1 isolates. Proc Natl Acad Sci USA 1990:87:6574-6578.
12. Moore JP. Coreceptors: implications for HIV pathogenesis and therapy. Science 1997;276:51-52.
This article is reprinted from the Journal of the International Association of Physicians in AIDS Care (June 1997, Vol 3, No 6) © 1997, Medical Publications Corp. For more information on the Journal, visit the International Association of Physicians in AIDS Care Web site or send e-mail to Journal@iapac.org.