Trying to make sense of treatment interruptions.

DO TREATMENT INTERRUPTIONS MAKE IMMUNOLOGIC SENSE?

Controlled treatment interruptions seemed like a reasonable idea a few years ago. From a historical viewpoint, by 1996, the message was, "Now that we have better and more effective drug regimens, we might actually be able to rid the body of HIV." Several years later, the hope (or hype) was gone and a new challenge arose. Remember, this is HIV--there is always a new challenge. The newer drugs contained HIV better, but they came with new or more blatant clinical side effects such as problems with let redistribution or worse. Not to mention the long-term expense of the medications.

At about the same time, immune studies showed that HIV-specific CD4 T cells could be found, especially in the newly infected and in those treated early after infection. These new assays were based on measuring antigen-specific cytokine production in both CD4 and CD8 T cells from HIV-infected people. Curiously, T cells that made cytokines after antigen exposure in vitro were detected most abundantly in those with fairly intact immune systems and some viremia. Previous assays for cellular immunity required antigen-specific T cells to proliferate and in the 1980s, numerous investigators showed that very few HIV antigen-specific CD4 T cells could be detected. However, the new, short-term, antigen-specific cytokine assays did not require such proliferation to demonstrate T cell function.

Therefore, based on the promise of heightened immune responses in the presence of viremia in some patients (as measured by these cytokine assays), some thought that perhaps a good strategy to bolster HIV immunity would be to provide an opportunity for small, controlled quantities of HIV to be made--ostensibly to provide antigen and to drive the immune response. This was to be done in a structured, monitored way so that if viremia spiked, the drugs could be re-instated. Many variations on this basic idea have been put into trials in the last 5 or 6 years. Early promising results suggested that in early infection, the immune response after such interruption of therapy seemed to contain HIV longer than thought possible. Indeed, a consensus is that if the immune system is still fairly intact, therapy interruption might indeed boost CD8 T cell responses, which likely contain HIV longer.

However, from an immunologist's perspective, the idea that antigen was needed to bolster the immune response never made much sense. Even when viral replication is halted, which vastly reduces the amount of circulating HIV, the amount of non-replicating HIV trapped in the lymph nodes all over the body should provide plenty of antigen for a continued immune response. In addition, if memory T cells had been formed (rather than short-term, cytokine-producing, HIV-specific T cells), then they should be capable of mediating ongoing HIV-specific immune responses, even in the absence of viremia. This notion is complicated because there is preferential HIV infection of HIV-specific CD4 T cells, perhaps crippling future HIV-specific antibody and/or CD8 responses. However, the reality is that most infected people make sufficient and even robust antibody and CD8 T cell responses to HIV. The quality of the response seems to be the problem and a rationale is that while many HIV-specific T cells are generated, too many do not become long-term memory cells. This eventually allows the virus to escape containment, indeed, recent data show that in those who have multi-drug resistant virus, therapy interruption is actually a bad idea because it allows HIV to mutate even further.

DO TREATMENT INTERRUPTIONS MAKE VIROLOGIC SENSE?

During acute or chronic infection, many HIV-infected individuals who are not on drug treatment have high levels of circulating virus that is ostensibly wild type; the virus has no detectable mutations that would make it resistant to drugs like tenofovir (Viread) or lopinavir/ritonavir (Kaletra). But don't be fooled: the virus is under tremendous selective pressure from the immune system and is trying (and largely succeeding) to stay at least a step ahead of neutralizing antibodies and CD8 T cells (and probably other aspects of the immune response). A tremendous amount of virus is produced and destroyed each day, and because of the errors made during replication of the virus, mutations arise constantly that are subject to selective pressures. Anyone infected with HIV is essentially a microcosm of Darwinian evolution. Each virus has a certain fitness, and the dynamics of the host immune response ensure that the virus needs to keep "moving."

Now throw antiretroviral drugs into the mixture. Clearly, the 23 antiretroviral medications approved thus far have changed the face of HIV disease and AIDS forever. The drugs, when prescribed and taken properly' dramatically and rapidly reduce the amount of circulating virus down to "undetectable" levels. In most circumstances, if one looks hard enough, there is evidence that the virus is still probably replicating at very low levels in immune-privileged sites such as the brain and testes and in secondary lymphoid organs. That means that the virus can (and does) still mutate, but at a much slower rate. That also means less chance of viral escape from the medications.

Now remove the drugs, either in a prescribed way (a treatment interruption) or because of lack of compliance. One little-considered fact is that drugs have different half-lives in the body: each is cleared at a different rate. Thus, during treatment interruption the virus may be exposed to only a single agent, perhaps just for a day or two. Because the virus replicates constantly, even such brief exposure can allow drug-resistant viral species to be selected. This has been shown in Africa, where pregnant women who were given a single dose of nevirapine (Viramune) during labor developed nevirapine-resistant virus. In the absence of antiretroviral drugs, the virus is free to mutate within the confines of a depleting immune system. But the more replication and mutation allowed, the greater chance that drug-resistant virus will develop. If that occurs, then the virus is already a step ahead when a drug regimen is restarted.

Surprisingly, drug-resistant viral mutants already exist in the virus population (ie, the virus does not suddenly decide to develop resistance to a drug it encounters). However, the mutants are present at such low levels that we cannot detect them, and in general they are less fit than wild-type virus and so do not replicate to the same extent. Sometimes such viral mutants are simply "archived" in a resting T cell or macrophage, only to appear when that cell is activated. But once the virus is exposed to drug, selective pressure favors the drug-resistant virus (even though it may be less fit). A clinical example of this is with viral resistance to lamivudine (Epivir): the emergent virus has the reverse transcriptase mutation M184V, rendering it less fit than wild-type virus and, therefore, less prone to cause a decline in CD4 T cells.

We all know that interruptions in treatment give the patient a break from complex regimens and the increasing number of side effects and toxicities. But we also know that treatment interruptions give the virus a break as well, allowing increased replication and favoring the development of drug resistance that only becomes apparent after the medications are restarted. That is why more patients are failing treatment interruptions in terms of rapidly rising viral loads and falling CD4 counts, and why more clinicians are looking away from STIs toward other creative treatment options.

By Dorothy E. Lewis, PhD, and Richard E. Sutton, MD, PhD Baylor College of Medicine, Houston, Texas