Absence Makes the HAART Grow Fonder

Most people have heard of the "Berlin patient" by now. First reported in Treatment Issues in November 1997, he was later featured in The New York Times Magazine on June 21, 1998. This man, a patient of Dr. Heiko Jessen in Berlin, with the code name AA, has the first case of HIV infection publicly pronounced as "in remission" as a result of treatment. His experience has touched off a growing discussion over new strategies to break the current stalemate between the virus and highly active antiretroviral therapy (HAART). New hints on how he was able to bring HIV under control emerged during February's 6th Conference on Retroviruses and Opportunistic Infections (see J. Lisziewicz et al., poster 351).

As related earlier, the patient came to Dr. Jessen early in the infection process, before he was producing antibodies to HIV. The man's viral load was about 90,000 copies/ml, and his CD4 cell count was 370 cell/mm3. Dr. Jessen prescribed a combination of ddI, hydroxyurea and the protease inhibitor indinavir, and the patient's viral load almost immediately fell to unquantifiable levels (below 500 copies/ml). Measurable viral loads temporarily returned between days 15 to 22 when the patient interrupted treatment during a testicular infection, for which he was hospitalized. Later, during an episode of hepatitis A, the patient again stopped treatment, this time between 121 and 137 days after he originally started. There was no uptick in the viral load this time, nor did measurable HIV reappear in the blood when patient AA permanently went off therapy on day 176. In a lymph node sample removed at day 564, replication-competent HIV appeared at very low levels, in fewer than one per 10 million CD4 cells.

Now, analyses of AA's immune responses have been conducted on blood samples taken up to two years after treatment first began. While his cell counts have normalized, cell activation is still higher than in uninfected individuals. What is most remarkable is the continued presence of CD4 cells that respond to HIV p24 core protein. Such cells frequently disappear during HIV primary infection. Accompanying this CD4 population is a small surviving population of CD8 cytotoxic lymphocytes (CTLs) primed to attack and kill HIV-infected cells. HIV-neutralizing antibodies, on the other hand, were never detected.

Dr. Jessen's patient joins a few other known cases in which HIV spontaneously disappeared from a person's body (see, for example, D. Schwartz et al., 9th International Conference on AIDS, Berlin, June 6-11, 1993, abstract WS-A21-6). At the 6th Retrovirus Conference, a University of Washington group described a cohort of 37 men who tested negative for HIV antibodies (the standard ELISA assay) but who had a history of high-risk sex and had CTLs responsive to HIV. When these men's resting CD4 cells were checked for HIV infection, viral genes were found in one to two cells per 10 million. This level remained steady over 32 months of observation. Adding cell stimulants to induce this virus to replicate met with at most only slight success, indicating that most of the remaining HIV was defective and permanently quiescent. Interestingly, the recovered HIV genes diverged from those of the men's main sexual partner. Evidently they had acquired an aborted infection through contact with someone else and were now somehow protected from repeated exposures to a different HIV isolate.

The implications for vaccine research are obvious, but the findings also underscore the possibility that HIV infection need not be completely cured. Eradication of the virus is the Holy Grail that has so far eluded our best efforts at drug therapy, and researchers have lowered their hopes of achieving this goal with present techniques. Cells infected with dormant HIV are not susceptible to drug therapy, nor are certain tissues that current drugs poorly penetrate. Some people do not absorb enough drugs to guarantee that sufficient amounts are always present in the bloodstream even if they rigorously adhere to dosing schedules.

Each of these factors probably contributes to a small amount of stubborn residual replication that threatens to escalate out of control should therapy stop. But just maybe, a prepared immune defense could curtail HIV in the event of therapy cessation. Putting HIV in "remission" would mirror what happens with other viruses that infect humans and are never entirely eliminated, for example, those of the herpes family and human papilloma virus.

The Perils of Treatment Interruption

Most experience with treatment interruptions have been negative, with HIV levels immediately rebounding, sometimes to levels that are higher than pretreatment ones. Unlike with patient AA, the HIV went out of control. A study last year by the British Columbia Center for Excellence in HIV/AIDS, for example, followed a set of patients who had had viral loads of less than 50 copies for one to three years when they stopped therapy. Viral loads doubled every day for the first three weeks after stopping (see Richard Harrigan et al., 12th World AIDS Conference, abstract 11152).

Tony Fauci, M.D., who, as head of the National Institute of Allergy and Infectious Disease, directs most of the NIH's HIV research, notes that the sporadic interruptions of therapy can have definite dangers. He said, "The strategy needs to be tested. The stop-and-go game can lead to drug resistance even if it looks so far like the wild type strain remains. But suppose hundreds of thousands of people are stopping and starting therapy. Less common things may happen routinely. A few resistant clones could emerge at the first interruption, but when people go back on drug most of the HIV goes down. While on therapy, the drug-resistant clones will preferentially survive. Next time, there will be a higher proportion of resistant virus, and the viral load won't go back down when therapy is resumed. Close monitoring in a trial would limit the HIV expansion each time and reduce the dangers."

Dr. Joseph Wong of the University of California San Diego, who was one of the first to report on the consequences of interrupting highly active antiviral therapy (see J. Wong et al., Proceeding of the National Academy of Science, November 1997, pages 1274-79), warns about another possible danger. "Our individuals who had rebounded for a month had HIV in their lymph nodes comparable to untreated patients," he noted. The reservoir of latent HIV slowly decays when successfully suppressed. When the drugs are stopped that reservoir is quickly restored, setting the treatment clock back to time zero.

Therapy Interruption after Early Treatment

If HIV generally rebounds when therapy is stopped, what are the salient features of Dr. Jessen's patient AA that enabled him to successfully enter remission? If these can be duplicated in other patients, then they too might be able to stop therapy. One possibility is that patient AA was in the first months of infection when he commenced therapy, before his anti-HIV cellular immune responses had been attenuated by the CD4 cell death associated with HIV. Since such responses generally do exist in the initial, acute phase of HIV infection, other people whose HIV is successfully suppressed during that period might preserve those responses and control the small amounts of persistent HIV after treatment stops. In this, they might resemble the repeatedly exposed seronegative men in the University of Washington study.

The first test of HAART regimens in people with acute HIV was at the Aaron Diamond Research Center in New York. This study enrolled 38 men between July 1995 and February 1997. All received AZT/3TC plus either ritonavir, indinavir or saquinavir/ritonavir combined. Eleven people have stopped taking study drugs or have dropped out of the study. (Seven of the remaining 27 have had occasional slight viral load "blips" -- to a median of about 100 copies/ml -- and are dubbed "partial suppressors.")

Of the 11 drug discontinuations, four are particularly interesting because of the intermittent nature in which they gradually stopped (6th Retrovirus Conference, slide presentation 256 and poster 636). After several stops and restarts, two have now been off therapy for two years, and their viral loads continue to be very low. Like patient AA, both have a vigorous anti-HIV CTL population, and this protective immune response increased during the first drug interruption. Their HIV levels bounce between 50 and 500 copies/ml in one case and 50 and 2,000 copies/ml in the other. These are not particularly exceptional values; they are sometimes encountered in people who received no prior therapy. It is premature to make generalizations based on these two experiences alone.

The other two volunteers discontinued drugs twice, and on the second occasion experienced very rapid HIV rebound to their pretreatment levels. One of these individuals, who at first had strong anti-HIV CTLs, took four months to exhibit viral rebound after the first discontinuation. During that period, the CTL response dwindled away, and the second drug discontinuation was followed by an immediate return of measurable HIV.

But a Massachusetts General Hospital team led by Bruce Walker, M.D., found that tenacious immune defenses are no guarantee that therapy can be safely terminated. (See Eric Rosenberg, 6th Retrovirus Conference, lecture S41.) The Massachusetts General researchers recruited and aggressively treated 20 newly infected individuals. Rather than dissipating, the CD4 cells responding to HIV progressively increased their activity as the time since contracting HIV lengthened. Among the five study participants who have been treated for at least a year, all have CD4 cells active against HIV (as measured by the extent that they proliferate when exposed to HIV antigens in culture). If anything, their responses increased over the second six months.

One of the five has now stopped therapy, after 17 months with HIV below 50 copies/ml, but the result was not promising. For the first three weeks, everything was fine. Despite robust anti-HIV immune defenses, though, measurable HIV levels then reappeared and gradually increased. The patient was put back on therapy after six weeks, when the viral load hit 10,000. It sank back below 50 copies/ml at week 9.

The Hydroxyurea Factor

Franco Lori, M.D., of the RIGHT Institute at Georgetown University and Pavia, Italy, is conducting experiments with monkeys and humans to demonstrate that copying patient AA's situation more exactly does in principle lead to viral control (6th Retrovirus Conference late breaker presentation LB5). For Dr. Lori, the three essential elements are early treatment, regular therapy interruptions and hydroxyurea, which improves the activity of anti-HIV nucleoside analogs by interfering with cells' production of natural nucleotides, the building blocks of DNA and RNA. Hydroxyurea, an old anticancer chemotherapy, also interferes with cell activation and proliferation, and by this mechanism may both down-regulate HIV production in infected cells and make uninfected cells less vulnerable to the virus.

The three monkeys tested so far were first injected with SIV, the simian version of HIV, and therapy with ddI, hydroxyurea and PMPA immediately commenced. PMPA is an experimental nucleotide analog similar in function to, but probably more active than, ddI (Videx). Like ddI, its effect is especially potentiated by hydroxyurea. PMPA is noted for a study published several years ago (Che-Chung Tsai et al., Science, November 17, 1995, pages 1197-9) in which it alone prevented SIV infection when given either before or just after inoculation with SIV. (Treatment continued for four weeks after the injection.)

Dr. Lori's monkeys were treated with the ddI/hydroxyurea/PMPA regimen for four weeks. Treatment was then stopped for six weeks, until viral loads reached 5,000 copies/ml. Treatment was reinstituted for six weeks and then stopped again. After the second interruption, the monkeys' viral loads had remained below 200 copies/ml for 100 days and were still unquantifiable at the time of Dr. Lori's presentation.

In a parallel test, Dr. Lori's associates studied three humans with chronic HIV (who now tested positive on the standard HIV antibody test but who were within a year of contracting HIV). The three were treated with a combination of ddI, d4T, hydroxyurea and indinavir. After three weeks of therapy, viral loads had declined to below the level of quantification (here, 400 copies/ml) -- a result frequently attainable using a regular three-drug combination without hydroxyurea. The volunteers stopped the drugs for the first time at this point. Viral loads rebounded to 5,000 copies/ml by day seven, and therapy resumed. After six months, the volunteers went off drugs again and were able to go 11 to 19 days before reaching the 5,000 copies/ml threshold. A third interruption, after another six months of treatment, lasted 27 to 48 days.

According to Dr. Lori, "the trick is how to teach the immune system to fight HIV using hydroxyurea -- the fourth drug." In another study, Dr. Lori treated ten volunteers within two months of contracting HIV (6th Retrovirus Conference, poster 401). The regimen used was a combination of ddI, hydroxyurea and indinavir. The viral loads of all ten had sunk below 50 copies/ml by the sixteenth week. After 25 to 67 weeks on therapy, nine of the ten had a CD4 subpopulation that responded vigorously when presented with HIV p24 core protein. But these results are not exceptional even without the hydroxyurea, judging by the early treatment observations made at Massachusetts General Hospital.

More notably, six of 12 chronically infected volunteers treated for an average of 122 weeks with ddI plus hydroxyurea alone possessed an energetic CD4 cell response to p24 (6th Retrovirus Conference, poster 401). HIV levels declined very slowly with this regimen: From a mean of 52,000 copies/ml at baseline, viral loads declined to 1,850 at week 40 and 186 at week 122. The fact that the ddI/hydroxyurea combination could maintain viral loads at low, but not undetectable, amounts may be the key to the immune response observed.

According to Dr. Lori's theory, the immune system was exposed to HIV enough to stimulate a response without the infection ever becoming great enough to decimate that response -- a quasi-vaccine, as it were. The RIGHT Institute is now initiating larger trials to test this theory that will involve 30 monkeys and 40 to 80 humans.

The Future of Remission

Viruses that are held in check by the immune system but never completely eliminated have the nasty habit of coming back when the immune system is weakened by age, malnutrition, acquired immune deficiency or any other stress. Chicken pox, for example, recurs later in life as shingles, whereas Epstein-Barr virus is associated with lymphoma. What, then, will an HIV breakthrough look like -- assuming that the strategies to provoke protective anti-HIV immunity succeed in putting the virus in remission? The answer is, no one knows. It might look like primary HIV infection or long-lasting HIV might ultimately trigger malignant changes in cells. Either way, drug therapy could be reinstituted and might work as well as it did the first time. Or maybe not, depending on the status of the body at the time. The future of remission is entirely speculative, but putting HIV infection in a resting state does not mean that it can be forgotten.

Back to the GMHC Treatment Issues February 1999 contents page.