The overarching AIDS image from this year's Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC, held in Toronto from September 28 to October 1) was a huge slide projection of a cell bursting forth with HIV displayed during Robert Siliciano's keynote address. This was not just any HIV-infected CD4 cell. It came from a patient who had been on triple combination therapy for 30 months and had viral loads below the limit of quantification since the first two months of therapy. The CD4 lymphocyte was a "resting" or quiescent cell that was latently infected with HIV all that time, it seems. When the cell was stimulated in the lab, the HIV genes within it were activated along with the cell's, and active viral replication commenced. The photo was a warning that HIV will not be banished from the body so easily. It will take years and years.
(Note that the existence of latent HIV in individual cells differs from the now debunked concept of "clinical latency" in which HIV supposedly lays dormant in the body for years during the asymptomatic period of infection. In truth, there are enormous amounts of HIV replication in untreated individuals at all stages of infection.)
Dr. Siliciano (pronounced "Silicano") and his colleagues from John Hopkins University in Baltimore have now published their work in the journal Science (Nov. 14, 1997, pages 1295-1300). An accompanying article by Joseph Wong and his associates at the University of California San Diego also appeared on the same topic (pages 1291-4). In the following week's Proceedings of the National Academy of Science, Drs. Tae-Wook Chun, Tony Fauci and others at the NIH weighed in with their own examples of latently infected CD4 cells found in patients after months of highly successful triple combination therapy for HIV (PNAS, Nov. 25, 1997, pages 13193-7).
The Johns Hopkins researchers looked at the peripheral blood cells taken from 22 patients who had been on triple or quadruple drug therapy (typically two nucleoside analogs plus a protease inhibitor) for 5 to 30 months. Pretreatment viral loads ranged from 10,000 to over a million HIV RNA copies per milliliter of plasma. In each case the viral load had fallen below the 200 copy/ml limit of quantification of the assay in use by the second or third month of treatment and remained so through the time the blood was drawn for testing for latently infected cells. Detection of such cells took place in a highly artificial environment: CD8 cells, B cells, monocytes, natural killer cells and activated CD4 cells were almost entirely removed, such that no immune protective mechanism remained that could suppress replicating HIV. Special stimulants induced the purified cell cultures to proliferate and produce HIV if infected by competent virus. To augment whatever HIV was produced, the stimulated cells were mixed with cells taken from an HIV-negative donor. In 4 of the 22 study participants, not enough resting CD4 cells could be isolated to run this assay, but replicating HIV was elicited from all the others.
Similar techniques were used by researchers at the University of California San Diego. Blood samples were taken from six individuals who had received AZT/3TC/indinavir for two years as part of a trial (Merck's protocol 035). All six had had HIV loads of less than 50 RNA copies/ml (the quantification limit for the new ultrasensitive PCR test) for over a year. Nonetheless, replicating HIV appeared in every case under the highly supportive conditions in which study participants' CD4 cells were cultured.
And lastly, the Chun and Fauci group of investigators extended the findings in Science with its own observations of 13 individuals. All 13 were on triple or quadruple combination therapy but otherwise presented a mixed bag: Four were initially treatment naïve and one had received only 3TC before. CD4 counts ranged from 15 to 806 prior to combination therapy, whose duration had ranged from 5.5 to 13 months before blood samples were drawn for this study. Nine of the 13 had viral loads below 500, which is the limit of quantification of the bDNA assay used here (and which gives readings about half as large as the PCR assays used in the other studies). In these nine, the frequency of resting CD4 cells carrying infectious HIV was about 1.5 per million cells. For the four volunteers with detectable virus, the frequency was somewhat higher. These figures square with the Johns Hopkins analyses.
Does Residual Virus Equal Residual Replication?
The NIH investigators raised an important issue: Is the pool of latently infected cells slowly turning over, i.e., are the cells slowly dying and being replaced by newly infected ones? Or are the latently infected cells very long-lived, like uninfected resting CD4 cells? They would then represent a viral reservoir left over from the pretreatment period when HIV was running rampant. The NIH investigators did find significant quantities of "unintegrated" viral DNA that had not yet merged with the chromosomes in the surrounding cell. Because the strands of unintegrated DNA are unstable, they presumably derive from new HIV that has recently penetrated the cells containing them.
The Johns Hopkins group noted in their paper that in the one individual with undetectable viral load who was tested, the frequency of activated CD4 cells releasing infectious HIV was a miniscule but detectable 0.0002%. (That incidence was ten times higher in patients with viral loads of 200 to 1,000 -- just over the quantification threshold.) Nonetheless, most researchers are confident that even if a few cells do release HIV in patients considered "fully suppressed," those virions are going nowhere because the antiviral drugs are blocking them. Dr. Joseph Wong of the University of California San Diego argued, "The lack of emergence of drug resistance [in these patients' HIV] indicates a vanishingly low rate of replication. Since the genetic makeup of the pre- and posttherapy virus is very similar, the latent HIV probably was there from the beginning."
Still, the evidence that a few cells are producing HIV demonstrates that the highly artificial conditions that promoted the activation of latent HIV in the lab does have some relevance to the processes in the body. Presumably as different populations of resting cells are activated to fight new infections during the natural course of events, any cells that harbor competent, integrated HIV genes will start churning out virus. This possibility raises the prospect that antiviral therapy will have to be continued even in patients with undetectable viral loads until all the latently infected resting cells are cleared from the body.
But how long is that? No one seems to know at this point. It is of course widely remarked that people who interrupt therapy in the first couple of years have very rapid HIV rebounds (see below). This resurgence is probably mainly due to chronically producing cells such as infected macrophages. The present findings extend the treatment horizon by adding a new long-lived HIV reservoir. The new reservoir greatly postpones the point at which HIV eradication might be achieved even in people whose viral loads approach zero under treatment.
Draining the Reservoir
David Ho, director of New York's Aaron Diamond Center, has made a number of reports over the past two years concerning how soon treatment could be terminated -- his most recent estimate being about three years. His estimates were based on the observed decay rate of HIV in the bloodstream and could not account for minor hidden reservoirs. In reviewing the recent papers (he was a coauthor with Robert Siliciano and his Johns Hopkins colleagues), Dr. Ho now says, "Since the papers showed no new drug-resistance developing, the drugs are doing what they are supposed to, which is to suppress viral replication by stopping the infection of new cells. [The latently infected resting cell reservoir] was something we hadn't analyzed. Now we have to figure out now how to remove the problem safely."
Anthony Fauci, head of the Laboratory of Immunoregulation and senior author of the NIH paper, voiced the most common proposal about how to speed elimination of this latent HIV reservoir: "You may want to find a way of flushing out the virus by stimulating cell replication and then suppressing it with a combination of the antiviral drugs already on the market." Tumor Necrosis Factor (TNF) and several other immune messenger cytokines are candidates for testing, but Robert Siliciano pointed out the dangers of this procedure: "Things that activate latent cells activate the immune system in general, causing tremendous syndromes like toxic shock."
One way to alleviate this threat is to activate only a small percentage of the immune cell population at a time, by sequentially inoculating patients with one or a few antigens. This would be a terribly slow process. Even if some safe, reasonably speedy technique could be found, there is something counterintuitive about trying to stimulate virus on the one hand while trying to suppress it on the other. It's true that HIV does not seem to develop drug resistance in the fully suppressed patients in these studies. But stimulating the latent HIV will have the effect of accelerating any possible evolution of resistant mutants.
An alternative approach would be to discourage cell and HIV activation with anti-inflammatory or immune suppressive agents. Resting CD4 T-cells would stay quiescent until they eventually died. But that can be counterproductive too. Essentially, the immune suppressants would tend to reestablish the immune suppression that the antiviral drugs alleviated by eliminating the active HIV. A possible way around this conundrum would be to find agents that specifically inhibit the creation of new viral RNA and proteins from the HIV DNA genetic template lodged in infected cells. One potential candidate is a small molecule created by researchers at the Swiss pharmaceutical company Novartis. Last spring, they reported on an agent that blocked the TAR region at the beginning of the HIV genome. In cells, the HIV tat protein binds to this region to help start the gene transcription machinery that leads to new virions. In the test tube, very low concentrations of the Novartis compound were highly effective at interrupting the HIV life cycle. According to the researchers, this compound is the first example of an antiviral compound that selectively inhibits a protein-gene interaction.
Betting on the Immune System
It is not clear, though, that any particular therapy is needed for latent cells. It is possible that as the immune system recovers in the nearly HIV-free environment, anti-HIV cytotoxic lymphocytes (CTLs) will be able to quickly kill the few cells with active HIV infection and prevent spread of the virus. Automatic containment by the immune system would be the most desirable scenario. In the words of David Ho, though, "At any given time the number of infected cells turning on might be small enough for the immune system to handle. But no one would bet on that."
Another new report in Science points out a remarkable weakness in the immune defense against HIV (Nov. 21, 1997, pages 1447-50). According to the authors, who were from the Harvard Medical School, the HIV-specific CD4 T-helper cell proliferation that promotes the CTL response to the virus disappears quite early, apparently quashed by the consequences of the massive HIV levels occurring during the initial acute phase of HIV infection. Once potent therapy brings HIV levels down, it might be possible to activate and proliferate whatever anti-HIV cells happen to exist by means of a vaccine. Such a "therapeutic" vaccine could use new biotech methods such as non-HIV viral vectors to implant an incomplete set of HIV genes into cells. The result would be to benignly mimic viral infection and promote the proper immune response. This technique was suggested by Dr. Siliciano in an interview and by the Harvard Medical School authors of the just cited November 21 article.
An alternative vaccine approach is the Immune Response Corporation's inoculum consisting of killed, envelope-depleted HIV particles stripped of their envelopes and mixed with a special adjuvant. As it happens, a preliminary trial is already underway to observe the immune responses generated from periodically vaccinating the IRC product into recipients of potent anti-HIV combination therapy. Results are not expected for another year. (For more on this trial and on therapeutic vaccines in general, see Treatment Issues, Dec. 1996.)
The Mad Scientists Go to Work
If triggering the anti-HIV immunity in the body proves elusive, it might be possible to breed anti-HIV immune cells in the test tube and then infuse them into patients. There have been several attempts at the simplest approach: isolate the highly sensitive CTL cells that target and kill HIV-infected cells in patients. Once multiplied in the test tube, it would seem a simple matter to send these killer cells to work in the patient from which they originally came. The results have been poor, unfortunately. In one report at the NIH-sponsored conference, "New Opportunities for HIV Therapy," held last June, Dr. Philip Greenberg of the University of Washington showed beautiful slides of the CTL clones homing in on HIV-infected cells within biopsied lymph node tissue. Viral loads did go down somewhat, too, but the effect of the CTL infusion proved temporary. Dr. Greenberg blamed the loss of response on lack of HIV-specific CD4 T-helper cells. He now proposes to isolate and breed such CD4 clones and include them when infusing the CTLs. Such a strategy has a chance if potent antiviral drugs can protect the new cells from HIV.
So as to go beyond the body's natural responses, Cell Genesys, a small biotech company, has developed bioengineered CTL cells that contain a CD4 receptor. These cells can latch onto and kill HIV-infected cells bearing the virus' gp120 envelope protein on their surface. This approach works well in the test tube (see Otto Yang et al., Proceedings of the National Academy of Science. Oct. 1997, pages 11478-83), but it did not exhibit a distinguishable benefit over antiviral therapy in a preliminary trial. Cell Genesys is now conducting several trials that combine the genetically modified CTLs with similarly modified CD4 T-helper cells to enhance and prolong the CTLs' activity.
In yet a further step, it might be possible to replace the immune system entirely with special bioengineered viruses that would infect and kill cells as the HIV within them becomes activated. There have been three recent reports using viruses manipulated to contain on their surfaces the CD4 receptor plus one of the chemokine receptors that various strains of HIV use to latch onto and fuse with new cells to start the infection process. When already infected cells start producing HIV, the viral envelope protein gp120 appears on their surface, and these synthetic viruses can bind to it and enter the cell, sowing death via their metabolic processes.
The new reports describe preliminary experiments based on this strategy. Two of the studies involved viruses that were rendered unable to replicate or kill cells but did selectively infect HIV-producing cells. In one study, a version of rabies attacked lymphocyte-tropic HIV, which has affinity to the CXCR4 chemokine receptor. Another employed a version of HIV itself, with CD4 and CCR5 substituted for its natural envelope so as to target macrophage-tropic HIV. (See respectively Teshome Mebatsion et al., Cell, Sept. 5, 1997, pages 841-7, and Michael J. Enders et al., Science, Nov. 21, 1997, pages 1462-4.)
The third study (Matthias J. Schnell et al., Cell, Sept. 5, 1997, pages 849-57) went a step further. It started with a virus known as VSV that disrupts and kills animal cells within hours of infection but is harmless to humans. VSV's envelope was exchanged for a CD4-CXCR4 construct, and researchers ended up with a virus that could indeed multiply within the HIV-infected cells it had entered. In lab cell cultures, it killed these cells as it multiplied and then ferreted out other HIV-producing cells, reducing HIV levels by 300- to 10,000-fold. But the dually infected HIV-VSV cell cultures always allowed some low-level HIV production to remain. VSV is a rapid cell killer, but apparently not rapid enough. Many safety issues also are raised here. For one, there is a slight possibility that a gp120-expressing VSV could be created in a cell infected with both viruses. The result might be a new, highly pathogenic human viral disease.
A lot more experimental work obviously is needed before reengineered cytopathic viruses can be accepted as a valid therapeutic strategy. Still, the concept suggests a number of more conservative alternatives: Incompetent or benign viruses bearing CD4 and chemokine receptors could be used to flag HIV-infected cells so that they could be killed by CTLs responding to cells infected by the engineered virus rather than HIV. Another idea is to construct tiny drug-containing fat globules (liposomes) that would bear the proper receptor complexes and latch onto and immediately destroy cells with active HIV replication.
Don't Walk Off in a Huff
Marty Markowitz of the Aaron Diamond Center in New York and one of Dr. Siliciano's co-authors worries, "I hear that patients are giving up on therapy after hearing about [the latent HIV] studies because they think that treatment has to be endless." As pointed out above, it is not clear that treatment has to be forever. Further study will define the necessary span of treatment, which will certainly need to last for many years. What is already clear from present studies is that stopping treatment prematurely rapidly sets the clock back to zero and promotes the rise of drug resistance. Having stopped, a patient will not be able to simply resume at some future date.
In an article published in the Nov. 11, 1997 issue of the Proceedings of the National Academy of Sciences (pages 12574-9), San Diego investigators reported on the one year blood and lymph node HIV levels of ten persons taking AZT/3TC/indinavir as part of a trial (Merck protocol 035). Three of the individuals interrupted their therapy in the second half of the year. One, who stopped for 29 days, had an immediate viral rebound to pretreatment levels, with the decline in viral load recommencing upon resumption of treatment. The other two individuals, who stopped therapy for only a few days, also saw their plasma viral loads rebound, and this rebound was accompanied by the appearance of HIV resistance mutations against 3TC and indinavir. These two were switched to different regimens.
In its pilot triple drug therapy trial (AZT/3TC/nelfinavir) in treatment-naïve patients, the Aaron Diamond Center can point to a case in which an individual's nonadherence to dosing schedules led to a reappearance of detectable HIV after a period of undetectability. But two study participants who conscientiously adhered to treatment schedules also experienced viral rebounds at 9 to 12 months. The return of HIV in these cases has been ascribed to 3TC and nelfinavir resistance mutations that were first detected after therapy began. Both patients were characterized by very low baseline CD4 counts (and therefore poor potential for controlling any HIV replication that persisted during treatment). They also had very high viral loads at the start of treatment. Some of the HIV in these two either accidentally possessed suitable resistance-conferring mutations prior to therapy or drug resistance was able to evolve during the residual replication in the months after starting therapy.
Even under the best conditions, then, current therapies exhibit a 25% failure rate (if one can infer anything exact from such a small study). Dr. Markowitz says the lesson from these cases is that "we need four drugs, not three." Indeed, the two study participants were switched to different, four or five drug combinations (d4T/ddI/ritonavir/saquinavir plus nevirapine in one case), and their slowly rising viral loads sank from under 10,000 back down to unquantifiable levels (below 400 copies of HIV RNA/ml). Now, one of these individuals is off therapy after coming down with acute infectious hepatitis.
You can keep adding drugs, but the current types have too many problems to expect even successful patients to take them forever. There is a need for more drugs, but it seems that innovative practical strategies are needed to confine the long-lasting viral reservoirs. Otherwise, like the never-ending series of Alien movies, the HIV saga will keep coming back with new episodes.
This article was provided by Gay Men's Health Crisis. It is a part of the publication GMHC Treatment Issues. Visit GMHC's website to find out more about their activities, publications and services.