International Association of Physicians in AIDS Care, Journal: March 1998 - Volume 4, Number 3
Mark Mascolini

Introduction

Chicago's a tough town. And you don't have to read Nelson Algren to find out. They'll arrest you for leaning against subway doors in Chicago--a threat the motorman incants with menacing regularity. Chicago's sports clubs favor names that start with in-your-face plosives--Bears, Bulls, Blackhawks. Even the wimpy-sounding White Sox are the only team ever caught cheating in a World Series. Two blocks beyond the tinseled facades of Michigan Avenue's "Miracle Mile," scrappy dogs lead scruffy masters along dim byways. Neither notice that keen, fitful gusts off the lake threaten to whipsaw the most Bunyanesque out-of-towners.

Yet few AIDS docs griped when organizers scheduled the 5th Conference on Retroviruses and Opportunistic Infections for a sepulchral February week in Chicago. And when Pittsburgh's John Mellors, MD, interrupted his plenary talk with a plug for Orlando in 1999, few cheered.

No surprise there, probably because the crowd that does the AIDS meeting circuit admits few sybarites. Their skins have thickened--to winter's blasts and each other's barbs--maybe because they've long since agreed on at least one thing: HIV's a tough disease. Tough to have, tough to figure, tough to treat. "Tough to have" is more than most meetings dare explore. The Retroviruses Conference, in fact, makes no bones about eschewing social, political, and nearly all economic issues. But the Chicago meeting spotlighted "tough to figure" and "tough to treat" as few conferences ever have:

• What trips the immune dysregulating switch, and how soon?
• Can dysregulation be permanently reversed?
• If so, is it ever too late to turn things around?
• If it's not too late, how long will it take?
• Can antiretrovirals do the job themselves, or do they need help from immune modulators?
• How soon should drugs be switched when viral load starts rising?
• What do undetectable--and rising--viral loads mean anyway?

Despite progress toward resolving such riddles, at meeting's end HIV seemed a tougher opponent than many may have imagined when leaving this conference a year ago. The cogs of its dysregulating machine, scientists said in Chicago, begin meshing in the first days of infection, even before seroconversion. And, although highly active antiretroviral therapy--HAART--often slows, sometimes stops, and occasionally reverses the wheels of immune destruction, no one knows whether the wheel of misfortune can ever make a full circuit, even in the luckiest few.

When HIV makes a new host its home, it does two things right away. For one, said Robert Siliciano, MD, PhD, of Johns Hopkins, it stakes its claim to memory T cells with dizzying alacrity. This piddling residue of infected resting cells--the pebble that brought down a HAART-wielding Goliath of eradication¹--can be detected even before an infected person seroconverts, Siliciano disclosed. Another thing that begins in the first weeks of infection, reported Bruce Walker, MD, of Massachusetts General, is the decimation of CD4⁺ T helper cells primed to respond to HIV.² When people in the throes of acute infection get treated, Walker found, this elite corps of infection fighters stays intact (abstract L4^*). But once HIV infection enters its chronic phase, the number of HIV-specific CD4⁺ helpers wanes dramatically, and the number of such cells correlates inversely with plasma viremia. HAART does not restore HIV specificity to CD4⁺ cells that have lost it, at least not in the first few years of therapy.

A third early trick of HIV is homing to cozy brain cells, where it can evolve differently from its peripheral forebears and so raise troubling questions of drug resistance. Evidence for such precocious breaching of the central nervous system (CNS) comes from SIV-infected monkeys, inwhich Janice Clemens, PhD, of Johns Hopkins, has been tracking retroviral escapades (abstract S18). Infected lymphocytes and macrophages, both vulnerable to so-called macrophage-tropic virus, can slip into the CNS during the first weeks of SIV infection, she reported. To make matters worse, infected cells of both types may later meander back into the periphery, and so possibly reintroduce early strains to the circulation, or strains that have evolved differently from peripheral virus. And the news gets worse, at least for monkeys. Working with University of Pennsylvania chemokine receptor expert Robert Doms, PhD, Clemens showed that SIV can bypass the CD4 receptor entirely to infect endothelial cells of CNS microvessels. Instead, SIV can make CCR5, the principal coreceptor on macrophages, its primary receptor.

But these gloomy tidings were mitigated by happier news, some of it equally astonishing.

• Working in the lab of Ashley Haase, MD, at the University of Minnesota, Zhi-Qiang Zhang, PhD, found that lymphoid architecture razed beyond recognition by HIV can gather itself into an identifiable semblance of it former self just months after HAART begins (abstract LB13). One person who managed this bravura rebuilding program started treatment with nary a measurable CD4⁺ T cell.
• Two of three patients studied by Duke University's Barton Haynes, MD, enjoyed hefty rebounds of naive CD4⁺ cells with HAART (abstract S44). What made these people exceptional is that they did it without a thymus. All had a thymectomy years earlier because of myasthenia gravis. (But, Haynes wondered, what does that say about the "naive" cells counted in HAART-helped people with a thymus?)
• Several groups reported the regular replenishment of (apparently) naive CD4⁺ T cells during HAART, banishment of manic activated CD8⁺ T cells, diversification of the T-cell repertoire, and--perhaps the payoff--plunging rates of opportunistic infections and neoplasms. A case-control study in New York City seemed to confirm that protease inhibitor regimens account directly for prolonged survival with HIV infection (abstract LB7).

To be sure, even the most optimistic reports came freighted with caveats. Many a speaker left attendees with a list of "unresolved questions" worthy of a Sphinx not pestered by an Oedipus. And no one promised more than a continuing contest with the virus, a contest--like a street fight--with few if any rules.

Return to Top of Page

Caution: New Immune System Under Construction

Two HAARTy years of intense study have established that the grail of immune reconstitution--variously defined--can at least be grasped. But not by all. What distinguishes those who claw their way back to a semblance of health from those whom HAART fails? Age is a certain factor, said Ashley Haase at a day-one press conference. He and others estimate that adults, whether infected with HIV or inflicted with chemotherapy, can crank out new T cells at 1 percent the rate enjoyed by infants. And how severely HIV cripples the immune system before effective therapy begins also figures in the equation.

Is there a point of no return past which even the strongest therapies avail one nothing? "I'm sure that's the case," said Haase. But that point may not be defined by CD4⁺ counts. As already noted, one person with the lowest CD4⁺ count possible--0 cells/mm³--managed to begin piecing together a network of follicular dendritic cells from the utterly ruined debris left by the virus, and he did so within two months of starting therapy. But, Haase cautioned, "this is not a uniform response." Biopsies of others with countable CD4⁺ cells showed lymphoid tissues still in shambles.

In Haase's analysis of nine persons treated for up to 14 months with zidovudine (ZDV), lamivudine (3TC), and ritonavir, CD4⁺ T cells returned to lymphoid tissue (in this study, tonsils) at a rate of 110 million daily. T cells with a naive (CD₄₅RA⁺) phenotype accounted for 80 million of this daily total--exactly the return rate found when cancer chemotherapy nearly wipes out lymphocytes in young adults.³ One might be forgiven for speculating that these coinciding numbers are more than coincidence. At any rate, Haase proposed, the result "argues that whatever the source of these cells is, that mechanism has not been irreversibly damaged even in relatively late disease." (The nine treatment-naive study participants had a mean baseline CD4⁺ count of 239 cells/mm³ when therapy began.) This study, spearheaded by Zhi-Qiang Zhang (abstract LB13) and published during the conference,⁴ also encouraged the investigators to note that "CD4⁺ T cell populations can be partially restored by control of active replication without eradication of HIV."

Overall, the average percentage of CD₄₅RA⁺ CD4⁺ cells in lymphoid tissue climbed from 24 percent of total CD4⁺ cells in tissue before treatment to 36 percent a year into therapy. But Zhang noted in his conference presentation that "the kinetics and extent of the [T-cell] regeneration vary from individual to individual." Similarly cautious, Haase observed at a press briefing that CD₄₅RA and CD₄₅RO are mere immunohistochemical markers thought to reflect, respectively, naive and memory cells. "These aren't functional tests," he stressed. And although the dazzling reconstruction of dendritic-cell networks in some patients limns the structure of uninfected lymph nodes, Haase suggested that resemblance may have all the depth of a Hollywood facade.

While Haase's team was sizing up T-cell repopulation in lymphoid tissues, Dutch investigators led by Frank Miedema, MD, PhD, were studying individuals from the same trial of ZDV/3TC/ritonavir to gauge naive and memory cell rebounds in blood (abstract 20). As one would hope, T-cell dynamics in blood closely resembled those in tissues. (Miedema's results, like Haase's, were published around the time of the meeting.⁵) Three weeks after therapy began, the Dutch group found, a surge of memory CD4⁺ and CD8⁺ cells into the peripheral circulation essentially ended. After that, all gains came from naive cells. Miedema calculated that naive CD4⁺ cells swarm back into blood at a slow but steady rate of 57 million daily, close to Haase's estimate of 80 million new naive CD4⁺ cells daily in lymphoid tissue. The Dutch team concluded optimistically that these "presumably de novo produced naive T cells" may rebuild the full T-cell receptor repertoire. (A separate study by Brigitte Autran, MD, and Patrice Debr, MD, published back-to-back with Miedema's report, directly confirmed "normalization" of the CD4⁺ T-cell receptor repertoire in eight individuals who responded to HAART within the first three to six months of treatment.⁶)

Return to Top of Page

What happens in the first weeks of infection?

As some researchers slowly define the timing and scope of immune reconstruction, others say they already know how soon the deconstruction starts: quickly. Massachusetts General's Bruce Walker took attendees through the details of an ongoing study showing that CD4⁺ cells lose the capacity to proliferate robustly when exposed to HIV proteins soon after seroconversion in all but a smattering of fate-favored long-term nonprogressors² (abstract L4). But it's a different story when people with acute HIV infection get treated promptly. In seven individuals with primary infection, Walker and coworkers saw no evidence of HIV-specific T-helper responses when they first looked in peripheral blood, probably, they infer, because HIV destroys T-helper cells as soon as they're activated. But six weeks to three months after HAART began, p24-specific and gp160-specific responses returned. And although plasma viremia remained undetectable in all seven of these individuals--for 11 months in the longest-treated man--HIV-specific CD4⁺ T-cell responses persisted and, in fact, grew.

Next Walker and colleagues blindly assessed proliferative responses in four people who began treatment during acute infection and in HIV-positive untreated controls. He correctly singled out the treated patients by identifying their strong virus-specific proliferative responses. So, Walker concluded in Chicago, 11 of 11 individuals treated in early infection all generated these responses. "If you don't treat people in acute infection," he said, "you don't get these responses." Even if HAART starts one year after acute infection, virus-specific proliferative responses can't be detected for up to two years of therapy. "There clearly appears to be a window of opportunity early on," Walker argued. "We think we're basically restoring an immune response in these individuals that is analogous to what one sees in long-term nonprogressors. . . . We hypothesize that if you come in with antiviral therapy [before chronic infection], you protect these activated [CD4⁺] T cells. . . . This allows for maintenance of CTL [cytotoxic T-lymphocyte] responses, and we would predict that we'll end up now with a situation of nonprogression."

Another look at T-cell behavior in early infection came from a collaboration between the San Diego group of Douglas Richman, MD, and Oxford's Angela McLean, PhD (abstract 587). Like Walker, they found that HIV quickly deranges the T-lymphocyte armies, in this study within six to 35 days after acute retroviral illness began in eight individuals. But HAART did not start setting straight the parameters Richman and McLean measured, unless the initial damage was slight. Therapy failed to remedy completely either the profound initial drop in CD4⁺ cells or the reciprocal expansion of CD8⁺ cells. People whose CD4⁺ contingent took the biggest early hit had the weakest rebounds, even when therapy induced a virologic response as robust as those in people who lost fewer CD4⁺ cells. And despite sustained suppression of viral replication from 100 to 450 days after therapy began, increases in memory (CD₄₅RO⁺) T cells and decreases in naive (CD₄₅RA⁺) lymphocytes remained uncorrected. HAART did begin to corral a stampede of activated (CD₃₈⁺ HLA-DR⁺) CD8⁺ cells, but when one person stopped treatment, the stampede started all over again.

Return to Top of Page

Immunohistomysteries

The caution Ashley Haase voiced over reading too much into immunohistochemical tags on T cells cannot be lightly dismissed. Careful listeners at the 5th Conference on Retroviruses surely noted such studied locutions as "RA-positive, 62L-positive CD4s, thought to represent naive cells" and "CD₄₅RO-positive CD4 cells, that is, cells with a memory phenotype."

Demonstration of the thin line that separates one cellular phenotype from another came in a February report from the Amsterdam group of Frank Miedema⁵ (see above). In this analysis of T-cell comings and goings in people taking ZDV/3TC/ritonavir, Miedema classified as "naive" those cells expressing CD₄₅RA and CD₆₂L and as "memory" those expressing CD₄₅RO. Both designations are standard these days. But these investigators also earmarked CD₄₅RA⁺ cells lacking CD₆₂L as memory lymphocytes, whereas other groups call any CD₄₅RA⁺ cell naive. So there can be little surprise if phenotypic results vary from study to study, even if one writes off other looming confounders such as differences in study populations and drug regimens.

Yet no one can deny that assessment of a lymphocyte's molecular mantle, imperfect as it may be, adds an important dimension to the study of antiretroviral efficacy. Pioneers in this field, such as UCLA's Janis Giorgi, PhD, have indeed correlated the ebb and flow of immune cell markers with clinical outcome. She presented one such report at the Retroviruses conference (abstract 110). This cohort study compared 11 individuals who died within six months of a first CD4⁺ count below 50 cells/mm³ with 26 people who survived at least 18 months after reaching that CD4⁺ benchmark. Markers of CD8⁺ cell activation--CD₃₈ and HLA DR--turned out to be much stronger predictors of death than plasma viral load in these people with advanced disease. The fast progressors differed significantly from the slow progressors in CD8⁺ cell expression of CD₃₈ (P=0.001) and in percentage of HLA DR⁺ CD₃₈- CD8⁺ cells (P=0.002). At this stage of HIV infection, neither relative reserves of naive and memory T cells nor chemokine receptor use correlated with survival. Giorgi and her colleagues suggested three explanations for the predictive power of CD8⁺ cell activation markers in late disease: "(1) Activated lymphocytes are functionally anergic. (2) Heightened lymphocyte activation and turnover lead to immunologic exhaustion. (3) Activated CD8⁺ T cells produce factors that enhance viral replication."

What happens to these activation markers when late-stage patients begin treatment with protease inhibitors? Their expression falls dramatically, according to Richard Hengel, MD, of Emory University, and colleagues at the CDC (abstract 166). These investigators logged changes in CD₃₈, HLA DR, and other signals of cellular behavior in 10 people with baseline CD4⁺ counts above 100 cells/mm³ and in 10 with counts below 100 cells/mm³. Even for the patients with more advanced disease, expression of CD₃₈ and HLA DR on CD8⁺ cells dropped after 16 weeks of therapy. The CD₃₈ decreases were most impressive, as fluorescence intensity of that marker waned 45 percent on CD4⁺ cells and 56 percent on CD8⁺ cells. But these researchers cautioned that "more investigation is needed to attribute functionality to observed phenotypic changes within lymphocyte subsets."

At the Chicago meeting, the most exhaustive analysis of immunologic changes during HAART came from multiple presentations of ACTG 315 results. For 48 weeks this study tracked immunologic changes in 53 people with CD4⁺ counts between 100 and 300 cells/mm³ who were naive to protease inhibitors and 3TC when they began treatment with ZDV/3TC/ ritonavir (abstracts LB14, S30, and 160). Like the Emory-CDC group (preceding paragraph), ACTG 315 investigators found about a 50 percent decrease in expression of the CD₃₈ activation marker on both CD4⁺ and CD8⁺ cells. And, with one important exception, this study confirmed the findings published by Brigitte Autran, who studied eight treatment-naive individuals with a mean baseline CD4⁺ count of 162 cells/mm³, who took ZDV, zalcitabine (ddC), and ritonavir for a year.⁷

Among the 34 people who completed 48 weeks of therapy in ACTG 315 (most dropouts were attributed to intolerance or toxicity), the median baseline CD4⁺ count was 192 cells/mm³ and the median plasma RNA about 83,000 copies/mL. As in Autran's study, and in many others, the first jump in CD4⁺ cells could be attributed to the memory subset. Elizabeth Connick, MD, from the University of Colorado, reported a steep rise in memory (CD₄₅RO⁺ CD₄₅RA-) CD4⁺ cells in the first two weeks of treatment, but no substantial gain in memory cells from then through week 48. After the second week of treatment, the ACTG 315 team attributed the continuing overall rise in CD4⁺ cells to naive (CD₄₅RA⁺ ₆₂L⁺) lymphocytes. Again echoing Autran's findings, ACTG 315 showed an initial spike in CD8⁺ T cells, then an overall decline after week 12. Disappearance of memory CD8⁺ cells accounted for this fall, as the number of CD8⁺ cells with a naive phenotype slowly rose. Results of ACTG 315 also confirmed Bruce Walker's finding that HIV-specific proliferative responses cannot be expected after a prolonged virologic response to HAART. Connick reported no significant increase in such responses to HIV p24, p66, or gp120 after 48 weeks of treatment.

But ACTG 315 did not confirm one felicitous finding in Autran's study--recovery of CD4⁺ proliferation to recall antigens, specifically to cytomegalovirus (CMV) and Mycobacterium tuberculosis antigen, in three of six individuals.⁷ In contrast, Connick found no proliferative responses to tetanus toxoid, Candida, streptokinase, or alloantigens. Perhaps, the ACTG team speculated, Autran did not detect actual reconstitution of lost responses; instead, reactivation upon re-exposure to familiar antigens may have perked up proliferative responses. Failure to recreate responses to HIV proteins (confirming Walker's results) or to recall antigens (contradicting Autran's results) led ACTG 315 investigators to propose that "in moderately advanced HIV disease these responses may not be reconstituted despite potent antiviral therapy."

But another study, reported by the University of Pittsburgh's Charles Rinaldo, PhD, sided with Autran in the recall antigen debate (abstract 169). This analysis looked at the question of reconstitution from a perspective different from ACTG 315 and Autran's published report: The protease inhibitor was indinavir instead of ritonavir, and the follow-up lasted almost two years, instead of one. Rinaldo and colleagues studied nine individuals from each arm of the Merck 035 trial, which compared ZDV/3TC/indinavir with ZDV/3TC or indinavir alone in people naive to 3TC and protease inhibitors but heavily pretreated with ZDV. People in the triple-therapy group enjoyed 12-fold increases in CD4⁺ cell proliferative responses to pokeweed mitogen, 10,000-fold increases to Candida, 3500-fold increases to tetanus toxoid, and 2700-fold increases to streptokinase. People in the ZDV/3TC and indinavir monotherapy arms had only moderate and transient responses, even when given open-label triple therapy after six months of the blinded comparison. Once more, though, these investigators found no substantial CD4⁺ T-cell reactivity to HIV antigens.

And Autran's group weighed in again on this question, this time in a study of 20 individuals (seven naive to protease inhibitors and 13 experienced) with CD4⁺ counts below 250 cells/mm³ but no AIDS-defining illnesses except cutaneous Kaposi's sarcoma (KS) (abstract 172). Six in this cohort from Paris's Hôpital Pitié-Salpétrière took ritonavir plus two nucleosides, and 14 took indinavir plus two nucleosides. After 12 months, the treatment-naive patients scored better gains in CD4⁺ cells (P=0.03) and deeper drops in plasma viremia (P<0.001) than did the experienced patients. Autran classified six of seven naive individuals and four of 13 experienced individuals as "immune responders." Among responders, she reported, CD4⁺ cell reactivity to CMV and tuberculin antigens increased from 20 percent at baseline to 75 percent at month 3 and was sustained through month 12. Factors that influenced this response were duration and amplitude of viral load decrease and absolute gains in total CD4⁺ cells and memory CD4⁺ cells.

Another French group, from the Hôpital de Bicêtre, also documented improved reactivity to recall antigens with HAART, in this case ZDV/3TC/indinavir in 20 ZDV-experienced patients with a baseline CD4⁺ count between 100 and 250 cells/mm³ whose courses were followed for 18 months (abstract 163). In this cohort, the response to recall antigens could be detected by month 3, persisted through month 18, and correlated with decreases in viral load.

If one scored studies of proliferative responses to recall antigens as one scores soccer games, the positive-response team drubs the no-response team 4 to 1. Unfortunately, scrutiny of textbooks on statistical methods fails to validate the soccer scoring method. So readers are left to weigh the ACTG 315 investigators' proposal that positive findings by others represent not true reconstitution but mere reactivation. Or, as ACTG 315's principal investigator, Michael Lederman, MD, suggested, the varying results may just boil down to different regimens and different patient populations studied by different labs. One way or the other, response to recall antigens is unlikely to be the critical determinant when deciding whether to start protease inhibitor therapy, especially in light of the consistent reports of plunging opportunistic disease rates among people taking HAART (see Good OI Trends Continue).

Three study teams did agree on a point that may be more easily translated into clinical practice: The immunologic benefits of HAART appear to persist despite rebounds in plasma viral load. That tentative conclusion was sounded by Case Western's Michael Lederman in his analysis of ACTG 315 results (abstract S30). His proposal was endorsed in poster presentations by Emory's Richard Hengel and his CDC colleagues (abstract 157) and by a group from San Francisco General Hospital (abstract 419). As already noted, the Emory-CDC team kept tabs on a cohort of 10 treatment-experienced and 10 treatment-naive individuals for 16 weeks. In 10 of these individuals, Hengel and colleagues could detect no naive (CD₄₅RA⁺ v62L⁺) CD4⁺ cells before HAART began. Five of these 10 began to reconstitute this cellular population within 16 weeks, even though three of those five still had detectable plasma RNA at that point. And of the five patients who showed no gain in naive CD4⁺ cells at 16 weeks, four had no detectable viral load. At least in the short term, then, a rebound in naive CD4⁺ cells does not always correlate with undetectable viremia, which in this study meant fewer than 400 copies/mL. That result may shed light on the report by Stephen Deeks, MD, who found that virologic breakthroughs on HAART did not lead to clinical deterioration in a large majority of protease inhibitor patients at San Francisco General (abstract 419), at least not right away (see What Does Virologic Failure Mean?).

Return to Top of Page

Who needs a thymus anyway?

The news that naive CD4⁺ and CD8⁺ cells return--however slowly--during HAART has raised hopes that even people with advanced HIV disease can rebuild a T-cell repertoire diverse enough to ward off the insults of AIDS. But, in adults with HIV infection, where are these virgin cells coming from? Are they trickling through a vestigial thymus that drowsily brands each with its trademark T? Or have they found their way into the circulation via some ill-mapped "thymus-independent pathway," originating perhaps in gut or liver? A third alternative--that the apparently new T cells merely reflect antigen-driven expansion of mature T cells already abroad in the periphery--is a more dire interpretation. In that case, few if any of the new T cells would truly be naive. They would be copies of whatever T cells were left when therapy began, in other words, mostly memory cells whose receptor repertoire depends on disease stage. Two intruiging presentations at the Chicago conference grappled with these slippery issues.

The National Cancer Institute's Crystal Mackall, MD, a veteran tracker of T-cell regeneration in cancer patients, affirmed the bad news HIV docs didn't want to hear: "Humans are generally very poor regenerators of T cells," she said, opening her plenary lecture (abstract L7). Adult bone marrow transplant patients, for example, never restore T cells to pretransplant levels. The reasons for this poor performance are straightforward, according to Mackall: age, disease, and--at least in people with cancer--therapy itself.

The difference between the regenerative capacities of children and adults is striking. In a three-year-old who undergoes intensive chemotherapy, Mackall found,³ the T-cell population eight months later looks essentially the same as the population before therapy. Plenty of naive cells are out and about. But in a similarly treated 23-year-old, she could detect virtually no naive T cells six months after therapy began. Only after 15 months did a stunted set of naive cells stumble into view. The cutoff for robust regeneration appears to lie somewhere below 16 years, Mackall believes. The reason for this difference is no mystery. That three-year-old still has a red-hot thymus, while the crusty 23-year-old must stoke the thymic-independent furnace to forge new lymphocytes.

That's bad news, Mackall found, because "reliance on thymic-independent pathways of T-cell regeneration leads to a pattern of immune dysregulation." That pattern has four baneful features: an apparent limit on how many new cells can be generated, a predominance of memory over naive cells and CD8⁺ over CD4⁺ cells, a heightened propensity for programmed cell death (apoptosis), and blunting of T-cell receptor diversity.

But Mackall's insights into T-cell regeneration were not entirely bleak. For starters, she said, "we should be careful not to assume that the aged thymus is doing nothing." And if that organ retains some function, that may be better news for people with HIV infection than for those who submit to T-cell devastation via chemotherapy or bone marrow transplantation. Mackall explained her idea of T-cell equilibrium with that homely model, the kitchen sink. In a healthy person, the sink holds a full complement of T lymphocytes. Some drip in daily via thymic or nonthymic faucets, and an equivalent number goes down the apoptotic drain. All-out chemotherapy or bone marrow transplantation knocks the bottom out of the sink. When fixed, that sink will take a long, long time to refill, if it ever does.

In HIV infection, Mackall figures, the sink merely springs a leak, so the loss of diverse T cells may not be as complete as it is after cancer chemotherapy, depending on when the leak is plugged. Indeed, in studies of cancer patients undergoing less than blitzkrieg chemotherapy, she found evidence of "a slow but persistent rise in the RA [naive] subset." That, of course, is exactly what Michael Lederman, Brigitte Autran, Frank Miedema, and others see in HIV-positive people who start therapy with about 100 to 300 CD4⁺ cells.

Mackall asked, but did not answer, another intriguing question: Exactly what does the thymus do in everyone past the crotchety age of 16 years? Is the dilatory training of T cells in adults all that necessary? After all, adults who undergo thymectomy generally do just fine without this dilapidated organ.

That very question was met head-on in a study by Duke University's Barton Haynes, MD (abstract S44). And his study involved not people with cancer, but people with HIV infection. What, he wondered, would happen if an HIV-infected person without a thymus began taking HAART? Apparently by dint of great effort, Haynes managed to track down three HIV-positive people who had thymectomies because of myasthenia gravis. (One of the three actually tracked down Haynes after reading about his work on the Internet.) The first person in this tiny cohort underwent thymectomy about 11 months after acute HIV infection. Within three years, during which she took sequential nucleosides, her CD4⁺ count plunged from 1247 to 99 cells/mm³, and she died of AIDS. The stories of the other two individuals are happier, indeed, astonishing.

One apparently became infected with HIV in 1989, years after undergoing thymectomy. Triple therapy began in 1996, and his naive (CD₄₅RA⁺ v62L⁺) subset of CD4⁺ cells "rose in a manner indistinguishable from nonthymectomized HIV patients." The third individual, who was probably infected between 1983 and 1985, again years after thymectomy, also responded well to HAART. Now, after 26 years without a thymus and at least 13 years with HIV, this man has a CD4⁺ count in the 500s with a growing complement of CD₄₅RA⁺ v62L⁺ T cells. Despite earlier HIV-related complications, he is now asymptomatic.

Thymectomy, Haynes concluded, "did not preclude a long-term survivor state or prevent postantiviral T-cell rises with CD4-positive and CD8-positive . . . naive phenotype cells." At least in these two individuals, he figures, most of the regeneration of circulating T cells during HAART must have resulted from proliferation, or redistribution and expansion, of peripheral T cells. Because they both have CD₄₅RA⁺ T cells a decade or more after thymectomy, he said, "it is unlikely that expression of the RA isoform of CD₄₅ signifies only recent thymic immigrants." Haynes suspects, therefore, that the so-called naive T lymphocytes counted in HAART-treated patients who still have a thymus may also represent primarily proliferation or redistribution of existing T cells. If so, however, explaining the T-cell repertoire diversification reported by Autran and colleagues becomes something of a challenge.⁶ No one said this would be easy. At any rate, Haynes believes that "normal-appearing thymus tissue on chest CT scan clearly does not indicate the presence of a functional thymus." If that's true, he continued, the only way to effect true immune reconstitution is to teach the old thymus an old trick: the generation of T cells from bone marrow precursors. But Haynes did not speculate on how that might be done.

Return to Top of Page

Good OI trends continue

For readers who wish they had never heard of CD₄₅RA⁺ v62L⁺ CD4⁺ cells, there is, to be sure, a more direct way to gauge the effects of HAART: How many people taking HAART get new opportunistic diseases or worsening symptoms of existing diseases? The answer is a nearly unanimous "lots fewer." Why only nearly? Because HAART may exacerbate the symptoms of some conditions--notably KS, tuberculosis, and Mycobacterium avium complex (MAC) infection, according to Judith Currier, MD, of the University of Southern California (see Other OI Offerings). By and large, though, conference reports linked consistent decreases in new opportunistic diseases with the arrival of HAART.

The largest study, a French population-based analysis of 59,256 individuals, compared the incidence of opportunistic infections and malignancies in the first half of 1996 and the first half of 1997 (abstract 182). The decreases proved so large and so consistent (Table 1), these French investigators maintained, that confounding nontherapeutic factors are unlikely. The conditions with less-than-profound drops in incidence--TB, bacterial pneumonia, lymphoma, and progressive multifocal leukoencephalopathy (PML)--may reflect onset at relatively higher CD4⁺ counts (TB and pneumonia) or poorer antiretroviral penetration of brain (lymphoma and PML). The latter rationale may be tenuous, though, because new cases of encephalopathy dropped 67 percent and others report lower rates of dementia (see below). Any population study has inherent limitations, of course, and these investigators drew attention to one in particular: They estimated that about 40 percent of the patients surveyed were taking triple combination therapy, while another 40 percent were taking only two nucleosides. So the gratifying declines in disease incidence might be attributed to double-nucleoside therapy as justly as to triple combinations.

A survey of more than 2800 HIV-positive people treated for Pneumocystis carinii pneumonia (PCP), CMV, MAC, or Cryptococcus neoformans infection at San Francisco General Hospital from 1994 through 1997 uncovered similarly steep drops in incidence over those four years (abstract 183). The rate of C neoformans meningitis slid 63 percent, PCP incidence fell 71 percent, new cases of disseminated MAC dropped 84 percent, and CMV incidence plunged 94 percent. More than 50 percent of each decrease came in 1996 and 1997, so the investigators believe a 30 percent increase in protease inhibitor use during those years must figure in the decreased OI incidence. But clearly, they added, more use of PCP prophylaxis and the introduction of MAC prophylaxis contributed to the declines.

A similar study at Johns Hopkins in Baltimore confirmed a precipitous downward trend in opportunistic disease incidence from 1993 to 1997 (abstract 184). Here, even diseases with more modest incidence declines in the French study dropped dramatically when measured in cases per 100 patient years. Rates of bacterial pneumonia fell 64 percent, PML 62 percent, and lymphoma 54 percent. Rates of dementia plummeted 83 percent at Johns Hopkins. The Baltimore team directly calculated the relative risk for an opportunistic disease among people who took a protease inhibitor (0.39, P<0.01), had a plasma viral load below 400 copies/mL (0.15, P<0.001), or had a CD4⁺ count below 50 cells/mm³ (5.6, P<0.001). No one in the Baltimore cohort, 78 percent of whom were African Americans and 50 percent of whom were injecting drug users, had a new opportunistic disease with a CD4⁺ count above 200 cells/mm³.

More good news on the neuropsychological front came from investigators at Cornell University in New York City, who linked use of protease inhibitors with less neuropsych impairment (abstract 186). Comparing 69 gay men taking a protease inhibitor with 61 gay men who were not, they found that significantly fewer in the protease inhibitor group (15 men or 22 percent) were neuropsychologically impaired than in the nonprotease group (33 men or 54 percent), according to a four-test battery (P<0.0001). And the better neuropsychological performance in the protease inhibitor group did not reflect less advanced HIV disease, since those taking protease drugs had a lower mean CD4⁺ count (264 cells/mm³) than those not taking protease inhibitors (350 cells/mm³, P<0.05). The investigators proposed that (1) protease inhibitors may cross the blood-brain barrier better than previously thought, (2) protease inhibitors may be crossing a compromised blood-brain barrier in some individuals in this study, or (3) protease inhibitors may lower peripheral concentrations of HIV enough to have a neuropsychological benefit. However, they noted that other researchers failed to correlate such benefits with lower plasma HIV RNA loads.

The first detailed analysis of opportunistic diseases in ACTG 320 (ZDV/3TC vs. ZDV/3TC/indinavir) confirmed that people taking this triple regimen enjoyed considerable clinical benefit compared with those taking only ZDV/3TC (abstract 257). Rates per 100 patient years in the two-drug and three-drug arms were 4.07 versus 1.39 for PCP, 3.58 versus 1.39 for CMV, and 14.3 versus 6.91 overall, all significant differences. Rates of MAC were higher in the three-drug arm (2.09 versus 1.9), but this difference was not significant. Judith Currier reported that all but one case of PCP, CMV, or MAC came at CD4⁺ counts below 200 cells/mm³. The outlier was a case of PCP at a CD4⁺ count of 608 cells/mm³. CD4⁺ count changes at week 4 or 8 strongly predicted occurrence of an opportunistic infection. If the CD4⁺ count increased by no more than 10 cells/mm³ at eight weeks, the OI rate was 21.1 per 100 patient years, compared with rates of 7.3 if the CD4⁺ count rose between 10 and 50 cells/mm³ and 4.5 if it rose more than 50 cells/mm³.

Work by Robert Lyles, MD, from Johns Hopkins, and other Multicenter AIDS Cohort Study investigators also correlated CD4⁺ counts with occurrence of PCP, CMV, or MAC disease (abstract 256). The study cohort included 734 AIDS-free men with a CD4⁺ count at or below 500 cells/mm³, and the analysis extended only until July 31, 1988 to avoid confounding results with the effects of antiretroviral therapy. No one in the analysis had taken an antiretroviral at that point. A CD4⁺ count between 200 and 350 cells/mm³ increased the risk for PCP 1.51 times (P<0.05), while a count below 200 cells/mm³ increased the PCP risk 3.72 times and upped the CMV risk 7.93 times (P<0.01 for both). Plasma viral load measured by the bDNA assay also proved highly predictive. A viral load between 30,000 and 60,000 copies/mL raised the risk for PCP 4.01 times, and a viral load between 60,000 and 90,000 copies/mL inflated the risk for PCP 8.08 times, for CMV 5.22 times, and for MAC 9.85 times. Viral loads above 90,000 copies/mL increased the risk for PCP 11.13 times, for CMV 8.80 times, and for MAC 14.97 times. All cited increased risks tied to viral load were statistically significant (P<0.01).

Return to Top of Page

Expect Eradication Work (For At Least Five Years

Frets over the ambiguities of immune reconstitution might be less troubling if the flickering promise of eradication is not quenched entirely. But this Retroviruses conference heard from a growing legion of experts willing to voice their skepticism over expunging HIV. Some even questioned whether ablating strategies should be pursued if long-term control of viral replication seems a more immediately attainable goal.

Yet the very scientists who showed that motes of replication-competent virus inside resting T cells elude current therapies refrained from dismissing the chance for an actual cure. Robert Siliciano, whose group at Johns Hopkins first suggested that viable viral DNA infests a few durable memory T cells, gave reporters the bad news that this dangerous, tiny pool appears even before a newly infected person seroconverts, and even if that person gets treated before seroconversion. "Available data," said Siliciano, "suggest this reservoir would not be eradicated for years." But, he added, that assumption may prove wrong, since data so far come from people who have been treated for only two years. Then again, these viral puddles may not dry up even under the heat of lifelong, fully suppressive HAART, Siliciano warned. Driblets of ongoing, though undetectable, replication may keep the pools deep enough to muddy lymphoid tissues of anyone taking current drug combinations.

Tae-Wook Chun, PhD, who piloted important early work in Siliciano's lab and now does the same for Anthony Fauci, MD, at NIAID, came up with his own estimate of how long eradication may take with antiretroviral therapy (abstract 515). Judging the half-life of latently infected CD4⁺ T lymphocytes to be 108 days, he figured that fully suppressive therapy would mop up the latent pool in 5.8, 6.8, or 7.7 years, depending on the size of the pool when therapy begins. Chun noted, though, that these estimates completely ignore the possibility of still-vibrant virus trapped in other compartments.

Using assays Chun helped develop, but plugging in different variables for pool size and half-life, Aaron Diamond chief David Ho, MD, proposed a wider range for time to eradication. The worst-case scenario, with a pool size of 100 thousand cells and a half-life of 18 months, sees two decades drifting by before that reservoir evaporates. A second scenario, which Ho rated "perhaps a bit more realistic," counts only five years to eradication if the pool includes 30,000 cells with a decay half-life of three to six months.

But, Ho argued, reducing the pool from thousands of cells to "less than one" may not be necessary. Suppose 200 latently infected memory cells get activated and spill new virions daily. The immune system counterattacks. "If 200 such battles are going on diffusely throughout the body," Ho proposed, "the chance of one breakthrough is quite real." But if therapy pares the number of activated cells to 100, 50, or 10 a day, chances of renewed replication dwindle accordingly. Still, no one knows at what point, if any, the immune system can control this seepage of new virus.

How much faith, though, can one put in any eradication estimate when the variables remain vaporous and reappraisals appear with each new meeting? Not much, suggested John Mellors at a press conference. "I would take these estimates lightly," he said, because a year from now they may be "debunked by new findings." But Ho insisted that, whether or not one cares to wager on the persistence of reservoirs, the quest for eradication should not be abandoned. "The road to HIV eradication is a bumpy one," he acknowledged. "But given how far we have traveled, it would be wrong to despair and abandon the journey in sight of our objective."

Return to Top of Page

A "new steady state": Slow, low, and dangerous?

What worries investigators like Tae-Wook Chun and David Ho more than the time it may take to wipe out HIV is whether today's drugs are even up to the challenge. Chun noted in his presentation, and in a paper published late last year,⁹ that detection of unintegrated viral DNA in people with undetectable plasma loads could bespeak still-replicating virus. Ho was similarly cautious in his Chicago presentation. "I think we have to bear in mind that [triple therapy] regimens may not be 100 percent suppressive," he said, "and a new steady state has been achieved at a very low level beyond our detection." The residue of detectable virus even after a year or more of apparently complete suppression, according to Ho, could be "a reflection of that steady state."

The same thought occurred to Joseph Wong, MD, who led a third team that scouted out a pool of latently infected memory cells in aviremic individuals.¹⁰ Virus recovered from these cells, he reported at the conference (abstract S19), shows no sign of replication if one looks only at drug-resistance mutations or changing nucleotide sequences. But Wong could demonstrate transcriptional activity in latently infected cells. Though he could not say whether that activity means ongoing low-level replication, neither could he exclude that possibility. And if low-level replication--Ho's "new steady state"--proves to be real, Wong asked whether it would eventually lead to drug resistance.

Chun's boss, Anthony Fauci, and Ho said they'd rather not wait to find out. Instead, each proposed a strategy for exposing virus in latently infected cells to antiretroviral onslaughts. Ho favors anti-CD3 monoclonal antibodies, already being used in cancer patients, to activate the resting T-cell pool. IL-2 might also send a wake-up call, he said, and that approach should be studied as well. But because resting T cells don't express IL-2 receptors "in the short term," according to Ho, using that cytokine could prove less speedy than using a monoclonal antibody.

Fauci, however, has some evidence that IL-2 may work well enough (abstract S43). Endogenous cytokines such as TNF-, IL-6, and IL-2 constantly bathe lymphoid tissues, activating resting cells. Such cytokines, argued Fauci, probably explain why full-force viral replication quickly follows withdrawal of suppressive therapy. His NIAID team cultured latently infected CD4⁺ cells sampled from people with HIV infection and exposed the cultures to IL-2 for six days. They also doused some of the cultures with ZDV, 3TC, and indinavir.

Their hypothesis--that the virus IL-2 prods from resting cells would infect fewer additional cells in HAART-treated cultures--proved plausible. Levels of viral replication in cell cultures were 100-fold lower after the six-day onslaught of IL-2 plus HAART than in cultures exposed to IL-2 alone. Fauci stressed, though, that the three antiretrovirals did not completely turn off replication in this in vitro model. But if the results hold true in vivo, he said, three antiretrovirals and IL-2 could substantially reduce--though possibly not eliminate--the memory cell pool. NIAID investigators are already testing the strategy in HIV-positive volunteers.

Return to Top of Page

Hydroxyurea front and center

Anthony Fauci and David Ho are hardly alone in proposing regimens to kick-start latently infected CD4⁺ cells and thereby expose this hidden HIV to antiretrovirals. Paradoxically, though, a regimen built around a drug that constrains T-cell proliferation--hydroxyurea--may have come closer than any to shutting down viral replication for good. No one, not even the drug's champion, Franco Lori, MD, PhD, claims that a hydroxyurea-containing regimen can eradicate HIV. But Lori appeared front and center at the Chicago conference, at a crowded poster (abstract 655) and a late-breaker session (abstract LB11), making his case for this previously dowdy, but still dirt-cheap drug.

Some virologists appear to harbor an inherent distrust of hydroxyurea. At a press conference, the University of Pittsburgh's John Mellors explained why. Effective antiretroviral combinations that don't include hydroxyurea usually lower viral load and boost CD4⁺ cell counts at the same time. But when you add hydroxyurea to the mix, typically with didanosine (ddI), T-cell counts may go nowhere. That makes Mellors think a viral load reduction induced by combos containing hydroxyurea does not equal a viral load reduction with standard regimens. "There seems to be a disconnect [between viral load and T-cell counts] when using hydroxyurea," he said, "and that suggests that hydroxyurea is having a major effect on outgrowth of lymphocytes."

But Lori yearns to turn this cytostatic sow's ear into an immune-suppressing silk purse. His thinking works this way: Replication of HIV in T cells triggers activation and proliferation of other T cells, and so the cell-consuming bonfire grows. CD8⁺ T cells, in particular, fuel the fires of immune activation, killing countless CD4⁺ cells and eventually exhausting their own stores--all without immolating HIV. "Since cell activation and proliferation are pivotal for the replication of HIV," Lori wrote in his poster presentation (abstract 655), "early use of drugs that inhibit immune cell activation might help the antiretroviral drugs inhibit HIV and also prevent/restore damage to the immune system." Cue hydroxyurea. If this drug's "cytostatic and potentially immunosuppressive properties" can be exploited early enough in HIV infection, those cell-consuming bonfires might be robbed of their fuel. Maybe the fire can't be doused completely, but it might be made to smolder harmlessly for years.

That may be what's happening in the now-storied patient from the practice of Heiko Jessen, MD, in Berlin. After hydroxyurea, ddI, and indinavir lowered this man's viral load to undetectable levels, he stopped taking the drugs during an acute bout of hepatitis A. Now, over a year later and still without drugs, his viral load has not rebounded. But Lori did detect minute traces of both HIV RNA and DNA in lymphoid tissue samples. Why it doesn't reignite rapid replication remains a mystery. Perhaps the level of virus is so low that the man's immune system can control it without drugs. Perhaps the DNA detected is unintegrated and harmless, suggested Robert Siliciano, who used his souped-up assays to spot traces of the virus in this patient. Perhaps, Mellors chipped in, controlling viral replication with the three drugs while depriving HIV of substrate with hydroxyurea results in "an aborted infection." But he hastened to label that hypothesis "very, very speculative" and one that should be addressed in studies of acute infection--first in animals.

Of note, CD4⁺ counts did climb a mean 116 cells/mm³ after five months of therapy in eight primary infection patients Jessen treated with hydroxyurea, ddI, and a protease inhibitor. So, at least in early HIV infection (the median starting CD4⁺ count was 684 cells/mm³ in Jessen's cohort), such a triple combination may pump up CD4⁺ counts.

Return to Top of Page

Bridge Washed Out in Drug Data Deluge

The intrepid conference attendee could, just by staying awake, come home with results from scores of antiretroviral drug trials. Somewhat more effort will be required to organize this data trove into a scheme that may inform current or future practice. One possibility (and, not coincidentally, the possibility realized in the balance of this article) is to begin with studies of combinations in naive individuals, or people with only nucleoside experience, then to proceed to trials involving folks with lots of drug experience, including failed protease inhibitors. Next one might group those noble experiments whose authors sought to simplify therapy. After a brief excursion into trials of two agents soon up for approval (abacavir and amprenavir), one could look at studies addressing two of therapy's thorniest issues: why drugs work and (a much longer section) why drugs fail.

At least 12 trials of note essayed the merits of protease inhibitor regimens in people who had taken only nucleosides or had taken no antiretrovirals at all. Table 2 summarizes these dozen studies. Especially interesting are abstracts 384, 385, and 394, whose authors unleashed quadruple and quintuple combinations. The investigators in these three studies, who come from two different sites, reached a similar conclusion: Piling on the drugs squelches plasma viremia even faster than triple combinations, so fast, in fact, that the authors suggested current estimates of viral turnover may be too low. No one should be surprised, though, that the most complicated regimen also caused the most side effects.

A fourth trial of quadruple therapy focused less on viral dynamics and more on the individuals treated, all of them women (abstract 722). At least in the first 16 weeks of follow-up, this study of stavudine (d4T), 3TC, hard-gel saquinavir, and nelfinavir turned up little difference between a twice-daily and a thrice-daily regimen, except in number of serious side effects. An ongoing trial of nelfinavir and soft-gel saquinavir showed much better 32-week results when the two protease inhibitors were added to nucleosides than when they were given without nucleosides (abstract 394b).

The Dutch CHEESE study, now at the 24-week mark, suggested equivalence between soft-gel saquinavir and indinavir when either is combined with ZDV/3TC (abstract 387b). One unexplained oddity in this trial is the paltry CD4⁺ cell increase in the indinavir arm. Still, some clinicians remain dubious about using the new soft-gel saquinavir as the sole protease inhibitor combined with reverse transcriptase inhibitors because of the high pill burden--six capsules three times a day. Combining saquinavir with another protease inhibitor typically halves that number and the cost as well.

Another intriguing study assessed the novel combination of d4T, nelfinavir, and the nonnucleoside nevirapine (abstract 350). This small trial looked first at the pharmacokinetics of the drugs by giving d4T and nelfinavir alone for seven days, then adding nevirapine. The bottom line of that analysis: none of the drugs had an undue impact on the others. At 21 weeks, nearly 90 percent of the participants, half of whom had nucleoside experience at baseline, had a viral load below 400 copies/mL.

Because nonnucleoside class resistance is a fact of life, many physicians who have not already given a person a nonnucleoside are holding out for approval of efavirenz (DMP 266). Efavirenz can be given once daily and, when combined with indinavir, packs a wallop that has kept HIV on the mat for 60 weeks (abstract 692). See Table 2 for further details of these studies.

Return to Top of Page

Hopes remain modest for most salvage regimens

People with HIV don't have long to wait before abacavir (1592), adefovir (bis-POM PMEA), efavirenz (DMP 266), and amprenavir (141) start filling pharmacy shelves. But leading investigators utter only words of caution when asked how effective these agents will be for people with ample drug experience. Scott Hammer, MD, of Beth Israel Deaconess in Boston, took his turn at the standard "new agents" lecture this year (abstract S48) and made the appropriately circumspect sounds. Researchers are testing all four agents in salvage regimens, Hammer said, "but expectations should be realistic because of the real threat of cross-resistance."

Building a regimen that works after one protease inhibitor failure remains particularly urgent. Numerous investigators at last fall's ICAAC and Hamburg conferences reported that double-protease inhibitor regimens proved less than awesome after a single protease inhibitor loses control of replication. Reflecting on those studies at a press conference in Chicago, Pittsburgh's John Mellors estimated that two protease inhibitors add up to a good rescue regimen in less than half of such cases. At the Retroviruses conference, two groups reported somewhat better six-month response rates, 63 to 68 percent, after switching to ritonavir/saquinavir when nelfinavir or indinavir faded. But ritonavir/saquinavir yielded only a 49 percent "undetectable" rate after 48 weeks in another, larger study. All three groups found it tough to predict who might respond well.

A small study that combined ritonavir and saquinavir with d4T/3TC after virologic rebounds on nelfinavir showed reasonable six-month results (abstract 510). Thirteen of 19 (68 percent) who switched to ritonavir/saquinavir reached a viral load below 500 copies/mL between four to 12 weeks and sustained that response through six months of follow-up. Viral load in about half of those responders sank below 40 copies/mL, reported Pablo Tebas, MD, of Washington University in St. Louis. Although a 68 percent sub-500 rate might be deemed subpar for treatment-naive individuals, people in this study had taken nelfinavir for an average 55 weeks before switching to the quadruple regimen, and many of them already had antiretroviral experience when they started taking nelfinavir.

What factors may predict a good response to the salvage regimen? In this study, it was hard to tell. People with a lower viral load at the time of the switch generally did better, but the predictive value of low viremia at the switch fell short of statistical significance (P=0.07). Two teams independently analyzed resistance mutations in this cohort. The first found the L90M mutation, associated primarily with resistance to saquinavir, in five of 16 isolates, and the second found it in seven of 18. But having L90M mutant virus before the switch to ritonavir/saquinavir did not predict failure of that regimen. Neither was resistance to 3TC predictive. Those results may give pause to anyone hoping to base second-line treatment decisions on genotypic analyses.

Although investigators must have been gratified to see that failure of one year's worth of nelfinavir (defined as two bDNA viral loads above 5000 copies/mL) is not the end of the line for most people, six months is a short stretch in the longer scheme of things. And Tebas adroitly declined suggesting that the study bolstered any particular drug sequencing strategy. Perhaps starting with a different protease inhibitor or switching to a different backup regimen would have worked better, he allowed, than starting with nelfinavir and switching to ritonavir/ saquinavir.

At Johns Hopkins, a chart review (abstract 427) of 21 individuals given ritonavir/saquinavir after virologic rebounds on indinavir or nelfinavir reached results similar to those of Tebas (abstract 510). Joel Gallant, MD, reported that 10 of 16 individuals (63 percent) attained the milestone sub-400 viral load after switching to the two protease inhibitors, including seven of 12 first treated with indinavir and three of four first treated with nelfinavir. This level of suppression persisted for an average 29 weeks in the indinavir group and 27 weeks in the nelfinavir group. Half of the responders maintained an undetectable viral load for 40 weeks or longer.

Despite scrutiny of potential predictive factors, however, Gallant could pinpoint none that decisively foretold success with the salvage regimen. Viral load when salvage therapy began, viral load as a percentage of baseline viral load when salvage began, and interval between virologic failure of the first protease inhibitor regimen and the backup regimen all failed as omens of virologic response. Gallant proposed, though, that the better response rate in his chart review than in earlier studies of double-protease inhibitor salvage may be explained by the relatively low viral load at the time of the switch, a median 13,500 copies/mL. He suggested that "early switching may be more effective and durable than continued therapy of the failing regimen followed by a delayed switch." Perhaps, but the average 41 to 44 weeks on indinavir or nelfinavir in Gallant's analysis, versus an average 55 weeks on nelfinavir in the Tebas study, did not yield a better salvage response rate in Gallant's study (63 percent versus 68 percent for Tebas).

Another observational study--this one larger and longer than either the Tebas or Gallant study--charted a 49 percent response rate (<400 copies/mL) to ritonavir/saquinavir plus two nucleosides after 48 weeks (abstract 390). Chris Duncombe, MD, and colleagues in Sydney reported that two thirds of the 56 people studied had taken a protease inhibitor before beginning the double protease inhibitor regimen, while the others had taken only nucleosides. After 24 weeks, there were more responders among those who had never taken a protease drug before. But, after 48 weeks, this difference between protease inhibitor-experienced and -naive individuals vanished. Baseline viral load (median 29,500 copies/mL) and baseline CD4⁺ count (median 191 cells/mm³) did not correlate with response to ritonavir/ saquinavir. Nor did response depend on whether the protease inhibitors were matched with ZDV/3TC (in 40 persons) or with d4T/3TC (in 16).

Two groups sized up the merits of indinavir plus the nonnucleoside nevirapine in experienced individuals, and one team also combined these agents in treatment-naive individuals with advanced HIV disease. The many differences between these two observational studies make it difficult to compare them, but those differences could explain why response rates didn't match.

The longer study, presented by Vancouver's Marianne Harris, MD, documented 50 percent attrition from a cohort of 22 individuals with advanced disease who began treatment with 3TC, indinavir, and nevirapine, all at standard doses (abstract 429a). None had taken a protease inhibitor, but 19 had taken 3TC, two had taken the nonnucleoside loviride, and all had taken some antiretroviral. By one year, 11 people had stopped the triple regimen, five because their viral load was above 5000 copies/mL, and six because of side effects and other reasons. Of the 11 still taking 3TC/indinavir/nevirapine, 10 have a viral load below 400 copies/mL, and seven are below 20 copies/mL. So, with the 400-copy benchmark, the one-year response rate is 45 percent. As in many other studies, the long-term responders all controlled viremia quickly. Harris reported a year ago that nine of 21 individuals then being studied (43 percent) were under the 400 mark.

That may bode well for 57 people being treated with indinavir and nelfinavir by Richard Beach, MD, and colleagues in Miami Beach (abstract 428). After 24 weeks, 31 of 36 treatment-experienced individuals (86 percent) had a viral load below 400 copies/mL, as did 16 of 17 who were treatment naive (94 percent). And these high response rates generally held at the 25-copy limit. Twenty-eight of 36 experienced people (78 percent) reached that mark, and the same 16 of 17 previously untreated patients did as well.

What could account for the better early response rate in Florida? Lots of things, even if one excludes the 17 antiretroviral-naive individuals in the Florida group. Beach's drug-experienced patients began treatment with much higher CD4⁺ counts, an average 184 cells/mm³ compared with 30 cells/mm³ in Vancouver. Even so, the baseline viral loads were the same in the two treatment-experienced cohorts, about 125,900 copies/mL. But the regimens differed. The Vancouver patients took standard doses of indinavir and nevirapine, whereas the Florida patients took an extra 200 mg of indinavir three times daily to offset nevirapine's effect on indinavir concentrations. And the Canadians added only one nucleoside, 3TC, even though 19 of 22 individuals had already taken it, whereas the people in Miami Beach added two nucleosides--ZDV/3TC or d4T/3TC.

An open-label extension of a phase I/II abacavir study suggested some benefit from adding this new nucleoside to other antiretrovirals (poster 659). But an unusual circumstance leading to this open-label follow-up, and the diverse other antiretrovirals being taken, muddy the results somewhat. The multicenter trial began as a pharmacokinetic and safety study of abacavir with or without ZDV in 75 treatment-naive individuals. After 12 weeks, though, these people had to stop taking abacavir because animal toxicology data were judged too slim. So they continued with other antiretroviral regimens. Then, after a median 36 weeks and enough rat results to satisfy the FDA, doctors told this possibly bewildered band they could start taking abacavir again. Forty-three did, 31 adding it to ongoing nucleoside-only regimens and 12 adding it to regimens including a protease inhibitor.

The gamble paid off for many. After 12 weeks, the median viral load skidded from a baseline of 63,100 copies/mL when restarting abacavir to below 400 copies/mL in both the nucleoside-only group and the abacavir-plus-protease inhibitor group. And the average CD4⁺ count climbed from 366 to 480 cells/mm³. But 48 weeks after starting abacavir again, a difference between the two groups emerged. Among those adding abacavir to other nucleosides, eight of 15 (53 percent) still have a viral load below 400 copies/mL, compared with nine of 10 (90 percent) who fortified a protease inhibitor regimen with abacavir. (For an analysis of abacavir cross-resistance, see Finding the Right Mates for Abacavirand Amperenavir.)

Gabriel Torres, MD, of St. Vincent's Medical Center in New York City, reported that no one in the study had the systemic drug reaction sometimes seen with abacavir. Investigators have reported this side effect in about 3 percent of people taking abacavir.¹¹ Although rash is a frequent component of the reaction, it is not the only one and may not even be noticed by the patient. Other symptoms are fever, nausea, and malaise. All symptoms are reversible if the person stops taking abacavir, but rechallenge should never be attempted because symptoms return within hours and more severely than before.

Return to Top of Page

Eight ways to simplify antiretroviral therapy

Rarely do clinicians make bold to suggest what writers should write, but that's exactly what a frustrated AIDS doc from Albany did when she tracked down this reporter and two others at a Chicago poster session. Don't treat two-drug therapies like a joke, she pleaded. I'm lucky if I can get some of my patients to take one drug, never mind three. There may be good arguments against giving just two nucleosides, she continued, but giving ZDV and 3TC together in the same pill is sometimes the best I can do.

Many clinicians surely share those sentiments. They must have been gladdened by the raft of conference reports on kinder, gentler combinations that may tempt patients horrified at the clock watching and dietary derring-do required by some regimens. First in these physicians' hearts surely come the new antiretrovirals that seem to work well when given once a day. Three studies essayed the efficacy of drugs that can be downed in a single daily sitting--efavirenz, PMPA, and FTC.

Efavirenz plus two nucleosides would be an attractive regimen for treatment-naive people, argued a multicenter clutch of investigators, because it saves protease inhibitors for a second-line regimen (abstract 698). After 24 weeks of ZDV, 3TC, and efavirenz (at doses of 200, 400, or 600 mg daily), more than 90 percent of 68 people had viral loads below 400 copies/mL, compared with about 65 percent of 23 people taking ZDV/3TC and placebo. And the viral loads of many of those taking efavirenz also fell below 40 copies/mL when measured at 16 weeks: 25 of 30 (83 percent) taking 200 mg, 19 of 29 (68 percent) taking 400 mg, and 18 of 27 (67 percent) taking 600 mg. All of these outcomes differed significantly from the four of 26 in the placebo group below the 40-copy mark (P<0.017), reported David Haas, MD, of Vanderbilt University. Mean CD4⁺ counts soared by nearly 175 cells/mm³ in the 200- and 400-mg groups at 24 weeks, and by more than 100 cells/mm³ in the 600-mg group. All results bettered the CD4⁺ gains in the placebo group, which added efavirenz and indinavir to ZDV/3TC at week 16.

The adverse event rate topped 90 percent in all four treatment arms at 16 weeks, but there were few significant differences between the efavirenz group and the placebo group. More than 44 percent of those taking 600 mg of efavirenz reported dizziness, compared with 18 percent in the placebo group (P<0.05). In the 400 mg arm, 53 percent had headaches compared with 24 percent taking only ZDV/3TC (P<0.05). And people taking 200 mg of efavirenz reported significantly more sinusitis and sweating than people in the placebo arm (P<0.05). The maculopapular rash that often irritates people taking nonnucleosides (and occasionally becomes severe) appeared in 11 percent of those taking efavirenz in this study. The rash was not dose related, and it contributed to one withdrawal from the 400-mg arm and one from the 600-mg arm. (The 60-week study of efavirenz and indinavir [see Table 2] used 200 mg of the nonnucleoside daily.)

As efavirenz nears approval, two other drugs farther back in the pipeline also look like once-a-day possibilities. Both were previewed at the late-breaker session.

PMPA, the nucleotide analog that shut down early infection in monkeys, showed dose-related antiviral activity in humans when given in an oral prodrug formulation called bis-poc PMPA (abstract LB8). A dose-ranging study in 28 people with a median CD4⁺ count of 375 cells/mm³ showed 35-day viral load reductions of 0.32, 0.44, and 1.22 logs in the 75-, 150-, and 300-mg daily groups. But San Francisco General's Steven Deeks said there were no clear T-cell trends during the month-long study. Five individuals had baseline 3TC-associated mutations at position 184, but their viral load responses did not differ from those without the mutation. (The trial included some treatment-experienced patients who began taking bis-poc PMPA after a washout.) No further detectable mutations arose during the study. Five of 28 people taking this drug (18 percent) had elevations of creatine kinase that resolved when treatment stopped. One person with a history of peripheral neuropathy reported reoccurrence of neuropathy, which also bothered one of eight people in a placebo group.

FTC has four to 10 times the in vitro anti-HIV activity of its cousin 3TC. A dose-ranging and safety study presented by Franck Rousseau, MD, of Triangle Pharmaceuticals showed that it may prove more powerful than 3TC in humans as well (abstract LB9). After investigators gave FTC for 15 days to 10 people naive to 3TC and abacavir, the median viral load dropped 1.4 logs in a 25-mg twice-daily group and 2.1 logs in a 200-mg once-daily group. Everyone in the high-dose group attained at least a 2-log reduction in plasma viremia. Two people, both in the low-dose group, reported side effects during the two-week study. Like 3TC, FTC also thwarts replication of hepatitis B virus.

For clinicians itching to simplify regimens including already-licensed agents, good news was not hard to find. Three studies of twice-daily versus thrice-daily nelfinavir suggested equivalence between the two regimens (abstracts 373, 387a, and 722). And an ongoing trial of bid versus tid indinavir reached the same conclusion. At weeks 24 through 32 weeks in the first two nelfinavir studies, about 90 percent of patients taking 1250 mg of nelfinavir twice daily (with d4T/3TC or ZDV/3TC) had viral loads below 400 copies/mL. A substudy (abstract 373) found that 78 percent of those under the 400-copy mark at 24 weeks had fewer than 50 copies/mL. A 32-week trial of indinavir plus ZDV/3TC showed that 75 percent of people taking 1000 or 1200 mg of indinavir twice daily had a (serum) viral load below 500 copies/mL, compared with 54 percent of those taking the standard dose of 800 mg three times daily (abstract 514). For the two twice-daily groups, these percentages also held for an assay with a 50-copy cutoff in serum. (For a summary of abstract 722, see Table 2.)

How about combining twice-daily indinavir and twice-daily nelfinavir? Don't try this one yet, because investigators from three sites, headed by San Diego's Diane Havlir, MD, are still hashing out the pharmacokinetics (abstract 393). When they tried 1000 mg of indinavir and 750 mg of nelfinavir, both twice daily, trough concentrations of nelfinavir were half those of 750 mg of nelfinavir given three times a day. Boosting the nelfinavir dose to 1000 mg twice daily brought the trough back up by 35 percent--better, but still substantially below the tid 750-mg nelfinavir trough. As might be expected, these troughs varied widely in the five individuals taking twice-daily nelfinavir, from a low of 0.131 mg/L to a high of 4.78 mg/L at the 1000-mg bid dose. But "outlier" must be the middle name of the person who brought home the 4.78 score. If that trough is kicked out, minimal concentrations in the remaining four people ranged from 0.131 to 0.618 mg/L, way under the 1.5 mg/L these investigators reported as nelfinavir's tid monotherapy trough.

Of the 21 study participants, 10 reached a viral load below 400 copies/mL, and only six were below 50 copies/mL, after 12 to 32 weeks of treatment. Two dropped out because of virologic failure, two because of noncompliance, and one because of rash and back pain. Three added nucleosides to their regimen, either d4T or ddI/d4T. No one had taken a protease inhibitor when the study began, and 12 had never taken any antiretroviral. Median baseline measures were 50,500 HIV RNA copies/mL and 259 cells/mm³.

Well, then, how about twice-a-day ritonavir and nelfinavir? Better results here, at least in the early going, thanks no doubt to the tight clamp ritonavir puts on metabolism of other protease inhibitors (abstract 394a). Joel Gallant, MD, and colleagues from Johns Hopkins added 400 mg of ritonavir twice daily to either 500 or 750 mg of nelfinavir twice daily. Of the 10 individuals in each group, half had taken only nucleosides and half had taken no antiretrovirals. They began the double protease inhibitor regimens with a median viral load of 32,500 copies/mL and a median CD4⁺ count of 325 cells/mm³.

At 12 weeks, 12 of 19 had reached a viral load below 400 copies/mL, and five of these 12 were under the 20-copy mark. Virologic results were almost identical in the 500- and 750-mg groups at 12 weeks, but then things took a bad turn in the 500-mg group. Four people taking that dose of nelfinavir withdrew, whereas none taking the 750-mg dose dropped out. Nine of 20 had moderate to severe diarrhea, but no one withdrew because of side effects.

Charles Flexner, MD, is still analyzing pharmacokinetic results from this study, but he did single out an intriguing case in a plenary lecture (abstract S47). One of the dropouts had the highest nelfinavir concentrations in the cohort, but nonetheless endured a quick virologic rebound. So what went wrong? One possible explanation is that, while this person's nelfinavir levels were high, levels of the drug's active metabolite were low. Or maybe that person simply started treatment with a virus partially resistant to nelfinavir or ritonavir, Flexner speculated, even though no study participants had taken a protease inhibitor before. "Life is complicated," he advised. "Pharmacology makes it more so."

Robert Kaspar, MD, and his colleagues in Austin put together a twice-daily non-protease inhibitor regimen and offered it to 25 treatment-naive people whose viral loads ranged from 7000 to 1,500,000 copies/mL (abstract 696). The regimen consisted of standard doses of d4T, 3TC, and nevirapine. Patients treated for 21 to 32 weeks had a uniform response: 20 of 23 individuals saw their viral loads sink below 400 copies/mL. One of the remaining three, one had a viral load of 6000 copies/mL (down from 76,000 copies/mL), one had a viral load of 4000 copies/mL (down from 18,000 copies), and one stopped taking all medications. The average CD4⁺ count increased by about 200 cells/mm³.

Clinicians from Vienna, led by A. Rieger, MD, took this logic one stop farther: They fashioned a once-daily regimen from ddI (400 mg), 3TC (300 mg), and nevirapine (400 mg) and began giving this combination to 18 people who were unwilling or unable to take drugs two or three times a day (abstract 697). Starting viral loads ranged from 2700 to 550,000 copies/mL. No one had taken a nonnucleoside before, but 12 had some nucleoside or protease inhibitor experience. After 24 weeks of follow-up, 10 of 18 (56 percent) had a viral load below 200 copies/mL, and five of these were below the 50-copy mark. Among the eight nonresponders, four had taken 3TC before and had six-digit baseline viral loads.

Return to Top of Page

Induction/maintenance fizzles in first studies

One grand plan to simplify therapy--the induction/maintenance scheme--fell flat in two large trials. Investigators hoped to steal a page from the book of cancer chemotherapy, unleashing a first strike deadly enough to let subsequent scaled-back attacks keep the disease in check. But it didn't work out that way for HIV. The researchers took pains to argue that their dismal results don't sound the death knell for the concept, only that these two particular strategies, with these particular drugs, flopped.

The first induction/maintenance maven to the lectern was France's F. Raffi, MD, who spelled out results of the TRILEGE Trial (abstract LB15). This randomized, open-label study gave ZDV/3TC/indinavir for three months to antiretroviral-naive individuals with a CD4⁺ count below 600 cells/mm³ and a viral load between 3500 and 100,000 copies/mL. Those whose viral load sank below 500 copies/mL at month three either continued triple therapy or eased back to ZDV/3TC or ZDV/indinavir. Tracking the number of participants whose viral load then climbed back above 500 copies/mL in two consecutive measurements, the French team counted six of 92 (7 percent) in the triple-drug group, 22 of 92 (24 percent) in the ZDV/3TC group (P=0.0018 vs. triple therapy), and 16 of 93 (17 percent) in the ZDV/indinavir group (P=0.034 vs. triple therapy).

Raffi reported that people who attained a viral load below 50 copies/mL during the three-month induction phase were less likely to have a virologic breakthrough during a mean 8.5 months of follow-up. Of 191 individuals who reached the sub-50 mark, 22 (12 percent) had a viral load rebound, compared with 20 of 74 (27 percent) whose viral load never got under 50 copies/mL (P=0.01). The French believe compliance complicates interpretation of the results: 22 percent of study participants interrupted therapy for more than one day.

The ACTG 343 team adopted a sterner strategy with the same three drugs (abstract LB16). People with more than 200 CD4⁺ cells/mm³ and a viral load above 1000 copies/mL took ZDV/3TC/indinavir for 24 weeks, instead of three months as in the French study. All were naive to protease inhibitors and 3TC, but they could have taken ZDV. San Diego's Diane Havlir reported that 309 graduated to the maintenance phase of the trial, after scoring sub-200 viral loads (instead of sub-500s in the French study) at weeks 16, 20, and 24 of induction. The ACTG investigators defined failure as a rebound above 200 copies/mL (instead of 500 copies) during maintenance. And the maintenance arms differed from the French study. People who didn't continue triple therapy in ACTG 343 got either ZDV/3TC or indinavir alone (instead of ZDV/indinavir).

Once again, more people who continued triple therapy maintained control of viral replication. Three of 105 (3 percent) in the three-drug group had a viral load above 200 copies/mL during maintenance, compared with 24 of 104 (23 percent) taking ZDV/3TC (P<0.0001 vs. triple therapy) and 23 of 100 (23 percent) taking indinavir (P<0.0001 vs. triple therapy). Those randomized to indinavir maintenance had a 7-fold increased risk of failure compared with the three-drug group (P=0.011). People randomized to ZDV/3TC had a 52-fold greater risk of failure if they had the ZDV-associated mutation at position 215 (P<0.0001) and a 5-fold greater risk if their virus was wild-type at position 215. People whose viral load lingered above 200 copies/mL at week 4, but who eventually got under that mark, had twice the risk of failure as those whose viremia was under 200 copies/mL at week 4, but this discriminator was not statistically significant (P=0.07).

Baseline CD4⁺ count did not independently predict failure in ACTG 343. Ironically, though, study participants whose T-cell counts increased most during induction therapy were more likely to have virologic breakthroughs during maintenance. Invoking the predator-prey model of viral dynamics, Havlir proposed that the profusion of CD4⁺ cells in the peripheral circulation after the induction phase merely provided more food for HIV when ZDV/3TC or indinavir maintenance lost control of replication. She concluded that the induction/maintenance strategy can prove successful only with stronger regimens.

While the first whacks at mellower maintenance have failed, other antiretroviral experts agree that less stressful therapies may yet find a place in managing HIV infection. Charles Flexner of Johns Hopkins went so far as to ask whether any anti-infectives, with the possible exception of aminoglycosides, require either perfect adherence or continuous plasma concentrations above some ordained threshold (abstract S47). Then he ventured an answer: "I think there are very few anti-infective drugs that meet these two criteria, including antiretrovirals."

Before the Grand Inquisitor swooped down to dissect these heretical notions, Flexner offered his apologia, by way of the sacred texts on TB therapy. Not long ago, he argued, the mycobacterial bible called for once- or twice-daily drugs for a year. That pace has slowed to two or three times a week (or, as noted above, from everyday therapy for a year to everyday therapy for two months). "Obviously tuberculosis is a very different disease from HIV," Flexner conceded. "It's a disease that can be eradicated. It's an organism that has a much longer generation time in the body." But one can still take a moral from the story: "When we start therapy for chronic infectious diseases," he concluded, "we're often engaged in a battle of overkill, using more drugs more often than we probably need to get long-term benefit." If this Flexnerian exegesis proves correct, only the drug sellers stand to lose.

Return to Top of Page

Finding the right mates for abacavir and amprenavir

When Glaxo's nucleoside abacavir (1592) and its protease inhibitor amprenavir (141) hit the market, how will clinicians combine them with other antiretrovirals? Four studies reviewed at the conference addressed that question.

John Mellors of the University of Pittsburgh detailed 16-week results of an ongoing open-label trial combining abacavir with protease inhibitors (abstract 4). All 80 participants were naive to antiretrovirals, had a viral load exceeding 5000 copies/mL (median 55,000 copies/mL), and had a CD4⁺ count above 100 cells/mm³ (median 349 cells/mm³). Most, but not all, patients attained a viral load below 400 copies/mL by week 12. Although Table 3 suggests a trend toward a poorer response with soft-gel saquinavir plus abacavir, Mellors emphasized that the small numbers in the study and the preliminary nature of the results cannot sustain that conclusion. However, of the five individuals whose viral load has rebounded from its nadir, four were taking saquinavir. The study will continue through 48 weeks.

An open-label multicenter European study, presented by Frankfurt's Schlomo Staszewski, MD, showed continuing drops in viral load and gains in CD4⁺ cell counts when clinicians added ZDV and 3TC to abacavir monotherapy (abstract 658). During the randomized phase of the trial, patients with CD4 counts generally in the high 300s took 100, 300, or 600 mg of abacavir twice daily. After 24 weeks, or if they had an inadequate virologic or CD4⁺ cell response earlier, 55 individuals switched to open-label abacavir (300 mg twice daily) plus ZDV/3TC. Three also added a protease inhibitor, one added d4T instead of ZDV, and one stayed with abacavir alone.

After a median half-log drop in viral load during abacavir monotherapy, adding ZDV/3TC induced a further median 2-log decrease. After 48 weeks, between 60 and 70 percent of the 46 persons still being studied had a viral load below 400 copies/mL. Between 40 and 50 percent of these 46 had a viral load below 50 copies/mL. The median CD4⁺ cell count rose an additional 100 cells/mm³ after people added ZDV/3TC to abacavir. Nausea and vomiting were the most common adverse events among the 55 persons who entered the open-label study extension. Staszewski reported no cases of the dangerous systemic reaction to abacavir. (See above. For studies of resistance to abacavir, see Rundown On Resistance: Little Good News.)

ACTG 347 assessed amprenavir alone or combined with ZDV/3TC in 46 naive and 46 experienced patients with a median CD4⁺ count of 304 cells/mm³ and a median viral load of 30,700 copies/mL (abstract 512). None of the study participants had taken 3TC or a protease inhibitor, and none had clinically defined AIDS. The randomized double-blind phase of the study ended abruptly after a median 88 days of follow-up when nine individuals, all in the monotherapy arm, reached a virologic endpoint (P=0.0009), defined as (1) a viral load above baseline, (2) a viral load 1 log above the treatment-induced nadir, or (3) a viral load above 500 copies/mL at week 16. An updated analysis determined that one of 39 persons in the triple therapy group and 15 of 26 in the monotherapy group met a virologic failure criterion during the first months of the study. But after six months, about one third of the people taking ZDV/3TC/ amprenavir had a viral load above 500 copies/mL. Virologic rebound in that arm has been linked most often to emergence of the I50V mutation in the protease gene. Principal investigator Robert Murphy, MD, of Northwestern University also reported that patients in the two treatment arms experienced similar average T-cell gains, about 50 to 60 cells/mm³, through the first three months of study. That increase persisted among people who continued the triple-therapy regimen. Of 34 individuals who began the trial with amprenavir alone then switched to d4T/3TC/indinavir/nevirapine, 32 quickly attained a viral load below 500 copies/mL.

Adding soft-gel saquinavir, indinavir, or nelfinavir to amprenavir yielded 1.5- to 2-log viral load reductions at two weeks in 33 individuals naive to protease inhibitors (abstract 6). At that point, patients who started the study taking amprenavir alone added ZDV/3TC. After 16 weeks, average viral load reductions measured 2.94 logs for saquinavir/amprenavir, 3.75 logs for indinavir/amprenavir, 1.84 logs for nelfinavir/amprenavir, and 2.79 logs for ZDV/3TC/amprenavir. Patients began the study with a median CD4⁺ count of 393 cells/mm³ and a median viral load of 43,650 copies/mL. Adverse events were equally distributed among arms, reported Joseph Eron, MD, of the University of North Carolina, and included perioral numbness even though no one took ritonavir.

Return to Top of Page

Go Slow: Dangerous Curves Ahead

To the surprise of none, clinical research reviewed in Chicago failed to define a simple strategy for sweeping HIV cleanly from the dusky side streets of lymphoid tissue, or even for backing it permanently into corporeal cul-de-sacs. But evidence from several trials showed signs of cohering into answers to two questions:

• What makes drugs work?
• What makes drugs fail?

Alas, answers to the second question handily outnumbered answers to the first. Still, the attentive observer could not fail to discern a few repeated motifs in studies of more successful regimens. Even the inattentive observer who showed up for the last day's symposium could not escape Chicago without a deeper appreciation of at least one theme, because the University of Pittsburgh's John Mellors sounded it clearly in his lecture (abstract S46): Deeper drops in plasma viremia lead to longer responses.

Credit for discerning this simple-sounding principle goes to Abbott's Dale Kempf, PhD, who formulated the concept in a poster at last year's Retroviruses conference. Studying people who endured virologic breakthroughs after beginning treatment with ritonavir, Kempf found that rebounds greater than 0.6 log above the lowest mark reached correlated neither with pretreatment viral load nor with how far from baseline that viral load sank.¹² Rather, the nadir achieved, no matter where the viral load started from, foretold durability. The median duration of virologic suppression was 199 days among people who attained a viral load below 200 copies/mL, compared with 128 days for people whose viral loads fell only into the 200- to 1000-copy range, and 60 days for those whose viral load never dropped below 1000 copies (P<0.01). Evidence supporting Kempf's theorem surfaced in Chicago. And these studies extend his idea by discriminating between patients whose viral loads fall below 200 copies/mL and those who attain even greater reductions measured by so-called ultrasensitive assays. As already noted, the French TRILEGE investigators correlated a viral load below 50 copies/mL with more durable suppression than a viral load merely under 500 copies/mL (abstract LB15). Three other studies looked specifically at the nadir hypothesis.

Jacques Izopet, MD, and colleagues from Toulouse studied 349 consecutive patients beginning a protease inhibitor regimen between April and October 1996 (abstract 326). After three months of therapy, 129 reached a viral load below 200 copies/mL and 36 had fewer than 20 copies/mL. Then, over a median 15 months of follow-up, Izopet counted how many from each group suffered a virologic rebound above 1000 copies/mL: Among those in the 20-to-200 group at three months, 34 percent had a subsequent rebound, compared with 7 percent of those in the sub-20 group (P<0.05). The French team concluded that "the durability of a three-drug therapy was related to initial maximal RNA suppression" and that "measuring plasma HIV-1 RNA concentrations by an ultrasensitive assay could be a useful tool for optimizing therapy."

A small retrospective analysis of stored samples by investigators at Oslo's Ullevaal University Hospital reached the same conclusions (abstract 327). A. Maeland, MD, and colleagues went to their refrigerators and pulled out 20 paired samples from individuals bled twice at a three-month interval. They picked 10 patients who had a viral load below 50 copies/mL at the first time point and 10 whose viral load measured between 50 and 400 copies/mL at that point. Nine of the 10 with sub-50 readouts remained under the 50-copy mark for three months. The investigators suspect that the person whose sub-50 viral load rebounded to about 1000 copies/mL did not adhere to his regimen. Among the 10 people in the 50-to-400 group at time 0, seven had higher viral loads three months later. The three whose viral loads fell in that period had all intensified their antiretroviral regimens. The Norwegian team bluntly concluded that "a level of RNA between 50 and 400 copies/mL indicates insufficient and unstable suppression."

Researchers studying the nonnucleoside efavirenz turned up similar findings (abstract 692). In a study of people randomized to take efavirenz plus indinavir or indinavir alone (plus d4T and efavirenz after a spine-tingling 12 weeks of indinavir monotherapy), they figured how many individuals reached a viral load below 400 copies/mL and how many had assays with PCR amplification signals below the optical density of the assay--which they designated "<1 copy/mL." Multivariate analysis found three factors that predicted subsequent virologic breakthrough to a level above 400 copies/mL: Every added year of age increased the risk of rebound 1.15 times (P=0.005); every log increase in baseline viral load raised the risk of rebound 4.31 times (P=0.035) (a finding Kempf did not make when using a 200-copy cutoff); and reaching a viral load between 1 and 400 copies/mL, but not "<1 copy/mL," increased the risk of rebound 7.59 times (P<0.001). The multicenter team, headed by James Kahn, MD, of San Francisco General, concluded that "there is a biological difference in terms of virologic treatment failure in the subset of individuals who achieve truly undetectable viral loads ["<1 copy/mL"] versus those who have residual detectable viral replication" between 1 and 400 copies/mL.

What might that "biological difference" be? Douglas Richman and his group at the University of California, San Diego, think they've spotted several such differences. At an ACTG meeting last December, Richman summarized the danger signs apparent in viral populations of people whose plasma viremia sinks below 400 copies/mL, but not below 50 copies/mL: (1) Drug resistance mutations appear, although slowly. (2) Viral evolution, marked by nucleotide sequence divergence, occurs. (3) Virus can be isolated from PBMCs by standard techniques. (4) Multiply spliced HIV RNA transcripts, indicating ongoing replication, can be detected in lymph nodes. Because of these findings, said Richman, "I personally think that the ultrasensitive [<50 copies/mL] assay . . . should be the standard of practice for patients" receiving fully suppressive therapy.

Still, one must bear in mind that viral loads below 50, or 20, or maybe even "1" may not equal total absence of viral evolution. As Joseph Wong in Richman's lab and other researchers observed in Chicago (see A "New Steady State": Slow, Low, and Dangerous?), wispy hints of evolution drift into view even when plasma seems squeaky clean. But, given the choice, no one will have a tough time deciding between a viral load under 50 and one that still rings the bell on an ultrasensitive assay.

Some evidence also indicates that how quickly the viral load drops--not just how far it drops--predicts the durability of response. Analysis of results from ACTG 343, the US induction/maintenance trial (see Induction/Maintenance Fizzles in First Studies), showed that people whose viral load remained above 200 copies/mL after four weeks of triple therapy had twice the chance of an ultimate virologic breakthrough compared with faster suppressors (abstract LB16). Speed of suppression proved not to be a statistically significant predictor of therapeutic success in this study, but it came close (P=0.07).

Such findings suggest that more frequent viral load monitoring might improve disease management. That was the hunch of the California Collaborative Treatment Group when it compared virologic trends among (1) patients who had their CD4⁺ cells and RNA copies counted every two months and at every treatment switch and (2) patients who had CD4⁺ counts every two months but only two RNA assays during the year (abstract 513). Experience with antiretrovirals and baseline CD4⁺ and RNA levels were equivalent in this cohort of 159 individuals before they were randomized to the RNA group (viral loads every two months) or the CD4⁺ group (two viral loads yearly). And everyone had at least two unused drugs available for treatment switches. After follow-up of nine months or longer, reported San Diego's Richard Haubrich, MD, the average viral load fell 0.47 log in the CD4⁺ group and 0.93 log in the RNA group. The mean "area about the change from baseline" viral load in the CD4⁺ group was -2.8 logs, compared with -5.6 logs in the RNA group. Both differences were statistically significant (P=0.002).

So, should everyone who can afford it start getting their RNA copies counted every two months and be ready to switch regimens at the first uptick? Some may lean in that direction, including the Aaron Diamond investigators who rescued three individuals from failing ZDV, 3TC, and nelfinavir and attributed that success to "an early switch in those demonstrating drug intolerance or virologic failure" (Table 2, abstract 371), and Johns Hopkins clinicians who surmised that "early switching" gave their patients a better response to double protease inhibitor therapy than others have achieved (see above, abstract 427).

Few, though, seem likely to advocate intense monitoring and lightening-like switches as everyday therapeutic policy. One reason is that no one knows how big a rebound in viral load should trigger a drug change. Any rebound from the nadir? A half-long rebound? A viral load of 51 copies/mL, or one above 5000 or 10,000 copies/mL, the levels some suggest for starting therapy? Any of these proposals can be called frivolous when one considers that plenty of people with rebounding viral loads have nonetheless remained as healthy as they were in their undetectable days (see What Does Virologic Failure Mean?). Yet staying with a weakening regimen clearly has its own risks.

At a press conference, Harvard's Scott Hammer mulled this conundrum. "Premature switching is one of the worst things you could do," he warned. When viral loads start climbing, Hammer suggested thinking about several things before declaring a regimen useless and risking new toxicities. First, whenever an assay detects viremia in a person whose viral load has been undetectable, that assay should be confirmed by a second one. Then, if drug failure looks like a possibility, try to exclude all other reasons for the rebound. Principal among these other reasons, Hammer believes, is poor adherence. Next consider how much drug experience that patient has. Is this person on his or her first regimen, or on a second- or third-line regimen? Clearly, the tendency to switch will be greater when a patient has more options.

Still, Hammer came back to the work in Douglas Richman's lab indicating "a difference pathophysiologically in what's happening in the lymph nodes in patients who are undetectable with ultrasensative assays under 50 [copies] and those with just a few hundred copies." The need to stop viral evolution is real, he stressed, but "practical circumstances, options, and limitations" obviously must weigh in any therapeutic decision.

Steven Deeks of San Francisco General Hospital picked his way through this same briar patch in a postconference interview with John James of AIDS Treatment News.¹³ What should one do, Deeks wondered, when the viral load falls below 400 copies/mL but not below 50 copies? Change the regimen completely? Intensify therapy? Wait and watch? "The problem is these new [ultrasensitive] tests provide you with powerful information," Deeks lamented, "but we do not know how to use it." Furthermore, he observed, if a report by Keith Henry, MD, from St. Paul holds true for other practices, a lot of people fall into the 50-to-400 gray zone (abstract 529). Henry figured that a third of his patients sit in that virologic shadowland.

Return to Top of Page

What does virologic failure mean?

While at the conference, Deeks shone a light on the other side of "undetectability," the side where all those 400- or 500-plus "virologic failures" make do. In fact, he reported, most of them do just fine for a year or more after pulling an F in plasma hygiene (abstract 419). Deeks and his San Francisco General colleagues kept track of 143 people who had two consecutive viral load measurements above 500 copies/mL after starting a protease inhibitor regimen. A median 18 months after therapy began, only three in this cohort have died from AIDS and only eight had a new AIDS-related illness. When Deeks tallied CD4⁺ count changes in a subgroup of 108 people who had viral load measurements before protease inhibitor treatment began, he found a median CD4⁺ increase of about 100 cells/mm³--even though the 108 included 22 individuals who never had as much as a 1-log reduction in plasma viremia.

"These observations," Deeks concluded in his poster, "suggest that failure of a [protease inhibitor]-containing regimen, defined virologically, may not be predictive of subsequent CD4 T-cell decline (over the following 12 months)." What this retrospective study does not mean, he stressed in the interview with James,¹³ is that "if you fail protease inhibitor therapies virologically then you should stay on that particular regimen." In fact, about half of the San Francisco General cohort continued the same regimen and half switched to another (although the half that switched didn't get much virologic benefit). And Deeks observed that delaying the switch to a second regimen boosts the odds that resistance will emerge to the first. (Nowhere was this point made with more startling impact than in two reports of cross-resistance, abstracts 395 and 405. See the first two items in Rundown on Resistance.)

Deeks made another piquant observation.¹³ Most of the eight people who had a new opportunistic disease after the virologic breakthrough never gained many CD4⁺ cells when they started protease inhibitor therapy, even if they had a virologic response. That trend, he noted, mirrors one reported at the conference by USC's Judith Currier in an analysis of ACTG 320 (abstract 257, see above): People in that study who registered even a modest T-cell gain after the first eight weeks of therapy increased their chance of a favorable clinical outcome.

In another Chicago presentation, Currier spotlighted a related trend (abstract S29): Despite the many reports of virologic breakthroughs during protease inhibitor therapy in the first half of 1997, overall rates of opportunistic diseases have not yet rebounded in tandem. Or, as other San Francisco General clinicians observed in their tally of HAART era OIs (abstract 183, see above), the high rate of virologic failures coupled with plummets in OI rates and hospital admissions "potentially call[s] into question the meanings of clinical and virologic failure of HAART regimens and how they relate to the clinical outcome of patients with HIV."

Why the long lag between viral breakthrough and clinical progression? Maybe even a transient virologic response from a half-HAARTed regimen has a disproportionately positive clinical effect, Deeks speculated in his interview with John James.¹³ (That speculation appears to be backed by an Emory-CDC finding of proliferating naive T cells in people with detectable viremia. See above.) Or maybe virus resistant to protease inhibitors sacrifices a fair amount of fitness and slower progression results. (That second possibility is precisely what the Pasteur Institute team of François Clavel, MD, claimed it found in a study of viral isolates resistant to protease inhibitors [abstract 543].) Either way, a year and a half of follow-up in the San Francisco General chart review is unlikely to breed too much confidence in any theory, and Deeks appropriately avoided coining any take-home message.

In the end, the Chicago conference had much more to offer on why drugs fail than on why they may be working. HIV resistance to antiretrovirals, which John Mellors guesstimated accounts for 50 to 85 percent of therapeutic failures, garnered the usual attention. (See Rundown On Resistance.) But two less heralded adversaries--flagging phosphorylation and plain old intolerance--snatched a few headlines.

Return to Top of Page

An example of "cellular resistance"?

A rough-and-ready pharmacologic study, patched together almost overnight by the Jean-Pierre Sommadossi, PharmD, PhD, showed that resistance is not the only thing that can torpedo antiviral therapy (abstract 3). When ACTG 290 investigators discovered that adding d4T to ZDV did nothing to hamper viral replication, they called in the University of Alabama pharmacologist to find out why. Sommadossi and others already suspected that ZDV impedes d4T's intracellular phosphorylation--the steps nucleosides take to produce their active metabolites. To test that hypothesis, Sommadossi scrutinized plasma samples from two individuals in the ZDV/d4T arm of ACTG 290 and from four in the d4T monotherapy arm.

First he showed that d4T and ZDV plasma concentrations stayed within normal ranges for these drugs in the two treatment groups, so low plasma levels of either drug could not explain the paltry effect of ZDV/d4T. However, mean levels of d4T triphosphate--the active metabolite at the end of d4T's phosphorylation pathway--were 6.5 times higher in the patients taking d4T alone than in those taking ZDV/d4T. Meanwhile, average levels of ZDV triphosphate were normal in the people taking ZDV plus d4T. One person who stopped taking ZDV but continued d4T after taking both for 42 weeks still had subnormal d4T triphosphate levels six weeks after stopping ZDV. The results, Sommadossi concluded, are consistent with pharmacologic antagonism between ZDV and d4T.

Do the results also mean that people who have taken ZDV will get less punch from d4T--and maybe even from other antiretrovirals? As a first step toward finding out, Sommadossi measured nucleoside phosphorylation in naive and experienced people enrolled in the ALTIS trial of d4T/3TC (abstract 362). In ALTIS, the average viral load of naive patients was 1 log lower than the average load of experienced people after 24 weeks of d4T/3TC. Sommadossi's substudy, called ALTIPHAR, divided 18 ALTIS patients into three groups: (1) 10 treatment-naive individuals who had more than a 1-log drop in plasma viremia after 24 weeks of d4T/3TC, (2) six nucleoside-experienced individuals who had taken ZDV for an average 27.5 months and had less than a 1-log viral load decrease on d4T/3TC, and (3) one individual who had taken long-term ddI and two who had taken short-term ZDV, all three of whom had more than a 1-log viral load drop on d4T/3TC.

Multivariate analysis showed that specific mutations associated with resistance to ZDV, ddI, or ddC could not explain the different virologic response between the naive and experienced patients in the substudy. And plasma levels of d4T and 3TC were within normal ranges in the three ALTIPHAR groups. Once again, though, ZDV experience proved a decisive factor. The long-term ZDV-experienced patients (group 2) had significantly lower levels of d4T triphosphate (P=0.01) and 3TC triphosphate (P=0.03) than the ZDV-naive individuals (group 1). And the long-term ZDV-experienced group also had significantly lower levels of d4T triphosphate (P=0.03) and 3TC triphosphate (P=0.004) than the ddI-experienced or short-term ZDV-experienced patients (group 3). Sommadossi concluded that intracellular levels of d4T and 3TC triphosphates correlated with virologic response in ALTIS, and that long-term ZDV therapy appeared to decrease phosphorylation of the two other nucleosides.

"These are provocative data," said Johns Hopkins pharmacologist Charles Flexner. "I think they make a lot of sense." But Flexner, and Sommadossi as well, urged continued study and caution about applying these early results in the clinic. "Beware of pharmacologists (and anyone)," Flexner counseled, "bearing limited data sets." Then he rattled off a list of questions to encourage such wariness. Exactly what causes decreased phosphorylation of d4T after ZDV? Does the blame lie with thymidine kinase (for which ZDV and d4T are analogs), or does it lie with some other enzyme or even some other still-unimagined process? Might some individuals be genetically predisposed to poor phosphorylation of d4T? Will stifled phosphorylation follow treatment only with ZDV, or with other nucleosides as well? Can ZDV's effect on phosphorylation of other nucleosides happen very early in the course of ZDV therapy, or strictly after long-term therapy?

Return to Top of Page

Lipodystrophy, or "the syndrome"

As much as researchers have learned about phosphorylation and resistance, such concepts remain arcane to most people with HIV. There's nothing mysterious, though, about another reason for drug failure: Everyone understands that drugs stop working if people stop taking them. Wobbly adherence is one thing. As Charles Flexner noted, it's hard to pin down how much a few missed doses may sabotage therapy. But when someone stops taking drugs completely, or never starts, now that's poor adherence. And the temptation to become utterly noncompliant must arise with regularity in a fair share of people taking antiretrovirals.

Indeed, events over the past few months may be testing the resolve of more than a few individuals. Eradication schedules seem to recede to a vanishing point perceptible only in Mantegna's paintings. Virus holed up in brain and resting T cells, or quite pleased to slosh about in semen, appears to set its own schedule. And more than one speaker in Chicago hinted at some Lilliputian steady state, down below 50 copies per milliliter, where replication spins a thread just stout enough to entwine trespassing giants. Still, such nightmares mean nothing to antiretroviral takers with a more palpable concern: What are these drugs doing to my body?

It was a similar question that Andrew Carr, MD, and his Sydney colleagues asked their protease inhibitor patients (abstract 410). And doctors packed 10-deep at Carr's poster in Chicago struggled with the answer: Two thirds of those patients said something awfully odd had happened to their bodies. Carr called them in for physical exams, cranked up his dual-energy x-ray absorptiometry (DEXA) machine, and found that 72 of 116 people taking protease inhibitors had become unusually paunchy, while fat was falling from their faces and limbs. Although they had gained weight around the gut, their overall weight loss measured 0.51 kg/month compared with people taking protease inhibitors who did not have this problem (P=0.0005). If Carr excluded abdominal weight, affected individuals suffered an average fat loss of 5.5 kg.

Scary enough just in its physical effects, this lipodystrophy syndrome bespeaks deeper, perhaps more dangerous metabolic changes. When Carr compared people with and without lipodystrophy, he found that those with the syndrome had significantly higher levels of triglycerides, cholesterol, C peptide, and insulin, as well as higher insulin resistance scores. Only three of the people with lipodystrophy, however, had new or worsening diabetes.

The syndrome was more common, and began more quickly, among patients taking ritonavir/saquinavir than among those taking indinavir, though it affected both groups. Carr told the Journal that he rarely uses saquinavir alone, but he saw no reason to suspect that nelfinavir or amprenavir may not cause these problems. He reported finding no association between lipodystrophy and family history of diabetes, duration of HIV infection, prior AIDS diagnoses, weight before taking protease inhibitors, CD4⁺ count or viral load before or while taking protease inhibitors, other antiretroviral agents, or drugs for opportunistic diseases.

All this was stern stuff, especially when compared with others' reports of such abnormalities in much smaller percentages of people taking protease inhibitors. Several posters clustered around Carr's detailed specific aspects of what sounds like part of the same syndrome. But none of these other reports suggested anything like a 64 percent incidence. A group from Cornell University in New York City, for example, found that five of 72 patients taking protease inhibitors (7 percent) had what they called "protease paunch" (abstract 408). A report from the FDA tallied merely nine cases of "abnormal fat distribution" that could be linked confidently to protease inhibitors (abstract 412). Clinicians at San Francisco General confirmed significantly higher levels of glucose, triglycerides, and cholesterol among people taking protease inhibitors than among those taking nonprotease regimens and untreated controls (abstract 414). But, after a mean follow-up of 4.4 months--compared with 14 months in Sydney--their DEXA analysis turned up no unusual fat distribution. On the other hand, a survey of 417 people taking protease inhibitors at Johns Hopkins discovered only eight cases of hyperglycemia or diabetes for a rate of 2 percent (abstract 415).

Why the big difference? It seems inconceivable that those 72 Australians should differ so much from people studied in the United States. One attempt to sort out these discrepancies comes from Donald Kotler, MD, at St. Luke's-Roosevelt in New York City. In a concise review available on the Web,¹⁴ Kotler enumerates all the physical and lab abnormalities that may be part of what he simply calls "the syndrome": truncal obesity accompanied by wasting of the face and limbs, fat accumulation at the back of the neck, hypertriglyceridemia, glucose intolerance, mild elevations in cortisol secretion, and subnormal serum testosterone concentrations.

Kotler posits two possible explanations for these changes: They may be a direct effect of protease inhibitors, as Carr worries; or they may represent a syndrome unmasked by an effect of protease inhibitors, such as suppressed viral replication. But evidence to date, Kotler writes, makes it impossible to chose between these alternatives. The side effects hypothesis is bolstered by the Sydney group's finding that an amino acid sequence of HIV protease looks like the sequence of a protein resembling a low-density lipoprotein receptor, which scavenges circulating lipids. But Kotler notes that "the syndrome" affects people taking antiretroviral regimens excluding protease inhibitors, as well as a few untreated HIV-negative individuals. Other preliminary evidence suggests that the syndrome predates the protease era. All that would weigh against--but not sink--the side effect theory. In fact, Kotler proposes, "it is . . . possible that both mechanisms are operative." Evincing a knack for understatement, Kotler adds that varying reports of the incidence of the syndrome (from 64 percent in Sydney to under 10 everywhere else) "tell us that more data are needed."

But Carr thinks there may be simple explanations for failure to recognize this syndrome in most patients taking protease inhibitors. He enumerated them for the Journal. (1) Early protease inhibitor studies usually involved people with advanced disease, so investigators may have written off fat abnormalities as AIDS-related "wasting." (2) More recently, other researchers have tended to zero in on discrete elements of the overall syndrome, the belly, for example. In Carr's cohort, truncal obesity did not affect everyone found to have this syndrome. (3) The Sydney team has tracked these cases longer than anyone else, an average 14 months. (4) Perhaps clinicians, "dazzled by the RNA responses" to protease inhibitors, have subconsciously avoided digging for problems. "It's the second law of the House of God," Carr suggested. "Don't take a temperature and you won't find a fever."

In Chicago, more than one clinician of merit agreed that the time has come to start paying more attention. Carr said his group now measures fasting lipids, C peptide, and insulin levels in anyone beginning therapy with protease inhibitors, and then every three to six months. Have his findings changed the way he prescribes these drugs? "My tendency to prescribe PIs is a little more muted," Carr allowed, "and I have always been a fairly aggressive treater." He told the Journal he still has no qualms about starting protease inhibitors when a person has AIDS symptoms or a CD4⁺ count below 350 cells/mm³. But for people with higher CD4⁺ counts or "low to moderate" viral loads, he is now more likely to consider nucleosides plus a nonnucleoside, or no therapy.

A common rebuttal to Carr's more cautious therapeutics, voiced by several virologists in Chicago, is that a more dire scenario--eventual diabetes or heart disease among people taking protease inhibitors--may be worth the risk if these drugs are sparing people from AIDS. But whether or not one agrees with Carr's reappraisal of protease inhibitor use--or even that "the syndrome" is more widespread than appreciated--the Sydney report offers another reminder that controlling HIV is a tough assignment. Anyone who gets down into the mean streets with this virus had better be ready for blind alleys, signals stuck on red, and hypothesis-puncturing potholes.

Three years ago, the Pasteur Institute's Simon Wain-Hobson, PhD, used the same urban metaphor to describe emerging notions of viral dynamics. Except, for him, it was the immune system that looked mean. "That billions of virions and infected cells can be destroyed every day," he wrote, "vividly illustrates the very hostile environment created by the immune system--the meanest streets are nothing by comparison."¹⁵ Today, with the help of better antiretrovirals, the immune system seems to be living up to Wain-Hobson's billing. The question now is, what's meaner, the immune system or HIV?

Do we really know if a HAART-swept immune system, aburst with brand new CD45+ v62L⁺ (naive?) CD4⁺ T cells, can enforce those just-posted "Do Not Enter" signs? For now, Wain-Hobson advised in a missive to the Journal, don't get distracted by all the scientific newspeak. "The ultimate thing is not scientific but medical--how many stay healthy and for how long," he wrote. "Always was, always will be."

Return to Top of Page

*Abstracts will be available online for one year after the February 1Ð5 meeting on the Web site of the Retroviruses conference (http://www.retroconference.org). Some abstracts are not cited in the text because a few symposia speakers and state-of-the-art lecturers did not submit written abstracts. But slides and audio of most such presentations can be accessed at the conference web site.

References

1. Finzi D, Hermankova M, Pierson T, et al. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science 1997;278:1291-1295.

2. Rosenberg ES, Billingsley JM, Caliendo AM, et al. Vigorous HIV-1-specific CD4⁺ T cell responses associated with control of viremia. Science 1997;278:1447-1450.

3.  Mackall CL, Fleisher TA, Brown MR, et al. Age, thymopoiesis and CD4⁺ T-lymphocyte regeneration after intensive chemotherapy. N Engl J Med 1995;332:143-149.

4.  Zhang Z-Q, Notermans DW, Sedgewick G, et al. Kinetics of CD4⁺ T cell repopulation of lymphoid tissues after treatment of HIV-1 infection. Proc Natl Acad Sci USA 1998;95:1154-1159.

5. Pakker NG, Notermans DW, de Boer RJ, et al. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: a composite of redistribution and proliferation. Nature Med 1998;4:208-214.

6. Gorochov G, Neumann AU, Kereveur A, et al. Perturbation of CD4⁺ and CD8⁺ T-cell repertoires during progression to AIDS and regulation of the CD4⁺ repertoire during antiviral therapy. Nature Med 1998;4:215-221.

7. Autran B, Carcelain G, Li TS, et al. Positive effects of combined antiretroviral therapy on CD4⁺ T cell homeostasis and function in advanced HIV disease. Science 1997;277:112-116.

8. Race EM, Adelson-Mitty J, Kriegel GR, et al. Focal mycobacterial lymphadenitis following initiation of protease-inhibitor therapy in patients with advanced HIV-1 disease. Lancet 1998;351:252-255.

9.  Chun TW, Stuyver L, Mizell SB, et al. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc Natl Acad Sci USA 1997;94:13193-13197.

10. Wong JK, Hezareh M, Gunthard HF, et al. Recovery of replication-competent HIV despite prolonged suppression of plasma viremia. Science 1997;278:1291-1295.

11. James JS. 1592 (abacavir): Do not rechallenge after hypersensitivity reaction. AIDS Treatment News 1997;285:1-2.

12. Kempf D, Rode R, Xu Y, et al. The durability of response to protease inhibitor therapy is predicted by viral load. Antiviral Ther 1997;2(suppl 5):63.

13. James JS. Clinical implications of virological "failure": interview with Steven Deeks, M.D., San Francisco General Hospital. AIDS Treatment News 1998;289:1-6.

14. Kotler DP. Truncal obesity, "Crix-belly"--Is it what it appears to be . . . or something else? Clinical Care Options for HIV. 1998;February (http://www.medscape.com/medscape/HIV/journal/1998/v04.n01/expert0298/expert0298.html).

15. Wain-Hobson S. Virological mayhem. Nature 1995;373:102.

980301
IA980302

Copyright © 1998 - International Association of Physicians in AIDS Care. All rights reserved. http://www.iapac.org