'Underwhelmingly innovative''

The International AIDS Society inaugurated its new program of summer research meetings with a sleeper of a gathering in the southern hemisphere port city of Buenos Aires. Mark Harrington attended on behalf of TAG and sifted the new from the old to prepare this report for TAGline.

At midwinter in Buenos Aires the days are short and the weather is cool, like summer in San Francisco. Here the International AIDS Society (IAS) has decided to initiate its pathogenesis and treatment conference, which will take place every other year. Thus the IAS is setting up an international counterpart to the Retrovirus conference. Cleverly, however, virtually the entire leadership of that meeting has been invited—so as not to appear that the two are in competition.

In an attempt to demonstrate their commitment to conducting meetings in developing countries, the IAS organizers chose Argentina for the first meeting of its kind. Now Buenos Aires is not exactly Durban, and the next two summer IAS meetings are slated for Barcelona (2002) and Paris (2003).

Buenos Aires is essentially a first world city within a third world country. Mired in recession with no turn around in sight, the country's fate appears to be in the hands of the International Monetary Fund, which is to say, U.S. Treasury Secretary Paul O'Neill, U.S. President George W. Bush and a handful of big lenders at Citibank, Bank of America and J.P. Morgan/Chase. Local newspaper headlines are filled with ominous warnings of imminent collapse, while those Argentines with the opportunity scramble to spirit their pesos into dollars and their dollars out of the country, before the nearly inevitable devaluation-and default. Each new refinancing buys only a few more months of hardship and uncertainty.

Still, the IAS meeting opened with great pomp at a conference center conveniently located under a noisy jetway, where every five minutes another plane took off with a great roar. The introductions were flowery, extensive and bilingual. The presentations contained more than a bit of data vu.

Hit even harder
David Ho treated us to a review of the highlights of his work on viral dynamics over the five years since Vancouver. In doing so he proved that, as Prince put it on a song from 1990's Graffiti Bridge, there is indeed joy in repetition. At the end of his talk he showed us that a five-drug regimen consisting of lopinavir/r with 3TC, efavirenz and tenofovir can accelerate the first phase of viral decay from fourteen days (seen with AZT/3TC/ritonavir back at Vancouver) to just seven days. It's too soon to say whether this regimen will accelerate the second phase of decay, which David attributes to virus produced from macrophages.

It's remarkable how little research has been done in this second compartment. Macrophages are terminally differentiated and found in tissue, not blood, and thus harder to sample than the latently infected CD4 T cells about which so much has been made. Maybe the five drug regimen will also curtail the second and third phases of viral decay. If so, perhaps we need take HAART (without a single slip up) for just thirty years, not sixty, in order to eradicate HIV. What fun.

Hit even later
Next Julio Montaner, updating us on the British Columbia HIV cohort, showed that baseline CD4 cell count has an immediate impact on hospitalization, AIDS, and survival, where a baseline CD4 count below 200 is bad and one below 50 very bad indeed.

Adjusting for adherence, however, shows that people starting with a CD4 count below 50 who were highly adherent—about two-thirds of them, unlike groups at higher levels, who may have been less motivated—had survival results virtually indistinguishable from those entering with higher CD4 counts. So you can indeed rescue individuals even with quite low CD4 counts. (This we knew from the Merck 060 study in 1996, but it's nice to have it corroborated.)

John Mellors just hates this data, although it doesn't in the least undermine the prime lesson from his analysis of the prognostic level of viral load from banked 1985 Multicenter AIDS Cohort Study (MACS) blood samples.

The Mellors data showed that viral load shortly after infection strongly predicts rate of CD4 cell loss and hence rate of disease progression. The Montaner data show that in the short term, CD4 count, especially at low levels, is more predictive of clinical disease than is viral load. Montaner cautioned that differences may appear over longer periods of follow-up. These data don't prove that it's better to start later, but neither do they support the notion of hitting earlier than 200 CD4 cells. (That shredding sound? That's five years of misguided guidelines being ripped to pieces.)

Hit intermittently
Tony Fauci kicked off the conference proper with an update on long cycle (60 days on, 30 days off) and short cycle (7 days on, 7 days off) structured intermittent therapy (SIT), with data on the latter out to 32 cycles (64 weeks).

The long cycle was less than optimal since viral load rebounded, did not always decline below detection after retreatment, and in some cases appeared with signs of reduced susceptibility to drug.

The short cycle SIT, out 32 cycles, looked pretty good. Ten people who had achieved good viral control enrolled and switched from their prior regimens to d4T/3TC/indinavir/ritonavir twice daily. Their median pre-HAART CD4 count was 406 and before the short-cycle SIT was 786. The median pre-HAART viral load was 55,000; the median pre-study was < 500. Two of the ten went off study (one after rebounding), while the other eight are all still being followed with good viral control. (One participant had a viral blip, but he had gone on vacation and stayed off therapy ten-rather than seven-days. Upon rechallenge, he resuppressed his virus.)

Over 60-64 weeks, there was no change in absolute CD4 count, CD4 percentage, CD8 count, activation markers, plasma HIV RNA, cellular HIV RNA, proviral HIV RNA, or number of latently infected CD4 cells. The data were "obvious and monotonous" in these respects, commented Fauci. Lymph node biopsies were similarly unremarkable and unchanged. There was no increase in HIV-specific CD4 or CD8 cell responses. No drug resistance mutations appeared. So far, so good.

The really interesting part was that triglycerides, total cholesterol and LDL cholesterol all dropped significantly from week 0 to week 24 and continued dropping out to week 52. So, at least in this handful of patients, the short-cycle SIT appears to preserve antiviral efficacy while reducing common and potentially serious toxicity. Larger randomized, controlled studies are urgently needed.

What's in the pipeline?
Despite prophecies of imminent doom periodically uttered by paranoid activists-some of whom recently claimed that activist pressure to lower antiretroviral prices in developing countries was leading drug companies to flee the HIV field-the IAS conference reflected a rather healthy, if underwhelmingly inventive or innovative, drug pipeline (see table). There were a number of presentations on new drugs early in development ("deep pipeline") and some moving towards potential approval ("late pipeline").

Nucleoside (and nucleotide) analogue RTIs
Farthest along in the pipeline is Gilead Sciences' tenofovir disoproxil fumarate (Viread), a potent nucleotide (not nucleoside-it's already got one phosphate group attached) analogue which the FDA will consider for approval on October 3. Data were presented out to 96 weeks showing the effect of three different doses of tenofovir plus other antiretrovirals among 189 treatment-experienced individuals. By week 24 the 300 mg daily dose group showed more potent antiviral activity than lower doses. At 48 weeks the 300 mg dose was significantly better than placebo. By week 96 just 21/135 individuals in the trial extension phase experienced a viral rebound.

One advantage of tenofovir is its once-daily dosing. Another is that it appears active against some nucleoside resistant HIV strains. At therapeutic concentrations in vitro the compound has little effect on DNA polymerase gamma, indicating its potential for mitochondrial DNA toxicity is likely to be low. Gilead should be commended for studying this drug in heavily pretreated individuals, something many companies have shied away from.

FTC (emtricitabine) is a once daily nucleoside analogue chemically related to 3TC. Franck Rousseau of the sponsor, Triangle Pharmaceuticals, showed that of 47 individuals who experienced virologic failure in a comparison of daily FTC with twice-daily 3TC (plus nevirapine and d4T), 31% had wild-type virus (indicating possible lack of adherence). Eighty-eight percent had at least one detectable drug mutation in the 3TC group, compared with just 56.7% in the FTC group. (For the M184V mutation, 58.8% of breakthroughs in the 3TC group had it versus just 16.7% of those in the FTC group.) Of course, such subset analyses must be taken with a dash of salt.

Unfortunately, the clinical development of FTC has been delayed due to the unanticipated rate of severe fulminant hepatitis which occurred in the pivotal FTC-302 study in South Africa, leading to the study's termination. (Apparently the hepatitis cases were concentrated among people receiving nevirapine as part of the study regimen.) Since the South African fiasco, two new studies of FTC have started, one by Triangle and one in France.

No news was reported in Buenos Aires on Triangle's other nucleoside analogue, dAPD (or amdoxovir), a prodrug which is converted into dioxolane guanosine (DXG) in vivo. HIV strains resistant to AZT, 3TC and FTC are responsive to amdoxovir in vitro, which has already been studied in a phase one dose-ranging study previously reported.

Another paper reported on Biochem Pharma's experimental nucleoside analogue (-)dOTC-also known as BCH-10618-which appears active in vitro against some wild-type and drug-resistant HIV strains. This drug, like tenofovir, exhibits little mitochondrial toxicity-at least in vitro. It reduced HIV RNA by about one log in a phase one dose-ranging study. However, some monkeys died after receiving the drug for three months. Thus, clinical development of Biochem's second dOTC appears to be on hold. (BCH-10652 was dropped earlier, also due to problems of toxicity.)

Non-nucleoside reverse transcriptase inhibitors
Belgium's Tibotec/Virco is moving forward with the development of its non-nucleoside TMC-120, which at Retrovirus was reported to induce a two log reduction in viral load over seven days among 43 individuals taking it at 100 mg twice daily. Tibotec has another non-nuke, tagged TMC-125, as well as a protease inhibitor, TMC-126, in early stages of development.

DuPont has two experimental non-nukes, DPC-961 and DPC-963, whose further development presumably awaits approval by the U.S. Federal Trade Commission of the proposed merger between DuPont and Bristol-Myers Squibb (BMS).

Agouron's capravirine is a non-nuke reported active against some HIV strains resistant to efavirenz, nevirapine and delavirdine. The drug's development was put on hold by the FDA in January due to some cases of vasculitis (an inflammation of the blood vessels) in canine trials, despite the fact that capravirine was already being studied in 650 HIV-infected individuals. Little of note has been reported since the FDA hold.

Another non-nuke whose development appears to be at an end is the compound known variously as GW420867 or HBY1293, whose development began with Bayer, continued with GSK, and now apparently has been terminated in spiteof intriguing preclinical data.

Protease inhibitors
Bristol-Myers Squibb (BMS)'s once-daily protease inhibitor, atazanavir, formerly known as BMS-232632, is moving slowly but deliberately through phase two. In addition to its convenient once-daily dosing, this drug appears not to perturb cholesterol levels, unlike the nelfinavir used in the control arm of the pivotal efficacy study. Two doses are being compared with nelfinavir (background nukes are 3TC and d4T). At 24 weeks the differences in lipid profiles were striking.

After a long hiatus, tipranavir has been transferred from Pharmacia & Upjohn, which left the HIV field, to Boehringer-Ingelheim (which passed on P&U's other antiretroviral, delavirdine). Tipranavir is structurally different from other protease inhibitors and hence exhibits activity against most PI-resistant HIV. On the downside, tipranavir has a daunting pill count.

The tipranavir data presented in Argentina were somewhat confusing, as the formulation changed mid-study. In addition, tipranavir was given in combination with low-dose ritonavir in order to boost the drug's less than overwhelming pharmacokinetics.

Among 41 individuals who had failed at least two protease regimens, but remained NNRTI-naïve, tipranavir/r plus efavirenz and at least one new nuke (when possible) reduced HIV RNA by 2.35 logs at the lower dose and by 1.71 logs at the higher dose. (Paradoxically, high-dose tipranavir® achieved lower blood levels of tipranavir than did the lower dose.) The side effects were not trivial: diarrhea (59%), nausea (31%), elevated LFTs (30%) and vomiting (17%).

Boehringer held a spirited community meeting where they were questioned aggressively on these results, although some of the problems with the development plan were the legacy of the previous sponsor. For example, the dose-ranging study included efavirenz, which is likely to affect the drug's pharmacokinetics. Despite its intriguing resistance profile, the combination of high pill count, bizarre pharmacokinetics and significant toxicity poses a trinity of major obstacles for this drug.

Mozenavir, also known as DMP-450, is a protease inhibitor being tested by its sponsor, Dupont, against indinavir in a 50-person phase II dose-ranging study. The drug appeared equipotent to indinavir (800 mg three times daily) at all three doses (70% below 50 HIV copies/mL on indinavir versus 67-77% on mozenavir). There was a lot of diarrhea (50-70% on mozenavir) but, hearteningly, no cardiac arrhythmias as was feared due to preclinical toxicity seen (at a 15-fold higher dose in dogs). Triangle reports that "DMP-450 [mozenavir] remains on partial clinical hold."

We haven't heard much of late about the Vertex/Glaxo amprenavir prodrug VX175/ GW433908. Similarly, little news has emerged concerning DuPont's experimental protease inhibitors DPC681 and DPC684, or Merck's indinavir follow-up (reputed to be a PK boosting drug like ritonavir), L-756,423.

Fusion and binding inhibitors, and other antiviral targets
Despite evidence of significant antiviral activity against drug-resistant HIV and relatively little toxicity, Trimeris Pharmaceuticals' two fusion inhibitors T-20 (pentafuside) and T-1249 appear to be moving forward at what can only be described as glacial speed.

Repeating a presentation from last fall's Glasgow conference, a team from Sweden's Karolinska Institute and Tripep AB reported on GPG-NH2 or glycyl-prolyl-glycine-amide, a tripeptide which inhibits HIV capsid formation in vitro and can be given orally. Previous results suggested a not-too-impressive 0.4 log reduction in HIV RNA among 9/15 individuals who received a consistent dose, though two people experienced a 1.0 log reduction.

Schering's new CCR5 blocker, SCH-C, has recently been released from a clinical hold and appears to be moving towards phase two. A meeting with the company late last month, however, did little to dispel a lack of confidence in the its understanding of AIDS drug development. Schering also has an apparently more potent follow-up compound, imaginatively dubbed SCH-D.

Pfizer too is reputed to be moving its experimental CCR5 inhibitor UK-427,857 into phase one soon. A meeting with the company later this this month promises to shed some light on the status of UK-427,857 and other HIV drugs in the Pfizer/Agouron/Parke-Davis pipeline. Unfortunately, two CXCR4 blockers-ALX40-4C from Allelix and AMD3100 from AnorMed-have been dropped due to formulation difficulties, toxicities, and lack of efficacy. AnorMed, however, may have a back-up candidate or two.

Meanwhile, several drug and biotech companies continue to attempt the difficult task of turning potential integrase and zinc finger inhibitors from lead compounds in vitro into potential drug candidates in vivo.

It's been quite awhile since we've heard anything new about Aronex Pharmaceuticals' zintevir [now Antigenics, subsequent to a summer takeover], formerly known as AR-177, a compound whose safety profile was reported back in 1996. Reports from Cornell/New York Hospital, where the drug was to be tested in a phase I study, are that zintevir has been scrapped. Cornell however, as well as Columbia Presbyterian and UA/Birmingham, are currently recruiting for a PK study of a different integrase inhibitor, that of Shionogi Pharmaceuticals, S-1360.

In an up-coming report, TAG will take a look at the obstacles-both scientific and institutional-to the discovery and development of chemical entities active against new antiviral or host factor targets. ¤

• Related table with extensive links: Drug development pipeline in 2002

Thanks to Ben Cheng and Yvette Delph for useful comments on this report, and to Keith Alcorn and Peter Staley for useful, up-to-date on-line information.

'Six years of lost data'

David Barr revisits the changes to the federal when-to-start recommendations announced earlier this year and urges us to consider both the roots and the ramifications of this extraordinary development. While acknowleding the predicament of a Monday morning quarterback, his words are frank and hard-hitting. "How many people have only suffered from the side effects of treatment—and not from HIV infection?" he asks. And "where is the research infrastructure to study this chronic, if not quite manageable, disease?"

Where is the outcry from AIDS advocates following the recent change in the U.S. government's adult HIV treatment guidelines? The guidelines panel sponsored by the National Institutes of Health (NIH) changed its recommendation from starting antiretroviral therapy at T cell counts below 500 to waiting to start until T cell counts fall below 350. New British guidelines have gone even further, recommending treatment not start until T cell counts fall below 200.

These changes are based on two central premises: 1) that the development of drug resistance and side effects are leaving individuals with fewer treatment options over time, and 2) that data from multiple (mostly European) cohort studies indicate no significant difference in response to treatment in people starting at 500, 350, or 200 CD4 T cells.

Perhaps jaded AIDS advocates think the guidelines change is not significant. But this change affects the largest group of HIV-infected people in the U.S.-the untreated-and has implications for treatment strategies worldwide. The guidelines are still an important tool for many, if not most, physicians, treatment educators, and patients. While the magnitude of this change may undermine their faith in the guidelines process, what it should really do is question the priorities of the research process.

The change indicates that the original guidelines were wrong. How many people with HIV were hurt because they followed those original guidelines? How many otherwise asymptomatic patients are now multidrug resistant? How many suffer from irreversible side effects? How many of those people have only suffered from the effects of treatment-not HIV infection? How many HIV-infected individuals wasted the very valuable benefit that HAART offers because they followed guidelines based solely on expert opinion about both drugs and diagnostic technologies untested in clinical practice?

Perhaps it is unfair to criticize mistakes from information viewed in retrospect. Certainly drug resistance and the possibility of long-term side effects were recognized as potential problems in 1996. Yes, one can justify the mistake of prematurely recommending early use of AZT monotherapy in 1989, but by 1996, everyone should have known better.

In 1996, it was clear that there were new, powerful drugs that could radically alter the course of HIV infection. Recommendations could have been made for immediate use of these new therapies for those who needed them most and for whom we actually had efficacy data-people with AIDS. Meanwhile, researchers could have charted a course of research to better understand the long-term and most effective use of these drugs over time, including when to initiate treatment, what combinations to take, how to address treatment failure, and how to recognize and potentially treat side effects.

Instead, the drugs were recommended to the broadest range of patients, making further study more difficult-perhaps impossible. The guidelines panel members who advocated for the most aggressive treatment approach in 1996 were the same people responsible for developing a research agenda to learn about how to use these therapies effectively over the long term. Unfortunately, they were less aggressive in developing their research agenda than they were their treatment recommendations.

The clinical trials networks should have geared up in 1996 to undertake studies of the long-term effects of these treatments. Large, randomized, strategy studies should have been designed. New cohort studies should have been created. But attempts to address these issues by the research establishment were meager. Instead, the clinical trials networks continued to primarily emphasize new drug development and smaller, shorter studies based on viral load and T cell changes. Such studies still have an important role in HIV research, but in 1996, long-term clinical effectiveness research should have become a priority, particularly within government-sponsored research programs. So far, people with HIV have lost six years of important data collection.

In 1999, the National Institute of Allergy and Infectious Diseases (NIAID) recompeted its clinical trials networks. Here was a crucial opportunity to develop a research agenda to understand the strategic use of HAART over the long term. NIAID could have used this funding process to create a clinical trials infrastructure expressly designed for such research. That opportunity was wasted.

The ACTG, with a proven record for conducting state-of-the-art studies based primarily on surrogate endpoints, was fully funded. The ACTG was created at a time when there were few treatments for HIV and when drug development was the only real priority. That infrastructure is still important, but it has not met the challenge of developing long-term clinical effectiveness research-nor is it designed to do so. Although its ALERT protocol will follow patients from study to study, ALERT is neither a controlled study to test different strategies of antiretroviral use over long periods of time in large groups of patients, nor is it a comprehensive cohort to study the effects of HAART on HIV infection in heterogeneous health care delivery settings.

The Community Programs for Clinical Research on AIDS (CPCRA) seems genuinely interested in questions of long-term clinical effectiveness, as evidenced by the FIRST, SMART, and long-term monitoring studies that are currently under way. These studies will address important questions about what combinations of drugs to start and what to do when they fail. The CPCRA, however, is too small a network to adequately enroll large, long-term studies-and its record of long-term follow-up is not good. It could have joined forces with the Veterans Administration (VA) and created a research network with the largest pool of HIV-infected individuals ever, but it squandered the opportunity, opting to squabble over turf instead. Rather than figure out how to use effectively the largest provider of HIV care in the nation, NIAID criticized the VA for an admittedly over-ambitious research agenda and refused to fund them. Meanwhile, NIH still has no way of studying what is happening to the more than 25,000 people with HIV getting care through the VA. Thus, NIAID failed to develop the clinical trials network needed to understand the effects of the treatment strategy it was recommending in its guidelines.

Think of the important information learned from the Multicenter AIDS Cohort Study (MACS) over the years. Why wasn't a new and more diverse cohort started in 1996 to examine the long-term effects of HAART? If HAART turned HIV into a chronic (if not manageable) illness, then where are the research infrastructures to study a chronic disease? The change in the guidelines marks yet another turning point in the roller coaster of HIV treatment. Perhaps it is time someone came up with a way to chart what is clearly going to be a long ride. ¤

Thanks to RITA editor Thomas Gegeny and the Houston Center for AIDS (www.centerforaids.org) for permission to reprint David's thoughts in this issue of TAGline.