It is painfully evident that new antiretrovirals are needed to treat resistant virus when people are forced to construct a therapeutic regimen by recycling six, or worse yet, eight old drugs. And if there were enough new agents far enough along in clinical development, people would not have to resort to strategic treatment interruptions (STIs). The situation for many is truly desperate.

It is not enough for a drug to be "new," however. To outwit HIV, we need molecules that are significantly different from what has been used before. All too often, new drugs cause undue excitement. Pharmaceutical companies publish posters at these conferences that make claims on the basis of minor test tube experiments that rarely ever translate into clinical success. Next thing they know, activists are demanding expanded access.

In fact, very few of the antivirals in development show promise as rescue agents. Clearly those must be prioritized. The compounds that offer the most promise for patients with multidrug-resistant virus are those that attack entirely new targets in the viral replication cycle. There are chemokine receptor antagonists, integrase inhibitors, transcription blockers and other new classes of drugs in preclinical development, but these will be discussed in an upcoming issue. Nevertheless, even if an antiviral is not active against multidrug resistant virus, we still welcome its development if it leads to HAART regimens that are less cumbersome, simpler to adhere to or easier to tolerate.

Research on several of the drugs in development was presented at the 39th ICAAC conference in San Francisco and the Lisbon meeting. Unless otherwise noted the citations are to the ICAAC abstracts. Because this editor's special area of interest is drug discovery and development, we thought we would expand our coverage to tell you some of the story behind the posters. Treatment newsletter writers and activists interact very closely with the pharmaceutical industry, and we learn things about the companies -- their economic situation, corporate philosophies, history, goals and personalities -- that rarely get reported, but bear upon how a drug gets developed. For example, if Abbott decides it is in its best interest, the company has the experience and might to develop a new drug and get it into expanded access much faster than a baby biotech company with an employee force of 20 bench scientists and an executive director. It is not uncommon for an inexperienced company to conduct poorly designed studies that make its drug look worse than it is because the company is not familiar with clinical practice in HIV disease. In contrast, some (and we emphasize some) of the larger companies know how to conduct a study in such a way to put their drugs in the best possible light. But if any company has a useful drug, it is in everyone's best interest that its venture is a success.

It is important that each company's data are reviewed fairly and that a compound's potential is neither exaggerated (because hype will not suppress your viral load) nor dismissed prematurely. Both are crucial to the survival of people with AIDS who need all of the "true" salvage agents they can get.

Protease Inhibitors


Data from two phase II trials of ABT-378/r, a new protease inhibitor from Abbott that is combined with ritonavir (hence, the small "r"), were presented at a late breaker session at ICAAC. The studies evaluated ABT-378/r in both treatment naive and experienced populations.

In the naive study, patients were randomized into two blinded groups. The first group (n=32) received either 200/100 mg or 400/100 mg of ABT-378/r twice a day, while the second group (n=68) received either 400/100 mg or 400/200 mg ABT-378/r twice a day. Both groups also took d4T/3TC. The median baseline viral loads were 100,000 and 79,000 copies/ml, respectively. At week 36, an on-treatment analysis of both groups showed that slightly more than 95% of patients had HIV RNA <400 copies/ml and 89% had <50 copies/ml. On an intent-to-treat analysis, 81% of those in the first group and 87% of those in the second group had HIV RNA <400 copies/ml. CD4 counts increased by 201 cells in the first group and by 170 cells in the second.

In the treatment experienced study, 70 patients currently failing treatment were randomly assigned to receive 400/100 mg ABT-378/r or 400/200 mg ABT-378/r twice a day, plus nevirapine and two nucleoside analogs (NRTIs), at least one of which was new. Subjects had been receiving a protease inhibitor and two NRTIs for at least three months but no previous NNRTI therapy or previous dual protease inhibitor therapy. The median baseline viral load was 10,000 copies/ml.

By week 36, an on-treatment analysis found that 78% of all patients had HIV RNA <400 copies/ml (67% on an intent-to-treat analysis), and there was a mean CD4 cell count increase of 104 cells. In both groups, the most common adverse events were diarrhea, nausea, asthenia (weakness) and headache, but only two patients out of 170 (which includes both studies) discontinued due to adverse events.

Given the choices available to treatment-naive populations, the primary interest in ABT-378, like most new drugs, is what it offers people who are treatment experienced. Yet, the design of this study makes it difficult to tell what ABT-378 contributed to the treatment outcomes. Note that study participants were receiving a five-drug regimen that included at least three new drugs -- ABT-378, an NNRTI (nevirapine) and at least one new NRTI -- and they had never had dual protease inhibitor therapy before, which means that have never used a protease inhibitor with plasma concentrations enhanced by ritonavir. Thus, it is not clear that the ABT-378-based regimen was any more effective for these patients than a new protease inhibitor plus ritonavir, their first NNRTI, and at least one new NRTI. Even the same regimen except without ABT-378 might have performed equally well. Without head-to-head studies of such regimens, it is impossible to tease out the relative contribution of ABT-378/r. The addition of a member of an entirely different class of agents (an NNRTI), in particular, makes it impossible to say how effective ABT-378/r is in protease inhibitor-resistant patients. After all, it only takes a 1.4 log reduction in viral load to go from 10,000 to 400 copies/ml, which an NNRTI can do on its own.

Moreover, the study provides little information for those who have failed three or more HAART regimens -- precisely the situation facing most people who qualify for Abbott's limited early access program (see ABT-378/r Compassionate Use Program in this issue). Once again, there is no evidence that the ABT-378/r-based regimen would be more effective than any other five drug regimen based on another new protease inhibitor plus ritonavir. Given that the mechanisms of protease inhibitor cross-resistance remain unclear, people with established protease inhibitor resistance face other uncertainties. If they are constructing a rescue regimen with one or two other new drugs and adding ABT-378/r to it, they may be in for a major disappointment. Not only might their regimens fail, but they may become resistant to their remaining therapeutic options because the regimen was not potent enough. Such patients need proof that ABT-378/r is effective against multiply protease inhibitor-resistant virus.

Abbott hopes to answer some of these concerns in future trials. One study, recently filled, will test an ABT-378/r and efavirenz based regimen in multiple protease inhibitor experienced but again in NNRTI-naive patients. Another, planned for early next year, will test higher doses of ABT-378 in patients who have NNRTI and multiple protease inhibitor experience. Abbott will also be collecting data from the early access program. However, results from these trials are not expected until the World AIDS Conference next July.


New pharmacokinetic data on tipranavir, a much-awaited protease inhibitor from Pharmacia & Upjohn, were presented at ICAAC. In contrast to ABT-378, which is promising for protease inhibitor-experienced patients primarily by virtue of its high plasma concentrations which may overwhelm low-level resistance, tipranavir appears to be dramatically different in terms of its resistance profile. Some have described the compound, which is non-peptidic, as being more "flexible," or even "stickier," than the current crop of protease inhibitors, although it is not really clear whether this nomenclature is scientifically accurate. Nonetheless, in vitro and in a number of clinical case studies, this non-peptidic protease inhibitor has demonstrated activity against HIV isolates that show high level resistance to all of the available protease inhibitors.

Unfortunately, the drug's development has been plagued by formulation problems. The first clinical trial of the compound used doses up to 2000 mg tid (ten "horse" pills three times a day). Subsequent studies have used a better formulation, but pill counts remain high nonetheless.

Yet even though this compound is chemically unrelated to the other protease inhibitors on the market, it is metabolized via the same P450 CYP3A4 hepatic enzyme systems as ritonavir, and coadministration may substantially reduce pill burden with tipranavir. However, unlike the other protease inhibitors, it tends to induce rather than inhibit this system. Therefore, it reduces drug levels of most other drugs metabolized this way. For example, in one poster at ICAAC(1), tipranavir reduced plasma concentrations of ritonavir by 80% to 88% (AUC) and there was a 20-fold decrease in ritonavir trough levels. The poster described two pharmacokinetic studies on the coadministration of the two agents. In the first study, tipranavir 1350 mg bid was coadministered to patients taking ritonavir 500 mg bid. This may have been overkill, as ritonavir 500 mg bid caused a 12-fold increase in the AUC of tipranavir and a 45-fold increase in trough levels. In the second study, two different doses of tipranavir, 600 and 900 mg bid, were administered with either 100 or 500 mg ritonavir. The effect of ritonavir upon tipranavir was determined to be dose-related with the low dose boosting tipranavir trough levels by 7- to 9-fold.

Clearly, this is moving in the right direction -- from 30 horse pills down to 8 or 10 per day. But since P & U sold Agouron the rights to delavirdine last October, there is some question as to whether it plans to continue developing tipranavir. A company representative told us that P & U will continue to press forward but acknowledged the company is looking into a partnership or even selling the rights to the drug. If P & U decides to leave the HIV market, who would buy it? Of the remaining major pharmaceutical companies in the field, only DuPont has not produced a protease inhibitor -- yet.


In cooperation with Vertex, Glaxo-Wellcome presented data on VX-175/GW433908, a lower pill count, water-soluble phosphate prodrug of the protease inhibitor amprenavir. Because amprenavir has poor aqueous solubility, it is formulated with a substantial amount of water-soluble Vitamin E to aid in its absorption. Together, it all adds up to eight sizable pills that must be taken twice a day.

This is amprenavir's Achilles' heel. While there is no clear evidence that the formulation contributes to amprenavir's gastrointestinal toxicity (the nausea, vomiting and bloating), the volume of drug can't help. Furthermore, one of the drug's chief selling points was meant to be convenient dosing: twice daily, with or without food. But the pill count can make adherence more difficult for some patients. A number find it easier to spread the pills out over the course of an hour, but so much for convenience.

There are studies that suggest that pill count is not that important to adherence. All things being equal, this might be true, but if the medication is also making you nauseous (the most frequently reported toxicity on amprenavir), each successive pill can grow harder to swallow. This could have been the case for some of the patients in the PROAB 3006 study, amprenavir's head-to-head competition with indinavir, presented in Lisbon (see A Cornucopia of New Drugs in this issue). As noted in the study's executive summary, adherence was better on indinavir, which must be taken three times a day and on a mostly empty stomach. While the rates of virologic failure were not very different between amprenavir and indinavir, poor adherence and a high drop out rate due to toxicity hurt amprenavir in the intent-to-treat analysis. (To both companies' credit, at least they dared to go head-to-head with the gold standard protease inhibitor -- Agouron never did.) Even so, amprenavir worked well in those patients who could tolerate and adhere to it, and much of the toxicity abated after the first month. Nevertheless, initial tolerability and ease of adherence are factors that must be weighed when choosing a first-line or first protease inhibitor regimen. Clearly, the formulation has been hindering the drug's performance.

The solution could be VX-175/GW433908. The amprenavir prodrug is more soluble which makes the Vitamin E unnecessary. The improved formulation should allow for a significant reduction of both the size and number of pills.(2) Studies in rats and dogs showed that toxicology is similar to that of amprenavir.(3) A 24-hour multidose phase I study in healthy men, given with and without food, showed bioequivalence and dose proportionality to amprenavir. Cmax was about 27% lower (but Cmax is more often associated with toxicity than antiviral effect). A high-fat meal reduced AUC by 20% and Cmax by 41%. However, the most important parameter, the trough level of drug, was unchanged.

The next step in bringing VX-175/GW433908 to market will be to conduct a larger study in people with HIV demonstrating bioequivalency of this formulation with amprenavir. Depending upon their final size, this formulation could cut the pill burden down to three or four moderately sized tablets twice a day. Given the long halflife of amprenavir (about 10 hours), it may even be possible to administer the drug once-a-day without a ridiculously high pill count, either with low dose ritonavir or perhaps even without it.

Other Protease Inhibitors

We will have to wait until the Retrovirus Conference to hear news on BMS 232632, which may be cross-resistant with indinavir and ritonavir, but might be administered once a day; AG 1776, which promises activity against protease inhibitor-resistant strains; and Merck's second generation drug L-756,423.

Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTIs)

Emivirine (Coactinon, MKC 442)

The NNRTI furthest along in clinical development is Triangle's MKC 442, now known as emivirine (EMV). This compound is chemically similar to a nucleoside analog, but it functions as an NNRTI. Emivirine is not very well known in this country because, aside from data from a small dose-ranging study, little clinical information has been available on it, and its key clinical trials have been conducted abroad. Yet, emivirine will be the next new anti-HIV drug up for FDA review after adefovir. Preliminary data from one of the pivotal studies that will support this application were presented at ICAAC, and a second was ready for the Lisbon meeting.

The first study, conducted in South Africa, was open-label and divided 197 HIV-positive patients into four arms: (1) 500 mg EMV twice a day, (2) 750 mg EMV twice a day, (3) 250 mg twice a day for 3 days and then 750 mg twice a day and (4) 375 mg twice a day for three days and then 750 mg twice a day.(4) The rationale for arms (3) and (4) was that, like ritonavir, blood levels of emivirine are higher when it is first administered, and it was hoped that a lower lead-in dose might be better tolerated. In all four arms, patients also received d4T and ddI. Patients were NNRTI- and protease inhibitor-naive and had fewer than 28 days of prior NRTI exposure. Thirty-eight percent were women, the median baseline HIV RNA was 4.5 log (about 32,000 copies/ml) and the median baseline CD4 count was over 300 cells.

At week two, 89% of patients in all four arms demonstrated a one-log reduction in HIV RNA from baseline. Seventy-five percent of the individuals who reached week 12 experienced a viral load drop to below 400 copies/ml. On an intent-to-treat analysis at week 24, 64%, 57%, 43% and 52% of patients had HIV RNA <400 copies/ml in arms (1)-(4), respectively (54%, 51%, 30% and 38% were <50 copies/ml, respectively). However, not all subjects in arms (3) and (4) had reached week 24.

Seventy-two percent experienced mild adverse events, the most common of which were nausea, headache, transient dizziness and diarrhea. There were no life threatening (grade IV) adverse events, and grade III events were seen in two to four percent of the patients. Twenty-three percent experienced rash but most (83%) remained on study. The lead-in arms did not improve tolerability, so the company is going to evaluate longer lead-ins.

Given the design of the study, its locale and the preliminary nature of the data, one should not rush to compare these results to those from studies of other NNRTI-based regimens. For instance, here there is no control -- one can only evaluate which emivirine dose or dose-escalation regimen is better. Also, since a sizable proportion of the patients have not yet reached week 24, an intent-to-treat analysis is unfair. Finally, cross study comparisons are never valid, but particularly when comparing studies conducted in the US and Europe to a study conducted in a developing nation, where the standard of care, even in clinical trials, may be much different.

In Lisbon, results were presented from a second pivotal study(5) that compared emivirine plus d4T/3TC to d4T/3TC. The randomized, double blind 24-week study enrolled 162 treatment-naive volunteers, 45% female and 55% male, with median baseline viral loads of 4.3 log (just under 20,000 copies/ml) and median baseline CD4 counts of 418 cells. At Week 24, 40% of the patients on dual nucleoside therapy versus 83% on triple therapy had viral loads below 400 copies/ml (84% of the latter were also undetectable using the ultrasensitive assay).

The side effect profile was very similar to that of the emivirine/d4T/ddI trial. The most common adverse events on triple therapy were nausea (36%), headache (22%), dizziness (17%) and rash (22%). Most of the rash was mild to moderate. Close to 8% of the emivirine-treated patients experienced grade III laboratory abnormalities. Most adverse events were early onset, observed during the first month on therapy. Again, this is part of the rationale for a lead-in dose.

The genotypic and phenotypic resistance patterns in patients who failed on an emivirine regimen in both of these studies were discussed in a poster at ICAAC.(6) Twenty-four out of 138 patients experienced virological failure. Resistance analysis revealed that the most common mutations in these patients were at reverse transcriptase positions K103N (56%), Y181C (9%) and G190A (9%). The K103N mutation causes resistance to all of the marketed NNRTIs, but the company was quick to point out that nine of 24 (37.5%) were still sensitive to at least one other NNRTI. However, such in vitro findings do not mean that an emivirine-failing patient without detectable K103N-mutant virus is likely to experience a durable response to subsequent efavirenz treatment. Genotypic tests are notoriously insensitive, and the K103N mutation is so common that there is a danger that it may be present in all patients failing emivirine but at levels just below the sensitivity of the test. Before banking too much on these findings, Triangle should conduct a sequencing study to show that patients with this resistance pattern can indeed be salvaged by subsequent NNRTI therapy.

In a context where there are several marketed NNRTIs, and where people are cynical about the ability to sequence these agents, the company has to find a clear way to distinguish this product from its competitors. Those who are prone to use NNRTIs for first-line therapy are already used to, and generally comfortable with, the agents they are using. Furthermore, Triangle has decided to run its major studies overseas, so clinicians here have had little experience with the drug. When emivirine hits the market, it may prove a challenge to convince clinicians to choose a less-well characterized compound over one that they have often used.

But Triangle will have the marketing might of Abbott on its side. Several months back, the two companies entered into a pact to codevelop and comarket the antivirals in their portfolios. These include ritonavir, ABT-378/r, emivirine, FTC or emtricitabine, dADP and LFMAU (for HBV). Abbott's sales force is substantial, but it will be interesting to see whether they will expend much energy on marketing a protease inhibitor-sparing regimen when they have one protease inhibitor already on the market and ABT-378/r waiting in the wings.

DPC 961, 963, 082 and 083

DuPont is currently looking at four new "second generation" NNRTIs, DPC 961, 963, 082 and 083, for possible development. Data on their in vitro resistance and potency were presented in a poster session.(7) Researchers examined their potency against resistant strains of HIV as well as the availability of free drug levels. In terms of resistant virus, all four new compounds showed high activity against the K103N mutant (which is associated with resistance to approved NNRTIs) and maintained potency against virus with single mutations and common double mutations such as the K103N/Y181C. In addition, the four compounds show increased levels of free drug in human serum compared to efavirenz. Tests on non-human primates suggest that all four compounds could be used once a day in human trials. While the in vitro results look promising, the real test will come when one of these is tested in patients with NNRTI-resistant virus.


Glaxo-Wellcome is continuing its effort to cover all the antiretroviral bases with an NNRTI of its own, GW420876X, as it appears unlikely that abacavir will be able to compete for efavirenz' and nevirapine's growing market niche. GW420876X is a molecule related to HBY097, which was discontinued by Bayer a couple years ago. Introduced last year at the Fourth International Congress on Drug Therapy in HIV Infection meeting in Glasgow, thus far the drug appears to have a number of interesting properties. It has a very long half life (about 50 hours) so it can be given once a day -- with or without food or even with a high fat meal, according to one study presented at ICAAC.(8) In contrast to the other available NNRTIs, the compound is not metabolized by cytochrome P450 isoenzymes, so it may be free of the drug interactions seen with the other NNRTIs.

The first data on GW420876X in HIV-positive patients were presented at ICAAC.(9) The ongoing double blind trial randomized 60 patients into four arms that received a placebo or 50, 100 or 200 mg of GW420876X once a day. Participants were antiretroviral-naive, had viral loads between 10,000 and 1,000,000 and CD4 cell counts over 200. Baseline characteristics between the groups were similar, with the median viral loads for each study arm ranging between 4.4 to 4.9 log (approximately 25,000-79,500 copies/ml), and median CD4 counts were above 300 cells. More than a third of the patients were women, and 48% were non-Caucasian.

On the first day, subjects received a "loading dose" double their randomized dose. During the first week, the drug's potency as monotherapy was evaluated; for the following three weeks, it was given in combination with AZT/3TC. At day seven, the median drop in viral load in the arms receiving GW420876X was 1.5 log copies/ml; there was no significant difference between the treatment arms. At day 28, 70-86% of patients on treatment had HIV RNA <400 copies/ml. The most common adverse events (in order of frequency) were headache, fatigue, nausea, flatulence and sleep disorders. (GW420876X readily crosses the blood brain barrier in animals, so one question about the compound is whether it will cause neurological side effects.) Most adverse events were mild to moderate, although one person had a serious elevation in ALT, a possible sign of liver toxicity, which resolved when therapy was discontinued. The two higher dose groups showed a trend toward increased adverse events, while there was no difference between the placebo and 50 mg groups.


Samjin Pharmaceuticals presented a couple of posters on a new NNRTI that also has activity by blocking viral attachment to the cell.(10) While mutations that cause resistance to the other NNRTIs also inactivate this molecule's anti-reverse transcriptase activity, it still exerts an antiviral effect in vitro via its anti-attachment mode of action. Yet, compounds that were "attachment" inhibitors in the test tube were some of the first ever be studied in people with HIV -- and not one made it to market.

And The Rest ...

At least a handful of other NNRTIs that promise to have activity again K103N mutant virus are in development. Agouron's AG-1549 is active against virus with single mutations in the reverse transcriptase enzyme, but two or more common mutations may defeat it. The drug is in phase II studies, but no new data were reported at either conference. Bristol Myers Squibb has purchased the development rights to UC781, a compound that shows activity at unusually low doses in the test tube, from Uniroyal. Now that Pharmacia & Upjohn has sold delavirdine to Agouron, it is not clear what the company will do with its second generation NNRTI, PNU-142721. And Sarawak MediChem Pharmaceuticals is still in the running with its rainforest tree derivative, Calanolide A.

Nucleoside Reverse Transcriptase Inhibitors (NRTIs)

Lodenosine (FddA)

Lodenosine is a nucleoside analog much like ddI, with at least one significant advantage: It does not need to be formulated with an antacid buffer. In order to be absorbed into the bloodstream, ddI needs a buffer to help it survive transit though the stomach's acidic environment, and that buffer's ability to protect ddI is so tenuous that one has to be careful what else is in the stomach with it. Lodenosine, however, is impervious to stomach acid and can therefore be taken without dietary restrictions. Also in its favor, and frankly much more critical for people in need of salvage agents, it has demonstrated activity against nucleoside analog resistant virus, both in vitro and in clinical trials at the National Cancer Institute (NCI). Despite its promise, lodenosine's development at the NCI has been slow, though now it is in the hands of US Bioscience, a small company that developed trimetrexate for PCP. This company licensed the drug from the NCI and launched a moderate sized phase II randomized dose-comparative study about a year ago, but the company has nowhere near the resources of a larger pharmaceutical company.

Much to the surprise of the study's Data Safety and Monitoring Board (DSMB) the company released preliminary results from this on-going study at ICAAC. The members of the DSMB were not advised of this unscheduled analysis (which is highly irregular) and resigned, but the company's action can probably be chalked up more to lack of experience than anything else.

Initially, the data presented looked promising. The study was comparing three different doses (100, 200 or 300 mg twice a day) of lodenosine plus d4T/IDV to 3TC/d4T/IDV. Of 206 patients enrolled, 12 week data were available for 66, who had a mean baseline HIV RNA of 4.6 log (a little under 40,000 copies/ml) and a mean baseline CD4 count of 421 cells. At week 12, there was a 2.6 log decrease in HIV RNA for the lodenosine arms and a 2.3 log decrease in the 3TC arm, a difference that was not statistically significant. Fifty-nine percent of those taking 300 mg of lodenosine had HIV RNA <50 copies/ml, compared to 55% in the 200-mg arm, 50% in the 100-mg arm and 50% in the 3TC arm. CD4 cell counts increased by a mean of 53-117 cells in the lodenosine arms. The highest increase was seen on the lowest dose.

The adverse event rate was similar across all study groups. The most common were nausea, increased bilirubin and vomiting. Most of the toxicity was mild and not dose-related. Thirteen serious adverse events were reported, but at ICAAC, none were linked to lodenosine. Nevertheless, only a week later, US Bioscience announced that it was suspending the trial due to those serious adverse events and the death of a study participant.

Is it curtains for lodenosine? We are not sure. It is hard to say whether events that happen regularly in any trial spooked an inexperienced company or whether lodenosine has toxicity that had heretofore gone unnoticed. Some of the nervousness might stem from the fact that it is a fluorinated nucleoside analog like FIAU, which caused the sudden deaths of several patients in an investigational study for hepatitis. At the time, the deaths on FIAU came as a shock, but it has since become clear that there were many warning signs. One was that FIAU-treated woodchucks, which serve as an animal model for hepatitis B, were dying before the drug was ever put in humans.

But fluorination was not the cause of FIAU's morbidity, severe mitochondrial toxicity was. Since FIAU, the Food & Drug Administration requires evidence on a nucleoside analog's mitochondrial toxicity before it can go into humans. Lodenosine seemed very benign in preclinical trials, and there had been no reports of severe adverse events in the three years that the NCI had been studying the drug (which entered clinical trials as far back as 1996). In fact, the NCI has at least 12-month follow-up data on some patients.

At present, researchers are investigating what role, if any, lodenosine played in these adverse events. It is our hope that it played none and that this interruption in the development of lodenosine is will be brief. While it is important to take all necessary precautions, if lodenosine proves not to be the culprit behind these adverse events, its clinical development must resume.


Triangle Pharmaceuticals is concurrently developing a number of antivirals, led by emivirine (see Emivirine, above) and FTC, now called emtricitabine, a more potent and once a day version of 3TC (the only new data on FTC was in patients with Hepatitis B). Joining these now is dADP, a new guanosine (like abacavir) nucleoside analog. dADP is actually the prodrug of dioxolane guanosine (DXG), a compound with very potent activity against HIV and Hepatitis B virus (HBV). There were two posters on dADP/DXG at ICAAC.

The first presentation reported that DXG remains active against recombinant viruses and clinical isolates of HIV from people who have failed NRTI and NNRTI combination therapies.(11) The second poster further noted that DXG has a synergistic effect in combination with AZT, 3TC and nevirapine.(12)

Preclinical studies demonstrated that DXG had anti-HIV activity (EC50=30-290 nmol) at doses far below those toxic to cell cultures. Furthermore, no significant dose-limiting toxicity was observed in mice, rats and two species of monkeys administered doses up to 1200 mg/kg/day for one month. Of course, animals tend not to complain about pill burden, but the human equivalent of this dose for a 70 kg person (154 lbs.) would be 84 grams a day!

In vitro studies have noted that the development of a L74V mutation, which also occurs with ddI and abacavir, may lead to a four-fold shift in sensitivity to DXG, while a mutation at K65R (also seen with ddI, ddC and abacavir) can lead to an eight-fold shift in sensitivity. Given the drug's potency and wide therapeutic window, such shifts in sensitivity may not amount to much. However, a combination of mutations may lead to greater resistance and should probably be expected. Nevertheless, these mutations do not occur commonly in patients who take ddI or in the very infrequent patient who takes ddC. What happens in patients who take abacavir no one yet knows.

Triangle Pharmaceuticals is now conducting Phase I/II dose escalation studies of dAPD.


Biochem Pharma's dOTC is an NRTI that is similar to 3TC. Researchers assessed the pharmacokinetics, safety and antiretroviral activity in a double-blind, placebo-controlled phase I/II trial of dOTC monotherapy in HIV-positive patients.(13) Six groups of ten patients (eight given active drug, two placebo) were randomized into one of the following arms: 200, 300 or 400 mg of dOTC twice a day, or 400, 600 or 800 mg once a day. At baseline, mean HIV RNA was 4.35-4.93 log (approximately 22,400-85,000 copies/ml) and mean CD4 counts were 295-523 cells. The median change in HIV RNA at day 7 was -1.03 to -1.36 log copies/ml, and 63-100% of the patients had either a 1.0 log decrease in HIV RNA or <400 copies/ml. One grade IV adverse event, CPK elevation, was recorded, but it resolved itself during therapy. The most common adverse events were headache, lethargy and low-grade fever, and they were not dose-related.

While similar to 3TC, the M184V mutation appears to only decrease susceptibility to dOTC slightly (two- to threefold). Nevertheless, earlier studies have reported that some clinical isolates from 3TC-experienced patients are resistant to dOTC. What causes this resistance has yet to be worked out. It is could be that other common mutations, say to AZT or other nucleoside analogs, might render the drug impotent, as occurs with abacavir. Whatever the reason, it is important that the activity of dOTC be determined against virus that is resistant to multiple nucleoside analogs in order to see how useful this drug might be for salvage therapy.

Nucleotide Reverse Transcriptase Inhibitors (NtRTIs)


Gilead's adefovir diprivoxil (ADV), a nucleotide analog reverse transcriptase inhibitor (NtRTI), is presently under review by the FDA. Much of the data in support of adefovir's new drug application (NDA) were presented at ICAAC, only a few weeks before its FDA hearing. As we went to press, the FDA's Antiviral Advisory Committee voted not to recommend approval for adefovir at present, the first time an FDA advisory committee has ever said no to an anti-HIV drug's NDA. Will adefovir be relegated to the pharmaceutical dustbin? A misadventure on the road to tenofovir? More next month.

Tenofovir (Formerly PMPA)

With adefovir's fate in limbo, the success of tenofovir disoproxil fumarate (TDF) becomes even more crucial for its maker, Gilead Sciences. Fortunately for the company, not to mention the people awaiting new agents, in the most recent data presented in a late breaker at ICAAC, the drug has yet to evidence any of the problems that plagued its predecessor.

The data were drawn from Gilead 902, an ongoing 24 week, double-blind study in 189 treatment-experienced patients with detectable viral loads on antiretroviral therapy (LB 19). One of three doses of TDF (75, 150 or 300 mg) or a placebo were added to a stable regimen of four or fewer drugs that participants had taken for at least eight weeks. Baseline HIV RNA was approximately 3.7 log (just over 5,000 copies/ml) and baseline CD4 count was 376 cells. Participants were heavily pretreated with a mean antiretroviral experience of 55 months. Genotypic analysis revealed evidence of resistance to NRTIs in 94% of the subjects. Sixty-three percent had isolates with decreased susceptibility to one or more protease inhibitors.

All three treatment arms showed significant HIV RNA reductions when compared to the placebo arm. At week 24, the 300 mg arm achieved a mean reduction of 0.75 log copies/ml in viral load, compared to a 0.40 and 0.45 log drop in the 150 mg and 75 mg arms, respectively.

Safety data were available up to 48 weeks. Serious adverse events were reported in four patients (7%) in the 300 mg arm and for three (11%) on placebo. None of the kidney toxicity (proximal renal tubular dysfunction) seen with adefovir treatment was observed, nor did any patients experience elevations in serum creatinine of 0.5 mg/dl over baseline (a warning sign for severe nephrotoxicity seen on adefovir).

While a 0.75 log drop in viral load in such heavily treatment-experienced patients is impressive, tenofovir may have yet unrealized potential. Given the nephrotoxicity observed on the other nucleotides that Gilead has worked on, it is easy to see why the company may want to proceed more cautiously developing this drug, but the maximum tolerated dose of tenofovir has never been established, and higher doses may be even more active. In studies in macaques with SIV, PMPA injections achieved viral load reductions of greater than 2.0 log in under two weeks.(14)

Perhaps Gilead has compelling preclinical toxicology data suggesting that nephrotoxicity would emerge at higher doses. If not, we need to be certain that we are using the optimal doses in patients who are failing their third or fourth regimens because incompletely suppressing viral load will lead to resistance and eventual treatment failure. Remember that long before 48 weeks (after 24, in fact), there was clear evidence of kidney toxicity on adefovir. Here there is none. Higher doses should be evaluated for patients failing currently available therapy, particularly in light of the fact that a small number can now get the drug on expanded access (see Tenofovir Compassionate Use Program in this issue).

Fusion Inhibitors

T-20 (Pentafuside) and T-1249

T-20 is the first fusion inhibitor, an entirely new class of antiretrovirals, being developed by Trimeris Pharmaceuticals in partnership with Hoffmann-La Roche. The compound mimics one of two coil-like segments of HIV-1's gp41 envelope protein that play a critical role early in the process of infection. After the virus latches onto the CD4 receptor, these two coils become exposed and twist around each other, pulling the gp41 protein taut and erect so that it can penetrate the cell membrane. Acting as a decoy, T-20 interferes with this intertwining, and the flaccid gp41 protein is unable to inject the viral core into the target cell. The synthesized drug is a large protein, composed of 36 amino acids. While potent at low doses, it must be administered as a subcutaneous injection.

With such a distinct mechanism of action, the drug is active against virus that has developed high level resistance to each of the other classes of antiretrovirals currently available. Thus its development is a life and death issue for people who need salvage agents. To the credit of Trimeris, the company has made salvage therapy the focus of the drug's development; but T-20 is costly to make and until the recent collaboration with Roche, resources for the young company have been limited. Still, Trimeris has been able to mount a few studies.

Sixteen-week results from the most recent of these studies were presented during a late breaker session at ICAAC.(15) This is an ongoing single-arm, phase II study of T-20 (pentafuside), self-administered as a subcutaneous injection. The study enrolled 71 HIV-positive adults who were heavily pretreated: patients had a median of 11 previous antiretrovirals, 93% had exposure to the three available drug classes, 93% showed genotypic evidence of protease inhibitor resistance and 87% had mutations related to reverse transcriptase inhibitors. The median number of NRTI and protease inhibitor relevant mutations was four and five per patient, respectively. To top it off, all participants had previously been briefly exposed to T-20 therapy as well, in earlier phase I/II studies. In fact, access to T-20 in this study was in a sense a reward for participation in the shorter trials.

At baseline, the median HIV RNA was 4.9 log (about 79,500 copies/ml) and the median CD4 count was 70 cells. Subjects took 50 mg of T-20 twice a day in addition to an individualized regimen consisting of a median of four oral drugs chosen on the basis of each patient's history and genotypic results. At week 16, data were available for 55 of 71 patients. In an as-treated analysis, 60% (33 of 55) had a sustained decrease in viral load of >1.0 log or were below 400 copies/ml, 36% (20 of 55) were below 400 copies/ml and 20% were below 50 copies/ml. In an intent-to-treat analysis, 55% had a 1.0 log reduction in viral load, and 33% were below 400 copies/ml. The average decrease in viral load was greater than 90% over 16 weeks. Most adverse events were mild to moderate, and they included irritation at the site of injection (67%), fever and headache. This study will continue to 48 weeks. A phase III trial is expected to begin in 2000.

Without a control arm, it is impossible to measure the relative contribution of T-20 to the antiviral suppression in these patients. It is possible, though, that T-20 furnished the lion's share of antiviral effect, bearing in mind the extent of these patients' pretreatment. But are these responses sufficient? While 30% to 40% below 400 copies/ml might be what we have come to expect from "mega-HAART" regimens in salvage populations, this regimen included a very potent drug from a new class of agents. As prior therapy included T-20, either given as a monotherapy or added to stable failing HAART regimens, it is reasonable to assume that at least some patients had developed some resistance to T-20. Further study of this cohort should demonstrate the frequency of resistance.

Yet, much of the resistance might have been preventable with a different T-20 dose. In the earlier dose-ranging studies of T-20, there was a clear dose-response to the drug, which peaked at the highest dose tested, 100 mg bid -- twice the dose currently under evaluation. Why did Trimeris not proceed with the 100 mg bid dosage? For one, this is a costly agent to synthesize and will likely be quite expensive when it is marketed. Consider the cost of other recombinant proteins such as IL-2 or EPO, which cost tens of thousands per year. Aside from the issue of cost, there is a limit to how many injections a person can take. But, at the risk of sounding monotonous, there is a danger that the doses being used are suboptimal. In fact, even 100 mg bid may be inadequate. There is no way of telling since that was the highest dose ever tested.

Given Roche's experience with the suboptimal Invirase dosage and formulation of saquinavir, one would expect them to think twice before moving forward with another antiretroviral without determining the maximum tolerable dose, but this is exactly what the companies are doing. As it is presently formulated, each shot can only deliver 50 mg of T-20, so a 100 mg bid dose would require four shots a day. This may indeed be too much for a patient to tolerate. Still, no study has addressed the question. And if it is the cost of therapy that is indeed the dose-limiting factor, how much is cost-prohibitive? Is $15,000 too expensive? $30,000? Just how expensive are we talking? At present, neither company is anxious to say.

Regardless of the cost, according to Roche's International HIV Medical Director, Dr. Neil Buss, a higher "induction" dose trial is being considered. The question is, when? Right now, the plan is to wait until post-marketing.

Trimeris and Roche may be betting that T-20 will not be the last "shot" for failing patients in this study. The companies are developing a second fusion inhibitor, T-1249, that is active against T-20 resistant virus in vitro. A poster in Lisbon described preclinical studies with this slightly larger peptide.(16) Preliminary dosing studies in monkeys demonstrate a longer half-life for this compound, suggesting that a once-a-day subcutaneous regimen might be possible. A phase I dose-ranging study is now enrolling.


  1. Abstract 657.
  2. Abstract 916.
  3. Abstract 917.
  4. Abstract 502.
  5. Abstract 201.
  6. Abstract 1173.
  7. Abstract 1172.
  8. Abstract 332.
  9. Abstract 504.
  10. Abstracts 925 and 926.
  11. Abstract 924.
  12. Abstract 932.
  13. Abstract 503.
  14. Bischofberger N et al. 9th ICAR, Japan 30:A42, 1996.
  15. Abstract LB-18.
  16. Abstract 319.

Back to the GMHC Treatment Issues September/October 1999 contents page.