Somewhere over the Rainbow . . .by Dave Gilden
. . . lies the cure for AIDS. But you wouldn't find it at the fourth annual Conference on Retroviruses and Opportunistic Infections (held in Washington, D.C. on January 22-26). Although not heralding any new breakthroughs, the Conference reflected the gradual accretion of promising developments leading toward fulfilling the now years-long hope of making HIV infection a
"chronic manageable disease." Though the Conference did not announce that a cure was at hand, there was some good luck to celebrate at long last -- recent discoveries show that some potential problems are not as bad as they might be.
The Reservoirs Drain -- Slowly
The most opportune time for antiviral treatment continues to seem to be shortly after transmission occurs (see Treatment Issues, December, 1995, pages 1, 3-4). This is the period when the immune system is relatively intact. At this point, too, the virus tends to be genetically the most homogeneous -- there are few mutations that would allow it to escape either drug therapy or the immune system. Antiviral therapy can work with the immune response to maximum effect during primary infection, with the immune system mopping up any HIV that survives the drugs.
The most publicized trial (thanks to Time magazine) of therapy during acute HIV is taking place at the Aaron Diamond Center in New York City and the University of Geneva in Switzerland. In presentations at the conference Drs. David Ho, Luc Perrin and Marty Markowitz (abstracts S1, 238 and LB8, respectively) reported on the latest observations from this group which now numbers 36 volunteers on AZT/3TC plus either indinavir or ritonavir. All volunteers so far who have been on therapy for more than five months have seen their viral loads go down to undetectable levels (less than 100, or sometimes 25, HIV RNA copies/ml). These virus levels have not as yet rebounded (for as long as 15 months now), with one exception (see below). Anti-HIV immune responses, as measured by antibody and cytotoxic lymphocyte levels also are subsiding by one year, a further indication of HIV's clearance from the body.
A major worry has been the enormous amounts of HIV concentrated in the lymph nodes and other lymphoid tissue. Nearly all of this virus are free virions trapped on the surface of the follicular dendritic cells (FDCs). It is unknown how long the HIV particles remain viable on the FDCs or what is their potential for infecting new cells. There are also many chronically infected cells, which slowly produce new virus over a period of weeks to years, as well as latently infected cells, whose HIV conceivably could activate itself anytime during the cells' lifetimes.
The Aaron Diamond Center study is checking gut-associated lymphoid tissue from its acute infection trial participants and finding no sign of active HIV replication there. HIV DNA does persist within some cells, though, and in three of five individuals, HIV virions (or at least, full-length HIV RNA-based gene sets) were detected by PCR. Given the lack of detectable viral replication, Dr. Ho speculated that this HIV was virus trapped on the follicular dendritic cell. No culturable virus was present in any of the specimens.
The extent that potent therapy clears HIV from lymph tissue, including the FDC-associated viral reservoir, was further described in two other major reports at the Conference. One (abstract LB9) examined tonsil biopsies taken from participants in a Dutch study of the AZT/3TC/ritonavir combination. In ten volunteers treated for six months, the HIV population attached to the follicular dendritic cells fell 2,500-fold while the number of HIV-infected cells fell 200-fold. But even at six months, there still was residual HIV on the FDCs (about 1,000 HIV RNA copies/gram) in six of the volunteers; in five, chronically infected cells were still producing small amounts of HIV.
A second study (abstract LB10), which included five people followed for nearly a year in a trial of AZT/3TC/indinavir, had similar results. Detectable HIV remained in the lymph nodes of all five, including the two who had undetectable blood plasma HIV even by the most sensitive PCR assay (detectable limit of 20 copies/ml). The lymph node HIV, whose level correlated with that found in the blood, was able to replicate, but had developed no added drug resistance in the course of therapy. Two other participants who had interrupted treatment for short periods saw the HIV levels in their lymph nodes rebound sharply, with the accompanying appearance of drug-resistant mutant virus.
Ferreting out the Protected Sanctuaries
Aside from holding out in the lymph nodes due to sheer force of numbers, it has been speculated that HIV may be able to seek refuge in areas that drugs cannot penetrate. Among these are the brain and testes. In the Aaron Diamond acute infection study, white blood cells in the semen were negative for HIV activity by one year of treatment. Although latent HIV DNA was still detected in some cells, these HIV gene sets probably are defective and incapable of replicating. Seminal fluid in this study has yet to be analyzed, but in two other investigations (abstracts 202 and 726), free HIV in seminal plasma did indeed decrease dramatically in people put on highly active antiretroviral therapy (HAART).
(A strong word of caution: Measurement of seminal HIV levels is still in its infancy because seminal fluid interferes with the standard PCR test for viral load. A newer technique, known as NASBA ("Nucleic Acid Sequence-Based Amplification") accurately measures free HIV levels in semen, but only down to 1,000 copies/ml. Everyone warns that the infectivity of HIV-positive men on HAART has yet to be determined and that abandonment of safe sex is not justified. The same goes for HIV-positive women even though, according to preliminary results from one study (abstract 727), vaginal and cervical HIV levels may also respond to treatment in parallel with the levels in blood plasma.)
As far as the central nervous system goes, the Aaron Diamond study plans to check HIV cerebral spinal fluid (via lumbar punctures) as a kind of surrogate for the brain, which obviously would be very difficult to extract samples from. While awaiting these results, a study of AZT/3TC versus d4T/3TC has monitored CSF viral load changes (abstract LB7). In this study CSF viral loads went from detectable to undetectable in all 31 participants (initial plasma viral load averaging about 50,000) after 12 weeks of either regimen. In contrast, one poster at the Conference described a case in which a patient taking AZT/3TC/indinavir had an undetectable plasma viral load but a very high level in the CSF (abstract 233).
Duration of Treatment: Years or Decades?
In one report, Katherine Luzuriaga, M.D., of the University of Massachusetts described clearance of HIV in a set of fraternal twin infants who became infected with HIV at birth (abstract 725, which also described four other babies who did not fare as well). Both twins were treated with AZT/ddI/nevirapine, and their plasma viral loads fell to below 20 copies/ml (again, the limit of the ultrasensitive PCR). Unfortunately, one of the twins rebounded at 16 months -- after 11 months of undetectable HIV. He recovered after switching to AZT/ddI/ritonavir, but the question remains, how long must anti-HIV treatment persist before you can feel secure, let alone terminate therapy?
Dr. Ho now estimates that treatment must be continued for at least 2.5 years, with HIV completely suppressed the whole time, before patients and doctors can even think about stopping HIV regimens. (This is the point at which he will ask his trial participants to consider such a move.) His estimate is a prediction based on the decay of HIV levels so far. That decay is a two-phased affair that starts with a rapid drop-off for the first few weeks, apparently due to the rapid death of most HIV-producing cells. Few newly infected cells replace them because the potent drug therapy protects most cells from HIV. The quick dive in HIV is replaced by a gradual tapering off, as the virus trapped on the FDCs, latently infected cells and infected cells chronically producing low amounts of virus gradually peter out as sources of free HIV.
The latest amendments to Dr. Ho's estimate of HIV clearance, scientific as it may seem, is based on a mere eight treatment-naive trial participants receiving AZT/3TC/nelfinavir. Even within these eight there are significant variations, and looking at a larger number of patients might uncover some people whose HIV tapers off at an exceptionally slow rate, adding years to the estimate of how long treatment must last to be sure of eliminating HIV.
But to Keith Henry, M.D., a noted AIDS researcher who directs the AIDS clinic at St. Paul, Minnesota's public hospital, "These results are interesting, but they reflect the conditions in clinical trials, not those of the real world. People have complex treatment histories and don't always take their drugs on schedule. We may have to treat patients for decades, or forever, to keep HIV under control."
Drug Resistance and Outcome
One of the strongest determinants of eventual treatment outcome is drug resistance and its associated HIV mutations, which is influenced by a person's treatment history. (See in particular abstract 584 on the deleterious effect of AZT resistance on outcomes in ACTG trial 175.) The phenomenon of cross-resistance means that the resistance HIV develops to one compound may make it less sensitive to other antiviral compounds, even those it has yet to be exposed to.
One study (abstract 9) looked at the prevalence of drug resistance-conferring mutations in 99 Iowans whose HIV isolates were gathered from 1993 to 1996. Of 32 patients with a history of reverse transcriptase inhibitor therapy, 72% had resistance mutations to a range of nucleoside analogs (particularly AZT) and NNRTIs such as nevirapine. Only five of the 65 individuals without prior treatment had such resistance mutations (four out of five involved AZT resistance). Especially noteworthy was the fact that 26 of the 99 had one or more mutations that are known to facilitate resistance to protease inhibitors even though no one had taken any drug of this class. And of course, therapy, which was carried out without any knowledge of the genetic resistance profile of patients' personal HIV, frequently included drugs that might not work, or at least not work for long, in the individuals taking them.
The Retrovirus Conference had mixed news in regards to which drugs might succeed against HIV that has evolved resistance to particular antiviral agents. Abbott Laboratories is readying for human testing a protease inhibitor that, preliminary tests indicate, suffers comparatively little from cross-resistance with the current inhibitors (see page 8). Also, contrary to fears in some quarters, treatment with 3TC and the resulting resistance (generated by a mutation at reverse transcriptase gene 184) does not seem to reduce the activity of ddI or ddC (abstract 588). Conversely, there was a warning that multiple resistance mutations arising from exposure to other protease inhibitors can reduce the power of Agouron Pharmaceuticals' nelfinavir. This new protease inhibitor, now being considered by the FDA for marketing approval, had been touted as immune from such cross-resistance. Similarly, the experimental and very potent nucleoside analog 1592U89 from Glaxo turns out to share resistance mutations with ddI, although the significance of such cross-resistance in humans has yet to be established. High-level resistance to 192U89 has not yet been observed.
Reduction in Hospitalizations, Mortality
Whatever the long-term complexities, the immediate results of the new combination therapies seem good: several reports at the Retrovirus Conference detailed a reduction of hospitalizations or deaths occurring over the past year. For example at New York's St. Vincent's Hospital, which contains the largest AIDS clinic in the city (and whose director, Gabriel Torres, M.D., is Treatment Issues' medical consultant) AIDS hospitalizations are down 10% from a year ago (abstract 264). For New York as a whole, the daily number of deaths per day has fallen by half in the past year, from an average of 20.9 to 10.1 (abstract 376). French doctors also found a sharp decline in AIDS-defining illnesses in four medical centers that rapidly adopted HAART regimens as compared to two medical centers that have prescribed such treatment less frequently. The first four centers were estimated to be saving $650,000 monthly as a result (abstract LB12).
These declines in illness may be related to earlier diagnosis of HIV and better overall medical care for opportunistic infections as well as the introduction of more potent HIV medications. Nonetheless, they are in line with clinical trials that have documented similar OI reductions in people with advanced disease when their current antiviral regimen is strengthened. In particular, the one-year CAESAR trial (abstract 367), which enrolled people with CD4 counts of 250 or less, added 3TC to volunteers' prior treatment (AZT, AZT/ddI or AZT/ddC) and measured a 50% reduction in OIs and death during a year's follow-up. Likewise, a meta-analysis of 22 trials enrolling 13,280 previously treated volunteers (abstract 562) found that mortality rates were halved by therapies that included 3TC or a protease inhibitor, at least for the first year. ddI, ddC or d4T seemed to produce less benefit.
Is the cup half-full or half-empty? It all depends on which side of the 50% line you find yourself. In any case, the cup is leaky. For advanced patients, who are in immediate danger of serious opportunistic infections, long-term disease stability will depend on the immune system's unexplored ability to reconstruct itself once HIV levels are kept down. There were anecdotal reports at the Conference of individual resolution of some conditions after commencing HAART. Among these conditions were chronic cryptosporidiosis, progressive multifocal leukoencephalopathy (PML), Kaposi's sarcoma and wasting syndrome, but these diseases may depend on HIV or HIV-induced inflammation for stimulus.
The reduction in hospitalizations and OIs could well be a temporary blip. At this point the only prudent conclusion one can draw is that therapy against HIV has to continue for years with the present range of drugs. Any therapy you commence may lose effectiveness at some point, even if it appeared highly potent and tolerable, on average, in clinical trials. So if you start on drugs A, B and C, -- you should plan which drugs D, E and F you can switch to -- if at all possible.
Those with long, complex treatment histories may have used up their present options, and those with AIDS do not have the luxury of planning ahead. Both groups already depend on the development of new antiviral (and anti-OI) agents to keep them stable. For some of the major therapy reports at the Retrovirus Conference, see the following pages.
Commentary: A Retro Retrovirus Conferenceby Dave Gilden
The Fourth Retrovirus Conference was marked by authoritarianism and unpleasantness. A multitude of guards and rules greeted Conference goers. Seventeen activists who arrived unregistered were not allowed in; pharmaceutical marketing representatives who tried to give up their registrations in favor of community physicians found that that was not allowed. Any complaints were met by the argument that the restrictions were necessary to preserve the Conference's supposed scholarly and intimate atmosphere. To this end, the conference first and foremost limited attendance to 2,100 people "actively working in the field of basic science and clinical investigation." All in all, thousands who would have attended an open conference were disinvited from the major U.S. AIDS conference of the year.
Hardest hit by the exclusions was the AIDS community -- local healthcare providers, activists and people with HIV alike. The exceptions were 60 people who were allowed to register as "community press" and 20 people "affected by AIDS" given full scholarships (out of 150 who applied). Both categories were expected to spread the conference news to their constituencies. By obstructing access, the conference organizers are disrupting a long-standing natural alliance between the AIDS and scientific communities. The result of that alliance has been better funded and more rational AIDS research programs.
It was recognized years ago, after considerable struggle, that people with HIV have an intrinsic right to attend scientific conferences, whether they communicate the news to anyone else or not. The information they glean will help save their lives as they confront a health care system that is inept in the best of times. Scientists who don't feel an obligation to help those at the center of their research are out of touch with human concerns.
The great irony is that there was little discussion at the Conference. The oral presentations mainly consisted of review lectures conducted in large, overcrowded halls. And, actually, these reviews would have been very valuable to the practicing doctors, nurses and patients in little evidence at the Conference. Most of the real news was stuffed into two one-day poster sessions, in which hundreds of displays stood side-by-side in a large basement room. You can only absorb so many posters at a time: glazed eyes were epidemic.
In the end, the pharmaceutical companies controlled the wider public's perception of the Conference. Their tailored presentations went largely unchallenged -- community people, motivated as nobody else, frequently ask the most incisive questions at conferences. And the news media, starved for direct access, was forced to rely more than ever on company press releases.
Did any of this really bother prominent researcher Douglas Richman, chair of the Conference's program committee? We doubt it. Dr. Richman is frequently a star at much more commercialized affairs, most recently the Conference on Drug Therapy in HIV Infection last November in the U.K., with its flashy trade show and array of company-sponsored symposia.
Wherever he appears, Dr. Richman always argues forcefully for early, aggressive treatment of HIV. Can people with HIV trust such experts? Community members need to be at these scientific meetings -- in force -- to make their own voices heard.
Protease Inhibitors: Resistance, Resistance, Resistanceby Dave Gilden, with John Falkenberg and Gabriel Torres, M.D.
Nothing at the Retrovirus Conference changed the fact that all the HIV drugs now available or in advanced development operate by interfering with either HIV's reverse transcriptase or protease enzymes. Protease inhibitors are potent drugs in their own right, and combining them with reverse transcription inhibitors makes for a synergistic "one-two punch" against the virus. In examining the various combinations described below, bear in mind that it is very difficult to obtain greater than a 99% (2 log) average reduction in HIV viral load over the first month or two, or more than 99.7% (2.5 logs) over the first six months. The reasons for this lie in both the drugs' mode of action and the dynamics of HIV infection. All the present drugs stop the infection of new cells. None can kill or inhibit infected cells in any way, nor do they affect the HIV-driven metabolic processes within them. Since 99% of the new virus comes from actively producing cells with a short lifespan on average, viral load can drop quickly only by that percentage. Further, limits in the sensitivity of viral load assays (commonly considered to be 400 to 500, sometimes 200, HIV RNA copies/ml of plasma -- or 20 to 50 copies/ml for the new ultrasensitive tests) reduce the observable viral load reductions. It is impossible with these tests to tell directly how close to zero a person's viral load has fallen.
Similarly, the dynamics of CD4 cell proliferation tend to limit the average initial rises in CD4 counts to 100 to 150 cells/mm3 of blood. The differences between combinations are most observable in the long-term viral load endpoints and the stability of those endpoints.
The present HIV drugs are plagued by resistance. These drugs are very specific in their action, which helps to protect the body from side effects but also makes it possible for HIV to slightly alter its enzyme structures and render the drugs ineffective. Combinations of drugs require HIV to mutate at more points in its genes before it can resist all the drugs and rebound to previous viral loads. Besides a drug regimen's immediate impact on viral load, the ease with which resistance can develop is a crucial factor in determining durability of response. Reports at the conference proposed various ways to use protease inhibitor-containing combinations to build a sufficiently high "genetic barrier" against HIV breakthrough. Creating this barrier is easier in people with no prior treatment. Those with advanced disease and/or long treatment history present a more difficult problem. Their circumstances, including high viral loads, mean that the HIV within them likely has already spawned many mutants at least partially resistant to various drugs. Plus, large residual HIV levels after initiating therapy are a breeding ground for more resistance.
HAART Therapy in Advanced Disease
Nonetheless, potent HAART (for highly active antiretroviral therapy) combination therapies can be of benefit even to people with advanced AIDS and with extensive past exposure to anti-HIV therapy. This was confirmed by two studies at the Retrovirus Conference. In one (Merck protocol 039 -- abstract LB7), 320 volunteers with CD4 cell counts of less than 50 received either AZT/3TC/indinavir, indinavir alone or just AZT/3TC. All participants had at least six months of AZT but had never before taken a protease inhibitor or 3TC. The results are summarized in Table 1.
The 65 CD4 count rise at 24 weeks from indinavir monotherapy is surprising given the regimen's poor antiviral performance. The reason may be that for the first two weeks, the viral load drop from indinavir alone equaled the reduction achieved with the triple therapy, about 1.3 logs (95.0%). The median viral load in the triple therapy arm kept gradually falling at least until week 16, when it essentially stabilized. In those on indinavir alone, the median viral load rebounded after the second week, and by week 16 was only about 0.1 log (20.7%) below baseline. (The AZT/3TC arm's median viral load was stable from week 2 to week 24.) As for the CD4 counts, the two indinavir containing arms rose in parallel through the twelfth week after which the monotherapy arm declined slightly while those on triple therapy continued their rise. Other indinavir trials, too, have seen a long lag between viral load rebound and fall of CD4 count back to baseline in those on indinavir monotherapy. That process can take over a year. (Trial 039 is now continuing beyond 24 weeks with everyone on open label triple combination therapy.)
Another trial looking at people in advanced disease was a pilot study conducted in Vancouver, British Colombia (abstract 234). This trial followed 21 heavily pretreated patients who were failing or intolerant to their nucleoside analog therapy and seemed to have run out of options in this drug class. The 21, who were all naive to indinavir and nevirapine, received those two drugs in combination with 3TC. Baseline CD4 count averaged only 28 (everyone was under 50), and average baseline viral load was 135,000 HIV RNA copies/ml of plasma.
This combination of a potent protease inhibitor, a rather well tolerated non-nucleoside reverse transcriptase inhibitor and a relatively benign nucleoside analog had a surprisingly positive effect as salvage therapy. In the 12 participants who had reached 20 weeks of therapy, CD4 counts were up 100 above baseline and continuing to rise, and viral load had fallen more than 1,000-fold (99.9%). However, so far only two of the 12 had undetectable viral loads using the new ultrasensitive PCR assay. Seven (58%) were undetectable by the standard PCR test that goes down only to 500 copies/ml. Three of the participants have dropped out, two due to nausea or vomiting and one because of skin rash.
Other indinavir news included the latest results from protocol 035, a trial of AZT/3TC/indinavir versus AZT/3TC versus indinavir alone in 97 adults with at least six months of prior AZT therapy. Follow-up is now out to 68 weeks in two-thirds of the group. This trial included a population that was a little less advanced than the two above: a baseline median CD4 count of 142 (one eligibility requirement was a CD4 count between 50 and 400) and a baseline median viral load of 43,000. Participants also had to have had at least six months of prior AZT treatment. The actual average past exposure to AZT was 30 months, with 80% of the participants carrying AZT-resistant HIV (by genetic analysis). The group was therefore somewhat representative of a real population with advanced diseased, in whom HIV has developed resistance to AZT during past treatment.
Trial participants assigned to receive AZT/3TC/indinavir from the beginning have exhibited stable viral load reductions, a median of 2 logs (99%) or greater, out to a year or a year and a quarter so far. Of those on treatment for 68 weeks, 18 of 21 (86%) still have plasma viral loads below the test limit of 500 HIV RNA copies/ml and 10 of 14 (71%) checked with an ultrasensitive assay are also below 50 copies/ml.
Nelfinavir: What about Cross-Resistance?
Agouron Pharmaceuticals' protease inhibitor nelfinavir (brand name Viracept) is now up for FDA licensing review. Although rumors of its imminent approval are rampant, the company's "conservative" prediction is FDA approval by June of this year. Meanwhile, Agouron researchers presented a number of reports at the Retrovirus conference, including information on the compound's efficacy when combined with AZT/3TC or d4T. In contrast to the indinavir studies above, the AZT/3TC/nelfinavir trial (Agouron protocol 511) included only people with less than one month prior AZT and no other past treatment. Their CD4 counts could be any number, and in many cases were below ten. The d4T trial (Agouron protocol 506) included both treatment-naive and -experienced people (90 naive and 218 experienced with an average of 27 months prior therapy). The trial required that their CD4 counts exceed 50. Table 2 summarizes the data from these two trials.
It is risky to compare the studies' data, especially given the two populations' different treatment histories. Still, the triple therapy seems superior to the double d4T/nelfinavir combination, both in terms of viral load and CD4 cell response and in the stability of that response: viral load in the d4T/nelfinavir cohorts reached a peak decline of about 1.5 logs (96.8%) and then started gradually rebounding. The viral load curves of those on AZT/3TC/nelfinavir were flat to slightly downward from week eight onward.
Most of the trial results show no significant difference in response to the 750 mg and 500 mg doses, and Agouron's application for FDA approval stated that the standard nelfinavir dose should be 500 mg thrice daily. Notice, though, that in the AZT/3TC/nelfinavir trial, 81% of those receiving 750 mg had undetectable viral loads, whereas only 65% of those receiving 500 mg reached this level. As a result of this statistically significant difference, some HIV specialists are disputing Agouron's proposed dose, asking that the 750 mg dose be specified instead since there is a sign that it better suppresses HIV. This higher dose would provide more insurance against the emergence of drug resistance and HIV rebound. Countering this argument is nelfinavir's tendency to cause diarrhea. In the triple combination trial, those on 750 mg nelfinavir had a 54% greater incidence of developing diarrhea (12 of 96 at the lower dose versus 19 of 99 volunteers at the higher).
Aside from controversies over the dose, there has been considerable discussion about nelfinavir's resistance profile. Agouron has presented data to indicate that if nelfinavir is the first protease inhibitor taken, it can profitably be succeeded by others if nelfinavir resistance emerges. There have also been claims that the converse is true -- that people who fail on other protease inhibitors due to drug resistance can still get benefit from nelfinavir. Should nelfinavir prove to have a unique resistance profile that obviates cross-resistance, it would be a very valuable addition to the anti- HIV armamentarium indeed.
Such hopes are premature at best, as the available data is scant and ambiguous. Cross-resistance between nelfinavir and other protease inhibitors first came up a year ago, during the FDA hearings on approving Abbott Laboratories' ritonavir. During its presentation the company presented a graph indicating that HIV that had mutated to protect itself from ritonavir could also resist nelfinavir to a certain extent. Agouron countered last spring with its own lab culture studies purportedly indicating that a unique, single mutation at codon 30 on the protease gene is HIV's response to exposure to nelfinavir. (This mutation switches amino acid 30 on the HIV protease enzyme so that nelfinavir has difficulty binding to the molecule.) Codon 30 mutant HIV is still fully sensitive to other protease inhibitors. However, the same nelfinavir test tube experiments found that when HIV is cultured a little further in the presence of increasing concentrations of nelfinavir, a second mutation appears that grants HIV further immunity to nelfinavir. This is the same codon 84 mutation that reduces HIV's sensitivity to ritonavir, although it is not seen that often in HIV isolated from humans taking ritonavir. (There are other, more common sets of mutations not involving codon 84 that also make HIV unsusceptable to ritonavir).
At the Retrovirus Conference, Amy Patick, Ph.D., Agouron's drug resistance expert, presented the results of several studies that she has conducted (abstract 10). The HIV isolates involved all came from participants in either the combination studies described above or earlier monotherapy studies. In the nelfinavir monotherapy trials, 36 of 54 (56%) participants analyzed had HIV with the protease codon 30 mutation by week 16. By contrast, only three of 49 (6%) of the people on AZT/3TC/nelfinavir had this mutant HIV. Sixteen such individuals who were maintained on nelfinavir for another 28 weeks (44 weeks on nelfinavir in all) had no detectable codon 84 mutation, but other, supplementary mutations did tend to accumulate. Six viral clones isolated from five people with the 30 mutation were five to 93-fold less susceptible to nelfinavir than wild type virus (the range in resistance probably arises from the supplementary mutations). All of these isolates were fully susceptible to other protease inhibitors.
But, the question remains, if someone's HIV is resistant to nelfinavir, will their HIV more rapidly become resistant to other protease inhibitors to which that person might be switched? One specific cause of concern: the supplementary mutations at codons 36, 46 and 71 are frequently seen from exposure to other protease inhibitors as well as to nelfinavir. Their presence could facilitate the development of resistance to protease inhibitors that follow nelfinavir.
And what about nelfinavir as a second-line protease inhibitor? People on one of the currently approved compounds would like to know that they can switch to nelfinavir if they start failing their present therapy. Dr. Patick presented an analysis of 23 HIV isolates obtained from people failing on indinavir, ritonavir or saquinavir. Fourteen (61%) of these isolates were still susceptible to nelfinavir in culture. The catch is that the protease inhibitor resistant HIV that had only one identified mutation was still sensitive to nelfinavir (with one mild eight-fold exception), but the virus found with two mutations (most at codon 90 plus either 82 or 84) was seven- to 75-fold resistant. Indeed, multiple mutations are more common than single ones after prolonged protease inhibitor therapy.
These data led Dr. Patick to conclude (however tentatively), "Nelfinavir failures are treatable by other protease inhibitors. Failures on other protease inhibitors with one mutation are treatable with nelfinavir, but more than one mutation is a problem -- so there is a need to switch early."
Emilio Emini, M.D., who led the effort to develop indinavir at Merck would exercise greater caution about nelfinavir as either first- or second-line therapy: "You cannot say that if you start on protease inhibitor A and become resistant, the virus will still be susceptible to B, C or D," he declared. "You will get varying degrees of cross-resistance up to 100%. It all depends on the mutations generated. You don't get just one mutation. You alter the genetic background in ways that you can't see [in lab tests] because the mutations are present at too low a level. And it is from this background of minor mutations that resistance to new drugs appears."
Nelfinavir has shown itself to be a respectable protease inhibitor, but it is odd that at this late stage, with the drug before the FDA, the optimum dose is still controversial. Further, it is premature to judge how nelfinavir will be affected by cross-resistance. Certainly it is intemperate to draw conclusions on optimum protease inhibitor sequencing on the basis of 23 isolates. Further information on nelfinavir's place as follow-on protease inhibitor in the case of treatment failure may be available in the near future: Agouron is tracking the viral load response in a group of protease inhibitor treatment failures now enrolled in the nelfinavir expanded access programs.
And lest anyone think that there is no money to made from AIDS drugs: Agouron's stock price has skyrocketed from the high twenties to the mid-nineties in the past year, almost entirely on the perceived strength of nelfinavir.
Abbott Labs Tries Again
Abbott Laboratories in the meantime surprised the conference by unveiling its new protease inhibitor ABT-378 with much fanfare (abstracts 14 and 206-213). This compound has remained hidden from public view for some time (see Treatment Issues, May, 1996, page 5). It is highly active against HIV, about three or four times more so than Abbott's original protease inhibitor, ritonavir. The structure of ABT-378 also circumvents some of the drug-resistance mutations that afflict other protease inhibitors. For example, its activity is hardly impaired by a protease gene mutation at codon 82 that confers considerable resistance to both ritonavir and indinavir.
In general, though, ABT-378 seems to overwhelm drug resistance by its sheer potency. In combination with small amounts of ritonavir, which blocks the relatively rapid breakdown of ABT-378 by the liver, it is easy to obtain drug levels that are 50 times the levels that are effective against wild type (non-drug resistant) HIV and ten to 20 times higher than HIV with a single mutation like that at codon 82. It is nevertheless possible to generate HIV that is not inhibited by ABT-378 and is considerably less sensitive to other protease inhibitors as well. One mutant created in the laboratory by exposing HIV to increasingly high concentrations of ABT-378 had five mutations and was 338 times less sensitive to ABT-378. It also was about 20 times less sensitive to ritonavir and indinavir. Four of the five mutations also appear during analogous culture tests of Glaxo Wellcome's experimental protease inhibitor 141W94.
ABT-378 is an interesting new compound, but it does not represent a radical break with the earlier protease inhibitors. Its structure, mechanism of action and resistance mutations are comparable to those of other members of its class. A big remaining question is how feasible it is to achieve levels within humans that overwhelm any resistance on HIV's part. There are as yet no human trials to evaluate what drug levels can be attained in reality, either with ABT-378 alone or in combination with ritonavir. No information as to the new compound's toxicity is available, either. The data Abbott presented at the Retrovirus Conference at best come from rat studies, and much of the figures are derived from test-tube cell cultures or even straightforward chemical analyses in cell-free solutions.
Combining Protease Inhibitors
Although human trials of ABT-378 plus ritonavir have not yet occurred, a trial of saquinavir plus ritonavir has been continuing since last spring (unpublished presentation in session 84 by John Mellors, M.D.). The reason for combining ritonavir with saquinavir is similar to combining ritonavir with ABT-378. In both cases, ritonavir, which inhibits one of the enzyme pathways in the liver, greatly increases and stabilizes blood levels of a compound that otherwise would be rapidly metabolized. The quantities required for saquinavir are much larger than those suggested for ABT-378 and are large enough to have an anti-HIV effect of their own. This reinforcement helps to delay breakthrough of drug-resistant HIV, the more so because the two agents trigger divergent mutational patterns. Results from the ritonavir/saquinavir trial are now in hand for 20 to 24 weeks of treatment.
This trial includes four arms: 400 mg ritonavir twice daily plus 400 mg saquinavir twice daily; 600 mg ritonavir twice daily with either 400 or 600 mg saquinavir twice daily; and 400 mg three times a day of both ritonavir and saquinavir. Initial viral load averaged about 30,000 and average initial CD4 was about 300. (The trial enrolled 141 participants, all protease inhibitor naive.) For all four arms, median viral load reductions go below the limits of the viral load assay (here considered as 200 copies/ml) at four to eight weeks and stay there. This represents at least a 2 log (99%) median drop. At week 20 or 24, 60% to 85% of people on the different trial regimens had no detectable viral load. CD4 counts were up 75 to 125. There was no statistically significant difference in the effect of the various regimens, although by week 20 the group receiving ritonavir and saquinavir at 600 mg twice daily was lower than the other three in percent with undetectable viral load and in CD4 cell count rise. (This group may be less compliant due to more side effects and a more complicated dosing schedule.)
Seven participants have been allowed to add d4T/3TC after their viral loads failed to become undetectable. Six of these seven are now undetectable, following four to 16 weeks of four-drug therapy.
Another trial reported at the Retrovirus Conference (abstract 254) tried a similar four-drug approach in 32 people with refractory HIV. The participants had persistent viral loads above 5,000 copies or evidence of disease progression despite four months on two nucleoside analogs and one protease inhibitor. They received ritonavir (600 mg twice a day), saquinavir (400 mg twice a day) and two nucleoside analogs. The week 16 results were as follows: viral load down 2.2 logs (99.4% reduction from a baseline mean of 72,500 copies/ml), with 93% of participants having undetectable viral load (below 400 copies/ml), CD4 count up an average of 72 (from a baseline average 79).
Ritonavir/saquinavir by itself frequently fails to elicit a good response in advanced patients (see Treatment Issues, December, 1996, page 4). Throwing two more drugs into the mix may be helpful for many of these cases, assuming that a patient can absorb and tolerate extra drugs, that his or her HIV remains sensitive to these drugs, and that his or her body is still able to support at least the minimum functioning needed for survival. There are many factors that might eventually or immediately interfere, although in these two small groups, at least, treatment was initially successful for most individuals.
Should Abbott have lost interest in the ritonavir/saquinavir combination due to the arrival of ABT-378, the Hoffmann La Roche protease inhibitor may have found a new partner: nelfinavir. In a preliminary trial of 13 people on saquinavir plus nelfinavir (abstract 371), nelfinavir slowed liver metabolism of saquinavir to the extent that total exposure to the latter was raised five-fold (compared to a 20-fold average increase with ritonavir). Combination therapy included 750 mg of nelfinavir thrice daily (note the dose) plus 800 mg three times a day of the new, more absorbable version of saquinavir (the "soft gel capsule" containing saquinavir formulated with special lipids which quadruple the proportion of saquinavir absorbed by the intestines). At 12 weeks, viral loads were down 2.0 logs (99%), with eight of 13 participants (62%) at undetectable levels (below 500 copies/ml). CD4 counts were up by 105.
There are 12 possible double protease inhibitor combinations just counting the three protease inhibitors on the market and nelfinavir. Many of these may be feasible combinations in terms of tolerability, drug-drug metabolic interactions and lack of cross-resistance. Bear in mind that ritonavir works with saquinavir and possibly ABT-378 above all because of ritonavir's particular metabolic properties. By blocking the hepatic breakdown of the other protease inhibitors, ritonavir raises them to highly potent levels.
Conversely, there may be interference between protease inhibitors because they compete for the same binding sites on the protease enzyme. By getting in each other's way, two protease inhibitors would have a combined effect less than the sum of their individual effects (this is termed antagonism), never mind the hoped for synergy, which occurs when the combined effects are greater than the sum of the individual ones. At the Retrovirus Conference, Merck and Roche researchers reported preliminary results indicating that indinavir administered concurrently with saquinavir raises the latter's blood levels five to eight times (abstract 608), similar to what nelfinavir does. But a team from Harvard Medical School determined that in lab culture at least, the interactive effects of indinavir and saquinavir are antagonistic, except at low doses in wild type virus (abstracts 11 and 188).
Table 1: AZT/3TC/Indinavir in Advanced AIDS
*Lower limit of viral load assay = 500.
Table 2: Two Nelfinavir Trials
AZT/3TC w/wo Nelfinavir
d4T w/wo Nelfinavir
*TID = three times a day.
Lower limit of viral load assay = 500. (Using an assay with a lower limit pegged at 100 HIV RNA copies/ml yielded viral load drops at 24 weeks of 1.4 logs (96.0%), 2.3 logs (99.5%) and 2.5 logs (99.7%) for AZT/3TC, AZT/3TC/500 mg nelfinavir and AZT/3TC/750 mg nelfinavir, respectively.)
Dethroning AZTby Dave Gilden, with John Falkenberg and Gabriel Torres, M.D.
Several trials announced last year found that AZT added little when included in two drug regimens. Among these were ACTG 175 and the pediatric trial ACTG 152, both of which found ddI monotherapy and ddI plus AZT to be equivalent or superior to AZT alone. Company trials of indinavir or ritonavir with and without AZT (Merck protocols 028 and 033 and Abbott protocol 245, respectively) likewise recorded no extra or even a negative effect from adding AZT alone to protease inhibitors (at least over four to six months). On the individual level, many people cannot tolerate AZT, or their HIV has become resistant to it. And AZT-resistant HIV may be more virulent. Despite these considerations, AZT remains the most prescribed HIV therapy, whether alone or in combination with other drugs. The time has long been ripe for its displacement.
The major factor that is shoring up AZT's position these days is Glaxo Wellcome's other HIV treatment, 3TC. 3TC has the most potent antiviral effect of the nucleoside analogs currently on the market and has comparatively few side effects, but HIV rapidly develops resistance to it unless the agent is used in combination with other medications that further suppress HIV replication. Until recently, nearly all research on 3TC has involved the "incestuous" combination AZT/3TC, a combination that has been touted as having a special "magic." One of the main arguments for this "magic" was the in vitro indication that the mutation at codon 184 of the reverse transcription gene, a single point mutation that confers high level resistance to 3TC, counteracts the mutations giving rise to AZT resistance. In short, HIV supposedly cannot be resistant to both drugs at once, and adding 3TC to the regimen of someone failing on AZT might return that drug to effectiveness. This was proven mainly false some time ago, although the appearance of resistance to the dual combination is delayed compared to monotherapy with AZT (see Treatment Issues, September, 1995, page 6) and Retrovirus Conference abstract 580). Still, AZT/3TC was the most used nucleoside analog component of the protease inhibitor studies cited in this issue.
Substituting d4T/3TC for AZT/3TC
A popular solution for those failing or intolerant to AZT or AZT/3TC has been d4T/3TC. Until now, this combination has been relegated to backup therapy because there has been little trial data to confirm its worth. Now it appears that d4T/3TC is as respectable as AZT/3TC:
The ALTIS 1 and 2 trials were open-label protocols in which everyone received d4T/3TC (abstract LB4). ALTIS 1 participants had had no prior treatment history, whereas ALTIS 2 covered individuals who were treatment experienced (a median of three years), but who had had no previous d4T or 3TC. Table 1 contains the baseline characteristics and the results of up to six months treatment.
In the treatment-naive group (ALTIS 1), plasma viral load reduction to below 3,000 copies of HIV RNA/ml was strongly correlated with initial viral load. For the treatment-experienced people (ALTIS 2), the only variable associated with greater viral load reduction was previous use of AZT monotherapy. People with a history of using the AZT/ddI or AZT/ddC combinations did not fare as well as those who had taken only AZT.
One of the concerns about substituting d4T for AZT is that AZT is known to penetrate the blood/brain barrier, thus helping to prevent or treat AIDS-related dementia and also depriving the virus of a drug-free sanctuary. A Dutch study (abstract LB5) looked into this question while comparing the results of d4T/3TC and AZT/3TC in a blinded trial covering people without past treatment. The baseline characteristics and 12 week results are shown in Table 2.
Three other reports described similar findings to ALTIS 2 concerning the antiviral potency of d4T/3TC in people switching from other therapies -- see abstracts 556, 557 and 560. The first of these presentations, a retrospective study of patients' files by AmFAR's Community-Based Clinical Trials Network, notably found that d4T/3TC provided little viral load improvement to people previously on AZT/3TC.
Blood/brain penetration was checked by conducting lumbar punctures at weeks 0 and 12 and assaying the volunteers' cerebral spinal fluid (CSF). It turned out that while all 31 participants initially had detectable HIV in their cerebral spinal fluid, by week 12 HIV levels were undetectable in every case. The Dutch investigators found that although AZT had greater immediate penetration into the CSF (where it achieved 80% of blood plasma levels), both 3TC and d4T actually had more staying power. Over time, total CSF exposure to these two was much greater. The study's conclusion was that AZT/3TC and d4T/3TC were essentially equivalent and that both should be helpful for dementia.
d4T/ddI: Going Farther Afield
Instead of two Glaxo Wellcome drugs, why not try two from Bristol Myers Squibb? Two French studies (abstracts 553 and 554) found that the combination of full-dose d4T and ddI reduced viral loads by 80 to 90% (0.7 to 1 log), at 24 weeks, accompanied by a CD4 count rise of about 40. One fear about the d4T/ddI combination is the two drugs' overlapping neurotoxicities. Reversible neuropathy or paresthesia (tingling patches on the body) forced six of the 85 participants to discontinue therapy.
The current studies managed to produce quite respectable results for nucleoside analog therapy in a group that had advanced disease and was heavily pretreated (though no one had prior exposure to d4T and only three in one trial had previous ddI). Baseline CD4 counts in the two trials were low --116 and 217 on average -- and viral loads high -- an average of 200,000 and 100,000 copies/ml, respectively. The trial's outcomes are not nearly as good, though, as the ones from the American study that peaked interest at 1996's Retrovirus Conference and the International Conference on AIDS last summer (see Treatment Issues, February, 1996, page 7 and August, 1996, page 7).
That 79-person U.S. trial involved less advanced patients than were the case in the French studies. The Americans' entry CD4 counts averaged 343, and their initial viral load was 16,000. All were all treatment-naive.
By last summer, the trial as a whole had recorded only one serious case of peripheral neuropathy, but only 20% of participants were receiving full-dose d4T and ddI in what was considered a pilot safety test of the combination. One-year results for the first 18 participants also were presented last summer. These 18 exhibited on average a sustained CD4 count rise of 80 to 100 and a viral load drop of 96% (1.4 logs). (See abstract Th.B.293 from the Eleventh International Conference on AIDS).
Combining d4T/ddI with protease inhibitors presents problems. Drug-drug interactions in the gut and requirements that ddI and indinavir be taken on an empty stomach make for a very complicated dosing schedule when trying to administer d4T, ddI and indinavir concurrently (see Treatment Issues, November, 1996, pages 8-9). And the combined gastrointestinal side effects of ritonavir and ddI might be difficult to endure. A poster at the Retrovirus Conference reported on a Bristol Myers pilot study of d4T/ddI/nelfinavir (abstract 241), which could be the simplest combination. But that study, which included only people naive to all three drugs, was plagued by noncompliance with the regimen: Overall, 16 people had a disappointing viral load drop of 96% (1.4 logs) by week 12, but the six completely compliant participants in that group saw their viral loads shrink 99.4% (2.2.5 logs). Much of the noncompliance was attributed to "inconvenience."
Here is another strategy for triple combination therapy: Adding hydroxyurea to d4T/ddI considerably strengthens this combination's effects, according to one placebo-controlled trial presented at the Conference (abstract 550). This cancer chemotherapy blocks a cellular enzyme that helps produce some of the natural nucleosides used to build DNA. Reducing the natural nucleoside pool makes HIV more susceptible to the nucleoside analogs, especially ddI in this case (see Treatment Issues, September, 1995, page 1).
The Swiss trial enrolled 142 persons, 80% of whom had had no previous treatment. The rest had no more than six months of ddI and no prior exposure to d4T or hydroxyurea. The trial's baseline characteristics and results are summarized in Table 3.
Note that CD4 counts did not rise in the hydroxyurea arm, and total lymphocytes fell considerably. This is probably due to hydroxyurea's inhibitory effect on cell division. Also, 15 of 72 volunteers on hydroxyurea had withdrawn from the study by the twelfth week, compared to only 5 of 70 on placebo. Much of this difference was due to extra nausea and vomiting and other side effects in the hydroxyurea arm.
A 19-person pilot study at the University of Texas in which everyone received d4T/ddI/hydroxyurea (abstract 549) also observed a 99% (2 log) reduction in viral load by week 12 with no change in CD4 count. In this trial, 17 of the volunteers had been on either AZT, ddI or d4T monotherapy. One each were receiving AZT/ddI and d4T/ddI. Baseline CD4 count averaged 226, and average initial viral load was 81,000.
In a third study (abstracts 589 and 591), Franco Lori, M.D., and colleagues from the Research Institute of Genetic and Human Therapy in Maryland presented data on over 40 patients who received ddI and hydroxyurea combination therapy for more than one year with no evidence of viral load rebound. The combination of ddI and hydroxyurea was unable to prevent the emergence of mutations that confer ddI resistance, but the mutants were still sensitive to standard doses of ddI in the presence of hydroxyurea. This continued sensitivity might be due to the deprivation of natural, functional nucleosides induced by hydroxyurea.
A Rectification in DMP 266 Analysis
The Retrovirus Conference also contained a further "preliminary" report on a trial of DMP 266 (abstract LB2), DuPont Merck's new and powerful non-nucleoside reverse inhibitor (NNRTI). This trial was first described at the annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) last September. Subsequent to that presentation, Treatment Issues uncovered a hidden flaw in the way the DMP 266 plasma viral load data was analyzed: everyone with a viral load below 100 HIV RNA copies/ml (2 logs) was counted as having a viral load of one (0 logs). Since the PCR assay employed in this trial is commonly considered accurate only down to 400 copies/ml, and most trials have counted any measurement below that figure as "undetectable" and assign it a value of 200 -- or 400 or 500 to be more conservative. Because of its nonstandard analysis, DuPont Merck made DMP 266 appear to reduce viral loads up to 100 times greater than other agents -- reductions which in fact exceeded biologic possibility (see the introduction to the protease inhibitor article, page 5).
Happily, DuPont Merck has rethought its approach. The Retrovirus Conference account, delivered by Sharon Riddler, M.D., of the University of Pittsburgh, gave the results obtained when assigning a viral load of 200 to values falling below 400. The two-arm trial compared DMP 266 plus indinavir to indinavir alone for 24 weeks. (Treatment Issues has also complained about this protease inhibitor monotherapy, which represents substandard therapy that imperils volunteers by promoting drug-resistant virus -- see Treatment Issues, May, 1996. Several trials reported at the Retrovirus Conference contained similar arms, and others are still enrolling. In particular, the AIDS Clinical Trails Group is now commencing a trial of Glaxo Wellcome's 141W94 monotherapy versus, once again, Glaxo's AZT/3TC combination plus 141W94. DuPont Merck is not planning any more trials with monotherapy comparison groups. Neither should anyone else. )
At the end of the trial period, viral loads were down 2.2 logs (99.4%) in the combination arm and only 1.5 logs (96.8%) in the indinavir arm. (The mean baseline viral load was 99,000.) Eighty-two percent of the 21 volunteers on combination therapy and 38% of the nine on indinavir had viral loads below the 400 copy/ml limit. (If DuPont Merck's old method of calculating viral load had been used, the reductions in the two arms would have been 4.l and 2.4 logs, respectively.) CD4 counts, initially averaging 224, were up about 100 cells in both groups.
The results achieved with DMP 266/indinavir rival those achieved with indinavir plus AZT/3TC or any two nucleoside analogs, even with this more conventional analysis. DMP 266 is an impressive drug. There remains the issue of the proper dose of indinavir to use with this NNRTI, which stimulates indinavir's breakdown in the liver. Midway through the trial, the indinavir thrice daily dose was raised from 800 mg to 1,000 mg, which may still be suboptimum. The toxicity implications in raising the indinavir dose under the influence of DMP 266 (particularly concerning kidney stones) have yet to be fully explored, although a similar dose escalation is recommended when indinavir is used with nevirapine. The main side effects noted during the trial were rashes, headaches, diarrhea and abnormal liver function test results.
DMP 266 not only is potent but is taken only once a day. Along with such agents as ABT-378 and Glaxo Wellcome's experimental nucleoside analog 1592U89, it indicates there is still plenty of room for improving the activity, convenience and possibly the durability of current classes of anti-HIV agents.
Table 1: ALTIS Open Label Trial of d4T/3TC
Table 2: AZT/3TC vs. d4T/3TC
Table 3: d4T + ddI w/wo Hydroxyurea (500 mg twice daily)
*Limit of detection = 200. (Using an assay with a lower limit of 20 copies/ml yielded viral load drops at 12 weeks of 2.2 logs (99.4%) in the hydroxyurea arm and 1.8 logs (98.4%) in the placebo arm. The week 4 viral load values were the same with either limit.)
Nevirapine Protects Chimpanzeesby Gabriel Torres, M.D.
A study presented by Boeringher Ingelheim at the Retrovirus Conference (abstract 728) evaluated the protective effect of the company's NNRTI nevirapine (Viramune) in chimpanzees. Chimps are the only other primate that can be infected with HIV-1, although it does not make them ill. The experiment involved four chimpanzees. One untreated animal was challenged with an intravenous injection of HIV-1 and observed to seroconvert (start producing HIV antibodies). Virus was detected in its blood. The other chimpanzees received high doses of nevirapine (800 mg) at either six hours, 12 and 36 hours, or 12, 24 and 36 hours pre-challenge. Treatment continued for 10 or 20 days post-challenge.
None of the treated chimpanzees seroconverted. After five to six months, no sign of HIV could be found in their lymph nodes. All remained negative for all viral load markers with the exception of proviral HIV DNA within infected cells. That DNA, measured with a PCR assay, subsided and became undetectable for the three treated chimpanzees one to two years after exposure to HIV. It persisted the longest in the least treated chimp.
This experiment indicates that nevirapine can abort infection in the chimpanzee model of HIV infection. It has obvious implications for preventing HIV transmission in humans.
Searching for SynergyAn Interview with Erik De Clercq
Professor Erik De Clercq, M.D., Ph.D., is one of the pioneers of antiviral research. From his base at the Riga Institute of Medical research in Leuven, Belgium, he has directed a team that made many of the original discoveries in anti-HIV medications. Prof. De Clercq is best known for opening up the field of nucleotide analogs, which are the first broad-spectrum antiviral medications. Members of this class are active against a large number of viruses, including herpes viruses, papilloma viruses, hepatitis B and HIV. One of the nucleotide analogs, cidofovir, was approved last year for treating CMV retinitis. It was brought to the market by Gilead Sciences, a small biotech company. Another of Gilead's compounds, adefovir, is in advanced clinical trials for HIV.
Treatment Issues talked to Prof. De Clercq shortly before the Fourth Conference on Retroviruses and Opportunistic Infections. Among other topics, he mentioned the difficulties encountered in developing some of the nonnucleoside reverse transcriptase inhibitors and HIV-cell fusion that he has helped discover. These new compounds would be very helpful in putting together innovative combination therapies to combat HIV. Prof. De Clercq also brought up his continued enthusiasm for the nucleotide analogs, which are very slowly wending their way down the developmental pipeline.
Creative Therapeutic Combinations
Treatment Issues: Is there a need for new kinds of HIV drugs beyond what we have now, and if so, what?
De Clercq: We should keep in mind that monotherapy definitely belongs to the past, and that we should go for a combination of several compounds. That may include protease inhibitors, that's for sure, but it should also include nonnucleoside reverse transcriptase inhibitors [NNRTIs -- ed.], and nucleoside reverse transcriptase inhibitors [nucleoside analogs]. So a combination of at least these three.
If you look at the present time to what has been done, I don't think that this kind of combination of a nucleoside RT [reverse transcriptase -- ed.] inhibitor and nonnucleoside RT inhibitor, and then a protease inhibitor, has even been tried so far. [Editor's note: but see page 6 of this issue.] It's always the same combination -- AZT, 3TC, and then they add a protease inhibitor to that, or they add nevirapine. It's only nevirapine that has so far been used in that combination.
As you probably know, my task is to develop new compounds and I think we have been, let's say, quite helpful in coming up with a number of nonnucleoside RT inhibitors, with which I have the most experience. So most of the remarks I would give are based on that experience with these nonnucleoside RT inhibitors.
First, they are all able to completely suppress the virus when they are given at sufficiently high doses, and that you probably can keep the virus completely suppressed with these compounds -- certainly when you give them in combination with each other, or combined with a nucleoside analog, or in combination with a protease inhibitor.
TI: Do you think it's a viable thing -- to put two NNRTIs together?
De Clercq: That's certainly possible, yes. In all the combinations that we have tried of two nonnucleosides together, we have found synergistic activity. So even with two compounds that are going to the same target, we can still have synergistic action.
We have already an example of the very nice interaction between two protease inhibitors -- saquinavir and ritonavir -- that potentiate, or at least, ritonavir potentiates the activity of saquinavir. They have a different resistance profile, and the metabolism of saquinavir is reduced or delayed in the presence of ritonavir. So you have two good reasons for a combination of these two compounds.
Apparently, they work synergistically in the patient, too. So there is, in my opinion, a possibility of combining two protease inhibitors, and at the same time, two NNRTIs could also be combined. Even if they work at the same target, if their resistance profiles are different, again, you have good reason for combining.
Second Generation NNRTIs
TI: Isn't there a lot of cross-resistance between the various NNRTIs?
De Clercq: That is not necessarily true. I mean, that has always been the perception from the outside, and some of the compounds do, but some of them have a totally different resistance profile. In fact, I would go even one step further. If you use these compounds from the beginning at sufficiently high concentrations, the resistance does not even develop, particularly some of the newer compounds that are only at the initial state of development. I would mention, for instance, the carboxanilide compounds. We have found that some of these compounds are quite active against HIV resistant to other nonnucleoside RT inhibitors.
We also have very good experience with HBY 097, which we have studied intensively. That is the Hoechst Bayer compound; it is a very powerful compound. In fact, in our ranking, it's really within the top one or two, within the top few compounds that we have analyzed ourselves. I hope it will be developed. It's certainly one of the compounds that I would advocate for being used in combinations in the future. It's one of the top compounds.
TI: What about the UC compounds? How potent are they?
De Clercq: The Uniroyal compounds. Yes, well, these are carboxanilides. These are the two best compounds in our hands, based on the experience we have had with these compounds. [Editor's note: On the now four UC compounds' potency and ability to overcome resistance, see abstracts 569 and 570 from this year's Conference on Retroviruses and Opportunistic Infections.]
TI: Are these compounds being developed?
De Clercq: First of all, the compounds belong to the company Uniroyal, and that is more of a chemical company. So they have to find a larger company, a more pharmaceutical one. They are in the process now of trying to set up an arrangement with a larger company to do this, but it has not been decided yet. So we still have to wait. These two sets of compounds turned out to be so much more active than nevirapine. So that's the reason why we would be in favor of their further development.
TI: It's kind of funny that Uniroyal discovered them in the first place.
De Clercq: Originally, these compounds were designed to be herbicides or insecticides. I forget -- one of the two. This happens all the time. It also proves that in terms of chemical synthesis, it must be rather easy. So here, we have another example of compounds that must be less expensive than the current protease inhibitors. I mean, you can already tell, from the structure.
HIV-Cell Fusion Inhibitors
TI: Something else I wanted to ask you about was about fusion inhibitors, which block HIV binding and entering cells.
De Clercq: Well, I would say there's a lot of progress on the mechanism of fusion. There are already fusion inhibitors, but we do not know whether they act on the so-called second receptor, on the chemokine receptor.
As you probably know, we have developed some of the compounds that affect fusion. We described a few years ago a very interesting class of compounds, which are called the bicyclams.
TI: Do you know how they act?
De Clercq: We believe that they somehow penetrate into the interior of gp120 [HIV's envelope molecule -- ed.] and then buckle the stems of the [V3 and V4 ] loops together, so that during the process of fusion, the gp120 cannot unwind and open as it normally does. The bicyclams probably do not work on the chemokine receptor. That would have been too nice, but they certainly work at the fusion process.
TI: What kind of concentrations inhibit HIV?
De Clercq: In vitro, that is down to the nanogram per milliliter range, so they're very potent, and no toxicity whatsoever, up to 500 micrograms per milliliter. So they have selectivity indexes of more than 100,000. It's in the same range as the protease inhibitors and as the nonnucleoside RT inhibitors -- in the range of the best ones.
TI: How did the bicyclams get discovered?
De Clercq: Oh, discovered, like most other compounds, by accident. The purpose was, in fact, to design [ringed] compounds that were partly organic, partly inorganic, that would have an organic surrounding and then, in the middle, a metal. In the first experiments that we did, we looked to the activity of several monocyclams without metals in it and we found that monocyclams had no activity -- as we expected -- because the purpose was to build in a metal to make them active.
So we tested the monocyclams and we did not see activity, except for one preparation coming from a company in England. Then, in our studies with the Johnson Matthey company, we found that it was the bicyclam contaminant, in which the two cyclam rings were tethered together, that was responsible for the activity. From then on -- from the lead compound -- we, of course, improved on the potency.
Our hypothesis is that the zinc, which is present in cells, makes the link between the cyclams and the bottom of the loops in the gp120 and freezes it in a kind of an immobilized form so that it cannot unwind.
TI: What is the status of bicyclams at this point?
De Clercq: We were a little unlucky in the development for the following reason -- we gave these compounds in a license agreement to Sandoz, and then Sandoz, on a blue Monday, decided not to develop any antiviral compounds.
That was just when we were ready to go in the clinic with these compounds. So we have not been able to get information in humans. These are very good compounds, I mean, very strong, potent fusion inhibitors, and these compounds still deserve a chance.
Sandoz has returned the rights to Johnson Matthey, but Johnson Matthey is essentially a metal company. It has now spun off a smaller company, which is called Inormed. It's going to operate in Vancouver.
We lost several years, but the people at Inormed are absolutely dedicated to putting these compounds forward.
TI: Several antiviral drugs have gotten lost or delayed at Sandoz. [Editor's note: Sandoz has recently merged with Ciba-Geigy to form the giant "Novartis," which describes itself as "the world's leading life sciences company."]
De Clercq: Well, we hope it's only delayed. These belong to the most potent compounds we have ever had a chance to look at.
The Nucleotide Analogs: Unexpected Activity
TI: I was fascinated by some of the nucleotides and the range of viruses. It seems like we're going from an era where there was almost no treatments for viruses, to things like acyclovir to a few treatments for HIV, to some compounds that are really broadly active against many viruses now. Can you say something about how that came about and why these compounds are so broadly active?
De Clercq: I would like to stress that the origin of this work goes back to 1985, when we wanted to make hybrids -- a molecule between acyclic nucleosides [like acyclovir or ganciclovir] and a phosphonate, like Foscavir.
We kept the activity as we expected, against herpes viruses, and we found now all herpes viruses to be sensitive to [hybrid] compounds like cidofovir. It's active against all the herpes viruses, but in addition, we found it active against polyoma. That means that even diseases like PML become, now, in the realm of treatment.
We have a model that we can induce tumors with polyoma virus, in rats. They are also responsive to the compound. We have only limited experience with PML in patients, but based on the effect against polyoma, we can easily predict activity against PML, and we have some cases -- two cases [in PML patients]. There is a recent case, of which I do not know the full story, going on here in New York, apparently.
Adenoviruses [mostly respiratory infections -- ed.] are sensitive too, but also, even the pox viruses. This is, of course, not such an important virus anymore, but molluscum contagiosum, for instance, it is also sensitive. Then we also showed the activity against papilloma virus [the cause of warts and possibly cervical cancer -- ed.], and this, I think is one of the major breakthroughs in that area.
We found all herpes viruses to be sensitive to cidofovir. There's no exception. At the present time, there is absolutely nothing against Epstein-Barr virus [associated with lymphoma, mononucleosis and OHL -- ed.]. But Epstein-Barr virus-related tumors would be susceptible to treatment with cidofovir.
TI: What about KSHV, the herpes virus thought to trigger Kaposi's Sarcoma?
De Clercq: At the present time, we're doing Kaposi's, first in a cell line, and one of my coworkers is trying to set up in the system for Kaposi's in mice. I'm very curious. I wonder whether doing all these studies with CMV retinitis, they have been able to see something on Kaposi's.
[Editor's note: see abstract LB20 from the Fourth Conference on Retroviruses, which describes cidofovir's activity against KSHV in cell cultures. Acyclovir, ganciclovir and foscarnet had much less effect.]