Geneva "New Agents" Sessions Betray Dearth of Interest in Imaginative (or Costly) Drug Development
Everything new is old again
What is becoming clearer and clearer as these national and international meetings accumulate post-Vancouver is the need for new antiretroviral therapies and, especially, for new therapies with novel mechanisms of action. While the exciting potential of integrase inhibitors, nef antagonists and fusion inhibitors may be tossed around seductively at the more conceptual plenary sessions, a look at what's actually in current Phase I trials around the world betrays an ugly truth: there are virtually no novel agents in development. "Me-too" proteases, nucleosides and NNRTIs abound (see table) but are expected to convey little if any therapeutic benefit to the hundreds of thousands of HIV-infected individuals expected to fail combination antiretroviral therapy -- either for the first time or the last -- in the coming months. This is a critical short-coming little addressed at this summer's 12th World AIDS Congress in Geneva or even at last winter's Chicago Retrovirus conference. But in a feigned attempt at objectivity, TAGline checked in with AmFAR's new drug watchdog Theo Smart (who co-chaired the Geneva session on "Pre-Clinical Drug Development") and ACTG HIV-RAC co-chair Dr. Dan Kuritzkes (who gave "The Future of HIV Therapy" talk) for their takes on what we can expect from the new drug pipeline over the next 2-3 years. Afterall, they're paid to be optimistic.
The dearth of new therapies was perhaps most tangibly exemplified by the abysmal poster showings at both the Geneva and Chicago meetings this year. The Geneva poster session dedicated to "New targets/New approaches" was reportedly cancelled, and the week's two sessions for Phase I clinical trials of "new" antiretroviral agents (one oral, one print) included a total of five agents -- all merely variations on already existing therapies. "New drugs/New targets" showings in Chicago were equally bleak. Fortunately, there are a few agents in Phase I or pre-clinical development that haven't managed to make it to these international conferences, but true innovation is still mighty hard to come by. If the goal is merely to develop therapies easier to take and tolerate for future waves of HIV-infected individuals, then progress is indeed being made. If, however, the goal is to come up with additional therapeutic options for the hoards of "treatment experienced" individuals rapidly cycling through the menu of current agents, it remains unclear whether the compounds now in various stages of clinical development will be of any use.
Triangle's me-too nuke: FTC
FTC is a fluoronated version of 3TC, with somewhat greater potency than 3TC -- at least in the lab -- and the very same resistance profile. Yes, M184V brings it to its knees. The hook here is that it can be dosed once a day. (But then 3TC is already being dosed that way in Europe...) Results from a Phase I/II study were presented at the February Chicago meeting -- and again in Geneva -- with a mean plasma viral load reduction of some 2.0 logs at the higher (200 mg QD) dose.
U.S. Bioscience's me-too nuke: FddA (a.k.a. lodanosine)
If you can fluorinate 3TC, why not fluorinate ddI while you're at it? The real benefit here being that the fluorine group renders the once acid-phobic ddI acid stable. So it no longer needs all that yucky buffer and empty stomach requirements. FddA is said to get into cells better -- largely due to the fact that its activation is independent of cellular activation. But, at least in vitro, FddA appears less potent than AZT, with a mean viral load drop of 0.4 log in nuke experienced individuals.
BioChem Pharma's me-too nuke: dOTC (a.k.a. BCH 10652)
Member of a new class of "hetero-substituted" nucleoside analogues. Reported to have activity against clinical isolates resistant to 3TC and AZT. Resistance is said to be "slow to develop": no change in the IC50 and no evidence of RT mutations after 12 serial passages in the MT-4 cell line. Studies in animals suggest a good safety profile. Moving into Phase I/II.
Triangle's me-too nuke: dADP
Guanosine analogue. Yet to enter human study. Said to be a super nuke, "like abacavir." (Great.) Also has activity against hepatitis B. In the SCID-Hu mouse model, dADP is less potent than FTC. Little or no cytotoxicity thus far.
Ray Schinazi's me-too nuke: d-d4FC
A cytosine nucleoside (like 3TC and ddC) claimed to be active against multi-nucleoside resistant (MNR) HIV, at least in cell culture. The active metabolite is reported to be a more active chain terminator than that of 3TC, but it is reported to be less potent than FTC in the SCID-Hu mouse. Half life in monkeys of about 4 hours. No human data yet.
Triangle's me-too non-nuke: MKC-442
The nucleus of this non-nuke is actually a nucleoside analogue (which resembles a pyrimidine). Currently in Phase II/III, making it the RTI furthest along in development. One to one-and-a-half log drop in plasma viral load in a 15-day Phase I/II study. So far, it's being dosed twice daily. Unfortunately, it reportedly only takes one mutation -- the ubiquitous K103N -- to knock it out. That imposes serious limitations.
Agouron's me-too non-nuke: S-1153
Agouron recently acquired the rights to this NNRTI claimed to exhibit a "10-fold greater in vitro potency than either nevirapine or delavirdine." Phase I results were presented in Chicago and again in Geneva. In a subset of 11 study volunteers with baseline viral loads > 10,000 copies/mL, the mean viral load reduction with S-1153 was 1.74 logs. Said to be active against the K103N mutant and to require a minimum of 2-3 mutations in order to develop significant resistance. Good bio-availability with or without food. BID dosing possibilities.
Upjohn's me-too non-nuke: PNU-142721
A thiopyrimidine. Low protein binding, according to Dan Kuritzkes. BID or QD dosing. Said to be active against delavirdine resistant HIV.
Carboxanilide analogue NNRTIs (e.g., UC-781):
No Phase I data to be had. Upside? In general, they tend to have a very long half-life. UC-781 inhibits RT-dependent pyrophosphorolysis and purportedly restores the chain terminating activity of AZT against AZT-resistant virus (as does the A114S foscarnet resistance mutation). Downside? They currently require intravenous administration.
Sarawak Medichem's me-too non-nuke: calanolide-A
Calanolide-A is reported to have been isolated from the Malaysian rain forest and subsequently synthesized for clinical development with a plasma half-life of some 20 hours. To date, only Phase I PK and safety data are available.
Glaxo's me-too PI: amprenavir
Next in line for FDA licensure after efavirenz and abacavir. Many protease resistant folks (and their physicians) are betting on it for at least a temporary viral load reprieve. Fact or folly? While mutations at codons 46, 47 and 50 -- with 50 predominating -- are first to appear, subsequent cross-resistance mutations at the more familiar codons 82, 84, 10 and 36 have been reported in a significant number of patients failing amprenavir therapy in Phase II trials. Additional mutations were also detected at the p7/p1 and p1/p6 site of gag in many of the patients. So broad class resistance still seems to be the rule of thumb here.
Abbott's me-too PI: ABT-378
Why do they call these agents "second generation," anyway? Ostensibly because, as Dan Kuritzkes put it in his Geneva talk, they've been "developed to have potency against [protease] resistant virus." And, in fact, Abbott's 378 does appear to have an IC50 about 10-fold lower than its predecessor -- whether against wild-type HIV or protease resistant HIV. Like Upjohn's tipranavir, 378 is poorly absorbed and will likely require co-administration with low-dose ritonavir.
While the drug is undoubtedly potent, it is not clear that it will represent much of an option for people already resistant to other protease inhibitors. The key mutation which leads to resistance to ABT-378, at codon 84, is common in patients who have become resistant other protease inhibitors. (It is worth noting too that in none of the recombinant viruses used in late-breaker slide shows of 378's prowess was the critical 84 mutation included.)
Upjohn's me-too PI: tipranavir
The first of a new class of non-peptidic protease inhibitors, tipranavir is said to be effective against ritonavir and indinavir resistant HIV. This is due to its chemical structure which, according to Smart, seems to be a bit more flexible than the peptide-based inhibitors.
Early activity data on tipranavir are not particularly impressive: Mean log plasma viral load reductions at day 11 were 1.0, 1.3 and 1.2 -- which by week 4 had been whittled away to 0.5, 1.1 and 0.9 log, respectively. By week 12, mean plasma viral load reductions were 0.6, 0.6 and 0.7 log. At the highest dose (1,500 mg TID), there was an overall 1.3 log reduction from baseline which rebounded to 0.7 log below baseline by week 12.
Part of the problem (other than the abominable study design: adding a single agent to a failing double nuke regimen -- in 1997!) was the sheer number of pills to be taken. At the highest dose, a total of 10 horse pills three times a day. Diarrhea was a big problem as well. (Maybe they shat it all out?) A new formulation with better bio-availability (and thus a lower daily pill count) is said to be in the works.
Parke-Davis' me-too PI: PD178390
Another non-peptidic protease inhibitor said to exhibit activity against resistant HIV. PD178390 is also unique in that it is not metabolized through the cytochrome P450 system; that should cut down on drug-drug interactions, but will also disqualify it for potentiation by P450 inhibiting agents such as ritonavir. While Theo finds this "the most promising protease inhibitor in development," cross-resistance and protein binding may turn out to plague this hot-shot wanna-be just as they have plagued past early protease-inhibiting favorites. Mutations at amino acids 10 and 36 increased the IC50 of the drug at least 4-fold, and when human sera or alpha acid glycoprotein was added to the mix, the IC50 for PD178390 increased an additional 2 to 4-fold. The drug's half life in mice is a mere 2.2 hours.
Triangle's me-too PI: DMP-450
The third non-peptidic protease inhibitor (it's a cyclic urea compound) is also said to have different side effects from the other PIs since it binds differently to the enzyme -- and is metabolized differently. Also the second PI in development to steer clear of the cytochrome P450 system. Apparently simple and inexpensive to manufacture. And since it is not highly plasma protein bound, it should be active at fairly low doses. Its water solubility lends itself well to a pediatric formulation. Appears to share cross-resistance with ritonavir and indinavir, so again it is unclear whether it will represent much of an option for protease resistant folks.
Bristol's me-too PI: BMS-232632
An azapeptidic protease inhibitor. As is the case with DMP-450, the Phase I study is expected to be completed by year-end.
Japan Energy's me-too PI: JE2147
An allophenylnostatine protease inhibitor, also recently licensed to Agouron. Like DMP-450, it is not highly protein bound and therefore expected to be active at lower doses. Said to be active against protease resistant virus, but the ubiquitous 82 mutation (along with 10 and 47) pop up after in vitro serial passage. Plasma half-life of only 1.5 hours.
An up-coming TAGline will take a look at drug development efforts which focus on truly novel targets, but a word of caution: work to date with CCR blockers, zinc finger antagonists and integrase inhibitors has yielded mostly disappointing early results.