By David N. Leff

After an AIDS retrovirus breaks and enters the immune system cell it aims to hijack, the invading HIV hooks up its own DNA to the genome of its target cell. This maneuver looks a little like a train robber coupling his caboose to the locomotive and string of railroad cars he's commandeering.

Once the HIV virion connects its DNA to that of the victim's T cells or macrophages, it forces the locomotive driver to replicate the virus big time. Then this swarm of new viral progeny bud out of the cell, and spread out to infect more cells - and so on, ad infinitum.

From the time, circa 1980, when AIDS raised its pandemic head, it took virologists until the mid-1990s to come up with an effective drug regimen against HIV infection and AIDS pathology. The now-celebrated triple-drug cocktail - two reverse transcriptase inhibitors plus one protease inhibitor - has tamed HIV from a slow but lethal threat to a costly, cumbersome pill-popping ordeal. (See BioWorld Today, Oct. 4, 1999, p. 1.)

But like all human-disease pathogens, the AIDS virus is fighting back. That HAART (highly active retroviral therapy) has a fixed-term lease on anti-HIV chemotherapy, not a bill of sale. Sooner or later, the virologists know, drug resistance will foreclose HAART's title. Which is why the race goes on for new drugs to protect future HIV-positive patients as the present chemotherapy loses its clout.

Looking for an as-yet-untouched Achilles' heel to bring down HIV once more, drug designers have fixed on a viral enzyme called integrase. This is the compound that actually hitches the viral strands of DNA onto the target cell's genome. It looks like a logical site at which to cut the invading pathogen off at the pass. But the search for integrase inhibitors has its problems.

"All the efforts to date to identify inhibitors of integrase," pointed out clinical and research virologist Douglas Richman, "have identified compounds that affect the interaction of integrase with nucleic acid substrates. But they turn out to be nonspecific in terms of inhibiting HIV replication." Richman's laboratory at the University of California, San Diego, identified HIV drug resistance about 12 years ago.

Retrovirus Conclave To Hear Science Data

Now scientists at the Merck Research Laboratories (MRL) in West Point, Pa., are trying a new tack. Their interim report in today's Science, dated Jan. 28, 2000, bears the title: "Inhibitors of strand transfer that prevent integration, and inhibit HIV-1 replication in cells." On Sunday, its senior author, Daria Hazuda, who is a director in MRL's Department of Antiviral Research, will present this data to a special symposium on Novel Antiviral Therapies at the 7th International Retrovirus Conference in San Francisco. She is the Science article's senior author.

"This paper, and the presentation in San Francisco," MRL's vice president of public affairs, Larry Hirsch, told BioWorld Today, "describe what we think is very elegant molecular biology and biochemistry. It has identified, or validated, the integrase enzyme in HIV as a potential candidate for therapy."

After HIV's RNA has been reverse-transcribed into DNA and then multiplied, it's at that point where there is strand separation. That's where the integrase goes into action. It's inside the cell, responsible for preparing the dinucleotide pairs at the ends of the viral DNA strand. Subsequently the enzyme literally, physically, moves that strand of viral DNA into the host cell genome.

Hirsch explained: "The two MRL compounds reported in Science have been shown to be effective inhibitors not only of the enzyme, but of viral replication - which is really the important news. However," he added, contrarianwise, in the same breath, "we do not believe they are likely to come through as drug candidates. The paper is describing findings in cell culture, not even any experiments in animals, much less trials in humans.

"We would love to go into animals," he added, "but we believe that the compounds described in this article are unlikely to be successful as therapeutic agents or drugs. Their utility," he went on, "is rather as proof of concept. That's the bottom line here. Up until this point, there have been descriptions of molecules that inhibited viral integrase, but they have not up until now also shown inhibition of viral replication. That's the difference.

"These very early preclinical compounds," he continued, "have worked to inhibit viral replication, because they block the integrase strand transfer. That's the key finding."

Two Winners In 250,000-Compound Lottery

Hazuda and her co-authors winkled out the two integrase inhibitors from Merck's library of 250,000 compounds. They asked the robotic, computer-controlled, high-throughput screening facility to find all molecules capable of inhibiting strand transfer. It came up with two potential chemicals, which MRL designated as L-731,988 and L-708,906.

Strand transfer, the Science paper explains, is the third and final act of the multistep process by which, by covalent linkage, integrase integrates the viral DNA 3' ends into the cellular target DNA. "That is the key aspect of what these compounds do," Hirsch observed.

Calling the Merck work "a solid and clever piece of research," Richman told BioWorld Today, "The news is that they've dissected and characterized the integrase reaction, so they could assay for one specific critical function of the enzyme, and get rid of the activities that were nonspecific, nonselective - that people always ended up looking at when they were measuring inhibition. So it's a nice piece of enzymology, and a critical path for drug discovery."