Combination therapy wasn’t on the minds of Abbott, Hoffman–La Roche, or Merck as they moved their most promising protease inhibitor candidates into full-scale clinical trials. They wanted blockbusters, a billion-dollar seller. For AIDS, where the ultimate market size was thought to be limited, that meant monotherapy.
By the summer of 1993, Abbott Laboratories had already dropped its intravenously administered protease inhibitor that European physicians had presented in Berlin. Earlier that year, Kempf and his small chemistry team had come up with an analogue of the huge molecule that passed its initial test-tube experiments. Leonard quickly approved it for clinical trials. Its chief attraction was that it could be administered orally. While European clinics began testing this second candidate in patients, Kempf continued cooking up new possibilities.
Shortly after Labor Day, Kempf cautiously advised Leonard that his group had come up with a third protease inhibitor candidate. The senior official gathered the small group of Abbott scientists working on AIDS in a conference room to showcase the molecule. An air of expectancy filled the room. Kempf, a quiet, self-effacing bench scientist, rarely spoke up during their regularly scheduled brainstorming sessions. Asking for time on the agenda meant he was onto something.
Deploying a series of overhead slides, Kempf showed how he had reconfigured one side of his original symmetrical molecule to make it easier to absorb through the gut. The change had not knocked out its effectiveness in blocking protease action in the test tube. Best of all, Kempf’s preliminary tests in animals showed one of the analogues of this new molecule stayed in the bloodstream for hours rather than being quickly metabolized by the liver and excreted. “I want that drug,” Leonard said. A-87538, later known as ritonavir, moved almost immediately into human trials. The patent application, filed September 14, 1993, still contained the government rights clause since work on the molecule was completed while the company was still receiving government money. The five-year government research grant, which Abbott had used to provide support for Kempf, expired two weeks before the patent application.1
While Abbott had relied on European physicians to test its two previous candidates, Leonard was now prepared to return to the United States for clinical trials. His way was eased by the close relationship that Kempf had forged with David Ho, the director of the Aaron Diamond AIDS Research Center in New York City.
Kempf first met Ho in late 1991 at a Florida conference that brought together several hundred of the world’s leading AIDS researchers. Ho, an ambitious virologist, was eager to learn more about Abbott’s protease inhibitor project, which was revealed for the first time at the conference. Philanthropist Irene Diamond had shocked the AIDS research world when she chose Ho to run her well-endowed new center, and two years later he was still scrambling to establish its program. The two men struck up a conversation after bumping into each other in a car rental checkout line. Ho offered to run tests on Kempf’s newly synthesized chemicals for their antiretroviral activity. His offer was timely. Kempf’s superiors had just cut the firm’s ties to the National Cancer Institute (NCI) article, which had previously done those tests. Kempf needed a new partner. “They didn’t have a lot of virology in-house,” Ho recalled.2 Kempf began sending his drug candidates to New York instead of Bethesda.
Ho’s relationship with Abbott would eventually lead to the central theoretical breakthrough that underpinned the emergence of triple-cocktail therapy. The contribution would later earn Ho worldwide fame as Time magazine’s 1996 “Man of the Year,” even though George Shaw of the University of Alabama at Birmingham shares the credit in the scientific literature. For Ho, the breakthrough brought together the threads of his entire career and would not have been possible without crucial support from Los Alamos National Laboratory.
Born in mainland China in 1952, Ho was raised in the United States and graduated from the Massachusetts Institute of Technology and Harvard Medical School. He was a medical resident at UCLA’s Cedars-Sinai Medical Center in 1981 when doctors there reported the world’s first known AIDS cases. The budding virologist immediately gravitated to the new field. In 1982, Ho accepted a postdoctoral fellowship in Martin Hirsch’s lab at Massachusetts General Hospital, where he raced to become the first to discover the virus that caused AIDS. He lost that battle, but through his daily contact with AIDS patients he gleaned crucial insights into the dynamics of the disease. In 1985 he coauthored a paper that first described the flulike symptoms that accompanied a person’s initial HIV infection.3 But it would take the emergence of protease inhibitors nearly a decade later before he learned the deeper significance of his own observations. At the time, he, like most AIDS researchers, believed infected individuals built up antibodies to knock down the initial infection, and those antibodies kept the retrovirus in a near dormant state for years before losing their potency.
He carried those beliefs with him when he moved from Boston to the new Aaron Diamond Center, which provided the young scientist with an ideal setting for pursuing intriguing avenues of research. Irene Diamond, the wife of a wealthy New York real estate developer who was devoted to medical research and the arts, believed the city that was an epicenter of the AIDS epidemic needed to do more to combat the disease. Forging a partnership with the New York City Department of Health, her foundation created the specialized center in 1988 and pledged $220 million over the next ten years for its support. The center officially opened next door to Bellevue Hospital in 1991.
Often overlooked in policy debates, nonprofit institutions like the Aaron Diamond Center have long been major players in biomedical research. The Howard Hughes Medical Institute, for instance, which was endowed by the reclusive billionaire’s fortune and is located around the corner from NIH headquarters, pours more than a half billion dollars a year into basic and applied medical research. The research arm of the practice-oriented Mayo Clinic, which is headquartered in Rochester, Minnesota, spends more than $125 million a year of its own resources on applied research. Like most nonprofits, even the best endowed, the Aaron Diamond Center’s permanent staff also competes for government funds. Ho and his colleagues received major AIDS-related grants from both NIH and New York City. Coupled with the Diamond family’s generous bequest, Ho had the wherewithal to forge an independent relationship with drug companies that were looking for help in testing their new drugs but wanted to steer clear of the government’s AIDS Clinical Trials Group (ACTG).
In early 1994, Ho’s relationship with Kempf at Abbott blossomed. Leonard included the Aaron Diamond Center on his short list of clinics testing the company’s latest and most promising protease inhibitor. Ho and his colleague Martin Markowitz began testing ritonavir in twenty AIDS patients. In designing the trial, they relied on a brief first trial that had been conducted in Europe. Like all first trials, the drug had been tested alone to determine its safety and an appropriate dose. How much can be given before intolerable side effects kicked in? How much needed to be taken to have an effect on the virus? How often must it be taken to maintain an appropriate level in the blood throughout the day? Most early protease inhibitors had to be taken in large amounts two or three times a day because they were cleared by the liver soon after entering the bloodstream.
The patients in the second-stage trial at Aaron Diamond were given fairly high doses of ritonavir. A few months into the trial, the clinicians began pouring over the data to get the first hints of the drug’s effectiveness. They were aided by the recent arrival of the first tests that could measure the amount of HIV in the blood. Previous tests to gauge the progression of the disease had measured either the level of disease-fighting white blood cells or the antibodies that fought HIV. Now scientists could actually measure viral load.
The results from that first U.S. trial for ritonavir stunned Ho and his colleagues. None of the earlier AIDS drugs had shown such rapid and dramatic ability to clear the virus. “[Of] the first twenty patients we treated, each one had their viral load drop fifty to one-hundred-fold in the first two weeks,” he recalled. “That was a crucial moment for us in this whole field because it showed how potent protease inhibitors could be.”4 Abbott also tested ritonavir at the University of Alabama at Birmingham, where Shaw was turning up similar results. Danner in Amsterdam also participated. By April, the data were streaming into Abbott Park. The drug worked—spectacularly so. But hidden in the data were hints of disaster. Patients rapidly developed resistance to ritonavir, just as they had to AZT and the other nucleosides before it.
While Abbott massaged its data, Ho and his colleagues pondered the implications of their limited test results. Markowitz’s test-tube experiments on ritonavir had shown that HIV was capable of mutating around the protease inhibitor. Though it required several simultaneous mutations along the protease’s genome before it escaped the drug’s blocking action, it happened relatively quickly, the test-tube results showed. The Aaron Diamond Center team wasn’t surprised when the same thing happened in humans. What did surprise them was the rapid escalation in the levels of HIV in the blood once the mutant strain emerged. For years, scientists had believed the AIDS retrovirus hid in the bloodstream and the lymph nodes at fairly low levels until it somehow broke out and overwhelmed the immune system. But the rapid reemergence of HIV after an initial knockout punch delivered by the protease inhibitor suggested a very different scenario.
Ho, never a math whiz, sent his data to Alan Perelson, a biophysicist at Los Alamos in New Mexico. Ho had gotten his name from a scientist in his lab who had met Perelson at a scientific meeting, one of the chance encounters that often start significant scientific collaborations. After crunching Ho’s numbers on a spreadsheet, Perelson delivered his verdict. HIV-infected cells weren’t part of a guerrilla army hiding in the hills. They were waging a Gallipoli-style campaign in the bloodstream, sending wave after wave of the proliferating retrovirus against the immune system. Perelson estimated that HIV reproduced at a rate of 680 million viral particles a day.
The finding had two major implications. First, the immune systems of HIV-infected individuals waged a valiant and largely successful fight to clear the virus that went on for years. It was only when the immune system gave up the fight (for as yet unexplained reasons) that the HIV army overwhelmed the carrier. Second, since the virus was reproducing so quickly, drug-resistant strains were bound to emerge. Indeed, given HIV’s replication rate, they would emerge very quickly. Scientists studying the process of genetic transcription during reproduction have known since the early 1970s that every time a genome reproduces, the offspring has at least one mutation somewhere along the strand of chemical base pairs that makes up its genetic code. Most are benign since they either do not affect a gene’s function or occur at places on the genome that don’t have a function. But the HIV genome was only ten thousand base pairs long. With 680 million new viral particles entering the bloodstream every day (a number since raised to more than a billion), Perelson believed every base pair on the HIV genome was mistakenly transcribed not just once, but thousands of times a day. That meant that a mutant that could escape a single drug’s action and still maintain its essential function was almost a certainty. Indeed, it probably happened within hours of the drug’s introduction into the bloodstream.5
That summer, Ho attended another of the many annual scientific conferences aimed at AIDS researchers, this one in Seattle. Over a dinner arranged by Dani Bolognesi, a senior researcher at Duke University who had helped Burroughs Wellcome conduct the first trials on AZT in the mid-1980s, Ho and Shaw of Alabama compared notes on their single-drug trials. “We looked at each other and realized we had made the same observation,” Ho recalled. The retrovirus reproduced so fast and mutated so often that it was going to take a number of drugs to knock it out. A single drug aimed at a single target was doomed to failure. But hitting it with drugs that effectively blocked a number of targets simultaneously would make mutant survival extremely difficult. Indeed, according to Perelson’s computers, it was almost a mathematical impossibility. “The odds of making five or six mutations simultaneously are one in a trillion,” Ho said. “Those are odds we like.”6
Their job now was to tell the world. In the fall of 1994, the Interscience Conference on Antimicrobial Agents and Chemotherapy, the major annual conference of the American Society for Microbiology, gathered in Orlando, Florida. Ho and Shaw were both slated to give major addresses to a plenary session on emerging AIDS therapies, and both wanted to release their latest data. Abbott’s Pernet and Leonard flew in from Chicago. Ho asked if he could present his findings. Leonard hesitated. Abbott was still debating how to move forward with the final clinical trials needed for Food and Drug Administration (FDA) approval. The senior company scientist still wanted to know if ritonavir could work as monotherapy if given at a higher dose. Ho said no. “Everybody was looking for a home run, but we didn’t put a lot of stock in that,” Ho recalled. Abbott gave the go-ahead to release the study data at the meeting and began laying plans for ritonavir’s final clinical trials.
Companies with promising new drugs had one obvious incentive for working with the ACTG on clinical trials. It saved them a lot of money. In an ACTG trial, the company often did little more than supply the drug. The government-funded clinicians at academic medical centers recruited the patients, paid for their care and blood tests, and kept track of the records. These clinical tasks were by far the costliest component of clinical trials, and the AIDS program of the National Institute for Allergies and Infectious Diseases (NIAID) spent hundreds of millions of dollars a year supporting them. However, there were trade-offs for pharmaceutical firms that chose to cooperate with the government, which took place at both ends of the experiment. On the front end, the company didn’t have the final say in designing the trial (although as a practical matter, academic researchers worked closely with company doctors on trial design because they wanted to maintain good relations for future access to drug candidates coming out of their labs). On the back end, the company had no control over the release of a study’s conclusions.
Hoffman–La Roche’s clinicians gladly accepted those limitations, given top management’s decision to veto more money on the search for a better protease inhibitor candidate or new trials that might escalate the acceptable dose of their poorly absorbed drug, saquinavir. Indeed, by the time saquinavir entered clinical trials through the ACTG in 1993, the company had several years’ experience working with government-funded clinicians. Between 1990 and 1992, the ACTG conducted Roche’s trials for ddC, the nucleoside reverse transcriptase inhibitor that the company had licensed from NCI.
Those trials took place in an environment that had been radically transformed by nearly six years of AIDS activism, which dated from playwright Larry Kramer’s fiery speech in a Greenwich Village church in March 1987. Over those six years, angry AIDS activists had forced the most dramatic liberalization in FDA rules in the agency’s history. Most of these changes were welcomed by the drug industry since they hastened approval of new drugs. But liberalization had unwanted side effects. The activists simultaneously blasted open the opaque process for designing ACTG trials, won expanded access to drugs while in clinical trials, and held up every corporate decision that might affect the health of AIDS patients to public scrutiny. Those were changes the industry could do without.
As early as June 1987, FDA commissioner Frank Young had unveiled new regulations in direct response to the exploding AIDS crisis and the fact that dying AIDS patients across the country were clamoring to get into the recently concluded AZT trials. Young’s new rules encouraged companies that were testing experimental AIDS drugs to give them to the doctors of desperately sick patients who weren’t lucky enough to be part of the trials. The policy became known as compassionate use or expanded access and set off grumbling among traditionalists within the agency and outside safety advocates. Both groups feared that unproven and potentially deadly drugs would be unleashed on a population that was willing to swallow almost anything to stay alive.
Sixteen months later the agency loosened its rules another notch. When evaluating new drugs that tackled life-threatening diseases like AIDS, the FDA would grant so-called accelerated approval if a company presented positive results from a limited set of first- and second-stage trials. The only quid pro quo for a company seeking accelerated approval was a promise to conduct postapproval studies to document the drug’s impact once in general use.
The first ddC trial, headed by Margaret Fischl at the University of Miami and Douglas Richman at the University of California at San Diego, threw a third liberalization into the mix. ACTG 106, which was conducted during 1989 and 1990, was designed as an early-stage study to determine what doses to use when combining ddC with AZT. AIDS patients, especially those who had been on AZT and were failing, were exhilarated by early results from that trial, which were unveiled in San Francisco in June 1990. Patients on the two-drug combination who had entered the trial with severely compromised immune systems saw their count of infection-fighting CD4 cells double within days. John James, writing in AIDS Treatment News, an activist newsletter, noted that the trial “had established a de facto standard of care among many of the best-informed patients and physicians.”7
The problem from the FDA’s point of view was that measuring surrogate markers like CD4 cells didn’t tell the agency anything about the long-term effectiveness of the drug being tested. Did it delay death? Did it delay the onset of opportunistic infections? Surrogate markers were used elsewhere in medicine. Drugs that lowered blood pressure or cholesterol to reduce the risk of heart disease had been approved on that basis. But their approvals were based on long, costly trials that established the connection between the surrogate marker and the progression of the disease. Many AIDS activists wanted approval for ddC and ddI, Bristol-Myers Squibb’s nucleoside that was moving through the approval process at about the same time, based only on the improvement in the CD4 cell marker. Their position was endorsed by NIAID in February 1991 and given regulatory sanction that summer when the FDA, now under David Kessler, approved ddI for use after a divisive public hearing of the agency’s Antiviral Advisory Committee.
A year later the FDA approved Roche’s ddC using just surrogate marker data. However, the FDA overruled the company’s request that the new drug be approved for monotherapy, which Roche had requested. Roche failed to show that using it alone was any more effective than AZT alone. Still, the transformation of the drug approval landscape was complete. In late 1992, the FDA codified final accelerated approval regulations in the federal register. The rules stated a drug could be approved based on surrogate marker data (for AIDS drugs, the best surrogate marker at that time was the CD4 changes) if it was “reasonably likely” the changes would predict clinical benefit. “We cannot wait for all the evidence to come in when people are suffering and dying from these devastating diseases,” Kessler said.8
While activists were pushing the FDA to loosen its drug approval criteria, they were also attacking the NIAID and the ACTG for failing to get more drugs into the pipeline. Their repeated protests outside NIH headquarters in Bethesda convinced Fauci and Hoth to open up their grants process. In the early days of the AIDS fight, the ACTG had followed the traditional NIH pattern in approving new trials. Investigators proposed studies, and secret panels made up of their peers approved their funding. But in late 1989, Fauci agreed to open all ACTG meetings to the general public and put several AIDS activists, including Mark Harrington of the Treatment Access Group, a spin-off from the New York chapter of ACT UP, on the governing board. By the time Keith Bragman and Miklos Salgo of Roche approached the ACTG to conduct a second-stage trial on saquinavir, Harrington had emerged as the AIDS community’s most trenchant critic of drug companies’ efforts to manipulate the new rules to their own advantage. The Harvard-educated freelance scriptwriter, who was HIV-positive, had also learned enough science to become an astute observer of the medical implications of clinical trial design.
The Roche team faced an impossible situation as it entered negotiations with the ACTG for testing saquinavir. The first-stage studies in Europe had not established a maximum tolerated dose for the drug because Roche’s top management had cut off funding for further tests. Patients in those early trials were already taking eighteen capsules a day, yet the drug’s level in the blood remained low because of its poor absorption, estimated to be about 4 percent. Could patients tolerate even higher doses without crippling side effects? No one knew. It wasn’t a minor question. Giving someone a suboptimal dose of the drug might allow the rapid emergence of resistance. Nevertheless, the company proposed studying the new drug in combination with AZT and in combination with AZT and ddC, which by the fall of 1992 were the generally accepted standards of care for advanced AIDS patients.
The ACTG committee that approved clinical trials met in mid-1992 to discuss the proposed trial, which became known as ACTG 229. Harrington immediately began raising questions about the design of the study. Why was the dose so low? Would a patient who failed a therapy regimen containing saquinavir have bred a virus in his or her body that would be resistant to other protease inhibitors when they came along? The University of Washington’s Collier, who was the study’s principal investigator, briefly left the room with the Roche representative to discuss the company’s earlier European studies, whose results had not been fully released to the committee.9 “We wanted to make sure there’s an effective dose, shown to have activity,” said Collier. “A small group of leaders of the ACTG were shown confidential data that proved we had an adequate dose that had activity. In the end, there is a limit to the number of pills any one person can take. The dose was subjected to a lot of scrutiny—by me, by the ACTG, and by Roche itself.”10
The committee voted to allow ACTG 229 to move forward. Collier and her colleagues enrolled 302 patients between March and July 1993 for a twenty-four-week trial that used surrogate markers to evaluate the various drug regimens. New tests that had recently come on the market allowed the clinicians to measure not only the level of CD4 cells in the blood, but the patient’s level of HIV. These viral load tests, as they became known, were considered a superior indicator. The preliminary results weren’t released until May 1994 and showed slightly better results in the three-drug combination than with either AZT alone or in combination with ddC. “While the results weren’t fabulous compared to what came later, it was clearly better than we had at the time,” Collier later said.
Roche management’s failure to fund tests of higher doses of the drug may have saved the company money in the short run, but its long-term costs were enormous. Because ACTG 229 used a suboptimal dose of saquinavir, the scientific community lost a golden opportunity to obtain early knowledge of the potential effectiveness of triple-combination therapy.
It wasn’t just bad science, it was bad business since the company lost its first-mover advantage. In the drug industry, the first company to develop an entry for a new class of drugs—assuming its entry is effective—almost always gains the lion’s share of the market. But as subsequent tests would show, saquinavir in its original formulation was not a very effective drug because of its poor bioavailability.
Fearing Roche would use ACTG 229 to pursue accelerated approval at the FDA, a coalition of AIDS activist groups centered on the East Coast asked the agency to hold off on rapid approval for saquinavir. “We feel that such an approval would penalize people with AIDS/HIV by setting an inappropriately low standard of evidential requirements that would govern the regulation of this entire class of therapies,” they wrote in a joint letter to Kessler. Pointing to Roche’s failure to conduct follow-up studies on ddC after its rapid approval in 1992, the activists complained that “we have learned through difficult experience that we cannot depend on the goodwill of pharmaceutical industry sponsors to produce the information that is necessary to make life-or-death treatment decisions.”11
The letter set off a firestorm of controversy within the AIDS activist community, stoked in part by Roche’s decision to fax copies of the letter to other AIDS groups across the country. West Coast activists in particular accused the Treatment Access Group and its allies of denying lifesaving medications to people with AIDS, who were clamoring to get their hands on the new drug. The FDA, accustomed to vilification for going too slowly, was stunned by the opposite charge, especially after a headline appeared over an August 1994 cover story in Barron’s magazine asking, “Do We Have Too Many Drugs for AIDS?” The agency held a two-day hearing that fall, giving industry officials and anxious AIDS activists a chance to attack Harrington and his allies publicly. There would be no change in the FDA’s policy of accepting surrogate markers when approving new drugs for AIDS.
Kessler’s decision was defensible at the time and later proven scientifically correct. But politics played a huge role in the decision. That fall the American political scene underwent an upheaval the likes of which hadn’t been seen in several generations. Republicans led by Rep. Newt Gingrich of Georgia won control of the House of Representatives. One of the core planks in their “Contract with America” called for dismantling the regulatory burdens on business. The drug industry played a key role in fueling those changes. Its famous Harry and Louise ads—“Keep the government out of our medicine chest”—had helped turn public opinion against the Clinton administration’s health care plan, which included elements designed to control the rising cost of drugs.
The industry also engaged that fall in a massive lobbying campaign to convince NIH to eliminate its reasonable pricing clause on therapies developed with government funds. Academic researchers complained they were being cut off from the most promising new drugs being developed in industry’s laboratories. Responding to their pleas, a new NIH director, Harold Varmus, held two hearings on the subject in the second half of 1994. Industry representatives crowded the docket with testimony suggesting the mere threat of price controls was keeping its scientists from cooperating with their government-funded counterparts at NIH.
In April 1995, Varmus ordered his technology transfer office to eliminate all references to reasonable pricing in future NIH agreements with the pharmaceutical industry. “The pricing clause has driven industry away from potentially beneficial scientific collaborations with Public Health Service scientists,” Varmus said in a prepared statement. “One has to have a product to price before one can worry about how to price it.” Democrats decried a move they said would open the floodgates to price gouging. But there was nothing they could do about it. “No one is oblivious to the political climate,” commented Senator Ron Wyden, a Democrat from Orgeon.12
But the change came too late to corral Roche back into the government program. In the summer of 1994, after consulting with the FDA about the results from ACTG 229 and realizing it wasn’t going to get accelerated approval, the company began planning for the final third-stage trials. In two trials in the United States and Europe, company-paid physicians recruited more than four thousand AIDS patients for a final study of saquinavir’s efficacy. Patients in just one of the studies’ seven arms received three anti-AIDS drugs. By early 1995 the tests were underway. To assuage its critics in the AIDS community, the company agreed to a generous expanded access program, offering the drug to an additional four thousand patients through a lottery. “There have been improvements in the manufacturing and the number of steps required [has been reduced],” Jürgen Drews, Roche’s president of research, told the editors of Gay Men’s Health Crisis’ Treatment Issues. “We now think we can produce the drug in sufficient quantities and at a price that is feasible.”13
The company presented its preliminary results at the FDA’s antiviral advisory committee meeting on November 7, 1995. By this time, the AIDS field was well aware how quickly resistance could emerge when patients were treated with a single drug. Ho and Shaw had presented their data on viral replication at a major conference in Washington early in the year. Martin Markowitz of the Aaron Diamond Center presented data from his ritonavir trials showing the rapid emergence of resistance to the drug in monotherapy and, alarmingly, those strains’ widespread cross-resistance to other protease inhibitors then in development. The answer was plain: The drug had to be used in combination with other drugs. “One drug won’t work. Two drugs won’t work. Three drugs might work. Four drugs can cure,” Markowitz reported.14
To counter the mounting evidence against monotherapy, Roche began releasing information from its ongoing follow-up studies with patients in its earlier trials. At a July conference in Sardinia, Italy, a company physician suggested to AIDS activists that saquinavir did not cause resistance for at least a year after beginning monotherapy or in combination with other drugs, and those new strains of HIV were not cross-resistant to other protease inhibitors.15 Indeed, when the company went before the FDA advisory committee in early November, it did not present any data from its ongoing triple-combination study. It relied instead on the smaller of its two third-stage trials, which tested saquinavir as monotherapy and in combination with ddC, the company’s other AIDS drug. The company’s application asked for approval as either a monotherapy or in combination with other already approved anti-HIV drugs. “It was the first HIV protease inhibitor,” said Bragman. “It was the first of a new class of drugs. We really had no idea what was going to happen in that hearing.”16 The committee recommended against monotherapy.
On December 7, the FDA approved saquinavir, a little over three months after Roche had submitted its new drug application. It was the fastest turnaround in the agency’s history. But the agency, following its usual custom of hewing closely to the advisory committee’s advice, turned down Roche’s request for monotherapy. The approved label said the drug should only be used with one or more already approved nucleosides and preferably ones that a patient had never used before.
FDA’s interactions with pharmaceutical companies during a drug’s development are not public records since companies demand privacy to keep their test results hidden from potential competitors (and the prying eyes of Wall Street if tests go poorly). But drug approval letters often contain a review of their developmental history. The FDA reviewer of saquinavir dated the agency’s concern about the drug’s poor bioavailability from the initiation of human trials. The reviewer recalled how the agency had suggested that Roche try saquinavir at higher doses, which would probably require a new formulation. Roche had rejected those pleas as it moved into its final clinical trials. It told the federal agency that it was “not feasible to either commercialize or study greater doses” because of “manufacturing limitations.”17
Roche’s official comments to the FDA were radically at odds with its public posture. At the same time that it was kicking off those trials, Drews was telling AIDS activists that the company had overcome the manufacturing problems. And during 1995, the year of the final trials, the company had enough capsules of its drug to expand access to more than four thousand patients.
Technical difficulties and the cost of manufacturing had indeed been a major issue for drug companies when they started their research programs on protease inhibitors. The drugs are large, complicated molecules. But the companies rapidly mastered the art of how to produce protease inhibitors in large quantities and at reasonable costs. That there are no inherent difficulties in manufacturing the drug would be proven in the late 1990s when Cipla, a drug manufacturer in India, began producing its version of Merck’s indinavir for less than one thousand dollars a year.18
Yet those facts remained hidden from public view when Roche got its FDA approval letter for saquinavir. The next day, the company began wholesaling the drug at a price of $5,800 per year per patient. Martin Delaney, the head of San Francisco’s Project Inform, the AIDS activist group with much closer ties to the drug industry than its East Coast counterparts, told the Wall Street Journal he was “appalled” by Roche’s price, which was three times higher than any AIDS medicine then on the market.19
As it launched its final round of trials on ritonavir, Abbott continued to steer clear of the ACTG. It also refused to cooperate with other drug companies. In mid-1993, a consortium of drug companies pursuing AIDS-related research launched the Intercompany Collaboration (ICC). The idea was to test combinations of AIDS drugs. Sixteen drug companies eventually joined the group. Though it received a lot of press attention when launched, the ICC never made a significant contribution to the field. Neither Abbott’s nor Merck’s protease inhibitors were included in the ICC’s first major test, which was initiated in June 1994. Instead, the intercompany collaborators used the same inadequate dose of Roche’s saquinavir that was used in the triple-combination arm of Roche’s own trial.
The research officials at the companies involved in AIDS research were publicly committed to the idea that an effective AIDS regimen would require cooperation. But their competitive instincts and the internal imperative to pursue big-selling drugs (or turn every drug under development into a big seller) precluded meaningful collaboration. “That was an organization where pharmaceutical companies could look like they were collaborating because it looked like AIDS required collaboration,” Abbott’s Leonard said much later. “But people didn’t share the information that mattered because people are highly competitive. I never went to an ICC meeting. People who went from other companies told me it was a huge waste of time.”20
Instead, Abbott ran its independent third-stage ritonavir trials in various clinics around the world. The dose-ranging studies conducted in Europe during late 1994 had shown the drug had impressive antiviral activity. But three to four months after the onset of monotherapy, the viral loads in patients returned to their original levels, which suggested the virus had mutated around the drug. Despite those results, only Ho’s admonitions finally convinced the company it had to pursue combination therapy in addition to its preferred monotherapy trials.
The company’s final two trials tested nearly fifteen hundred patients with either ritonavir alone or in combination with the patient’s existing therapy, which for most meant AZT. The company also initiated a third study in France to test ritonavir in combination with two nucleosides, AZT and ddC. But they enrolled just thirty-two patients. This small trial provided the company, and eventually the world, with the first dramatic evidence of the success of triple combination therapy.
The company presented its preliminary results for the tiny French trial at the Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) meeting in San Francisco in the fall of 1995. The results wowed the AIDS activists who attended the meeting. The company that had broken free of the government and then labored in secrecy for more than two years went public with the most startling results in AIDS therapy to date. Daniel Norbeck, then only thirty-seven years old, presented the data. Norbeck, who had received his undergraduate chemistry degree from Wheaton College, an evangelical Christian school in the western suburbs of Chicago, before earning his doctorate from the California Institute of Technology, reported that more than half the patients in the trial had cleared 99.9 percent of the virus from their bloodstreams. “The reservoirs of virus are emptying,” Norbeck declared. The reaction among activists was a mixture of joy and disbelief. “It is debatable whether Dr. Norbeck can claim such success on the basis of one small study,” one activist said.21
Norbeck, who several years later would become head of Abbott’s pharmaceutical research division, had other good news for the meeting. What had been considered a major flaw in ritonavir—its propensity to impede the liver’s ability to clear many drugs from the body—actually might turn out to be a major plus. When mixed with saquinavir, a small amount of ritonavir (eventually given the trade name Norvir to honor Norbeck’s role in the latter stages of its development) kept the Roche drug in the bloodstream of rats long enough for it to have its potent antiviral effect. The company announced plans to begin testing the combination in humans immediately.
While that testing would eventually prove a boon to the marketing of ritonavir, it wouldn’t be part of the data that Abbott presented to the FDA’s antiviral advisory committee on the last day of February 1996. The pivotal trial presented to the committee added either ritonavir or a placebo to patients already on drug therapy. That was either AZT alone or AZT and ddC. The company did not report on the subgroup on triple-combination therapy. After twenty-four weeks, the death rate and onset of opportunistic infections among those patients given ritonavir alone was half that of the placebo group, Abbott officials told the federal reviewers. The second trial presented to the reviewers, which did not contain any patients on triple combination, had peculiar results. Patients on ritonavir alone did better than those taking both AZT and ritonavir. Abbott admitted that “patients on combination therapy may have been less compliant with their drug regimen,” the FDA’s reviewer wrote. “This may have explained the poorer performance of combination treatment in this study.”22
The next day, the FDA granted Abbott a license to sell ritonavir. The agency’s approved label said the drug could be given either as monotherapy or in combination with existing antiretrovirals. Two days later, Abbott began shipping the drug to pharmacies across the country. It cost sixty-five hundred dollars for a year’s supply. The retail price hit a stunning eight thousand dollars a year—a price not seen on a single AIDS drug since the early days of AZT. Combination therapy, the new coin of the AIDS therapy realm, soared to nearly fifteen thousand dollars a year.23
The dream of a monotherapy knockout punch died hard at Merck, too. The initial reports on indinavir’s first clinical trials, which took place in late 1993, sent hopes soaring at the company’s research headquarters in West Point, Pennsylvania. The drug had a “marked antiviral effect in monotherapy.” Its ability to lower viral loads immediately and its bioavailability suggested indinavir was going to be a superior drug to the competitive products being cooked up by Abbott and Roche. “We were delirious,” research director Edward Scolnick later noted. “We absolutely thought we had the cure for AIDS: monotherapy.”24
But through the early months of 1994, the handful of patients who were on indinavir began exhibiting the same pattern that had affected every AIDS drug that came before. A few months into therapy, viral loads began creeping back up to where they had been before the drug treatments began. Emini and Jeffrey Chodakewitz, the Merck physician in charge of clinical trials, suspected that the AIDS virus had once again figured out a way to mutate around the best efforts of scientists to control it. Yet there was one mysterious patient, dubbed patient 142, a “wise-cracking, forty-one-year-old law student from Philadelphia,” whose viral load had gone down and stayed down.25 Perhaps his story contained the key to understanding how indinavir knocked out the AIDS virus. Merck needed to keep the tests going.
Late one evening, just as Scolnick was sitting down to dinner with his wife, Emini called him with the bad news-good news story. The virus appeared to be mutating around their new drug. But one patient seemed to be doing well. Scolnick heard only the first part of the message. His dinner ruined, Scolnick immediately went over his own scientists’ heads to get a second opinion. He called Anthony Fauci, the head of the NIAID and the government’s leading AIDS expert. After explaining their results, Fauci confirmed the bad news. “Ed, you’ve got resistance,” he said.26 Merck’s research team once again faced the prospect that it had come up empty.
For weeks Merck scientists batted ideas back and forth about how to keep the protease inhibitor project moving forward. The fact that one patient was improving on indinavir alone suggested to some that the dose may not have been high enough. Although they didn’t have access to David Ho’s data, Merck scientists also realized that the rapid emergence of resistance meant the virus was reproducing at an extraordinarily fast rate. Perhaps, they finally admitted, they should hit it with two or more drugs at once. They won management approval for a second set of clinical trials to test the drug at higher doses and in combination with other antiretrovirals. “Because of 142, some of us argued that [indinavir] was doing something right,” Emini later told the Wall Street Journal. “We finally decided to go forward, to try some new things, to up the dosage and combine [indinavir] with other drugs. If not for 142, it’s possible we might have ditched the whole thing.”27
These second-stage trials began recruiting patients at nine sites around the United States in mid-1994. Company officials met with AIDS activists that summer to bring them up to date on the program. Merck sent a clear signal at the meeting that they now had modest hopes for the drug. The company revealed it wouldn’t pursue accelerated approval under the new FDA guidelines. They also rejected pleas for an expanded access program for desperately ill patients. “The company’s position is that, at this time, [indinavir’s] potential does not appear great enough to merit what it claims is the substantial investment in rapidly expanding production,” an activist newsletter reported.28
Merck was a company in transition by the time these second-stage trials got underway. Roy Vagelos, who had spent a decade at NIH and in academia before moving to Merck, was slated to step down from his chief executive post in June. Research director Scolnick would soon be reporting to a new man.
Fears swept through the AIDS activist community that the company with a social conscience—the one that had been the most open to their incessant demands for access and information—would soon be cutting back on its research programs as a result of the changeover. In April, outgoing chief executive officer Vagelos gave an exclusive interview to Gay Men’s Health Crisis’ newsletter to diffuse the issue. The mixed early results from the indinavir studies must have weighed heavily on his mind. The company had gotten involved in AIDS research in 1987 with a promise to develop a breakthrough within five years, he said. Yet here it was, seven years later, and the company still had nothing to show for the “several hundreds of millions of dollars” invested in AIDS research. But that didn’t matter. “Our company is dedicated to it. It doesn’t matter who succeeds me,” he told the newsletter editors. “We don’t like to quit just because things aren’t going well.”29
His replacement, Raymond V. Gilmartin, would make the critical “go/no-go” decision on indinavir. Gilmartin, an engineer with a master’s of business administration from Harvard, joined Merck after a successful career guiding Becton Dickinson to the top of the medical supply industry. The board of directors hired him for his reputation as a brilliant business strategist and because he was a professional manager who could keep an eye on the company’s stock price at the same time it pursued global markets, a new emphasis at the firm. Compared to his predecessors at the nation’s most admired drug company, he knew almost nothing about the intricacies of pharmaceutical research. What he did know about was pricing and profits, and like the rest of the drug industry, Merck’s officialdom was embroiled in 1994 in the ongoing war with the Clinton administration over health care cost containment. The pause in economic growth that spread across the economy that year had also slowed the growth of company sales. The company’s stock price was half of what it had been just two years earlier.
Fortunately for the AIDS research team, by the time Gilmartin began focusing on its project, Scolnick had better news to report. The early results from the two-drug combination showed 40 percent of patients had lowered their viral load below the level of detection. While this was nowhere near the results that would later be shown with triple-combination therapy, it was good enough to encourage the research staff to press for third-stage trials. These were the large-scale tests, often involving a few thousand patients, that when successful led to FDA approval. Few companies undertake third-stage trials unless they have a pretty good indication from earlier tests that larger trials will be successful.
Merck had its own approach to third-stage trials. The firm almost never launched one without building the facilities needed to produce the drug in bulk. Their reasoning—some called it arrogance—was that nothing they took that far would fail, so they might as well be prepared to hit the ground running when they got final approval. Merck engineers estimated costs would run anywhere from $100 million to $200 million to build the facilities needed to synthesize indinavir in bulk. Chodakewitz, the physician in charge of the clinical trials, presented the data from the second-stage trial at a January 1995 meeting that reassured the company’s top officials that the money wouldn’t be wasted. Gilmartin, who attended, gave the nod to move forward, both with the final trials and production of a new plant in Elkton, Virginia.
Between April and December 1995, Merck enrolled just ninety-seven patients in what turned out to be its crucial clinical trial. A few months earlier, Ho and Shaw, who were working with Abbott’s ritonavir, had stunned the AIDS research world with their findings about viral replication and its implications for combination therapy. Merck scientists, who claim they independently came to the same conclusion, designed the critical trial to evaluate indinavir alone against the most popular two-drug combination at the time—AZT and 3TC. It also added a wing to the trial that tested all three together. Dying AIDS patients across the country clamored to get into the trial, dubbed Merck 035. Barring that, they wanted access to the drug. The company, fearing it wouldn’t have enough of the drug for its other trials, which included a thousand-patient indinavir-AZT trial being conducted in Brazil, initially resisted. It invited AIDS activists to a presentation at its corporate headquarters, where they detailed their production concerns. But as the new plant got up to speed, those concerns melted away. By late summer, more than two thousand AIDS patients were taking the drug.
Merck, still pursuing traditional FDA approval, anticipated a long, leisurely rollout of the new drug. Their original target date was late 1996. The company knew its drug was vastly superior to the poorly absorbed saquinavir, and the early reports of toxicities and drug interactions from Abbott’s ritonavir led them to believe it wouldn’t be a serious contender. More important, the initial reports on Merck 035 suggested the nearly ten years that Merck scientists had spent on AIDS research had paid off in spectacular fashion. The patients in the triple-combination therapy wing of the trial had nearly undetectable viral load counts. Merck scientists left the meetings where the preliminary results were unveiled in tears. Some compared the moment to Merck’s famous triple-therapy trials in the late 1940s that had provided the first breakthrough against tuberculosis.30
It took outsiders to shake Merck out of its complacency. In the middle of September, Abbott went public with the stunning results from its triple-combination study. Abbott also indicated at the San Francisco ICAAC meeting that it would soon seek accelerated approval for ritonavir. Two months later, an FDA official asked Merck’s chief of regulatory affairs, “What are you guys waiting for?”31 Merck filed for accelerated approval for indinavir on January 31, 1996.
The day after filing, Emini gave the leadoff presentation to the annual conference devoted to the latest in AIDS research. He called for a “new paradigm” for treating the disease. Shutting down HIV replication could eliminate or delay the accumulation of mutations leading to drug resistance, he said, and three-drug combinations that included indinavir had succeeded in lowering viral loads below the level of detection. The FDA advisory committee meeting in early March was a mere formality. The company received its license to begin selling the drug on March 13. It was the fastest approval in agency history.
Though clinical tests showed Merck had the most effective drug with the least side effects, it had the disadvantage of being the third entrant into the protease marketplace. The company established its initial price below its rivals—around six thousand dollars per year retail. The company put a humanitarian spin on its pricing strategy. “We saw what Roche and Abbott did with pricing, and we felt that we had to do something different,” a public relations official said. “We had to make a statement that we recognized that access is an issue.”32
The next year reshaped the AIDS treatment landscape in ways unimaginable only a few short months before. The new drugs transformed the number-one killer of young male adults in America into a manageable disease. Time and Newsweek magazine covers heralded the end of the AIDS plague. The nation’s airwaves crackled with miraculous recovery stories, where young men that only a few months earlier had been on the brink of death were shown pumping iron to rebuild their once-wasted bodies.
The media lavished most of its praise for the breakthrough on the pharmaceutical companies that brought the new drugs to market. Michael Waldholz, the Pulitzer Prize-winning reporter for the Wall Street Journal who had followed the AIDS story longer and more closely than any reporter in the United States, set the tone. His front-page story in June 1996 emphasized the corporate mad dash to the finish line after “a decade of disappointment and frustration, unexpected product failures, intense corporate rivalries, and secrecy.” His account made no mention of the government’s significant scientific and financial contributions to basic and applied research that took place from the early 1980s through 1992, nor of its support for many of the clinical trials that took place in the two years leading up to the drugs’ approvals.33
The alleged cost of private industry’s research investment mushroomed almost as quickly as media’s coverage of the amazing breakthrough. Two years after Merck’s chairman had spoken about spending “several hundreds of millions” on all AIDS research, a Wall Street Journal reporter led a story about the marketing of indinavir by claiming the company spent more than $1 billion to develop the drug.34 The Washington Post Magazine a few months later raised the financial stakes even higher when it declared the protease inhibitor “drama was a decade-long, multibillion-dollar race pitting the world’s top drug companies against AIDS and one another—arguably the most time, money, and scientific manpower ever focused on a single medicinal target.”35
Large pharmaceutical companies do not report their research and development costs by individual drugs. They do not even break them down into areas of research, like, say, AIDS or cancer or heart disease. But the Merck example is instructive since it had by far the most comprehensive research program on HIV- and AIDS-related diseases and its officials have not been shy about releasing data about their expenditures. The company began a multifaceted research program on AIDS in the mid-1980s, which pursued vaccines, non-nucleoside reverse transcriptase inhibitors (NNRTIs), protease inhibitors, as well as novel pathways for attacking the virus. The company conducted a number of early-stage clinical trials on possible drug candidates and by 1995 had pushed indinavir to the point where it was ready for a final, multicenter clinical trial involving at least a thousand patients. The company also deployed its renowned vaccine research expertise against HIV. It is fair to say that no company devoted more of its resources to the fight against AIDS, and it is likely that no other company spent even half of Merck’s total.
Business historian Louis Galambos of Johns Hopkins University, who had access to internal Merck memos and interviewed all its key players for an in-house history of indinavir’s development, wrote that officials told him that Merck had spent $400 million on all of its AIDS research programs by early 1995. This was not just for the development of indinavir. It was for all of its research and development into AIDS: drugs, both the one that worked and the several that failed; vaccines; and the manufacturing facilities for the drugs sent into clinical trials. At that point they finally had a drug—indinavir—ready for its final clinical trials, and they were faced with the “go/no-go” decision on building a $200-million plant to produce the new drug in sufficient quantities for both the trials and its eventual marketing. It got built. They also began their final clinical trials.
Clinical trial costs have been well documented, both by the government, which ran the ACTG, and by outside groups, who are interested in promoting research into new drugs to combat the diseases devastating the developing world. The Global Alliance for TB Drug Development, which operates on major grants from the Rockefeller Foundation and the Bill and Melinda Gates Foundation, during 2000 asked consultants from private industry to put together a study on the cost to develop new drugs to combat the disease. The Global Alliance sponsors research into new drugs to fight virulent strains of drug-resistant tuberculosis that have emerged in many parts of the world, since the pharmaceutical industry has abandoned the field. Their report, issued in October 2001, estimated that all phases of clinical research for a new drug cost about $26.6 million, with the lion’s share of that—$22.6 million—coming in the third and final set of trials.36
By taking the final clinical trial-stage estimate—$22.6 million, as of 2000—and adding that to Merck’s total expenditures up to the time it entered the final stages of testing indinavir, one can arrive at a ballpark estimate for Merck’s total AIDS research and development investment before the company succeeded in bringing one antiviral drug to market. That total would have been about $623 million.
As the preceding chapters showed, Merck spent far more than any other drug company in its efforts to combat AIDS. Indeed, Merck spent more than twice as much as any other company involved in the hunt. Abbott consciously operated its program on a shoestring. Its one reported comment on costs came in early 1997, when John Leonard told the Washington Post that by late 1993 the company had spent $75 million on its AIDS program.37 Given that the company at that point was entering its major clinical trial phase, it is unlikely the company spent much more than $100 million on AIDS-related research by the time its successful product—ritonavir—entered the marketplace. Roche, which sharply curtailed its AIDS-related research program in 1993, probably spent about the same amount.
There is another way to estimate the typical research budget for a company investigating AIDS-related drugs. Within three years of the first three protease inhibitors entering the marketplace, the two biotechnology companies whose efforts received significant press coverage over the preceding decade succeeded in getting their drugs approved by the FDA. Unlike the major drug companies, the research and development efforts of these firms are easily documented.
Agouron Pharmaceuticals Inc. was launched in the mid-1980s in La Jolla, California, for the express purpose of using X-ray crystallography and structure-based drug design techniques to develop drugs. It focused the company’s limited staff on two major programs. One was an experimental drug that might inhibit the HIV protease. The other was a chemical agent designed to inhibit solid, malignant tumors in several major cancers. Its protease inhibitor, generically known as nelfinavir, gained FDA approval in March 1997. Its annual Securities and Exchange Commission (SEC) report, filed four months later, showed the firm through June of that year had spent $299 million on all research-and-development projects since its inception. Since both of its major programs had drugs in clinical trials, it is reasonable to assume the company spent about half that, or $150 million, on AIDS-related research to bring nelfinavir to market.
A year later, Agouron told the SEC that during its first year on the market, about 85,000 people in the United States and 120,000 worldwide were taking nelfinavir (the trade name is Viracept). Those AIDS patients generated $358 million in sales for the company. A year later, Pfizer Inc. bought the biotech firm, and separate reports for its research-and-development budgets ceased. But based on the earlier filings, it is safe to assume that Agouron spent less than $200 million on research and development for a drug that in its first year on the market generated nearly twice that in sales, of which more than half (57 percent) was profit.38 In other words, the company earned back its research investment in the first year.
Vertex Pharmaceuticals Inc., the Cambridge-based biotech company headed by Merck exile Joshua Boger, received more journalistic attention during the protease inhibitor hunt than any other firm except Merck. It finally brought a successful product to market in April 1999. The company licensed it to GlaxoSmithKline (the product of several mergers), which began selling it under the trade name Agenerase. Vertex, which ignored HIV research during its formative years and didn’t return to it until the early 1990s, went public in 1992. From 1992 to 1999, the company spent $328.2 million on all its drug research projects, which included major investments not just in AIDS-related drugs but in drugs to combat multidrug resistant cancers, rheumatoid arthritis, and psoriasis.39 It is reasonable to assume that Vertex—like Agouron—spent less than $200 million on developing Agenerase, a total that included the final clinical trials.
A number of companies entered the hunt for a protease inhibitor but dropped out along the way. Only G. D. Searle, later a unit of Monsanto, got far enough along in its program that its drug candidate entered second-stage human trials. But even if a half dozen companies spent $100 million each on protease inhibitor programs, every company directly involved in searching for what turned out to be the breakthrough product that turned AIDS into a manageable disease spent a combined total of less than $2 billion.
It is almost impossible to calculate a grand total for all AIDS-related research by private pharmaceutical and biotechnology firms. But it is reasonable to assume—based on estimates by Merck, which was the outlier in the industry—that the total through the end of 1996 was significantly less than $5 billion. The government, by comparison, spent a shade under $10 billion on AIDS-related research in the decade leading up to the breakthrough discovery. Taxpayer-financed projects ranged from pure science to drug discovery to clinical trials. NIH bureaucrats and university-based peer review panels undoubtedly wasted some of that money—as press accounts based on the angry protests of AIDS activists would occasionally point out. But there can be no doubt that the government program was the driving force behind the private sector’s eventual development of a treatment for the disease.
Was the investment worth it for the drug companies? Did the risk of getting involved in the fight against a seemingly intractable disease require the high prices that the successful companies ultimately set on their drugs? The emergence of protease inhibitors in 1996 ended the era where HIV treatment was a desperate attempt to extend life for a few months or years at the end of the tragic course of AIDS. Triple-drug therapy enabled the majority of people with AIDS to live relatively normal and productive lives as long as they took their medicine regularly. As the annual death toll plunged, the number of people taking HIV antivirals soared, reaching more than a half million by the end of the 1990s.
Not only did sales for protease inhibitors soar, but so did the sales of the existing drugs on the market, whose efficacy had been wondrously enhanced by the addition of the new class of drugs. A successful therapy turned AIDS into a chronic disease and created a permanent market. All the drugs in the cocktail saw their sales soar. By 2000, the protease inhibitor market in the United States was a shade under $1 billion a year, according to IMS Health, the market research firm that closely tracks sales of prescription drugs in the United States. The total market for antivirals reached nearly $3 billion a year. Worldwide sales of protease inhibitors totaled about $2.2 billion, which was again about a third of what had become a $6.5-billion market for anti-HIV drugs.
Profits from selling HIV drugs are easily derived since there are several small biotechnology companies that generate most of their sales from those drugs. If the profit margins of the major companies are the same as the biotech firm whose only sales come from selling HIV drugs, substantially more than half of those total sales fell to the bottom line as profit. In other words, the drug industry each and every year earns before taxes nearly as much as it invested in all of its anti-AIDS research in all of the years leading up to the major breakthrough.
FDA approval of the first three protease inhibitors did not end the fight against AIDS. The HIV strain inside a small fraction of patients remained resistant to the drugs, and over time that fraction was growing. Once again, the wily virus was outwitting the best efforts of medical science to stamp it out.
Doctors recognized several patterns among these therapy resistant patients. Some HIV carriers who did not respond to triple-cocktail therapy had been on single- or double-drug therapies in the mid-1990s. By the time the protease inhibitors came along, they had already developed viral strains inside their bodies that were resistant to the reverse transcriptase inhibiting drugs in the cocktail.
But the biggest part of the problem came from poor patient adherence to the complicated pill-popping regimens associated with the new drugs. In some manufacturers’ formulations, rows of pills had to be taken three times a day, and some people had violent gastrointestinal reactions after taking them. Yet anything less than 95 percent fealty to the regimen allowed mutant strains of the rapidly multiplying virus to emerge. Alarmingly, these new strains not only resisted the action of the protease inhibitor they had been subjected to, but they were resistant to all the protease inhibitors on the market. A government study released in December 2001 estimated nearly half of all HIV carriers had drug-resistant strains inside their bodies.40
John Erickson, who had left Abbott in 1991 to join NCI in Frederick, Maryland, decided to devote his government X-ray crystallography lab to this resistance problem. He had collaborated with his former colleagues at Abbott until management cut the ties. He then stayed in touch with their program by talking to the physicians at the Aaron Diamond Center, who were doing the initial resistance studies on ritonavir. “I knew from the beginning, resistance was going to be a big problem from these drugs,” he told me over a cup of coffee at his home in the Catoctin foothills outside Frederick, Maryland. “I didn’t know the extent to how complex it was going to be. You struggle for years to develop a drug candidate, and the virus mutates around it within days. It was humbling.”
The mild-mannered researcher resumed probing of the scissors mechanism of the HIV protease, looking for the key to the resistance problem. During viral replication, the enzyme cut the long protein chain produced by the HIV genome into small pieces, which would eventually make up the new virus. Protease inhibitors blocked that scissors action by fitting themselves into the folds of the enzyme where the cutting took place. Yet mutant strains emerged that still got the job done. There must be a part of mutant protease that was common to every protease—the fragment that actually did the work. If he could only find that common component, Erickson reasoned, he might be able to design a drug that blocked the protease action in all strains of HIV.
Throughout the mid-1990s, Erickson’s lab at NCI worked with chemists at the University of Illinois at Chicago to synthesize new compounds capable of jamming up the scissors in all strains of the HIV protease enzyme. In December 1994, they came up with their first molecule that worked against every test-tube strain of the virus they had. After filing a patent, they turned to the NIH’s Office of Technology Transfer, whose job under the 1980 Bayh-Dole Act was to move government-discovered technologies from the laboratory to the private sector. Tibotec, a Belgian biotechnology company that was later bought by Johnson and Johnson, was interested in conducting the human trials. After a year of negotiations, NIH licensed Erickson’s invention to the European start-up. And in mid-1998, Erickson went with it, moving his entire lab to Tibotec.
The results of the first human trials on what Tibotec called TMC 126 were released at the eighth annual Retrovirus Conference held in Chicago in early February 2001. Erickson called a press conference to describe his new discovery. He glowingly described how his flexible molecule could jam up the mechanism of mutant protease scissors by bending to fit into the new shape of the active part of the enzyme, no matter what shape it took. He called the molecule “resistance repellant.” His longtime colleague at the Aaron Diamond Center agreed. “We have for the first time a very, very powerful protease inhibitor that could suppress resistant virus,” said David Ho. “That’s pretty impressive.”41
But by the following summer, Erickson was not a happy man. As he paced the kitchen of his modest home, he said he couldn’t talk about his experiences with Tibotec. The company had just shut down his lab, laying him off, along with the two dozen colleagues that had moved with him from NCI. The company kept his equipment, his lab notes, and his molecule, the products of a decade and a half of methodical work that had brought him to the brink of his second major triumph. “They had to cut costs to keep the drug development going, so they cut early-stage drug development,” he told me dejectedly. Erickson, the idealist who has now twice failed to find a congenial home in the private sector, said he hoped to reconstitute his lab, perhaps as a nonprofit, where he could research new molecular entities to combat drug-resistant strains of tropical diseases ravaging the developing world.42
His departure did not stop Tibotec from pursuing his drug. At the December 2001 ICAAC Conference in Chicago, Tibotec reported on its latest trials. But TMC 126 was no longer in the picture. Using Erickson’s original molecule, company chemists synthesized seven hundred analogues and used an automated screening system to pick a more promising drug candidate, which they dubbed TMC 114. “Single doses of this novel candidate protease inhibitor were safe and well tolerated at all tested dosage levels,” a company spokeswoman said. It would soon move into second-stage clinical trials. Nowhere in the accompanying press release did the company as much as hint that TMC 114 was the direct descendant of a drug that had been invented in a government-funded lab.