Four days after scientists infused trillions of genetically engineered viruses into Jesse Gelsinger's liver as part of a novel gene therapy experiment, the 18-year-old lay dying in a hospital bed at the University of Pennsylvania.

His liver had failed, and the teenager's blood was thickening like jelly and clogging key vessels while his kidneys, brain and other organs shut down.

It was a rare and irreversible blood reaction, but it wasn't the first time the researchers had seen it. Unbeknown to Gelsinger, who had signed up for the experimental treatment for a rare and often fatal liver disorder, monkeys that the Penn team had similarly treated had succumbed in very much the same way.

The team had moved forward with the human experiment despite the monkey deaths, and despite criticism from other researchers who thought it was too dangerous, because they believed that a new version of genetically altered virus they had developed was safer than the one that had killed the monkeys.

"Now I have to think, 'Was I blind?' " said James Wilson, head of Penn's Institute for Human Gene Therapy and a lead scientist in the trial. "I'm asking myself that question a lot these days."

So far, the Penn team's investigation into Gelsinger's death in September--thought to be the first fatality from gene therapy--has yielded no clear indication of what triggered the fatal reaction, so it's too soon to say whether it should have been predicted and prevented. Besides, experimental therapies by definition often are dangerous, and the distinction between laudable perseverance and unjustifiable risk-taking can be blurry.

Gelsinger's father remains highly supportive of the Penn team. And Gelsinger himself was eager to participate, despite the risks.

But a close look at the Penn research provides a rare snapshot of that subtle scientific and ethical landscape known as the cutting edge of medicine. And it resurrects old questions about whether the Penn experiment ever should have begun.

Many experts wonder, for example, whether an impatient Penn team overlooked the study's pitfalls out of eagerness to win a nine-year-old race to produce the world's first gene-based cure. As scientists work to cure diseases by giving people new, healthy genes, some are asking whether federal regulators' enthusiasm for the hot new field--and their enchantment with Wilson's stellar reputation--may have led them to give Wilson the benefit of the doubt at too many key decision points along the way.

Also anxious for success were the corporate investors who have funneled tens of millions of dollars into a gene therapy company Wilson founded and into his lab at Penn on the bet that he'd push the field forward.

Wilson adamantly denies that he was influenced by financial concerns. But such tight ties between university researchers and private industry are worrisome to ethicists.

A detailed examination of research records, along with interviews with scientists and government officials, has revealed:

* The Penn experiment was the first to shoot such a heavy dose of gene-altered viruses directly into the bloodstream of patients with a genetic disease, even as researchers acknowledged that it had no chance of curing that disease and despite widespread uncertainty in the field about toxic side effects.

* Contrary to most high-risk research, the Penn study experimented with the healthiest rather than sickest segment of the patient population--people such as Gelsinger, who had the disease under control with conventional drugs and diet, or people who had no symptoms at all.

* Researchers were not deterred by early indications of toxicity as they gradually increased doses in their experiment, including an especially severe case of liver damage in one participant. And they discounted or missed evidence of serious side effects in their own and other animal and human studies.

* Volunteers were recruited in ways that federal officials had explicitly precluded as being too potentially coercive, with direct appeals on a patient advocacy Web site that heralded "promising" early results from the clinical trial and said the experiment used "very low doses" when in fact they were relatively high.

* The original consent form, reviewed publicly by the National Institutes of Health, clearly notified prospective participants that monkeys had died from a related treatment, but the final version given to patients eliminated any mention of the deaths.

* Wilson has a financial interest in a private company he founded, Genovo Inc. of Sharon Hill, Pa., which has rights to discoveries made by Wilson at his lab on the Penn campus and which has a substantial financial stake in seeing liver-directed gene therapy succeed.

Gelsinger's death may never be fully explained. But already it has revitalized longstanding criticisms that gene therapy researchers, whose treatments have disappointed more than 3,000 patients since 1990, are experimenting on too many people too soon. It may be hard to argue against rushing to test novel treatments for incurable diseases, experts said. But moving ahead too quickly could do more than just place patients at risk. It could undermine the public trust that is so crucial to gene therapy, a field long seen as controversial because of its goal of altering people's fundamental genetic makeup.

"People will say they cannot cure patients and now they're killing them," said Guenter Cichon of Berlin's Humboldt University, one of several researchers who have documented the potentially deadly nature of the viruses that were given to Gelsinger in the Penn study.

Focusing on OTC Disease

Gelsinger suffered from a disease called ornithine transcarbamylase (OTC) deficiency. Victims are born with a genetic mutation that leaves their livers unable to break down ammonia, a normal byproduct of metabolism. High ammonia levels can quickly become fatal, and the disease affects newborn boys most seriously, killing about half of them soon after birth.

An estimated 1 in 40,000 babies is born with a defective OTC gene, a tiny population of patients for a major scientific effort. But Wilson, one of the country's leading geneticists, saw an opportunity to tackle a horrific disease while perfecting his technique for delivering genes to the liver, which he considered the key to curing many other more common diseases that have their roots in that organ.

In discussions with colleague Mark Batshaw, who pioneered a drug-and-diet regimen for OTC survivors that has saved numerous lives, Wilson became convinced that OTC deficiency was the perfect disease to prove the potential of gene therapy for the liver. Together, Wilson and Batshaw, with surgeon Steven Raper from Penn, developed a plan to use gene-altered viruses to deliver healthy copies of the OTC gene to affected newborns as soon as they were diagnosed--generally when they were on the brink of death from skyrocketing ammonia levels soon after birth.

The scientists knew the therapy was potentially dangerous, because the gene-altered virus they would use to get the new genes into the newborns' cells can trigger life-threatening reactions. They also knew the therapy would be short-lived, because the babies' immune systems would shut down the new genes within a few days or weeks.

But in such a desperately ill population, for whom no effective treatment exists, the risk of the new therapy would be justified, the researchers reasoned. And once these infants made it through their first crisis, they could be placed on Batshaw's diet-and-drug regimen and perhaps live fairly normal lives.

It seemed like medicine at its best: trying to save babies from an incurable disease. So the team was shocked when Penn's own ethicists rejected the proposal. Parents whose children are so close to death cannot be counted on to make rational decisions about whether to enter those children into a new, potentially dangerous and unproved experiment, they said.

Thus began a change of focus that, in many experts' eyes, triggered an unconscionable shift in the study's balance of risks and benefits: Rather than drop their focus on OTC altogether, the researchers decided to experiment on healthy or stable adult OTC patients such as Gelsinger, who had survived to adulthood because they had milder forms of the disease.

If the treatment proved safe, the researchers could bolster their case for testing its usefulness in newborns. But first they would have to test its safety in adults for whom safe and conventional treatments already existed, whose livers were already stressed because of their disease, and who stood no chance of getting any lasting benefit from the experimental approach.

That shift in focus stirred intense debate in 1995 when the study was reviewed by federal officials. But despite concerns that the researchers still needed to answer some basic scientific questions before moving ahead, the plan--like many other gene therapy protocols then and today--was approved by federal regulators amid a wave of optimism that scientists were on the brink of a breakthrough that would forever change the face of medicine.

"We had to select one disease to move forward in for gene therapy to the liver," Wilson recalled recently. "I thought the severity of the disease would warrant and justify trying this therapy. It's such a compelling story."

In the Business of Science

Wilson also had financial incentives to stay focused on the liver, although he denies that they influenced him in any way.

Since Wilson founded Genovo in 1992, the company has attracted two major corporate investors. While neither of those companies is interested in OTC specifically--it afflicts too few people to be commercially attractive--both are interested in gene delivery to the liver and both were drawn to Genovo's access to Wilson's discoveries.

Biogen Inc. of Cambridge, Mass., has paid Genovo $37 million since 1995 for the right to eventually market various liver- and lung-related genetic therapies developed by Genovo. The deal, which is up for renewal next year, called for Genovo to make progress in moving gene therapy toward a marketable product, said Genovo President Eric Aguiar. Under the agreement, Genovo must share the Biogen money with Wilson's institute at Penn, which today depends on that arrangement for about 20 percent of its budget.

In August, Genovo sealed an additional deal with Genzyme Corp., another biotechnology company in Cambridge, to develop liver-directed gene therapy for metabolic disorders.

Genovo has a direct stake in the genetically engineered adenovirus that Wilson developed for the OTC trial, according to Wilson and Aguiar. If Wilson and his colleagues can demonstrate that the virus is a good vehicle for ferrying genes into the body, the company and Wilson could benefit financially.

In a posting on the Genovo Web site, Aguiar boasts that the relationship with Wilson not only provides the company with access to Wilson's discoveries, but also minimizes business risks, because the company can wait until Wilson's lab tests new treatments on humans before deciding whether to invest in them.

Wilson said he went to great pains to ensure that his business interests would not influence his judgment during the OTC adenovirus trial. Although he was a senior scientist, for example, he gave Raper control over medical and patient care decisions.

"To suggest that I acted or was influenced by money is really offensive to me," he said. "I don't think about how my doing this work is going to make me rich. It's about leadership and notoriety and accomplishment. Publishing in first-rate journals. That's what turns us on. You've got to be on the cutting edge and take risks if you're going to stay on top."

Nevertheless, Wilson's own financial disclosure statement says Wilson and Genovo "have a financial interest in a successful outcome from the research involved in this study." Wilson acknowledged that the ties with Genovo are tight enough to require him to include that statement on research papers and the consent forms that patients sign when entering his clinical trials, including the OTC experiment.

Academic researchers increasingly are setting up their own companies or business deals on the side. Forty percent of the gene therapy protocols approved in the past three years have had corporate sponsors. Wilson and others argue that sponsorship provides an important source of funding for research and an eventual pipeline to get cures to the public.

Yet the business-academia pipeline has been the subject of much criticism in recent years, because it may sometimes force scientists to choose between good science and good business.

Aguiar said Wilson is too much a scientist to compromise. But Wilson is also very plugged in to the company, Aguiar acknowledged, and calls him on the telephone frequently during the week.

One of those calls from Wilson stands out in both men's memories. It was in September, and the talk was about business. When Aguiar asked how things were going, Wilson said he was having a stressful day. Gelsinger had reacted badly to the treatment and was heading into multiple organ failure in the hospital intensive care unit. "He was really upset about it," Aguiar said.

Problematic Genetic Messenger

In their worst-case scenarios, Wilson, Batshaw and Raper thought they might see an inflamed liver in their patients. But the researchers thought they had licked earlier problems with a new, safer brew of genetically engineered adenovirus.

Adenoviruses, a class of viruses that cause the common cold and conjunctivitis, or pinkeye, are extraordinarily efficient at infecting many kinds of human cells. Because of that, they have become popular with gene therapists as a way of delivering helpful genes to sick people. But there are downsides: Adenoviruses trigger intense immune system reactions and can prompt a life-threatening inflammatory response.

Because of those possible problems, many scientists have stopped using adenoviruses to treat genetic diseases and are looking for other, more promising gene delivery systems.

But Wilson's team had worked feverishly during the early 1990s to develop a less inflammatory adenovirus. Over several years, the researchers methodically deleted different combinations of the virus's genes until they had one that seemed to be safe enough to use in the fairly high doses needed to be effective.

The team also spliced copies of the OTC gene into the virus--the payload to be delivered to OTC-deficient patients.

While all four rhesus monkeys that had been given high doses of the first-generation virus had died in previous experiments, subsequent tests of the later viruses on monkeys and mice seemed to confirm that it was less toxic, although it still triggered liver inflammation. By cutting back the dose two hundredfold, the scientists and the FDA concluded that they could deliver it safely to humans with no or minimal side effects, especially because humans are so much bigger than monkeys. That is a presumption that federal regulators now question.

But at the time, the Penn researchers were confident in their plan, which included monitoring for early signs of trouble as they gradually increased doses and treating unexpected crises with a proven backup treatment. They commenced treatments in 1997.

Some scientists today remain supportive of Wilson's decision to go forward. "I don't know anyone in gene therapy who has done more animal studies before starting in people," said A. Dusty Miller, a gene therapist at the Fred Hutchinson Cancer Research Center in Seattle.

But while the new virus appeared to be safer, there remained wide disagreement over how much safer.

And between the start of the clinical trial and Gelsinger's death in September, new research elsewhere in the field provided further evidence of toxicity and the need for caution in using adenoviruses in the liver.

"Not many people consider it appropriate for treating genetic diseases," said NIH investigator Richard Morgan.

The NIH tested a closely related adenovirus on the livers of three macaque monkeys, which fell seriously ill with symptoms similar to those that killed Gelsinger. They recovered, but one suffered permanent liver damage. A German study involving similar adenoviruses caused acute, toxic responses in rabbits that also resembled those that killed Gelsinger. Cichon, the Berlin researcher who led that study, concluded that adenoviruses should be used only in dire circumstances, such as when the only other alternative is a liver transplant. It is a standard that other scientists say they have adopted.

To give adenoviruses to patients like Gelsinger "would never be justified," Cichon said. "And I am not the only one who thinks this way. We do not understand why [researchers] are taking these risks."

Meanwhile, Schering-Plough Corp. of Madison, N.J., and the University of California at San Francisco had begun a clinical trial that infused high doses of adenovirus into the livers of dying cancer patients, a situation in which greater risks are allowed because patients have little other hope. The company lowered the dose when two early participants experienced serious drops in blood pressure. But even at that lower dose--which was lower than the total dose eventually given to Gelsinger--two patients suffered serious stroke-like attacks.

Wilson said he was only vaguely aware of those studies, none of which had been published in peer-reviewed journals until after Gelsinger's death. However, the information was widely disseminated in other, less formal ways, such as postings on Web sites and presentations at scientific meetings.

Some scientists at a leading gene therapy meeting that Wilson attended in Washington in June saw disturbing similarities in the NIH and the Penn team's preliminary results, which described some drops in white blood cell counts among participants and evidence of increasing liver damage with each increase in dosage.

Wilson said no one approached him at the conference with concerns. But neither did Wilson reveal in his written report at the meeting that one patient had suffered an especially serious reaction to the treatment--a reaction that threatened hopes of getting Food and Drug Administration approval to move up to the next dosage level.

Ultimately, the team tested its virus on an additional monkey. "We did it and the animal got hepatitis but it lived," Wilson said. "That was reassuring." The FDA was aware of all the data, Wilson said. If the study was too risky, he asked, why didn't federal regulators stop it?

The Regulation Process

In fact, it hadn't been easy for Wilson and his team to convince federal regulators that they should be allowed to try their approach in people. In 1995, when the researchers first pitched their plan to the NIH committee of scientists and ethicists that then reviewed all gene therapy proposals in advance, several committee members expressed strong reservations. The approach looked too dangerous, they said. And because the immune system would quickly destroy the genetically repaired cells, the treatment would not have any lasting benefit.

But according to several scientists and some NIH committee members, Wilson's charismatic style and his good reputation as a scientist won the day. "Wilson said that you should let people be heroes if they want to be," said Robert Erickson, a committee member at the time and a University of Arizona scientist. "In retrospect, I wish I hadn't been convinced."

The committee approved the protocol but insisted on two changes. The viruses should be infused into a distant blood vessel, not directly into the liver, hopefully reducing the trauma to that already diseased organ. And the researchers should recruit their subjects only through physicians, not by making direct appeals to patients, who might be swayed too easily into participating in the potentially dangerous study.

The researchers agreed, but that's not what happened. First, the FDA became convinced that direct infusion to the liver was preferable and made the team switch back. Penn is now investigating whether that direct infusion contributed to Gelsinger's reaction.

The committee never had a chance to review the change because in 1996, shortly after approving the Penn trial, its powers were greatly reduced by the NIH, under pressure from biotechnology companies seeking relief from federal regulations that the industry deemed overly burdensome.

The NIH committee members say they also never got word of a significant change in the consent form. Although the final form included a perfunctory clause stating that the experiment could result in injury or death, it dropped mention of the original fatal animal studies. Neither the FDA nor the Penn team can explain today why the reference to monkey deaths was dropped.

The Penn team also broke its assurance to the NIH committee that it would recruit participants only through physicians.

In the summer and fall of 1997, Batshaw wrote pieces that appeared in the National Urea Cycle Disorders Foundation newsletter and on its Web site, seeking volunteers. "Obviously, the faster we can complete the Phase I study," he wrote, "the sooner we can move on to the treatment phase in children."

When asked about his call for volunteers, Batshaw said: "We did recruit through the foundation newsletter. It was passive recruitment in that it just appeared there. I wouldn't want to say it was advertised." The possibility of volunteering was merely "posed," he said. As for his comments that some participants had experienced some "correction," he now concedes that might be an overstatement. "Perhaps it could have been stated better."

Volunteering 'to Help the Babies'

There's little question that Gelsinger was a willing and eager participant in the Penn experiment.

He had a mild form of the disease that could be controlled through diet and drugs. He didn't always follow the grueling regimen, and a year ago nearly died from an ammonia attack. But since then, he'd felt better than ever, thanks to a new regimen developed by Batshaw.

After hearing about the experiment from his doctor, he tried to volunteer at age 17--too young for the protocol--then returned as soon as he turned 18. The research team made it clear that the experiment wouldn't cure him, and the teenager also knew there was a small chance it could hurt him. But he was swayed by the scientists' dream that the treatment might someday help severely stricken newborns.

"He wanted to help the babies," said his father, Paul Gelsinger, in a recent interview. "My son had the purest intent."

Paul Gelsinger also remains an outspoken fan of the researchers, whom he considers "very ethical men."

"Kids are dying all the time from these orphan diseases," he said. "How long do you wait to do something?"

Raper, Batshaw and Wilson are deep into their investigation of Gelsinger's death and will present their results to the NIH Dec. 8-10.

"As a scientist it's very difficult to do the tests, because of what they might show," Wilson said. However, he added, "If a mistake was made, we've got to own up to it and learn from it. Ultimately, the tragedy of Jesse's death would be if we don't learn anything."

Staff researcher Alice Crites contributed to this report.

The New Biology

Human cloning, once thought impossible, today seems possible, and perhaps even probable. Growing spare body parts in the laboratory, once the stuff of science fiction, is being pursued by scientists everywhere. Embryos are being genetically selected to be healthy, or male; "designer babies" may not be too far behind. Plants genetically engineered to produce their own pesticides, and even certain types of plastics, are growing in fields. Science is entering a new world where the once-unthinkable is suddenly doable. The Washington Post is examining this scientific revolution in a series of occasional articles about this new biology exploding at the turn of the century.

OTC Gene Therapy

Genetically engineered viruses inject potentially curative genes into a patient's liver cells.

A mutated gene blocks a metabolic pathway in patients with OTC deficiency. A dangerous buildup of ammonia develops.

Genetically engineered adenovirus (with some toxic genes removed) infects liver, injecting normal OTC genes into liver cell DNA.

Altered liver cells engage in normal OTC metabolism, breaking down ammonia.

CAPTION: Jesse Gelsinger, an 18-year-old with a rare liver disorder, died during the gene therapy study.

CAPTION: James Wilson, front, surgeon Steven Raper, at the overhead projector, and another University of Pennsylvania research team member knew the risks of their study but thought they were justified.

CAPTION: Wilson on the phone with FDA officials early in the study. If it was too risky, he wonders now, why wasn't it stopped?

CAPTION: Jesse Gelsinger so wanted to help other patients that he tried to get into the study before he turned 18.