RIVAL SCIENTISTS QUESTION RESEARCH ON AIDS LESIONS
Gallo critics say tests of drug may be invalid

By John Crewdson

Chicago Tribune 9 June 1994

A major medical journal has taken the highly unusual step of publishing an article questioning the veracity of research appearing in a rival journal, raising tempers as well as new questions about what had appeared to be one of the few hopeful developments regarding AIDS.

The article, published in Thursday's issue of the Journal of the American Medical Association (JAMA) by a little-known research group from the University of Arizona, is critical both of the journal Science, where the original article appeared, and of its senior author, Dr. Robert C. Gallo, one of the biggest names in AIDS research.

The JAMA article contains some of the most pointed language ever seen in the normally well-mannered scientific prints, dismissing a central conclusion of the Gallo piece as "highly dubious" and declaring that "the validity of the peer review process and self-correcting nature of scientific inquiry are also called into question."

Gallo's Science article, published two years ago, concerns Kaposi's sarcoma, or KS, an often-fatal, cancer-like skin condition whose origins and frequent appearance in gay men (but not women or heterosexual men) with AIDS represent one of the enduring mysteries of that disease.

Gallo's report, that a drug named SP-PG appeared to inhibit the development of Kaposi's sarcoma in mice, was seen as welcome news by KS patients and by physicians desperate for better AIDS treatments. "It really prevents the development of lesions," Gallo said at the time.

The Arizona group states flatly, "Serious systematic errors and omissions flaw the original study, and we cannot replicate some of the pivotal findings." The group emphasized, however, that its findings did not necessarily mean the drug is ineffective against KS.

Two weeks before the Science article was submitted, in July 1992, Gallo was ordered by Dr. Samuel Broder, director of the National Cancer Institute, to review all "primary data" for every manuscript published by Gallo's laboratory.

A spokesman declined to say whether Gallo had done so in this case, saying only that Gallo had "personally reviewed" the manuscript before it was submitted to Science. He said Gallo planned to send JAMA a response to the article "within the next few days."

The National Cancer Institute, where Gallo heads a major research laboratory, struck a commercial development deal with Daiichi Pharmaceuticals, the Japanese company that owns the U.S. patent on SP-PG, which quickly began laying plans to try the drug on Americans with AIDS.

The NCI declined to disclose the amount of money Gallo's laboratory received from Daiichi to support Gallo's research on SP-PG, but sources put the figure at $75,000. After the deal was announced, Gallo told the Wall Street Journal he did not stand to earn anything personally.

"As to what Daiichi could earn, I don't know," Gallo said. "I'm not an economist. But if it helps patients, that's what I'll get out of it."

Daiichi's stock gained 6 percent on the Tokyo exchange the day the Gallo article appeared in Science.

Despite the initial fanfare, the SP-PG compound, produced by a strain of bacteria found in soil, currently is being tested in only seven patients at the University of Southern California. The researchers in charge of those trials reported recently that several subjects had experienced unexpected bleeding and other side effects, and that only two of the seven appeared to have improved since taking the drug.

"We pursued this because it's clinically important," said Dr. Marlys H. Witte, a professor of surgery at the University of Arizona College of Medicine and the principal author of the JAMA article, which also is signed by her husband, Dr. Charles L. Witte, and four co-workers.

"This article seemed to be offering some drug treatment that might be useful. When we read it, many important questions were raised. The answers that we were given were not legitimate or relevant."

Marlys Witte explained in the article that she and her colleagues had turned to the Chicago-based JAMA only after Science and Gallo refused to publicly acknowledge their criticism.

"More than 2 years have passed since the ( Gallo) article was published and our efforts to address the validity of the data and conclusions in an open scientific forum were initiated," she wrote. "No follow-up studies by these authors have appeared in print that might clarify the points in question."

Daniel E. Koshland Jr., the editor of Science, said in a statement that he still believed "our actions were justified" in not publishing the Witte group's original criticism of Gallo. Koshland also noted that the JAMA article contained new information reflecting the Witte group's unsuccessful attempts to reproduce Gallo's results.

The deputy editor of JAMA, Dr. Drummond Rennie, said he had made the unusual decision to publish an article critical of work appearing in a rival journal only after Witte's group "provided us with ample evidence that Science had absolutely finished with them."

Rennie added that the Witte group's questions about Gallo's research "seemed valid" to him, and that an earlier letter Witte sent to Science raising those questions "had been rejected for reasons that didn't hold scientific water."

The third major biomedical journal in this country, the New England Journal of Medicine, suffered the same treatment last year when JAMA published a critique of one of its more important research articles. "I think it is an odd thing to do, and it's certainly something we wouldn't do," said Dr. Jerome Kassirer, editor of the New England Journal.

When the Science article appeared it was viewed with particular interest by Witte and her Arizona group, one of the few in the country then concentrating on Kaposi's sarcoma. In particular, Witte's attention was focused on photographs the article contained of four mice, their skins splayed and pinned back to reveal their insides.

A central point of the Gallo article was the ability of SP- PG to prevent whatever "growth factors" might cause Kaposi's sarcoma lesions from leaking out of the blood vessels and forming lesions in the surrounding tissues. Hoping to demonstrate this, Gallo's assistants inoculated some of the mice with varying doses of SP-PG and others with different drugs or none at all.

They next infected the mice with Kaposi's sarcoma cells, waited a few hours, then injected them with blue dye through a vein in the tail. When the amount of blue dye that had leaked from the blood vessels in each mouse was measured, the Science article said, it was greatest in those mice that received no SP- PG and least in those that got the most.

When the Wittes examined the pictures closely they saw something they thought odd: the tails of the untreated mice in which the most dye apparently had leaked through the vessel walls were an intense blue color, while the tails of mice treated with SP-PG were not.

To the Wittes, the difference suggested that Gallo's assistants had made a serious mistake, injecting dye into the tissues of the untreated mice as well as into the tail vein. If the Wittes were right, the increased leakage in the mice that did not receive SP-PG would actually have been caused by a botched injection.

A month after the Gallo article appeared in Science, the Wittes and their co-workers sent a letter to Science. Entitled "The Tell-Tale Blue Tail," the letter described how the Arizona group had achieved an effect similar to one reported by the Gallo group via just such a "botched" injection of blue dye into the tail vein of a mouse.

The letter also questioned Gallo's claim that a sizable amount of dye-containing mouse blood had leaked into a small Kaposi's lesion in just a few minutes. "Even a localized scalding burn of that tiny dimension," the Wittes wrote, would not cause the loss of so much blood in so short a time.

The Wittes also questioned Gallo's assertion that the blue dye, which they termed "notoriously difficult" to extract from soft tissue, had been removed by Gallo's assistants from the mouse lesion overnight.

They also expressed "surprise"-the scientific equivalent of the raised eyebrow-at the "smoothness" of the graph in the article showing that blood vessel leakage decreased in step with increases in the dosage of SP-PG.

Science sent the Wittes' letter to Gallo, who dismissed their criticism as "an extraordinary waste of time and effort." Gallo noted that the Science article contained other claims for the effectiveness of SP-PG that the Wittes had not addressed, and he included a picture of another mouse that he said disproved the Wittes' theories.

Science asked four other scientists, whose identities it did not disclose, to read the Wittes' letter and Gallo's response and decide whether they should be published in tandem.

All said no, even though two reviewers shared the Wittes' surprise that so much blood had leaked from the blood vessels in so short a time, and that Gallo's lab had been able to extract the blue dye so easily. The most important shortcoming in the Wittes' arguments, a third reviewer said, was their failure to have tried to reproduce Gallo's experiments themselves.

The Wittes were not satisfied. Gallo's response, the new picture and the comments of the anonymous reviewers, they told Science, "not only fail to answer our original questions but actually raise additional disturbing issues."

The Wittes sent Science a lengthy written analysis from a reviewer of their own, Dr. Hugh J. Carroll of the State University of New York, who found much merit in the Wittes' claims. "The editors of Science were remiss," concluded Carroll, "in foreclosing the opportunity for exploration and debate on a topic that I perceive to be still unsettled."

Science sent the expanding collection of criticisms and rebuttals to yet another scientist, identified only as "someone in whom we have the utmost confidence." Like previous reviewers, this one pointed out that while some of the Wittes' assumptions were reasonable, "for all their concern and criticism" the Wittes had not tried to test them by conducting the experiments described by Gallo.

In the fall of 1992, Witte and her group set out to do what their critics suggested, beginning from scratch to replicate Gallo's experiments with the mouse and dye, even persuading Gallo to send them some of the special Kaposi's sarcoma cells he used to generate the lesions in his mice.

"Perhaps the most serious aspect of the whole matter," the JAMA article concludes, "has been the reticence and obstacles encountered to public airing of our questions and the inability of the peer review process to correct itself once errors and inconsistencies were pointed out and bolstered by further experimentation." *