HARVEY Pass, the chief of thoracic surgery at the National Cancer Institute, in Bethesda, Maryland, was sitting in his laboratory one spring afternoon in 1993 when Michele Carbone, a wiry young Italian pathologist who was working as a researcher at the NCI, strode in with an unusual request. Pass had never before met Carbone, and had talked to him for the first time, on the telephone, only a few hours before. Now Carbone was asking Pass for his help in proving a controversial theory he had developed about the origins of mesothelioma, a deadly cancer that afflicts the mesothelial cells in the lining of the chest and the lung. Mesothelioma was virtually unheard of prior to 1950, but the incidence of the disease has risen steadily since then. Though it is considered rare -- accounting for the deaths of about 3,000 Americans a year, or about one half of one percent of all domestic cancer deaths -- the disease is particularly pernicious. Most patients die within eighteen months of diagnosis.

Pass, one of the world's leading mesothelioma surgeons, knew, like other scientists, that the disease was caused by asbestos exposure. But Carbone had a hunch he wanted to explore. He told Pass that he wondered if the cancer might also be caused by a virus -- a monkey virus, known as simian virus 40, or SV40, that had widely contaminated early doses of the polio vaccine, but that had long been presumed to be harmless to people.

Pass listened as Carbone described for him the history of the early polio vaccine. A breakthrough in the war against polio had come in the early 1950s, when Jonas Salk took advantage of a new discovery: monkey kidneys could be used to culture the abundant quantities of polio virus necessary to mass-produce a vaccine. But there were problems with the monkey kidneys. In 1960 Bernice Eddy, a government researcher, discovered that when she injected hamsters with the kidney mixture on which the vaccine was cultured, they developed tumors. Eddy's superiors tried to keep the discovery quiet, but Eddy presented her data at a cancer conference in New York. She was eventually demoted, and lost her laboratory. The cancer-causing virus was soon isolated by other scientists and dubbed SV40, because it was the fortieth simian virus discovered. Alarm spread through the scientific community as researchers realized that nearly every dose of the vaccine had been contaminated. In 1961 federal health officials ordered vaccine manufacturers to screen for the virus and eliminate it from the vaccine. Worried about creating a panic, they kept the discovery of SV40 under wraps and never recalled existing stocks. For two more years millions of additional people were needlessly exposed -- bringing the total to 98 million Americans from 1955 to 1963. But after a flurry of quick studies, health officials decided that the virus, thankfully, did not cause cancer in human beings.

After that the story of SV40 ceased to be anything more than a medical curiosity. Even though the virus became a widely used cancer-research tool, because it caused a variety of tumors so easily in laboratory animals, for the better part of four decades there was virtually no research on what SV40 might do to people.

Carbone had reviewed some old research papers on the contamination and some of the early tests on SV40. He had even reviewed the notes from a crucial 1963 epidemiological study, by Joseph Fraumeni, an NCI researcher, which had concluded that children inoculated with contaminated vaccine did not show increased mortality rates. The studies did not impress Carbone: no one had systematically searched for evidence of the virus in tumors, and, as Fraumeni himself noted, the epidemiological study was too short to have detected certain slow-developing cancers. (Mesothelioma can take twenty to forty years to develop.)

Carbone had just finished a series of experiments in which he had injected the virus into dozens of hamsters. Every one of them developed mesothelioma and died within three to seven months. The results made Carbone wonder if SV40 might also play a role in human mesothelioma. He had come to see Pass because he had heard that the senior surgeon had meticulously saved tumor tissue from every one of the dozens of mesothelioma surgeries he had performed, and now had one of the largest collections of mesothelioma biopsies in the world. Carbone asked Pass if he could look for SV40 DNA in Pass's tumor-tissue samples, using a sophisticated molecular technique, known as polymerase chain reaction, or PCR, to extract tiny fragments of DNA from the frozen tissue and then amplify and characterize them.

As they talked, Pass became more and more impressed with Carbone. The young scientist was energetic and extremely self-confident -- something Pass attributed to Carbone's surgical patrimony. (Carbone's father is a well-known orthopedic surgeon in Italy.) When Carbone had finished describing his proposed experiment, Pass realized that the implications were potentially significant. Only a handful of viruses have been directly associated with human cancers, and none of them are simian in origin. If SV40 was linked to mesothelioma in people, might it also cause bone and brain cancers in human beings, as it had done in hamsters? What if the monkey virus could spread from person to person? And if the virus was cancer-causing, or oncogenic, what was one to make of the fact that millions of Americans had been exposed to it as part of a government-sponsored vaccination program?

"I thought to myself, He's got this wild-assed idea," Pass recalls. "If it's true, it's unbelievable. Even if it's not, I'm going to get a hell of an education in state-of-the-art molecular biology."

Others at the National Institutes of Health -- including some of the scientists who had been around at the time of the contamination scare -- were less receptive to the novel theory. They told Carbone that the last thing anyone wanted to hear was that the exalted polio vaccine was linked to cancer. Too much was at stake. Implicating a vaccine contaminant in cancer -- even if the contamination occurred some forty years ago -- might easily shake public confidence in vaccines in general. And besides, everyone knew that asbestos was the cause of mesothelioma.

Carbone sought the advice of two renowned pathologists, Umberto Saffiotti, the chief of the NCI's Laboratory of Experimental Pathology, and Harold L. Stewart, a former director of pathology at the NCI who was once the head of the American Association for Cancer Research. Both urged Carbone to follow his intuition. "Forget what people tell you," Stewart told Carbone. "They told me I was wrong all my life. If you want to do it, you should, or you will regret it." That spring afternoon in 1993, with Pass's mesothelioma samples in hand, Carbone called an old friend, Antonio Procopio, a professor of experimental pathology in Italy who had worked for three years at the NIH. "I asked him if he was willing to do this crazy project with me," Carbone says. "I told him I could not pay him or his expenses." A month later Procopio arrived in Bethesda. "We had no money," Carbone recalls. "He slept in my house for six months, and we worked day and night."

It turned out that Pass's samples were loaded with the monkey virus: 60 percent of the mesothelioma samples contained SV40 DNA; the nontumor tissues used as controls were negative. Moreover, Carbone found that in most of the positive samples he tested, the monkey virus was active, producing proteins -- suggesting to Carbone that the SV40 was not just an opportunistic "passenger virus" that had found a convenient hiding place in the malignant cells but was likely to have been involved in causing the cancer.

In 1994 Carbone, Pass, and Procopio published the results of their experiment in one of the world's leading cancer-research journals. They proposed SV40 as a possible co-carcinogen in human mesothelioma. It was the first time researchers had put forward hard evidence that the all-but-forgotten vaccine contaminant might cause cancer in human beings.

A Solution to an Enigma

MICHELE Carbone is almost stereotypically Italian: generous with his emotions, outspoken, and jovial. He is strikingly handsome, with large brown eyes and shoulder-length brown hair. Carbone grew up in a cultured home in Calabria, on the shores of the Mediterranean in southern Italy. As a youth he often spent hours poring over medical texts, some of them 300 years old, in the voluminous library started by the first of the seven generations of Carbone physicians to date. If his father gave him science, from his mother he may have inherited the strong intuition that is his distinguishing characteristic as a researcher. She is an accomplished artist whose work is exhibited widely in Europe.

Carbone graduated in 1984, at the top of his class, from the University of Rome Medical School, one of the largest in the world, and quickly won a coveted NIH doctoral fellowship. In 1993 he received a Ph.D. in human pathology. In less than a decade he has risen to the top of his profession. Today he is internationally recognized as an expert in mesothelioma.

Since 1994 Carbone has written more than twenty studies and reviews investigating SV40's link to human cancer. "There is no doubt that SV40 is a human carcinogen," he says. "SV40 is definitely something you don't want in your body." Carbone suggests that the virus works in tandem with asbestos or by itself to transform healthy mesothelial cells into cancerous ones.

Since he published his first study, scientists at seventeen major laboratories -- in the United States, Great Britain, France, Belgium, Italy, and New Zealand -- have confirmed Carbone's research with respect to the presence of SV40 in human mesothelioma. Their results point to a solution to an enigma that long puzzled researchers. At least 20 percent of mesothelioma victims report no asbestos exposure, and only 10 percent of people who have had heavy exposure to asbestos ever develop mesothelioma. The experiments suggest that SV40 may be another factor at work in the tumors.

Two very recent studies, from Finland and Turkey, found no SV40 in domestic mesothelioma samples but did find it, respectively, in American and Italian samples. The authors observe that their negative findings lend support to the theory that contaminated polio vaccine is associated with the disease: neither Turkey nor Finland used SV40-contaminated vaccines. Today Finland has one of the lowest rates of mesothelioma in the Western world.

The virus has also been located in other kinds of tumors. More than a dozen laboratories have found SV40 in various kinds of rare brain and bone tumors. In 1996 Carbone reported that he had found SV40 in a third of the osteosarcomas (bone cancers of a type that afflicts about 900 Americans a year) and nearly half of the other bone tumors he tested -- research that has since been confirmed by numerous laboratories. The virus has also been detected in pituitary and thyroid tumors.

The possibility of a link between SV40 and brain tumors is particularly intriguing. Like mesothelioma, brain tumors have become dramatically more common in recent years. Brain tumors will be diagnosed in about 3,000 children in the United States alone this year. In 1995 Janet Butel, the chairman of the department of molecular virology and microbiology at the Baylor College of Medicine, in Texas, and her chief collaborator, John Lednicky, also a Baylor virologist, reported that they had found SV40 in a number of children's brain tumors. Butel and Lednicky reported that DNA sequencing revealed that the virus was not a hybrid but rather authentic SV40 -- the same as the SV40 found in monkeys. In the fall of 1996 an Italian research team, led by Mauro Tognon, of the University of Ferrara, announced that it had found SV40 DNA in a large percentage of brain and neurological tumors, including glioblastomas, astrocytomas, ependymomas, and papillomas of the choroid plexus. The researchers suggested that SV40 may be a "viral cofactor" involved in the sharp rise in human brain tumors. Late last year an extensive study undertaken in China reinforced those results. The study examined sixty-five brain tumors, finding SV40 in each of the eight ependymomas and two choroid-plexus papillomas, common brain tumors among children. It also found the virus in 33 to 90 percent of five other kinds of brain tumor examined. The authors, writing in the November, 1999, issue of Cancer, noted that the virus was actively expressing proteins.

Recent research also indicates that SV40 has gained a secure foothold in the human species. In 1996 Tognon and his collaborators reported that they had also found the virus in 45 percent of the sperm samples and 23 percent of the blood samples they tested from healthy people, suggesting that the monkey virus could spread through sexual contact or unscreened blood products. In 1998 the presence of SV40 antibodies in human blood samples was reported by Butel, who tested several hundred American blood samples and found antibodies to SV40 in about 10 percent of them. Butel's laboratory also tested samples from children born from 1980 to 1995 -- decades after the contaminated vaccine was removed from the market. A surprising six percent tested positive -- offering evidence that the virus may now be spreading from person to person, including from mother to child.

The Critics Respond

THE presence of SV40 in human tumors has been reported on in more than forty independent research papers. But one molecular study that has had an enormous impact on the direction of SV40 research and funding was performed not by a virologist, like Butel, or a molecular pathologist, like Carbone, but by an epidemiologist named Howard Strickler. Strickler served as a senior clinical investigator in the NCI's Viral Epidemiology Branch for many years before he joined the Albert Einstein College of Medicine, in New York, last winter. He has been persistently skeptical of any association between the vaccine contaminant and tumors. Though he is no longer at the NCI, he remains instrumental in the government response.

In June of 1996 Strickler published a paper with Keerti Shah, of the School of Public Health at Johns Hopkins University, in Baltimore, in the journal Strickler and Shah reported that they had come up empty-handed in their search for SV40 in fifty mesothelioma samples. Their study and a 1999 British study are the only two published SV40 studies with negative results. These two papers, particularly Strickler's, are cited again and again by federal health officials as proof that the dozens of peer-reviewed papers reporting SV40's presence in human tumors are unpersuasive and that a major research effort on SV40 is unnecessary.

Strickler acknowledges that he has never done PCR himself (Shah was responsible for the PCR work for their 1996 collaboration), but he challenges the work of other labs that have found SV40 in human tumors. "I feel that the data are mixed regarding the detection of SV40 DNA in human tissues," Strickler says, citing his own negative study and the British study. Strickler also points out that when SV40 is found in tumor cells, it often occurs only at very low levels. Whereas human papilloma virus (HPV), which causes cervical cancer, can be detected at rates of fifty viruses per cancer cell, SV40 is sometimes found at a rate of one virus per cell. "I find it curious that even the laboratories that detect SV40 in the cancers report that the virus is present at such extremely low levels," Strickler says. John Lednicky, of Baylor, counters that HPV is very different from SV40. Strickler "is comparing an apple with an orange," he says. "SV40 is known to be far more tumorigenic than HPV in animals. One copy of SV40 per cell is enough to transform a cell."

Several SV40 researchers have criticized Strickler's 1996 study and the more recent British one, saying that they treated specimens in a manner that would not result in the efficient extraction of SV40 DNA. Bharat Jasani, the director of the molecular diagnostic unit at the University of Wales, in Cardiff, has found SV40 in British mesothelioma samples. He recently wrote a lengthy critique of the two studies that has not yet been published. In this critique Jasani concludes that the negative results "are explainable by the paucity of the diagnostic biopsy material used and/or insufficient sensitivity of the overall PCR methodology used." Jasani says that Strickler's PCR technique would have missed low levels of SV40.

Federal health officials are understandably concerned that any link between SV40 and human cancers could frighten people away from the polio vaccine and vaccination in general. They stress that before SV40 in the polio vaccine can be linked definitively to cancer, the proposition must clear important scientific hurdles. Carbone and others must prove that the SV40 they have found is not a laboratory contaminant. They must demonstrate that SV40 is responsible for the cellular damage that leads to cancer and is not just a benign "passenger" in human tumors. And they must show that it was introduced into human beings through the polio vaccine.

In assessing the research to date, Strickler is perplexed that the virus has been found in so many kinds of tumors. In addition to the confirmed research reporting the virus in more than a half dozen kinds of brain tumors and a similar number of bone tumors, researchers in new, isolated studies have reported finding the virus in Wilms tumors, which afflict the kidney, and adenosarcomas, rare cancers of the uterus. "It's not likely that a single virus causes ten thousand different diseases," Strickler says. "That's not how it works."

These anomalies have fueled Strickler's suspicion that many of the SV40 findings in human tumors may really be false positives resulting from laboratory contamination. He points out that SV40 is used for cancer research in so many laboratories around the world that almost any lab involved with tumor assays could conceivably harbor it. "Is it possible that SV40 is in human tumors and that SV40 is at some level circulating in the human population?" Strickler asks. "Could it be true? I can't exclude the possibility, but the studies to demonstrate it haven't really been done, and the data in our hands have been negative." Strickler's former boss, James Goedert, the chief of the NCI's Viral Epidemiology Branch, agrees. Although he says he has an open mind about SV40, he believes that contamination may lie behind the findings of Carbone, Butel, and others.

In 1997, largely in response to Strickler's study, the International Mesothelioma Interest Group set out to determine once and for all if the virus was present in human mesothelioma samples. The organization asked an internationally known molecular geneticist, Joseph R. Testa, the director of the Human Genetics Program at the Fox Chase Cancer Center, in Philadelphia, to oversee a study. Testa, who specializes in mesothelioma research, confesses that initially he doubted the idea that SV40 could be found in human mesotheliomas, because he believed it was well established that asbestos was the cause of the disease. "I'm a very careful person," Testa says. "I had a fair amount of skepticism about it." But the results of the investigation he led changed his mind. Four laboratories participated in the tightly controlled study, including Carbone's. All four found SV40 in at least nine out of the twelve mesothelioma samples they tested. Each laboratory's control samples tested negative, suggesting that the positive SV40 samples were not the result of laboratory contamination. The results were published in the journal Cancer Research in 1998.

Strickler believes that Testa's study "did not really move the ball forward" in determining whether contamination lies behind findings of SV40 in human tumors. He questions Testa's conclusions. "They are trying to make a large point out of the fact that results were reproduced," he says. But according to Strickler, that such a high percentage of tumors tested positive actually casts doubt on the study's reliability and raises the possibility that the labs merely exchanged contaminated samples. "The prevalence [of SV40-positive samples] was so high ... that you have no way to make the distinction between [contamination] and a true positive result," he says.

Carbone and some of the other scientists we have interviewed say that Strickler's contamination theory is a red herring. "We've documented that it is the case that this virus is present and is expressed in these tumors," Testa says. "I think the onus is on [federal health officials] to take this new research into consideration." Carbone, not surprisingly, is even more adamant. "The idea that these tumor samples, tested in laboratories all over the world, were all contaminated, while all the controls remained negative, is ridiculous," he says. "There is no scientific evidence in support of contamination, and plenty of evidence to the contrary. Moreover, many labs have demonstrated SV40 using techniques other than PCR."

Recently we asked several prominent scientists to evaluate the SV40 studies. George Klein, at the Karolinska Institute, in Stockholm, who chaired the Nobel Assembly, and is a longtime expert on SV40, read Testa's study. His conclusion was different from Strickler's. According to Klein, the Testa study is "quite convincing concerning the association between SV40 and mesothelioma," and "the evidence suggests that SV40 may contribute to the genesis of some human tumors, mesothelioma in particular."

Carlo Croce, the editor of Cancer Research and a member of the National Academy of Sciences, agreed. Not only is it indisputable that SV40 is present in human tumor samples, he told us, but "it looks like the presence of the virus contributes to the cause of mesothelioma."

Janet Rowley, the editor of the journal and a professor of molecular genetics and cell biology at the University of Chicago, is a pioneer in the study of chromosome abnormalities in cancer. Rowley's groundbreaking research was itself called into question for years. "People didn't believe that chromosome abnormalities had anything to do with leukemia," she recalled. "It took a long time to break down that prejudice." She told us that Carbone had faced the same kind of doubts that first greeted her. "Everybody had assumed that mesothelioma was associated with asbestos. One of the important things in medicine is not to let your assumptions and those generally accepted paradigms obscure the fact that maybe there's more." Rowley believes that Carbone and Testa's work strongly implicates SV40 as a causal factor in some mesotheliomas.

"Like Somebody Set Off a Bomb"

CARBONE'S office is tucked into a quiet second-floor corner of the glass-and-concrete Cardinal Bernardin Cancer Center, at Loyola University, in Maywood, Illinois. The center is just a few miles west of Chicago and about ten minutes by car from Oak Park, where Carbone lives in a stately Frank Lloyd Wright house, with his wife and two daughters. Carbone came to Loyola in 1996 after a two-year stint at the University of Chicago. Now an associate professor of pathology, he works with Paola Rizzo, his senior scientist and closest collaborator, and a handful of post-docs and lab assistants in a tidy laboratory just down the hall from his office.

The lab is lively. Carbone has recruited compatriots as some of his research assistants, and the whir of high-tech machinery is punctuated by good-natured banter in Italian. This afternoon Carbone is examining an SV40-infected cell-culture plate under a microscope. He speaks almost fondly of the virus he has studied for most of the past decade. SV40 is "the smallest perfect war machine ever," Carbone murmurs. "He's so small. But he's got everything he needs."

Magnified 50,000 times under an electron microscope, SV40 doesn't seem particularly menacing. It looks almost pretty -- bluish snowflakes, against a field of white. The virus consists of six proteins, three of which make up the twenty-sided triangular scaffolding that is the virus's protein skin. But one of the remaining proteins, called large T-antigen (for "tumor antigen"), is, according to Carbone, the most oncogenic protein ever discovered. It is unique, he says, in its ability to cause cancer when it is set loose inside a cell.

In 1997, in Nature Medicine, Carbone published the first in a series of papers that outlined how large T-antigen blocks crucial tumor-suppressor pathways in human mesothelial cells. Whenever a cell begins to divide, in the process known as mitosis, a small army of quality-control agents goes to work. Running up and down the cell's DNA, these genes and proteins work together to scrutinize the DNA's integrity. If at any stage of cell division they detect DNA abnormalities that cannot be repaired, mitosis is halted and the cell undergoes apoptosis, or cellular suicide. The principal in this elaborate regulatory dance is a gene called p53. Arnold Levine, the president of The Rockefeller University, in New York City, and the discoverer of p53, says that 60 percent of all cancers involve some sort of p53 damage, mutation, or inactivation. "The p53 gene is central to human cancers," he says, describing it as "the first line of defense against cancer formation."

Carbone's experiments have shown that in human mesotheliomas large T-antigen attacks p53, binding to it so that it cannot function properly. Large T-antigen also strangles a series of proteins called Rbs, which together serve as some of the final gatekeepers in cellular division.

No other cancer-causing virus uses just one protein to knock out two different regulatory pathways simultaneously. For example, human papilloma virus must produce two proteins, E6 and E7, to inactivate p53 and the Rbs respectively; SV40 does its damage in one stroke. Levine calls large T-antigen "a remarkable protein."

Large T-antigen's cancer-causing havoc isn't limited to disabling a cell's most important tumor suppressors. It can also damage chromosomes by adding or deleting whole sections of DNA or reshuffling the genes. Once the virus is finished with a cell, Joseph Testa says, "it looks like somebody set off a bomb inside the cell's nucleus, because of all these chromosome rearrangements." Carbone says that because SV40 binds to tumor-suppressor genes and also causes genetic damage, it "is one of the strongest carcinogens we know of."

Yet he emphasizes that most people who carry SV40 in their cells won't develop cancer, because a healthy immune system generally seeks out and destroys invading viruses. He points out that large T-antigen normally provokes a particularly strong immune response, unless a person has been exposed to asbestos, a known immunosuppressant. "Human beings," Carbone says, "have devised many mechanisms to defend themselves against cancer. This is one of the reasons that human beings live so long compared with other animals. Human cancer is usually the result of a number of unfortunate events that together cause a malignant cell to emerge."

But SV40 may have evolved other strategies to elude the immune system. In a recently published article Carbone writes that sometimes SV40 produces such small amounts of large T-antigen that the virus escapes detection. Paradoxically, in this hypothesis small amounts of the virus are even more dangerous than large amounts.

Other scientists suspect that SV40 can inflict damage and then disappear completely, in what is described as a "hit-and-run" attack. This analogy is lent credence by a recent German study in which rat cells were infected with SV40 and transformed into cancer cells. When scientists searched for large T-antigen, it was no longer present in some of the cells. Further, these cells appeared to be even more malignant than those that were still expressing the protein, because the immune system could no longer recognize them as a threat.

The new theory may explain how SV40 and perhaps other viruses can induce cancer and yet not be readily detectable once tumors start proliferating rapidly. But that notion runs counter to traditional scientific thinking about cancer. "As a geneticist, I would like to see every single cell have evidence of the virus," Testa says, noting that the hit-and-run theory must still be proved. But, Testa observes, "This is an area that's going to perhaps establish a new paradigm."

Although Carbone's T-antigen research has bolstered his contention that the SV40 found in human tumors is not simply a passenger virus, until recently he had no answer to a criticism commonly voiced by those skeptical that the polio vaccine could be linked to cancer: some of the SV40 he and others have isolated in human tumors has a crucial genetic difference from the virus that contaminated the polio vaccine. The SV40 that its discoverers isolated from the polio vaccine in 1960 had a genetic feature that allowed it to replicate more quickly than the SV40 subsequently found in human bone and brain cancers and in most monkeys. That led some to question the idea that the SV40 that researchers were finding in these tumors was related to the SV40 in the polio vaccine.

To settle the issue Carbone sought to examine old vaccine stocks. He was told by government and drug-company officials that they had thrown out all the old lots. Then, two years ago, Carbone found an elderly Chicago-area physician who had an unopened case of polio vaccine from 1955, which he had stored in his refrigerator for more than forty years. "I would have gone all the way to Alaska to find this stuff, and here it was three miles away," Carbone says. Last summer Carbone finally completed tests on the vintage vaccine. He found that the tiny vials contained SV40 genetically identical to the strains found in human bone and brain tumors and in monkeys. "This proves that the SV40 that was present in the polio vaccine is identical to the SV40 we are finding in these human tumors," he says. Why was the SV40 isolated from the 1960 vaccine the faster-growing version? Because, Carbone says, both kinds occurred in the monkey kidneys used to grow the vaccine. Carbone and Janet Butel say that the SV40 that grew more quickly might have had an advantage in cell cultures -- perhaps explaining why it was the strain originally isolated from the vaccine. However, the slower-growing virus would almost certainly have an advantage in tumor formation, because it would be less likely to be detected by the immune system.

Because he believed that the slower-growing SV40 was more likely to induce tumors, Carbone wanted to see if federally mandated vaccine-screening tests for viruses were adequate to detect it. Vaccine manufacturers are not required to use state-of-the-art molecular techniques -- PCR, for example -- for virus detection. Instead they rely on ordinary light-microscope examination to look for evidence of cellular damage by viral contaminants after fourteen-day cycles in tissue culture. Although the current screening protocols -- themselves forty years old -- are, according to Carbone, more than adequate to detect the faster-growing form of SV40, his tests found that the slower-growing SV40 took at least nineteen days to grow out, and thus wouldn't be detected in the fourteen-day screening cycles. Carbone says his experiments suggest that any slow-growing SV40 present in the vaccine after the early 1960s could have gone undetected.

Carbone recently tested six vials of polio vaccine manufactured in 1996, and found that they were negative for SV40. He concludes that the colonies of monkeys used today must be free of the virus, because if slow-growing strains were present, the tests used for routine screening would not detect them. (Today's injected vaccine is produced on monkey cell lines, and is therefore free of any viral contaminants, whereas the oral vaccine is still produced on actual kidneys. Under Centers for Disease Control regulations that went into effect last month, American children should now receive only injected vaccine.) In a paper on his tests of vaccines Carbone recommends conducting extensive molecular testing of polio-vaccine stocks from the 1960s, 1970s, and 1980s to look for the slower-growing SV40. The issue is more than academic: the results would help to establish whether SV40 is present in young children today as a result of continued exposure to contaminated vaccine or as a result of human-to-human transmission based on the original, 1955-1963 exposure.

Debbie Bookchin specializes in health and political issues. Her articles have appeared in The New York Times, and The Nation. Jim Schumacher is a freelance writer who lives in Vermont. His articles have appeared in Boston magazine, The Boston Globe, and Newsday.

Illustrations by Giacomo Marchesi.

The Atlantic Monthly; February 2000; The Virus and the Vaccine - 00.02 (Part Two); Volume 285, No. 2; page 68-80.

A Study Plagued by Strife

DESPITE the accumulating evidence of SV40's association with human tumors, the NCI has been preoccupied with determining whether the virus is even present in human tumors. For more than two years the NCI's chief focus with respect to SV40 was the design and administration of a multi-laboratory study whose stated purpose was to assess whether PCR was a reliable tool for identifying the presence of SV40 in human tissue. Critics of the study, including scientists at some of the participating labs, worried that other agendas were involved. The study was directed by Howard Strickler and overseen by James Goedert. Nine labs participated in the study, including those of Keerti Shah, at Johns Hopkins; Bharat Jasani, at the University of Wales; and Janet Butel, at Baylor, but not Carbone's. The study, which was planned and administered by the NCI's Viral Epidemiology Branch, had a fairly unusual design. Instead of just seeing whether different labs could replicate one another's work, as is usually done, the labs were asked to prove that they could replicate their own work. Each lab was given a variety of samples from unidentified human mesothelioma tissues and asked to see if it could find SV40 DNA. Then it was asked to find SV40 DNA again in masked samples from the same tumor tissue.

We asked Richard Klausner, the director of the NCI, about his views on SV40 and about the design of the experiment. Klausner said that the research to date hadn't quelled his doubts that SV40 is present in human tumor tissue, and he questioned the reliability of the techniques that Carbone and others have been using. "These sorts of molecular technologies are wonderful tools but very complicated and sometimes misleading to use," Klausner said. "I think there is very good reason to question whether there has been the development of adequate standards or probes, PCR probes," for detecting the virus.

Like Strickler and James Goedert, Klausner raised the possibility of contamination to explain the positive findings of dozens of laboratories. "I do not see any compelling molecular data" to support the association of SV40 with human tumors, he told us. "In the absence of compelling clinical or epidemiological data, it's very difficult to say this looks like a pressing problem." We asked him about the many molecular studies, from numerous independent laboratories around the world, that had identified SV40 in human tumors. "There's too much irreproducibility and too many good explanations for artifact," he said. Klausner told us that the NCI has taken "an open approach but a critical one" to the notion that SV40 is associated with human tumors, and he insisted that it is seriously studying the issue. Michele Carbone's work, for instance, has been funded by the NCI. (Carbone is also funded by the American Cancer Society.)

We asked Klausner to explain why the Viral Epidemiology Branch had directed the multi-laboratory molecular-biology study, especially given that neither Strickler nor the head of the branch, Goedert, has a strong background in the field. Why hadn't he tapped an NCI division with more expertise in DNA extraction, sequencing, and characterization? "Their expertise in viruses and virus-associated disease makes [the Viral Epidemiology Branch] really the right place to do it .... As an expert in doing this sort of work, I feel that I can make that decision and I feel very comfortable with the decision," Klausner said. "What we are trying to do is establish some agreed-upon probes and standards that independent laboratories could utilize to provide ways of either validating or not validating molecular findings."

On another issue, Klausner referred to an epidemiological study that Strickler had done to determine whether SV40 was linked to human cancer. That study appeared in 1998 in the Journal of the American Medical Association, and received extensive publicity upon its release. It concluded that the NCI's database on cancer incidence shows no statistically significant correlation between exposure to SV40-contaminated vaccine and rates of cancer, including rarer cancers such as mesotheliomas, ependymomas, and osteosarcomas.

Strickler did find elevated cancer rates among those exposed to SV40, including a threefold increase in mesothelioma. Susan Fisher, an associate professor of epidemiology and biostatistics at Loyola, says that although the correlation Strickler found did not achieve statistical significance, it was at least "scientifically interesting." Strickler's study was "technically correct," Fisher says, but "it's hard to look at these numbers and turn around and say there is no evidence to suggest an association."

Moreover, Fisher says, standard epidemiological techniques may be useless in determining whether SV40 exposure is linked to higher cancer rates. If the research of Janet Butel and others is correct and SV40 is now spreading among human beings, it may be impossible to assemble an experimental group that has never been exposed to SV40.

The multi-lab NCI study concluded with six of the nine laboratories detecting SV40 in some samples. However, only two of the labs got the same positive results on samples from the same tissues. Although the multi-lab study was completed at the end of 1998, at the time this article was written it had yet to be submitted for publication.

Memos sent to Strickler by some of the participating laboratories show that from its inception the study was plagued by considerable internal strife. (Participating laboratories we approached declined to share the memos or discuss them. We obtained them independently.) Two laboratories suggested that poor DNA-extraction techniques by the outside laboratory Strickler had chosen to provide the DNA samples were to blame for the largely negative results obtained. Their concerns were heightened when it was learned that the contractor had contaminated some of the negative controls.

They also complained that Strickler was wrongly using the study to imply that previous positive findings were caused by contamination. "It cannot be that all of these laboratories are contaminated and that contamination always happens in mesotheliomas, osteosarcomas and brain tumors, while the negative controls are always negative," a scientist from one of the laboratories wrote Strickler. "Contamination is a random event .... [The] flaws and unresolved scientific issues ... have become so cumulative as to outweigh any positive scientific benefit which might be derived from the publication of this study." From another laboratory came this objection: "We feel that our comments about data interpretation are being dismissed and ignored. Your intransigence about the interpretation of the data and the conclusions of the study have forced us to admit that the collegiality and the scientific collaboration that was the basis of this study is very strained." Both laboratories maintained that Strickler's draft manuscript summarizing the study results was wrong in asserting that contamination was the cause of previous SV40 findings.

An unlikely ally in the laboratories' cause has been William Egan, the acting head of the Food and Drug Administration's Office of Vaccines Research and Review. Egan believes there is no strong epidemiological proof that SV40 is associated with human cancers and emphasizes that the current polio vaccine is free of SV40. However, he says, there is evidence that the virus may well be present in some tumor samples. After he had reviewed Strickler's draft manuscript, last February, Egan wrote a lengthy letter to Strickler criticizing it. "I think that this paragraph, and the following paragraph, imply, unintentionally so, that the positive results [of SV40 in tumors] that have been reported are due to laboratory contamination; I do not think that this should be implied." Strickler responded, "This study would not have been conducted if there was not some doubt. That point must be made and made clearly."

Later Egan chided Strickler about another section of his draft, which stated, "This multi-institutional study failed to demonstrate the reproducible detection of SV40 in human mesotheliomas." Egan wrote,

More exactly, it failed to demonstrate SV40 sequences in this set of mesotheliomas. This is not inconsistent with SV40 being found by others previously. Indeed, the fact that laboratories that previously found SV40 in their samples do not now find SV40 in these samples (and get the study controls correct) only lends credence to their previous findings .... These laboratories are able to find SV40 when it is there, and do not find it when it is not there.

Frustrated by continuing objections, Goedert and Strickler considered publishing the study without the approval of the dissenting labs, but that plan was dropped. Last September an independent arbitrator was called in to rewrite Strickler's manuscript. The dissenters apparently gained some ground. The arbitrator made major changes in its tone and conclusions. The study now states that "laboratory contamination was unlikely to have been the source of SV40 DNA"found in human tumors in previous experiments (by Butel, Jasani, and the other participating labs).

The Search for a Vaccine

THIRTY miles north of Venice, in the seaside resort town of Lignano Sabbiadoro, 200 clinicians and researchers are gathered at the international Conference on Malignant Pleural Mesothelioma. At a similar conference in Paris five years ago Carbone startled his audience when he presented his first SV40 paper.

Today a significant portion of the conference is devoted to SV40's association with mesothelioma -- testament to a sea change among researchers regarding the simian virus. Brooke Mossman, the director of the environmental-pathology program at the University of Vermont, was the first scientist to tease out the complex molecular pathways by which asbestos disrupts cellular regulatory mechanisms and causes mesothelioma. She has been impressed by Carbone's work. At Lignano she and Carbone are co-chairing a panel on the molecular pathways employed by asbestos and SV40 which lead to tumor development. In another presentation Luciano Mutti, a researcher at the Salvatore Maugeri Foundation's Institute for Research and Care, in Pavia, will report that mesothelioma patients who test positive for SV40 have a shorter life-span than those who test negative.

At the moment the floor belongs to David Schrump, the new chief of thoracic surgery at the NCI. Schrump matter-of-factly announces the results of a series of experiments he has just completed. When he "turned off" SV40 large T-antigen, he says, human-mesothelioma cell cultures that contained the virus stopped proliferating and started to die. Schrump explains that he undertook the experiment partly because he was skeptical of SV40's role in the development of mesothelioma. He and his team assembled human mesotheliomas that tested positive for SV40 and then devised a genetic bullet, a strand of RNA called an "antisense," which would prevent the expression of SV40 large T-antigen.

Within days after the antisense was administered to the cancer cultures, Schrump found, the growth rates of mesotheliomas with SV40 in them dropped dramatically; the negative controls were unaffected. One important finding was that even very low levels of SV40 appeared to be biologically important -- a discovery that speaks to Strickler's objection about the low levels of SV40 often found in tumor tissue. Schrump's study was published late last year in Cancer Research.

Another study in that same issue also supports the idea that SV40 is actively involved in mesothelioma. Adi Gazdar is a professor of pathology and the deputy director of the Hamon Cancer Center, at the University of Texas Southwestern Medical Center. He originally doubted Carbone's work on SV40. "Here's a monkey virus suddenly popping up in a rare tumor -- I was skeptical of the data," he told us. So Gazdar devised an experiment that could determine at one stroke whether the SV40 found in tumors was a lab contaminant and whether the virus is involved in tumor formation. Gazdar used a technique called laser microdissection to separate cancerous cells from nearby noncancerous ones. He found SV40 in more than half of the mesothelioma tumors. He also found the virus in some adjacent pre-cancerous cells. Significantly, 98 percent of nearby noncancerous cells tested negative for SV40. "That rules out any contamination," Gazdar says, "because if a specimen was contaminated, the SV40 would be in all parts of the specimen -- it wouldn't be localized to the mesothelium alone." Moreover, Gazdar says, his study "suggests that the virus is in the right type of cells many years before they become malignant" -- evidence that SV40 contributes to the development of cancer. Gazdar says of Carbone's work, "I feel everything he's said, I've been able to confirm, and more."

Gazdar and other scientists believe that the time has come for a major federal funding initiative on SV40 to better understand who is infected, how the virus works, and what might be done to prevent disease. "There's still a lot we don't know about the basic biology of this virus in human infections, including what tissues it infects, how it is transmitted, and when people become infected with it," Janet Butel says. "Until more studies are done, we don't know if we're looking at the only types of cancers that have an association with SV40," she says of the lung, bone, and brain cancers with which SV40 has been associated most often. "Maybe these are just the ones we've recognized so far. There may be others people haven't run across." Gazdar says, "It's such a crucial issue. Possibly millions of people are sitting with this virus in their mesothelium or other tissues and are at risk for developing cancers." Cancers that were once rare "may suddenly become not so rare," he says. "I think it's an enormous potential health problem."

Arnold Levine, of The Rockefeller University, is not convinced that the virus causes cancer in human beings, but he, too, believes that the discovery of SV40 in human tumors warrants a serious federal response. "If it's part of the cause of a disease," he says, "it has a significance in public health and I think we ought to find that out. That's a good reason to spend taxpayers' money: to do science to find out whether the public health is really monitored here properly. I think that maybe there's enough evidence in the literature now that the National Cancer Institute ought to put out an RFA." The reference is to a Request for Applications, the formal process by which the federal agency identifies a major health-research initiative and invites scientists to apply for research funds. "That would stimulate people to come in and design experiments and replicate these things." Carbone made the same suggestion to federal health officials in 1997 but was rebuffed.

Like the NCI, the Atlanta-based Centers for Disease Control maintains a stance of neutrality with undertones of skepticism. In a four-page fact sheet called "Questions and Answers on Simian Virus 40 (SV40) and Polio Vaccine" the CDC notes that SV40 has been found in some tumors and adds that "more research is needed" to confirm a causative link with human disease. It also raises the possibility of contamination as an explanation. It cites Strickler's work by name but not that of Carbone, Butel, or Testa.

Some researchers plan to conduct screenings for the virus. Joseph Testa hopes to initiate a screening program at Fox Chase's new cancer-prevention pavilion that focuses on asbestos exposure. He is collaborating with officials from the Asbestos Workers Local 14, in Philadelphia, to identify people who are particularly at risk. Carbone applauds that effort. "If you test positive for this virus, you should not be anywhere near asbestos," he says. Bharat Jasani, who has found SV40 DNA in a high percentage of the British mesotheliomas he examined, has begun testing British and Canadian mesothelioma patients, at their request. He hopes they may be candidates for future SV40-targeted therapy.

Last year scientists reported that a vaccine they had developed targeting large T-antigen appeared to help prevent and reverse tumors expressing large T-antigen in mice. Carbone and Harvey Pass, who is now the chief of thoracic oncology at the Karmanos Cancer Institute, at Wayne State University, in Detroit, are collaborating with Martin Sanda and Michael Imperiale, of the University of Michigan at Ann Arbor, who are among the vaccine's developers. They hope soon to bring the experimental vaccine to Phase I clinical trials, in which it will be tested for its safety in human beings, though not yet for whether it works. Even if the vaccine eventually proves effective in human beings, years may well pass before it is widely available.

In an age of uncontrolled AIDS in Asia and Africa, rampant tuberculosis in Russia, and antibiotic-resistant microbes in American hospitals, does SV40 really warrant a significant public-health response? There is no doubt, Carbone says, that the virus is linked to some cancers. What's more, millions of Americans now have been exposed to the virus. Studying SV40 may teach us something about the dangers of cross-species infection at a time when the use of animal tissue for medical purposes is gaining acceptance.

Good science is ultimately about the exchange of ideas unfettered by presuppositions. Sometimes great breakthroughs come out of theories that at first seemed heretical or even nonsensical. "Can you think of anything more different on earth than asbestos and a monkey virus?" Carbone says. "Yet you stick them together and they work together to be more deadly than either one of them is alone." He goes on, "This research is important in so many different ways. It's not just about SV40 and mesothelioma. It helps us understand the whole picture of how viruses interact with environmental carcinogens. This research can help us understand how completely unrelated carcinogens can work together in causing disease -- a mystery we have barely begun to unravel."

The Atlantic Monthly; February 2000; The Virus and the Vaccine - 00.02 (Part Three); Volume 285, No. 2; page 68-80.