Dna: The Secret of Life
Page 35
One of the strongest arguments in favor of the mapping approach was that the work produced was useful even before the gene had finally been identified. The hunt for the Huntington disease and Duchenne muscular dystrophy genes yielded genetic markers that could be applied in diagnosis before the genes themselves were found. So it was too with one of the most prevalent genetic disorders, cystic fibrosis (CF). But the hunt for the CF gene would prove particularly notable for two reasons: it marked the first time that a company became involved in mapping a human disease gene, and the first instance of brutal competition among the scientists involved in such an endeavor.
In cystic fibrosis patients, a thick mucus accumulates in the lungs, making it difficult to breathe. The cells lining the tubes of the lungs can't clear out the mucus in which bacteria thrive, producing pulmonary infections. Before antibiotics those afflicted had a life expectancy of just ten years; today survival rates are substantially better. CF is also one of the most common genetic disorders, with about 1 in 2,500 individuals of northern European descent affected. It follows a recessive pattern of inheritance: you need two mutant versions of the gene to be affected. But since as many as 1 in 25 people of northern European descent carry a single mutant version (though are themselves protected because they have one normal copy), there is a relatively high risk that two carriers will get together and both pass it on to their children. It therefore became a medical priority to devise a diagnostic test as soon as this became a realistic goal.
Born in Shanghai, and raised and educated in Hong Kong, Lap-Chee Tsui came to the United States as a graduate student in 1974. Tsui learned his molecular genetics doing research on viruses before moving in 1981 to Manuel Buchwald's laboratory in Toronto to work on cystic fibrosis. Tsui is a quiet, pleasant man who is nevertheless intense and passionate about his goals. Planning to track down the gene via RFLP linkage analysis, he spent the first couple of years finding CF families before starting the painstaking process of testing their DNA with every RFLP he could lay his hands on. But the luck that had smiled upon Jim Gusella in his pursuit of the Huntington disease gene did not favor Tsui: after about a year all he had managed to do was eliminate a lot of RFLPs. He needed more, and was thrilled when Collaborative Research offered to share its RFLP markers with him.
Tsui's Toronto group was not alone in pursuing the CF gene: Bob Williamson in London, who had worked on DMD, also took up the hunt, as did Ray White, now in Utah, attracted by access to the very extensive pedigrees assembled by the Mormon church. These records, the Ancestral File, permit present-day members of the church to make provision for deceased forebears, who lived their lives outside the fold or died before the church was founded in 1830. The aim is to unite families for eternity. Seldom have the needs of religion and genetics been so happily aligned.
But it was the Toronto group that would notch the first success, when it found in 1985 a linkage between one of Collaborative Research's RFLPs and the CF gene. At the time, the location of that RFLP was unknown, but, seeing its potential as a golden egg, Collaborative Research quickly set about locating it. They soon determined that it was on chromosome 7 but did not immediately inform Tsui, their collaborator. Nor did they mention the chromosomal location when they announced the discovery in the November 22 issue of the prestigious journal Science. Clearly they were trying to preserve their monopoly on the new information, but secrecy and science often don't mix well: word soon spread on the grapevine that 7 was the place to be.
As Collaborative kept quiet, Williamson and White were just days away from the same discovery. Their own two papers, published in Nature, Science's British rival, both mentioned that the key RFLPs were on 7. Tsui was incensed: he was about to lose his claim to the linkage discovery thanks to his partners' shenanigans – in science there are no prizes for coming in second – but Helen Donis-Keller persuaded Nature to accept a paper from the Toronto-Collaborative team announcing the location. So it was that three papers appeared in the November 28 issue of Nature, along with an editorial explaining how it had all come about.
The Toronto-Collaborative partnership did not survive the clash of academic and commercial cultures. Collaborative Research would find that the academic world had become wary of collaborating with them, a situation hardly helped by the crass and not very sturdy claim made by Orrie Friedman, Collaborative's CEO, that "we own chromosome 7." Fortunately, this soap opera saw its final episode in December 1985, when all the research groups agreed to pool their resources in order to test 211 families for linkage to chromosome 7 RFLPs. The results were spectacular. The RFLPs were very close to the gene, within 1 million base pairs – which made them useful in diagnosis, one of the major goals of the CF research.
The next step promised to be even more difficult. Learning that New York is halfway between Washington, D.C., and Boston is better than merely knowing it is somewhere in the United States. But when one must set out on foot from Washington to Boston, looking yard by yard for a sign that reads "Welcome to New York," the clue may seem not so helpful after all. One million base pairs might be close by the standards of linkage analysis, but it is a very long way by the standards of gene cloners who analyze regions one base pair at a time. To go the distance from the two RFLPs nearest the CF gene, Tsui teamed up with Francis Collins, who was then at the University of Michigan and would later succeed me as director of the Human Genome Project.
Collins had developed "jumping" techniques to facilitate the cloning of a gene between a pair of known RFLPs, but he was under no more illusion than Tsui about the magnitude of the problems facing them. After two years of work, they managed to localize the CF gene to a 280,000-base-pair segment of DNA, within which they found the sequence of a gene known to play an important role in human sweat glands, which are dysfunctional in cystic fibrosis patients. It seemed the complete CF gene might finally have been corralled.
The only way to be sure they had got it right was to sequence the cDNA and search for the disease-causing mutations. Given a region 6,500 base pairs long, this was quite a challenge in 1989, and it had to be done twice: once using DNA from a CF patient and once with DNA from a healthy individual. The result, however, was clear-cut: the patient's DNA was missing a stretch of three base pairs, resulting in the absence of just one amino acid in the protein. This one mutation accounts for about 70 percent of CF cases, but over a thousand others found in the CF gene also cause the disease. This multiplicity of harmful variants has greatly complicated the task of DNA-based diagnosis.
Let us now return to Nancy Wexler, David Housman, Jim Gusella, and their colleagues, whom we left back in 1983 at the triumphant moment when a particular RFLP, G8, had been linked to the gene for Huntington disease. If it seemed up until then that they had enjoyed more than their collective share of good luck in locating the HD gene with astonishing speed, the gods were soon to redress the imbalance. Finding the gene had taken a mere three years; isolating it for detailed analysis would take ten years and an international team of 150 scientists. In this case, the region where the gene had been localized was 4 million base pairs long. The Huntington disease geneticists worked hard to narrow that window, but genetic mapping gets more difficult as the genetic distance gets smaller, and finally these efforts were rewarded only with ambiguous data. Imagine the foot journey from Washington to Boston, in search of New York. Now imagine arriving at an intersection in Philadelphia to find a signpost indicating New York in both directions.
Giving up on the contradictory linkage analysis, the Huntington gene hunters devised an alternative strategy, focusing on the region that was most similar among Huntington disease patients. This approach eventually reduced the region to only 500,000 base pairs, and the time had come to turn to gene-cloning techniques. The first results were disappointing: they found three genes in the right-hand half of the region, but none showed any abnormalities in patients with Huntington. Undaunted, they explored the left-hand side and found a single gene, with the prosaic name IT15. Finally, after ten
years and many losing lottery tickets, luck had begun to smile on them once more. The gene contained a short sequence, CAG, that repeated over and over again, like the short tandem repeats (STRs) used in DNA fingerprinting. It turned out that unaffected people have fewer than thirty-five CAGs, while people with more than forty will develop Huntington as adults; in the rare instance of more than sixty, a severe form of Huntington develops before the age of twenty. CAG is the genetic code for the amino acid glutamine, so each of the CAG repeats adds an extra glutamine to the protein. In the case of Huntington sufferers, the protein coded by the HD gene – the rather difficult-to-say huntingtin – contains extra glutamines. This difference likely affects the behavior of the protein in brain cells, probably by causing molecules to stick together in gluey lumps within the cell, somehow causing its death.
It had been a tremendous effort by all the laboratories in the Hereditary Disease Foundation's team, and in recognition that it was truly a collaboration, the only name appearing as the author of the article was that of the Huntington Disease Collaborative Research Group. The same strange type of mutation – repeats of the three-base-pair sequence – had already been implicated in three other disorders, remarkably all of them also neurological diseases. We now know of fourteen of these "trinucleotide repeat disorders," but we still are no closer to under standing why brain cells are so susceptible to this kind of mutation.
It may be depressing to know that despite the substantial time it has taken to hunt down their respective genes, these disorders – Huntington, Duchenne, and cystic fibrosis – are, by the standards of geneticists, "simple." They are caused by mutations in a single gene and not much affected by environment. If you have the three-base-pair deletion in both your cystic fibrosis genes or more than forty CAG repeats in one of your Huntington disease genes, you will develop those disorders no matter where you live or what you eat or drink. There is a large number of single-gene disorders – the current genetic disease database lists several thousand – but the majority are extremely rare, each occurring in just a few families.
Much more common are "complex" or "polygenic" disorders, which include many of our most common ills: asthma, schizophrenia, depression, congenital heart disease, hypertension, diabetes, and cancer. These are caused by the interaction of several – perhaps many – genes, each of which alone has only a small effect and perhaps no detectable effect at all. And typically in polygenic disorders, there is a further complication: these sets of interacting genes may create a predisposition to a particular disease, but whether you actually develop a case of it depends on environmental factors. Suppose that you have a set of gene variants that predisposes you to alcoholism. Whether or not you actually become an alcoholic depends on your exposure to the environmental trigger, alcohol. Your fate may be quite different growing up in a dry county in Texas as compared with Manhattan. The same principle holds for asthma; in a "good" summer, when the pollen and spore counts are low, you may develop no symptoms despite being genetically disposed to the disease.
The complex interplay of genes and environment is nowhere more evident than in cancer. Cancer is fundamentally a genetic disorder caused by mutations in several genes. Each mutation alters one more element in the cell's behavior until it acquires all the characteristics of a fully malignant cell. Cancer mutations arise in two ways. Some are inherited. We have all heard the phrase "it runs in the family," and while some traits described this way – Catholicism for one – are not necessarily heritable, some kinds of cancer are. Still, the disease is so lamentably common that it is not so unusual to have two or even three cases in one family even without a hereditary component. (Geneticists studying "cancer families" therefore apply very strict criteria in deciding whether a cancer is inherited.) Plenty of cancer mutations also arise in the normal course of living. DNA can become damaged owing to errors the enzymes make in the course of duplicating or repairing the genetic molecule, or as a consequence of the side effects of the normal chemical reactions within the cell. And many cancers arise thanks to our own foolishness. Ultraviolet rays in sunlight are potent mutagenic agents to which sun worshippers willingly expose themselves, and cigarettes are a very efficient way to deliver carcinogens straight into your lungs, where they cause lung cancer. Other environmental factors, for instance asbestos in the workplace, have also been shown to promote cancer. The point is that DNA can get damaged quite naturally, but it is up to us to minimize the damage through informed social and personal choices.
In 1974, Mary-Claire King (of human/chimpanzee and Las Abuelas fame) moved to UC San Francisco to work in a laboratory studying breast cancer, where she decided to commit herself to the hunt for a breast cancer gene. At the time, the RFLP linkage approach was still six years away, but King knew that there would be clues in pedigrees, so she set about collecting families. She looked for families in which members had developed breast cancer at an early age, and in which there was also ovarian cancer, reasoning the odds favored a hereditary culprit in such cases. The only genetic markers available to her were protein markers, and after a few years she published her first breast cancer paper, describing unsuccessful tests for linkage with cell surface proteins. This was followed by other papers showing similarly negative results. Naysayers were equally negative: breast cancer is too heavily affected by the environment to permit genetic analysis, they said, referring predictably to needles and haystacks. Undeterred, King continued to refine her data-set, and by 1988, with an analysis of 1,579 families, she thought she had good evidence for a breast cancer gene in these high-risk families.
The medical world was astonished when in 1990 she reported that she had found an RFLP on chromosome 17 linked to breast cancer in a subset of 23 of her families, involving a total of 146 cases of breast cancer over three generations. She checked factors that might have confounded the analysis – perhaps these women had been exposed to more X rays, or they differed from others in their pregnancy histories – but her data held up. There was a gene at chromosome location 17q21 that when mutated greatly increased a woman's risk. King's paper set off a race to isolate the gene itself, called BRCA1 (for Breast Cancer 1), and an ongoing controversy about the commercial exploitation of genes.
Isolating the BRCA1 gene would inevitably be a big event. Even if it was important only in a small subset of high-risk families (i.e., it would only be responsible for a small proportion of all breast cancers), the insights that might come from knowing what the gene did would be cause enough for excitement. King teamed up with Francis Collins, whose gene-hunting credentials were impeccable, but the pair had tough competition. Mark Skolnick, the Utah population geneticist involved in the RFLP linkage breakthrough, formed a company, Myriad Genetics, with Wally Gilbert, whose entrepreneurial spirit had survived his uneasy tenure at the helm of Biogen. Myriad's business plan was to use the power of the Mormon family pedigrees to map and clone genes, and BRCA1 came within their crosshairs very soon. In 1994, a consortium of geneticists from Myriad, the University of Utah, NIH, McGill University, and Eli Lilly beat the rest of the world, announcing what they rather coyly called a "strong candidate" for the BRCA1 gene. They had found it. Everyone involved filed for a patent (although Myriad initially saw to the exclusion of the NIH scientists). In 1997, Myriad's application was approved.
At the moment BRCA1 was being cloned, a different consortium of geneticists, including scientists from Myriad and the Institute for Cancer Research in England, reported they had located a second breast cancer gene, BRCA2, on human chromosome 13. Once again a race began, and within a year the English group claimed success in isolating BRCA2. They knew they had bagged their quarry once they had determined about two-thirds of the gene's DNA sequence and shown it to be defective in six different families. Not to be outdone, Myriad formed yet another consortium, this one comprising institutes in Canada and France; soon they would publish the complete sequence of BRCA2, a very large gene. Of course both Myriad and the Institute for Cancer Research filed patent claims.
It sounds like a worthy thing to have brought to market: a genetic test for a very serious disease, a means to help women make informed decisions about their health. Why, then, is Myriad frequently portrayed as exemplifying all that is wrong when commerce is married to science? Myriad now has nine U.S. patents covering BRCA1 and BRCA2, and in 2001 it was granted one in the European Union, one in New Zealand, four in Canada, and two in Australia. In effect, the company now enjoys a global monopoly on these genes and worldwide control over how they are used. It is entirely reasonable that Myriad should make money from testing for BRCA1 and BRCA2 mutations – the company provides a valuable service and has invested a great deal of money to develop the test. But how much money should the company reasonably be making? Today each test costs more than $2,700. At the same time, Myriad restricts academic researchers from using the BRCA gene sequences to develop alternative tests. And information about BRCA mutations gleaned from DNA sequencing among patients enrolled in academic research projects is withheld even from the patients themselves; to do otherwise would be a diagnostic clinical use, infringing the Myriad patents.