The method was tricky but entirely rewarding. Before applying the Giemsa stain, the cell samples needed to be pretreated with quinacrine mustard, a powerful fluorescent dye. When viewed under a fluorescent microscope, the sample was no longer monochromatic. A pattern of yellow-green stripes, varying in brightness, ran up and down each chromosome. It was as if, having always watched movies with the lights on, someone finally thought of turning them off. It was a sight to behold.
Now geneticists could learn the ins and outs of the forty-six chromosomes—twenty-three pairs, one set inherited from each parent—in unprecedented detail. The patterns of stripes within each chromosome arm differed, and after just a few months of practice, a person could know instantly which chromosome he or she was looking at. With banding, each chromosome became a fully realized picture, poised to reveal the inner workings of the human genome and the mysteries of genetic diseases. The looming challenge was to understand what part of the human story each chromosomal picture was telling.
As always, Rowley’s sharp intelligence served her well. She had the method perfected by the time she returned to Chicago. She eagerly resumed her work, armed with her new skills. She’d been studying a condition characterized by unexplained anemia and bone marrow abnormalities that is often a prelude to leukemia. Before her sabbatical, Rowley had found a genetic abnormality in people who developed this so-called preleukemia. She had also been looking at cells from patients with CML and had seen additional genetic abnormalities beyond the Philadelphia chromosome in samples from patients who’d entered the blast crisis stage, the culminating phase of the disease.
Using the new banding techniques, Rowley wanted to see if the assorted abnormalities she’d spotted among CML patients were in any way consistent. Did each person experience the same progression of mutations—the same branching off of twigs, as Nowell might have explained it—or did further mutations after the Philadelphia chromosome occur willy-nilly, with no predictable sequence?
Every day, Rowley would examine slides under the fluorescent microscope and take pictures with the attached camera. She could have the 35-mm film developed at any photo store, right alongside her family photos. Night after night she sat at her dining room table, cutting out photographs of chromosomes to study the patterns, her four children wondering who was paying her to play with paper dolls.
plate 8
All of her samples had an abnormally small chromosome 22, the Philadelphia chromosome. She could see that one of the two copies of chromosome 22 in CML cells was missing some portion that was present on chromosome 22 in normal cells. But then, in 1972, she noticed something else. The samples also had an abnormal chromosome 9. With the banding patterns now revealed, she could see that the stripes among CML patients were different from those among healthy individuals. In cells from CML patients, chromosome 9 was longer than in normal cells. The difference wasn’t much; without the banding techniques, she might never have spotted it. But with the fluorescent dyes working like invisible ink on a blank page, a new message had emerged.
She checked samples of CML that had been obtained earlier in the course of the disease. The alteration in chromosome 9 was already there. Even when none of the other abnormalities she’d documented were present, chromosomes 9 and 22 were altered. The chromosome 9 abnormality, appearing at exactly the same time as the shortened 22 mutation known as the Philadelphia chromosome, had been there all along. What’s more, Rowley could see that the banding patterns missing from chromosome 22 mirrored the addition she was seeing on chromosome 9. There was no way for her to be absolutely certain, and yet she knew it had to be true. The piece that was missing from 22 hadn’t disappeared at all. The vanished stub of chromosome 22 had migrated to chromosome 9. And banding that she could see on a normal chromosome 9 but that was absent in samples from CML patients was now on chromosome 22. The Philadelphia chromosome wasn’t a deletion, as so many scientists had thought. It was, in the terminology of geneticists, a reciprocal translocation. Genetic material from two chromosomes had switched places.
plate 10
The idea of a translocation, rather than a deletion, seemed crazy to many scientists at the time. Translocations had been spotted, but rarely in cancer. Just as Painter’s erroneous chromosome count had become engrained, so had the notion that the Philadelphia chromosome was a deletion. This time around, though, acceptance came much more quickly. Rowley had already reported another translocation in acute myeloid leukemia (between chromosomes 8 and 21), and her peers soon came around. When they banded the chromosomes in their own samples of CML cells, they could see that she was right.
Yet the paradigm that chromosome abnormalities were of no consequence still prevailed. “They were dismissed as wholly unimportant, the result of genetic instability,” Rowley would explain years later, her broad face framed by her loosely bunned hair. Alterations in chromosomes—additions, deletions, translocations—were clearly indicators for various diseases, but they were thought to be a neutral presence, a curious sideshow. As far as nearly everyone involved in the research and treatment of cancer was concerned, genetic abnormalities had nothing to do with the causes of this disease. Even as the evidence continued to mount, the thinking that had brushed aside Theodor Boveri’s assertion sixty years before still held its grip.
But when Rowley found a third translocation in chromosomes 15 and 17 in cells from patients with acute promyelocytic leukemia, a variety of AML, she knew that this thinking was wrong. The existence of three independent translocations in three separate cancers was undeniable evidence that there was some essential connection between these translocations and the development of cancer. Despite the lack of professional support, she clung to the notion that somehow these translocations were causing leukemia. In particular, the one involving the Philadelphia chromosome—referred to as t(9;22), the t for translocation and the two numbers indicating the relevant chromosomes—which appeared even before the cancer had taken full effect, had to be triggering this dangerously unregulated white blood cell growth. But how?
5
_______
THE SURPRISING SOURCE OF THE CHICKEN CANCER GENE
Beginning in 1970, just as Rowley was preparing for the sabbatical year that would enable her to learn about chromosome banding, Ray Erikson’s unrelated search for Src, the protein product coded for by the src gene, was getting under way.* Erikson knew that this gene was the one responsible for the ability of the Rous sarcoma virus to cause cancer. The temperature-sensitive mutant experiments had revealed that much. But for the finding to be of any use, scientists had to learn what that gene did to turn healthy cells malignant.
Years passed with nothing to show for their work. As Erikson and the young postdocs he mentored watched one Denver winter after another come and go, the experiments yielded almost nothing. In the meantime, his search was the butt of good-spirited jokes among his peers. “A lot of my colleagues in the field would say, ‘Ray Erikson only wants to cure cancer in chickens,’” Erikson recalled. After all, it was a hen from which Rous had surgically removed a tumor decades earlier, and RSV didn’t occur in people. Still, Erikson—and many others—cleaved to the idea that the mechanisms underlying the transformation from normal to malignant cells would have similar qualities across many species. “We weren’t concerned about the fact that we were working with chickens,” said Erikson.
CURIOSITY ABOUT SRC extended well beyond Erikson’s lab. While he tackled the question of the gene’s protein product, others turned their attention to the gene itself. Where had src come from? How had a virus that, everyone knew, could replicate just fine without the src gene come to contain this cancer-causing sequence? Knowing how it had gotten into the chicken could provide important clues about how cancer worked.
Around the time that Erikson had begun his Src protein search, a theory called the oncogene hypothesis (the prefix “onco” comes from the Greek word for “bulk” or “mass”) was gaining traction. Two scientists at the NIH had put forward the ide
a that cancer was caused by oncogenes—that is, genes specifically programmed to trigger this deadly disease—that had been deposited into the genome of vertebrates by viruses eons ago. These oncogenes, the theory went, would remain latent unless they were activated by some environmental carcinogen.
Over at the University of California–San Francisco, J. Michael Bishop, who ran a lab researching cancer genetics, found the theory dubious. “[It] didn’t make a lot of evolutionary sense,” Bishop explains. “A gene that has no purpose in a cell does not last very long.” Bishop figured that such a gene would not have remained in the DNA of the species in question because the gene had no role in survival. But the question remained: Where had the src gene come from? How had it become part of the viral genome? By altering the DNA, researchers had created variations of RSV that did not induce cancer. If a genetically mutated strain of RSV failed to transform cells from healthy to cancerous, then clearly there had to be a gene responsible for transformation. If the virus could lose the gene and still continue to replicate, then that cancer-inducing gene wasn’t inherent to the virus’s survival. Inducing cancer was a gratuitous trait that, though it may have somehow enhanced the virus, was not inextricably tied to its existence. Replicating was what viruses did. A feature unrelated to replication—such as causing cancer—had to be extra, added to the virus after it evolved. So where had the gene come from? Why was it part of the RSV genome, and how had it gotten there? Answering these questions set in motion a third crucial pathway of research. First was unraveling the Philadelphia chromosome, then searching for the Src protein, and now tracing the src gene back to its source.
Bishop knew that the gene had to have come from somewhere outside of the original viral genome. The work by Hanafusa, Martin, and others more than a decade earlier had also helped lay this foundation of understanding. But the question of where the gene had come from remained a stubborn riddle. The idea that a cancer-causing gene could originate in the animal that got the cancer simply did not exist. Like the invention of flight or the first notion of a round Earth, the concept would have seemed outlandish at the time.
For Bishop, as unlikely as the oncogene hypothesis seemed, it was as good a starting place as any. So he and Harold Varmus, a scientist who’d joined Bishop’s lab to learn about viruses, began searching for the src gene in chicken DNA. Today, this work would be simple: The viral genes would be cloned quickly, perhaps even by an outside company, and finding whether there was a match in the chicken genome would be easy. But Bishop and Varmus were sending out their search party in the era before recombinant DNA, and creating effective foot soldiers and flashlights was just as difficult as finding a matching sequence in the vast genetic forest.
plate 7
First, Bishop and Varmus took the RNA from the RSV genome and made a complementary strand of DNA, just as the virus does when it replicates inside the host. They rendered that DNA radioactive. Then, they took RNA from a mutant version of RSV that did not contain src. They zipped together the mutant RNA with the radioactive DNA. It was like pasting up two strips of wallpaper so that the pattern lined up correctly at each edge.
Because the radioactive DNA strand and the RNA strand were complementary, Bishop and Varmus knew they would match up just as DNA and RNA did in normal cell division. But because the RNA strand lacked the src gene, the code for the src gene on the radioactive DNA was left out of that match-up, leaving Bishop and Varmus with a piece of radioactive src. A piece of wallpaper was left trailing onto the carpet.
That radiating gene could be used as a probe, a way to poke and prod DNA from other sources for a matching gene sequence. They knew src had come from somewhere outside of the viral genome. The probe gave them a way to search for it in other genomes, because only an actual src gene would match up with the src probe. Like a detective with a photograph of a missing person, they could hold the probe up to DNA from one species or another and search for the gene.
It took them four years to create the probe. The next step was to send out the search, using their radioactive src to find the gene outside of the virus. The first obvious genome to investigate was the chicken. After all, that was where the virus had been found in the first place. “To my utter astonishment, there it was,” said Bishop. “And then we found it in duck DNA, and then we looked at more ancient birds, like ostrich, and it was there.”
The discovery, made in 1976, was absolutely startling. Here was a virus that had somehow taken on a gene that caused cancer in chickens. And where had the gene come from? The chicken. What’s more, the gene was present through a long trace of history. “Here it is preserved through all of evolution from the early metazoan on,” said Bishop, who is still at the University of California–San Francisco forty years later. “It’s a normal gene.” So it turned out that src was a legitimate gene present in many species, which, if transposed into the genome of a retrovirus, became an oncogene.
The attempt to find the original source of the src gene had started as part of the quest to understand how src turned normal cells cancerous. But in tracing the gene back to its source, Bishop and Varmus inadvertently transformed the entire field of oncology. No one had imagined that cancer-causing genes came from the cells of the same species that later got the cancer.
The surprising discovery gave birth to two crucial notions that quickly became fundamental to the research of cancer and its treatment, the wagon wheels on which the future of cancer medicine would be pioneered. First, Bishop and Varmus proved without a doubt that normal genes can change into cancer-causing genes. How and why that change occurred remained a mystery, but the fact itself was undeniable: A normal gene that goes about its business in the DNA of a healthy animal can, without notice, turn into an oncogene, somehow launching a series of cellular events that lead to cancer.
But that wasn’t all that the src probe research showed. More shocking was Bishop and Varmus’s proof that these normal genes with oncogenic potential—proto-oncogenes, as they became known—were part of the animal that got the cancer. Until this moment, most scientists had assumed that the oncogene had come from the scene of the crime: the moment RSV had infected the chicken. But Bishop and Varmus traced the criminal back to its roots, to the time when it was still innocent. The src in RSV was an oncogene, but this same gene existed in the normal cells of healthy chickens. The oncogene had originated not in a virus but in a mammal cell. At some point in its history, RSV had infected a chicken and picked up the src gene from the chicken as it replicated inside its host. But when src was incorporated into the viral genome, it had mutated into a cancer-causing form of itself. Thus was born the understanding of the cellular origin of oncogenes. The phenomenon of a virus taking on a proto-oncogene from its host was dubbed “oncogene capture.”
But was the normal src gene present in other mammals, such as mice and—the ultimate question—humans? As Bishop and Varmus widened their net to allow for more diversity with their probe, they found that normal, healthy, non-cancer-causing src was in the human genome. So a gene known to cause cancer in the animal from whence it had originally come was also present in human DNA. Might cancer in people be attributable to the same mechanism? Did humans have proto-oncogenes that could, somehow, turn oncogenic? If so, how would that change occur? In 1989, Bishop and Varmus would win the Nobel Prize for their discovery of “the cellular origin of retroviral oncogenes.” They had proved the seemingly impossible: Cancer-causing genes in retroviruses were once normal genes in normal cells. This work transformed all of cancer research to follow. Eventually, the journey to understand oncogenes would lead to the genetic roots of cancer.
THE IDEA—RATHER, the fact—that src was a normal cellular gene that had, at some point in history, been scooped up by RSV and, through that adoption, been rendered oncogenic, brought a whole new level of importance to Erikson’s search. Understanding the function of the gene was more important than ever. If cancer genes were altered forms of normal genes, then knowing what those genes normally did could
prove crucial. It might allow scientists to identify other proto-oncogenes. It might provide clues about why the alteration occurred. Maybe, eventually, it would allow for cancer to be prevented or at least treated more effectively. Pinpointing the cellular mechanism associated with a proto-oncogene could mean pinpointing the cellular mechanism that was hijacked to induce cancer.
Bishop, too, began looking for the protein product of src. The question was so obviously central to understanding cancer: If the transformation of this single gene was enough to cause cancer, then knowing what protein that gene encoded was vital. Understanding where the oncogene had come from was cause to celebrate. But why did the change from proto-oncogene to oncogene occur? And how did the oncogene do its bidding? That the gene encoded a protein was clear; most genes did, according to the science of the time. But no one had a clue about what type of protein was involved.
Proteins are made of strings of amino acids, a family of twenty-two molecules inherent to all life, most of which occur naturally in our bodies. Proteins shape how we look, determine the speed of our metabolism, and conduct every minute process constantly churning inside our cells. Discovered in the early 1800s by a German scientist named Gerrit Mulder, proteins—a word derived from the Greek protos, meaning “first”—are considered the basic substances of all living organisms; they translate our genetic code into a physical, tangible result. Egg whites were the first protein Mulder identified, and the catalog grew exponentially from there.
Knowing the pedigree of the src gene did little to aid the search for the protein it encoded. There were scores of proteins to choose from and still few clues about the criminal’s weapon of choice. In 1976, six years after Erikson had begun his search, he did not publish a single scientific paper, and he was beginning to wonder about the future of his lab. Without any publications to show for his work, grant money would grow harder to come by. “There were some very good students there,” recalled Erikson. “If I couldn’t acquire funds, they would have to be let go.”
The Philadelphia Chromosome Page 5
