Knudson, Alfred

Alfred G. Knudson Jr.

Fox Chase Cancer Center

Nowell, Peter

Peter C. Nowell

University of Pennsylvania School of Medicine

Rowley, Janet

Janet D. Rowley

University of Chicago Medical Center

For incisive studies in patient-oriented research that paved the way for identifying genetic alterations that cause cancer in humans and that allow for cancer diagnosis in patients at the molecular level.

In 1960, Peter Nowell discovered the “Philadelphia chromosome.” It was only four years earlier that the precise number of human chromosomes had been fixed at 46 and chromosome studies were, by today’s standards, quite primitive. After photographing chromosomes under a microscope, researchers literally cut them up, like paper dolls, and arranged them according to size, thereby producing a karyotype. Nowell, a tumor biologist in the pathology department at the University of Pennsylvania School of Medicine, was interested in the relationship between cancer and alterations in genes (although he had no proof there was one).

One day while “diddling around with leukemic cells in culture,” and rinsing them with tap water, Nowell noted that cells were dividing. Staining them with a special dye made the cells’ chromosomes more visible. Nowell collaborated with the late David Hungerford who, he says, “knew more about chromosomes than I did,” and together they made the startling observation that individuals suffering from chronic myelogenous leukemia (CML) had an abnormally small chromosome in the tumor cells.

At a time when the idea that cancer had a genetic basis was widely disbelieved, Nowell’s results provided the first clear evidence that a particular genetic defect in a single chromosome can lead to a population or clone of identical cells that accumulate in numbers to form a deadly malignancy. What made Nowell and Hungerford notice the Philadelphia chromosome, named after the city in which they worked, was its size. The tiny Philadelphia chromosome became a clear and consistent marker of CML, a cancer of the myeloid or bone marrow cells, with broad implications for diagnosis and prognosis of disease.

But even so, many researchers continued to believe that genetic aberrations were the result, not the cause, of malignancy. It would be more than a decade before other cancers were found to be associated with other, consistent chromosomal abnormalities. Likewise, more than a decade passed before scientists understood exactly why the Philadelphia chromosome was so small.

Janet Rowley, who has spent her entire professional career at the University of Chicago, would be the one who understood. In 1961, Rowley went to Oxford with her husband, who was on sabbatical. She got a grant to study chromosomes and, when she returned to Chicago, even though she had “no special interest in chromosome abnormalities in hematological diseases,” the course of her research was set by her ready response to clinical colleagues who frequently asked her to study their patients. “I came to realize that there were many questions about chromosome changes in patients that would be rewarding to study,” noted Rowley, and for the next decade she labored over the microscope looking at chromosomes in leukemic cells.

It is worth noting that in science nothing is quite as powerful as a prepared mind armed with new technology and in the early 1970s geneticists perfected the art of “banding,” a new way of visualizing chromosomes with great clarity. Rowley was ready. Using banding technology, she discovered that the tiny Philadelphia chromosome was missing a piece of itself.

In fact, she showed that in patients with CML, a crucial segment of chromosome 22 broke off and moved to chromosome 9, where it did not belong. Moreover, a tiny piece of chromosome 9, which carried an oncogene, had moved to the breakpoint on chromosome 22. Rowley had identified the first “translocation” in cancer, providing clear evidence that the cause of CML could be related to the fact that by moving from one chromosome to another, the aberrant segment of chromosome 22 was no longer sitting next to genes that controlled its behavior.

Rowley and her colleagues subsequently identified several other signal chromosome translocations, including one characteristic of acute myeloblastic leukemia. Quickly picking up on her lead that translocations contribute to malignancy, scientists around the world joined the search for chromosomes that either switched genetic material or, in some cases, lost it altogether in a process known as “deletion.” A whole new area of cancer genetics opened up.

Not content to rest on her laurels, Rowley is still in the forefront. Using yet newer techniques for detecting abnormal chromosomes (called spectral karyotyping), Rowley found a chromosomal rearrangement that characterizes one of the childhood leukemias, and her work continues.

In addition to its implications for accurate cancer diagnosis, understanding the genetics of cancer at the level of chromosomes and genes is now opening the door to the design of drug and radiation therapy that encourages the hope that very specific therapies will be developed for specific diseases.

Explaining why some tumors are hereditary and others appear to be “sporadic” was one of the great conundrums of cancer biology—until Alfred Knudson, Jr., came up with the “two-hit” hypothesis that provided a unifying model for understanding cancer that occurs in individuals who carry a “susceptibility gene,” and cancers that develop because of randomly induced mutations in otherwise normal genes. Like many significant conceptual leaps in science, Knudson’s two-hit hypothesis was met with skepticism when he first published it in 1971.

Knudson, who has been affiliated with the Fox Chase Cancer Center in Philadelphia since 1976, was studying children with retinoblastoma, a cancer of the eye, noting differences between the 40 percent of cases with heritable tumors and the 60 percent of non-heritable cases. “Most people assumed that retinoblastoma genes were inherited in a dominant fashion—that is, if you had the gene, you would get the cancer,” Knudson said. But he observed the variable number of tumors that develop in individuals who inherit one retinoblastoma gene, and proposed that a second mutation, after conception of the child, was necessary for a tumor to develop. The same gene, known as RB1, is involved in children with the non-hereditary form, but both mutations, or hits, occur after conception.

The hits can occur in many ways—from an environmental toxin, dietary factors, radiation, or the kind of random mutation that sometimes occurs during the intricate process of normal cell replication. Knudson proposed that retinoblastoma develops either because both copies of a key gene are lost, or because they are inactivated and unable to function.

In essence, Knudson, far ahead of his time (and ahead of his own hard data) hypothesized that some genes’ normal role in life is to behave as anticancer or tumor-suppressor genes that keep cell division under healthy control. At first, the strength of his hypothesis rested on a complex mathematical model, but was supported in 1976 when Knudson and others showed that some patients with hereditary retinoblastoma are missing a segment of chromosome 13 in all of their cells. In 1986, other scientists applied the tools of molecular technology to clone the gene, RB1, so that its function as a tumor-suppressor gene could be studied in detail.

One of the most significant achievements of molecular genetics in the past few years has been the identification of a number of tumor-suppressor genes that, when mutated, lose their ability to control cell division. Malignancy is the result. Although Knudson’s initial studies were directed at relatively rare tumors, including retinoblastoma and Wilms’ tumor (another childhood cancer with heritable components), it is now apparent that his two-hit hypothesis explains the etiology or origin of many common forms of cancer, and is one of many defining concepts behind all of modern cancer biology.

Key Publications of Janet Rowley

Rowley, J.D. (1973) A new consistent chromosomal abnormality in chronic myelogenous leukemia. Nature 243: 290-293.

Rowley, J.D. (1975) Nonrandom chromosomal abnormalities in hematologic disorders of man. Proc. Natl acad Sci USA 72: 152-156.

Rowley, J.D., Golomb, H.M., and Vardiman, J.W. (1977) Nonrandom chromosomal abnormalities in acute nonlymphocytic leukemia in patients treated for Hodgkin’s disease and non-Hodgkin lymphomas. Blood 50: 759-770.

Thirman, M.J., Gill, H.J., Burnett, R.C., Mbankollo, D., McCabe, N.R., Kobayashi, H., Ziemin-van der Poel, S., Kaneko, Y., Morgan, R., Sandberg, A.A., Chaganti, R.S.K., Larson, R.A., LeBeau, M.M., Diaz, M.O., Rowley, J.D. (1993) Rearrangement of the MLL gene in acute lymphoblastic and acute myeloid leukemias with 11q23 chromosomal translocations. New Engl J. Med. 329: 909-914.

Rowley, J.D., Reshmi, S., Sobulo, O., Musvee, T., Anastasi, J., Raimondi, S., Schneider, N.R., Barredo, J.C., Cantu, E.S. Schlegelberger, B., Behm, F., Doggett, N.A., Borrow, J., Zeleznik-Le, N. (1997) All patients with the t(11;16)(q23;p13.3) that involves MLL and CBP have treatment-related hematologic disorders. Blood 90: 535-541.

Award presentation by Joseph Goldstein

Dolly the Sheep may be the world’s most famous clone, but the most infamous clones are the ones that produce cancers in human beings. Cancer begins when one cell in the body undergoes a genetic change that endows that cell and its clonal descendants a growth advantage vis-a-vis other cells. Over time, the cancer clone accumulates other mutations that help it to grow. Two classes of genes are targets of the mutations that convert normal cells to cancer clones. One class comprises the proto-oncogenes discovered by Bishop and Varmus in 1976. These are cellular genes that normally stimulate cell growth. The second class of cancer-causing genes are the tumor suppressor genes. They have an opposite action: They produce proteins that normally inhibit growth. When the proto-oncogenes and the tumor suppressor genes operate normally, the cell cycle (about which Ira Herskowitz so eloquently spoke) is exquisitely controlled, cell division proceeds in an orderly fashion, and cancer does not occur. Under the normal conditions of cell growth, the proto-oncogenes are the accelerators of the cell cycle, and the tumor suppressor genes are the brakes. Or, in the parlance of Wall Street, the proto-oncogenes are the Bulls, and the tumor suppressor genes are the Bears.

Cancer occurs when mutations create an imbalance between the accelerating actions of the proto-oncogenes and the braking actions of the tumor suppressor genes. Mutation in a proto-oncogene acts in a dominant fashion and converts the normal version of the gene to an oncogenic form that produces a hyperactive growth-stimulating protein. This sequence of events was demonstrated by several scientific groups, including those led by Bob Weinberg and Mike Wigler in the early 1980s in classic experiments on Ras. Mutation in a tumor suppressor gene acts in the opposite way, causing cancer in a recessive fashion by inactivation of the tumor suppressor protein. This inactivation requires that both copies of the same gene be disrupted, an event that is referred to as “two hits.” In order for a single cell to evolve into a cancer clone, mutations must occur in various combinations, involving the dominant activation of three or four proto-oncogenes plus the recessive inactivation of three or four tumor suppressor genes. In all, scientists have identified over 100 genes that cause cancer in humans—75 proto-oncogenes and 25 tumor suppressor genes. It is now established beyond any doubt that alterations in our genes are the fundamental initiating event in human cancer. The genetic paradigm for cancer is here to stay.

This year’s Lasker Award in Clinical Research celebrates the accomplishments of three scientists who provided the first convincing evidence that human cancers arise from mutations in our genes, setting the stage for our present understanding of the genetic basis of cancer. The pioneering work of Peter Nowell, Janet Rowley, and Al Knudson was done in the 1960s and 1970s. Their insights were way ahead of their time, decades before the tools of molecular biology were available to confirm their hypotheses.

Our story begins 40 years ago at a time when our knowledge of cancer genetics was simply nonexistent. In 1958, Peter Nowell, then a newly minted instructor in pathology at the University of Pennsylvania, teamed up with the late David Hungerford to apply the primitive techniques of cytogenetics to the study of chromosomes in patients with leukemia. Within two years, Nowell and Hungerford discovered the first chromosomal abnormality in human malignancy—a piece of the smallest chromosome, no. 22, was missing in 9 out of 10 patients with chronic myelogenous leukemia. The important conceptual point was that this genetic change was present in all the cells of this neoplasm. This led Nowell and Hungerford to advance the audacious proposal that cancer could arise from a mutation in a single cell. In other words, cancers could be clonal. The discovery of the Philadelphia chromosome had immediate clinical application in the diagnosis and management of patients with leukemia, and it opened a new field of research cancer—cytogenetics. Nowell went on to study chromosomal changes in other tumors, and this led him to advance the theory of the clonal evolution of tumor progression. He proposed that carcinogenesis occurred in multiple steps. The progression of the ancestral tumor cell to a full-blown cancer results from the sequential acquisition of additional mutations that confer a selective growth advantage to the original clone. This powerful theory, originally published in 1976, is widely accepted today. It has been verified most elegantly by Vogelstein and colleagues in their molecular analysis of the multiple mutations in oncogenes and tumor suppressor genes that occur in human colon cancers.

One other brief comment about Peter Nowell, and this relates to his remarkable attachment to Philadelphia and Penn. He was born in Philadelphia, left home briefly to attend college for four years at Wesleyan University in Connecticut, returned to Philadelphia to enter medical school at Penn in 1948, and has remained there for the last 50 years. Such municipal and institutional fidelity is almost unequaled in this peripatetic age.

I say almost unequaled because our next honoree, Janet Rowley, has outdistanced Peter by four years in both municipal and institutional fidelity. Janet was born in Chicago, went to college at the University of Chicago where she read all the Great Books, obtained her MD degree from the University of Chicago, and did her internship in Chicago. Except for two sabbaticals at Oxford, she has been affiliated with the University of Chicago for 54 years. Despite this interminable love affair with Chicago, I’m told that Janet’s favorite Frank Sinatra song is not “Chicago,” but “New York, New York.” Janet, today New York honors you.

Stimulated by Nowell’s discovery of the Philadelphia chromosome, Janet began her scientific career in 1962 by analyzing the chromosomes from patients with different types of leukemias and lymphomas. She describes the 10-year period between 1962 and 1972 as “the era of chromosomal chaos.” The techniques for identifying chromosomes were primitive. Except for Nowell’s Philadelphia chromosome, no one had been able to discover a second consistent chromosomal change in any cancer. The Philadelphia chromosome stood alone as a medical curiosity for more than a decade, and many disillusioned biologists began to question whether it was really the cause of the myelogenous leukemia. Maybe it was an epiphenomenon that was secondary to the process of transformation.

Finally, in the early 1970s, new staining techniques were developed by Swedish scientists that allowed each human chromosome to be identified on the basis of its unique pattern of bands. Cytogeneticists could now bring order to the chromosomal chaos. Janet is the scientist who first saw the light in 1972 when she looked into the microscope and discovered that the Philadelphia chromosome was not a simple deletion of chromosome 22, as everyone supposed. Rather, it was a reciprocal translocation between pieces of chromosome 22 and chromosome 9. The result was that the DNA from the end of chromosome 9 moved to the end of chromosome 22, and conversely the DNA at the end of chromosome 22 moved to the end of chromosome 9. In the Philadelphia chromosome, this exchange of DNA creates a new hybrid oncogene that stimulates cell growth. This chromosomal rearrangement was the first somatic translocation to be identified in any malignant or nonmalignant disease in man or animals.

Janet speculated that nonrandom chromosomal rearrangements in tumor cells might provide the crucial signposts that point to the locations of new cancer-causing genes. This proposal, bold at the time, has turned out to be an understatement. More than 70 translocations are now known to cause cancers in humans. In each case the translocation breakpoint has been molecularly cloned and shown to produce an activated oncogene or an inactivated tumor suppressor gene. More than half of these 70 genes are new ones that were not previously known to the scientific community. Janet herself identified 10 of these 70 translocation breakpoints, including the translocation in follicular lymphoma that led to the cloning of the now-famous Bcl-2 gene that regulates the cell suicide program. When Bcl-2 is activated by a chromosomal translocation, lymphomas occur because the cells fail to die. Cancer cells must not only show uncontrolled growth; they must also find a way to avoid programmed cell death, which is the body’s way of eliminating cells that don’t obey the rules. This newly appreciated aspect of malignancy is now being unraveled by Stanley Korsmeyer in St. Louis and Suzanne Cory in Melbourne, thanks in large part to the careful cytogenetics of Janet Rowley.

Throughout her career, Janet led the world in applying the latest technologies of cytogenetics and molecular biology to clinical medicine. She pioneered the use of DNA-based techniques to diagnose patients, to follow their progress, and to develop more effective treatment protocols targeted to particular subgroups.

In 1985, Mike Bishop, the doyen of the tumor retrovirologist, noted in one of his refreshing review articles that studies of tumor chromosomes were no longer “mere amusements for the myopic microscopist,” because they often provide useful clues for the molecular biologist. Nowell pointed out to Bishop that “we microscopists had known all along that these were important clues, but had to wait for the retrograde retrovirologists to provide us with the means to exploit them.” Peter obviously made his point, and a decade later in another of his inimitable reviews Bishop wrote in reference to Janet Rowley’s classic 1973 paper on the Philadelphia translocation: “Those of you who persist in thinking of cytogenetics as simply peering through a microscope should read the crucial paper by Rowley: it is a gem of ingenuity.” No one could have said it better!

And now to Al Knudson. Al differs from Peter and Janet in two ways. First, he is not a cytogeneticist, and second, his strong point is not fidelity—at least in the municipal and institutional sense. Quite the opposite, in fact. For the first 44 years of his life, Al bounced back and forth like a ping pong ball between Los Angeles and New York. Born in Los Angeles. College at Caltech. Medical school in New York at Columbia P&S. Back to L.A. for an internship in Pasadena. Back to New York Hospital for a residency in pediatrics. Back to L.A. for a second residency at L.A. Children’s Hospital, followed by a PhD in genetics at Caltech, and chairmanship of the department of pediatrics at City of Hope in Duarte. And then back to New York as Dean at Stony Brook. Too bad for Al that there were no frequent flyer miles in those days.

In 1969, Al got tired of coast-to-coast travel and moved to the middle of the US where he could get to either coast in a hurry. I’m proud to say that he settled in Texas, where he became Associate Director of the MD Anderson Hospital and Professor of Biology at UT Medical Center in Houston. After 10 years, Al apparently decided that life in middle America was too middle-of-the road. He yearned for a seacoast. Resuming his peripatetic ways, he moved back to the east coast—this time to Philadelphia and to the Fox Chase Institute for Cancer Research. He has now spent 22 years in the same city at the same institution. In 30 years he’ll catch up to Peter and Janet.

During his many trips back and forth between L.A. and New York, Al had plenty of time to think, and he began to ponder the problem of pediatric cancers, a problem that had intrigued him since his early days as a pediatric resident at New York Hospital. He had seen and read about rare cases of retinoblastoma, neuroblastoma, and Wilms’ tumor that occurred in newborn babies. This early onset suggested to Al that the number of predisposing events in these pediatric tumors must be quite small. When Al moved to the MD Anderson hospital in 1969, he was no longer distracted by the folderol of L.A. Within two years he proposed a novel genetic theory, based on a complex mathematical model, to explain how tumors of the eye arose in retinoblastoma. As a card-carrying pediatrician and geneticist, Al knew that retinoblastoma appeared clinically in two forms —a familial form in which a tumor-causing gene is transmitted from parent to offspring and a sporadic form in which affected children have no family history. He proposed that retinoblastomas required two mutations that arise differently in the sporadic and familial forms. In children with the familial form, the first mutation, which Al called the first hit, was inherited; it was present in the germline and thus found in every cell in the body. The eye tumor did not develop until a somatic cell—in this case a retinal cell—underwent a second mutation, which he called the second hit. This second hit would be triggered by an environmental insult, such as radiation, a chemical, or some dietary factor. In children with sporadic retinoblastoma, there was no inherited mutation. The two hits must both occur in the same retinal cell in both copies of the same gene. The likelihood of the same retinal cell undergoing two independent mutations in the same gene would be extremely rare. This formulation immediately explained why children with the inherited form of the disease developed multiple tumors in both eyes at a very early age, whereas children with the sporadic form typically developed only one tumor in one eye at a much later age.

Knudson’s two-hit model, so elegant in its simplicity yet so powerful in its ability to predict, provided the first unifying explanation of how hereditary and sporadic forms of the same cancer could involve the same gene. The two-hit model of 1972 also predicted that malignancies can occur because of a loss or inactivation of both copies of a gene that normally functions to inhibit cell growth, which we now call a tumor suppressor gene. The idea that cancer in humans would be caused by a loss of gene function and that hereditary and sporadic forms of cancer could both involve the same gene were heretical. If 1962–72 was the decade of chromosomal chaos, then 1972–82 was the decade of hereditary heresy. Like a fine Bordeaux wine, Al’s new ideas had to ferment and age for many years before the molecular biologists were willing to taste and experiment.

Direct confirmation of the two-hit model came in 1986, 15 years after its formulation, with the cloning of the retinoblastoma gene by Dryja, Friend, and Weinberg. Today, we know that the retinoblastoma gene product, the RB protein, is a master regulator molecule of the cell cycle. The Knudson model has now been validated in thousands of experimental and clinical systems and provides the conceptual foundation for our current views of the crucial role of tumor suppressor genes in cancer. The discovery of tumor suppressor genes, such as p53, the two breast cancer genes BRCA1 and BRCA2, and the familial polyposis gene APC, is directly traceable to Knudson’s ideas. The most frequent tumors in human—cancer of the colon, breast, lung, and prostate—all involve mutations in tumor suppressor genes.

It is extremely rare in biomedical science for a single idea to be so influential in changing the direction of an entire field. It must be especially gratifying to Al—the consummate pediatrician, medical geneticist, and cancer biologist—to know that cancer centers throughout the world are routinely performing DNA-based tests of tumor suppressor genes for diagnosis and counseling of patients from cancer-prone families.

As a final footnote, it’s personally gratifying to me to note that all three clinical discoveries that we celebrate today were made by scientists trained in medicine whose initial stimulation came from their contact with patients. Their research accomplishments epitomize the dictum that “medicine is the tutor of biology.” In this sense, our awardees are superb models for young physicians who aspire to careers in creative patient-oriented research.

Interview with Alfred Knudson and Richard Klausner

Peter C. Nowell

Janet Rowley interviewed by Francis Collins

In an August 1998 interview with Francis Collins, director of the National Human Genome Research Institute, Dr. Janet Rowley discusses a career that spans several decades. She describes how she became interested in medicine while in college, and how that interest led her to make exciting discoveries in the field of cytogenetic research.

Part 1: An Early Start
At age sixteen Dr. Rowley was accepted to a four-year program at the University of Chicago’s Hutchins College. Here, she explains what effect the program had on her, and why she decided to pursue a career in medicine.

Collins: Well I am honored and delighted. Let me start off by saying congratulations to you for a well-deserved honor. I am just thrilled that the Lasker Foundation has made the right choice here.

Rowley: Well thank you. Obviously I’m on cloud nine and have been for quite a while.

Collins: When did you find out? Who…what were the circumstances of this revelation?

Rowley: Well I was actually in Germany with my husband at the Wilsede meeting on leukemia and Jordan Gutterman tracked me down there late one evening, June 26th or thereabouts.

Collins: Ah ha!

Rowley: So it was extremely exciting.

Collins: And you’re of course sworn to secrecy until some time in September when this all gets trotted out for the world to see.

Rowley: Well I thought that was the case, but then I got a number of phone calls from people all across the country and I decided it wasn’t such a well kept secret after all. But I think the main thing is to make sure that it’s not something published in a newspaper article. That is probably the only reasonable way to deal with this.

Collins: I see. I imagine they’d be a little upset to be preempted in some sort of public announcement. But the rest of us can certainly enjoy the rumors flying around. Well I think this is just great. I’m tickled to have the chance to talk with you in a format that I guess then is going to appear on the web site that Bradie Metheny runs for the Lasker Foundation. Because I think people are always curious to know how this came to pass and you had a very distinguished career, but I think particularly young scientists might be interested in knowing something about your earlier years. So maybe we could start off there.

In terms of your own training and the way in which you got involved in research, which I know from previous conversations is a little unusual compared to the sort of path that many folks follow. So can you run down that part of your life?

Rowley: Okay, well obviously the first thing is that my parents were very encouraging of me in any kind of intellectual activity. My mother always hoped that I would be a doctor. The critical component was my getting a scholarship to the University of Chicago four-year college when I was a junior in high school. This is the so-called Hutchins College.

Collins: Oh, oh.

Rowley: We were really treated as though we were college students, even though we were sixteen and seventeen and given a great deal of responsibility and dealt with as adults. It was an important experience for me because we were taught to question and to read primary materials not just what you’d see in a textbook. So we had almost no textbooks to study.

Collins: Yes. You were probably surrounded by a group of pretty bright peers as well.

Rowley: Absolutely. We were a class of sixty-five. So we weren’t very big, and we were separated from the standard University of Chicago college. Everyone was both smart and motivated and the teachers were outstanding. So that was important.

I found, as I took college biology classes that I enjoyed them. Initially I was going to go into physiology because that seemed to me to be such a dynamic field.

Collins: Yes.

Rowley: This was back in 1942. But all of my lab mates were pre-med. So I decided well I might just as well be pre-med along with them.

Collins: And make your mother happy.

Rowley: That’s right. So then I applied to medical school and the quota for women was filled, because it was three women out of a class of sixty-five.

Collins: And that was all they wanted?

Rowley: That’s right. And they’d already selected their three, so I had to wait nine months.

Collins: Oh, my God.

Rowley: But since I was only nineteen at the time, it wasn’t a great tragedy. So I started medical school at age twenty in 1945. I enjoyed medicine, and I always intended to be a clinician, but also, because I was married the day after I graduated from medical school, I intended to do medicine only part-time, because I wanted to take care of my family.

Collins: Yes.

Rowley: I ultimately had four sons. I worked, therefore, part-time in well-baby clinics and then began working at a clinic for retarded children. I was working in that clinic in the late fifties when Jerome Lejeune discovered that Down Syndrome was trisomy for chromosome twenty-one.

Collins: Yes.

Part 2: A Transition from the Clinic to the Laboratory
After practicing medicine in children’s clinics, Dr. Rowley traveled to Europe on an NIH fellowship. Here she describes this early research experience, and talks about her decision to continue her efforts in the United States. She also explains the rather unusual circumstances surrounding her initial research at the University of Chicago.

Rowley: My husband was going to Oxford to work with Lord Florey. So I applied for a special NIH fellowship, which allowed me training in Europe. I worked with Laszlo Lajtha and Marco Fracarro and learned cytogenetics in Oxford.

Collins: Now, was that your first foray into research?

Rowley: Really, it was. I did some research after I graduated from medical school because my husband was behind me in school. That was really pretty minor. But I worked with Laszlo Lajtha on the pattern of DNA synthesis in chromosomes. That was just at the time when people became aware of the late labeling “X” chromosome.

Collins: Yes.

Rowley: At this time, working with Laszlo, we didn’t know about any of the work of say Jim German or Grumbach or others. I could go to Sweden and get material from Jon Lindsten on patients with abnormal “X” chromosomes, including four “X’s” and a “Y” and ring “X’s”, etc.

Collins: Yes.

Rowley: I did autoradiography on this material, and we showed that in all of those patients with abnormal “X’s”, all of the “X’s” except for one were late labeling, therefore, presumably inactive. In cells with structurally abnormal “X’s” they were preferentially late labeling as compared with the normal “X”. Of course you have to recall that back in 1960 we couldn’t tell the “X” chromosome in the karyotype.

Collins: Right.

Rowley: It was clear that there was one “C” group chromosome, as they were called then, that was late labeling.

Collins: Yes. What was that like Janet? You had been primarily doing clinical work.

Rowley: Absolutely.

Collins: For almost a decade, I guess.

Rowley: Three days a week.

Collins: And then suddenly you’re put into a very different environment. Not only in terms of doing research instead of clinical work, but being over in Europe. It must have been quite a transition. Was that exhilarating? Was it a little unsettling? What was that like?

Rowley: Oh no. It was very exhilarating. I really enjoyed it and the challenge of trying to figure out what chromosome was involved, particularly when there was this one late labeling chromosome in female cells. I used myself as the donor of the peripheral blood.

Collins: Oh, in the long tradition of self-experimentation.

Rowley: That’s right. So then I used somebody, I don’t remember whom, who was a male and didn’t have an obviously very late chromosome, using tritiated thymidine. It was through Marco Fracarro, who was a good friend of Jon Lindsten that we knew of all these patients with abnormal “X’s”. Jon arranged to get blood on all of these patients when I came over to Sweden. I added the tritiated thymidine and then went home with the slides.

I got so excited about what I was doing, that when I came back to Chicago after the equivalent of a year sabbatical, it was clear to me that I didn’t want to go back to the clinic. I then approached Dr. Leon Jacobson, who was head of a DOE-funded large institute here at the University of Chicago, about the possibility of continuing my research.

It’s very important and instructive, that I approached Dr. Jacobson, in the sense, very naively.

Collins: Yes.

Rowley: Because I had no credentials, none whatsoever. And I asked him: A. Would he give me a job and pay me to work three days a week? Secondly, would he give me lab space so I could keep on studying all these slides that I’d made in Europe of these abnormal “X” chromosomes.

Collins: Yes.

Rowley: And Dr. Jacobson did that.

Collins: Boy that’s remarkable. How likely would it be today?

Rowley: Not at all.

Rowley: I mean I had a nice paper in “Nature” that I’d written with Laszlo and other people on the abnormal “X’s” and the labeling pattern. But that was a single paper and obviously coming out of a supportive environment. What was I going to do on my own when nobody at the university was doing anything like that?

Collins: That’s amazing.

Rowley: But Dr. Jacobson, whom I had known as a professor when I was a medical student here, did have resources. I got five thousand a year salary. I had no technician, but I could use the microscope and do the work myself.

He actually supported me all the time that I was here through this DOE Institute. It was really almost ten years before I did anything that was noteworthy. Certainly it was six before much started coming out that would even catch anybody’s attention. But he was very supportive right through that period of time.

Collins: Sounds like he was one of your heroes then, in terms of giving you a chance.

Rowley: Absolutely. Absolutely.

Part 3: An Interest in Leukemia
As her research at both the University of Chicago and Oxford progressed, Dr. Rowley became interested in the study of genetics as it relates to leukemia. Here she describes some of her early work in that field.

Collins: So how did you get interested in the side of genetics as it relates to leukemia?

Rowley: Well, Dr. Jacobson was a hematologist.

Collins: Oh.

Rowley: Actually he was a member of the National Academy because he was one of the people who proposed that there was a substance, erythropoietin, that led to a lot of the work of Gene Goldwasser, who was a colleague of his. So, I was in the section of hematology and oncology. So, my laboratory was in hematology.

Collins: Yes.

Rowley: And every so often Dr. Jacobson would have a patient who would have CML, and he’d want to see whether the Philadelphia Chromosome was present. So, while I was continuing and extending studies on the pattern of labeling of some of the other chromosomes, I would do a cytogenetic analysis for him. Then, as my own research in labeling patterns of chromosomes came to an end, there were all these interesting patients. Particularly those with what was then called pre-leukemia.

Collins: Yes.

Rowley: So I started studying their chromosomes and I found that some of them had gains of chromosomes and some of them had losses of chromosomes. But again, in the sixties it was not possible to tell whether they were the same chromosome or different chromosomes, etc.

Collins: Right.

Rowley: And then my husband took a second sabbatical 1970-71, again in Oxford at the Dunn School.

Collins: Yes.

Rowley: I arranged to work with Walter Bodmer. It was just when Walter went back to England to become the Professor of Genetics at Oxford, in 1970. In fact, the laboratory was gutted and being renovated. So I did my research work up at an MRC Unit with Peter Pearson.

Collins: Oh.

Rowley: And that was just when banding was coming in and Peter had a fluorescence microscope, so I could go over at night and work on his fluorescence microscope. What I was studying then were some of the cells and cell lines that Walter and Marcus Nabholz were using for gene mapping.

Collins: Yes.

Rowley: I showed that there were major rearrangements in NIH 3T3 cells. I also showed the association of dense hetero-chromatic regions using the technique of Gall and Pardue with the dull staining regions on quinacrine fluorescence. I could do the equivalent of “C” banding and “Q” banding on the same cells.

Collins: Oh, that was what you needed the fluorescent scope for was the “Q” banding.

Rowley: That’s right. Because I had the naive notion again that I was going to map them, karyotype the cells, and track through all the rearrangements in the NIH 3T3 cells. Fortunately I discarded that idea right away. But I was using that technique also to identify the human chromosomes in the hybrids based on their fluorescent banding pattern.

Collins: Okay.

Rowley: I worked on that project, and I didn’t do any work on human leukemia in Oxford. I did analyze material from patients whom I studied who appeared to have only forty-five chromosomes. They were male patients, and it wasn’t clear whether they were missing a “Y” chromosome or not. In fact, using Peter’s scope, I could show that they were all missing a “Y”.

Collins: Okay.

Rowley: So, a paper came out in “The British Journal of Hematology,” I guess in 1971, just showing that the chromosome that was missing in these older males was the “Y”. This was one of the first studies of leukemia cells using bonding.

Collins: Okay.

Part 4: An Exciting Discovery
Here, Dr. Rowley explains the process of confirming the first two translocations she discovered, and how these discoveries led to more complicated research.

Rowley: I came back to the university, and again, Dr. Jacobson helped find the resources for me to buy a fluorescence microscope.

Collins: Which were not widely available to a lot of people at that point.

Rowley: That’s right. So I got my fluorescence microscope in 1972.

Collins: Yes.

Rowley: And then I started looking at material that we had on various of our leukemia patients, particularly because I was interested in this question of the gain and loss of chromosomes and trying to define whether they were the same or different chromosomes.

Collins: Yes.

Rowley: I also looked at patients with some chromosome rearrangements. In the first group I looked at, were patients who I showed had an 8;21 translocation. That was the first translocation I discovered–probably June/July of 1972.

Collins: I didn’t realize that that was the first one you were sure of and…

Rowley: That’s right…

Collins: That depended on the “Q” banding to be sure you had the right partners?

Rowley: Absolutely. Because previously in the literature based on standard banding, they were called minus “C,” minus “G,” plus “D,” plus “E.”

Collins: Okay.

Rowley: That was because a piece of chromosome 8 was moved to 21.

Collins: Yes.

Rowley: So 21 looked like a “D,” and the 8 looked like a 16.

Collins: I got it. Was it realized up until then that this was actually a balanced rearrangement? Or was the perception that was much more complicated?

Rowley: It wasn’t clear.

Collins: Okay.

Rowley: Actually, I think Eric Engel proposed that maybe it was a translocation before.

Collins: But nobody was sure.

Rowley: It was not clear. So I first looked at one patient with “Q” banding, and then I had a second patient with this abnormality. I sent a letter to the “New England Journal of Medicine,” and they rejected it.

Collins: That wasn’t interesting?

Rowley: I sent it to “The Annales de Genetique”; Jean DeGrouchy was a good friend of mine, and he was editor. It’s apparently the most cited paper in the journal.

Collins: That’s fascinating. Now, was it well received? Or did people not believe it? Or?

Rowley: Well, I think it was just sort of “so what.” And that’s why everybody focuses on the Philadelphia Chromosome. And again, what I was looking for in patients who were in blast crisis, because they had very complicated karyotypes, was to identify what appeared to be extra “C” group chromosomes.

So, I was paying attention to trying to identify whether the “C” group chromosomes were the same and looking carefully at all the chromosomes. I noticed that one chromosome, 9, was too long and had this long piece of pale material at the end of it.

Collins: Again the “Q” banding told you that was a 9.

Rowley: That’s right. None of this could have happened without banding.

Collins: Yes.

Rowley: Banding was absolutely critical. I had cytogenetic material in the chronic phase from some of these patients when they were only Philadelphia positive, and again, the 9 had the long piece of material on it. So I decided that in fact the piece from 22 wasn’t missing but that this was another example of the translocation. I sent a paper to “Nature,” and after the usual sort of delay, it got published in “Nature” in 1973.

Collins: So what did that sort of experience feel like? People are always interested in knowing, what was the moment like? Either for the 8;21 or the 9;22. When you were sure that you had understood something that people hadn’t been clear about before. That this really was a balanced translocation. You knew what the partners were. Is that something that sort of came as a flash one afternoon? Or was it over a course of time, building up the evidence and convincing yourself?

Rowley: Once I saw it in the chronic phase, and I saw it in three or four patients, I went back and got peripheral blood on these patients and could show they had a normal karyotype. Thus this wasn’t some rare congenital translocation.

Collins: Right.

Rowley: I have to say I was very excited, but I was very perplexed. Because clearly translocations had been found before, and they were obviously known as a cause of Down Syndrome.

Collins: Sure.

Rowley: If you have a 14;21 translocation in a parent, then you have a risk of Down Syndrome. But I kept trying to figure out whether there was any precedence for this, and so I talked with a number of colleagues at the University of Chicago as well as Barbara McClintock, about what could break and rearrange two chromosomes so precisely.

Collins: Yes.

Rowley: I never got a satisfactory answer. In fact you can say there is no satisfactory answer now.

Collins: You could say that.

Rowley: That is because we don’t know the mechanisms for translocations, but at least of course now we know the….

Collins: Consequences.

Rowley: That’s right and the genes that are involved. Not too long after that I discovered the 15;17 translocation in acute promyelocytic leukemia. I also showed that the gains and the losses of chromosomes in other patients were nonrandom and often involved the same chromosomes. I became a believer that chromosome abnormalities in leukemia cells were central to the development of the malignant process.

Collins: Did you have a hypothesis in your mind at that point about how that might work? About how such recurrent translocations might play a positive role in the leukemic process?

Rowley: Well in the sense of a kind of sophisticated understanding we have now, the answer has to be no. All I knew was that these were critically important because every patient with APL has a 15;17 translocation.

Collins: Yes.

Rowley: And therefore it had to be central. Thinking of two genes being broken and fused, I have to say, that certainly didn’t occur to me. But it was clear that there were some critical genetic events that were the same in the same translocation. That I certainly believed in. So I’d go to hematology meetings. They had education sessions on Sunday morning, and I would just preach to these people that chromosomes were important and you as hematologists have to pay attention to them. Not too long after, in 1978, we found the 14;18 translocation in follicular lymphoma.

Collins: Yes.

Rowley: So recurring translocations were occurring in lymphomas as well as in leukemias, thus they were a central part of malignant transformation.

Collins: I guess at the time it would have been very difficult to imagine the precision of these rearrangements would turn out to be what it is. That these breaks, which under the microscope appear to be similar, would even when you got to the molecular level, be so closely spaced together in terms of the exact location of the break points–often times in the same intron of two different genes that would then have to be stitched together in a very precise way. I think…it would have been difficult for anybody to imagine that was going to be the outcome, and it sounds like that was not on a lot of people’s minds.

Rowley: I don’t think it was. The first translocation that was cloned was Burkitt lymphoma, and, in fact, the breaks are quite variable. They occur within the immunoglobulin gene, but as far as MYC is concerned, they can occur five prime, three prime of the gene.

Collins: Right.

Rowley: The translocation does not lead to a fusion protein. Rather, it leads to the abnormal expression of a normal MYC protein.

Collins: Right.

Rowley: The same is true for BCL 2.

Collins: Right.

Rowley: If you had asked scientists, even very very sophisticated, thoughtful people, back in the late 1970’s, I don’t think anybody would have imagined that there would be fusion genes. That would have not have been on anybody’s mind.

Part 5: Molecular Cytogenetics: Its Present and Future
Dr. Rowley began a laboratory dedicated to research into molecular genetics. Here, she explains the challenges associated with such a laboratory, and gives her thoughts on the future of molecular cytogenetic research.

Collins: Yes, yes. Who would have thought it? Well then you went on from there to actually identify the molecular level. The partners. And a significant number of these rearrangements that you had first described. So this must be a fairly satisfying span of experimental effort going from the initial observation to now saying at the molecular level exactly what the deal is. Reflect on that for a minute here, what it’s like to have that kind of perspective?

Rowley: I’m overjoyed at how all of this has worked out. I’m especially pleased that what we’ve managed to do is to co-opt outstanding scientists, including yourself through Paul Liu, into being interested in chromosome translocations and in cloning them and trying to figure out what the genes do. For such a long time, cytogenetics was considered by a number of eminent people as just so much stamp collecting. It is very rewarding. I did know from the late 1970’s that trying to identify what was going on was critical. And back then, of course, the techniques were pretty primitive. My own notion was to work with Tony Carrano to use chromosome separation in CML so that you could isolate the Philadelphia Chromosome and probably some chromosomal fragments. Maybe with techniques that were available then, you might be able to identify the Philadelphia Chromosome and the genes on 9 and 22. Obviously, with new techniques there were better ways to approach this.

Collins: Yes.

Rowley: It took me three years to be able to get the resources, and the space, and a person to come and work in the laboratory to do molecular genetics.

Collins: Yes. That’s quite a transition.

Rowley: At this point I felt that I was not in a position to become a molecular geneticist. After all I was more than fifty years old or some such. So it was a matter of recruiting somebody who could come and develop this program. Then Manuel Diaz came and started the molecular genetics laboratory. During the next fifteen years we’ve been able to recruit many different people, who’ve come and worked on cloning a number of these translocation break points.

Collins: And having visited your lab, I can certainly vouch for the fact that you were a lot more than just an overseer of the molecular efforts. You’ve got yourself very much involved. Don’t sell yourself too short.

Rowley: Well that’s true, but I had very dedicated teachers. Very patient teachers as well. But I was not satisfied with just sitting back and finding the translocations and letting somebody else have the fun.

Collins: Yes. Yes. Well that’s very clearly not the mode that you took. Where do you think this is all needing to go next in the future, as we sort of stand here at almost the turn of the century looking at the remarkable strides that have happened in molecular cytogenetics–a field which didn’t exist thirty or forty years ago. Through the efforts largely of yourself and others…a small number of people has become this very exciting, rapidly moving field. Where’s it going next?

Rowley: I think there are at least two major unanswered questions. One is what causes chromosome translocations? We and others are trying to study that. We’re using the fact that unfortunately some patients with cancer, anywhere from one to fifteen percent, who are given high doses of drugs that inhibit the function of topoisomerese II, will get secondary acute leukemia.

Collins: Yes.

Rowley: And in these patients a gene that we identified, as well as others, the MLL gene at chromosome 11 band q23, is very frequently involved in the leukemia.

That gives us a clue to look for things like topo II sensitive sites within this gene to see whether that might play a role in the translocations. In fact, we and others have evidence that in fact there is an in vivo topo II cleavage site, which we’ve mapped as being the general location of the breaks. We don’t have it down to the nucleotide level, but there is also a DNAse I hypersensitive site.

Collins: Yes.

Rowley: The other interesting fact is that if you look at infants who also have MLL translocations, they have breaks near to this topo II cleavage site.

Collins: Yes.

Rowley: The breaks are often up to several kilobases away from the topo II cleavage site, so I don’t want to make it sound too specific. But there is some tantalizing evidence that maybe we can begin to figure out what causes translocations.

The second major goal, and this of course is dear to me as a physician, is trying to figure out how we can develop genotype-specific therapy.

Collins: Yes.

Rowley: You can say that it was developed for acute promyelocytic leukemia, in a sense by accident, because it was the Chinese who were experimenting with various drugs, who found that all transretinoic acid, specifically induced remissions in patients with APL. When it was discovered that retinoic acid receptor alpha was one of the genes involved in the translocation, that made some sense.

We need the same kind of genotype-specific therapy for all the other translocations.

Collins: Yes.

Rowley: Again, naively, I once thought that antisense therapy would be effective, or ribozymes. But it’s clearly much more complicated than that.

Collins: What do you thing the chances are that we will make real strides in that area, the genotype specific therapies in the next decade? Are we perched on the brink of that becoming a reality? Or is this going to be a long hard slog?

Rowley: Well, I think ten years is probably too optimistic.

Collins: Yes.

Rowley: Again, scientists are working on an area about which I don’t know a whole lot, but I do know that you can intercalate a third strand of DNA in the double helix and that this can somehow perturb the replication and the function of the target gene.

Given how quickly one can actually do genome sequencing, it would not be a problem to get the exact DNA sequence at the site of a translocation break point in an individual patient. If you could really get a third strand of DNA somehow specifically intercalated at that translocation junction and have it bind to the DNA, so first it can’t be expressed but second, the cells can’t replicate, that could be the type of genotype specific therapy that might be effective.

Collins: Yes.

Rowley: The other approach we thought of was the fact that the BCR ABL fusion protein is a unique protein in these cells.

Collins: Yes.

Rowley: Could you target it with antibodies? But I think that many of these fusion proteins are not expressed on the surface of the cells. So then you’re going to have to figure out, how can you target a fusion protein that may be intercellular?

Collins: Right. Or maybe internuclear as many of them seem to be.

Rowley: That’s right. So I think those are things to work hard on in the future.

Collins: But they’re not right around the corner.

Rowley: That’s right.

Part 6: Taking the Long View in Life
Dr. Rowley talks about the importance of balance between work and family. Working only part time when her children were young allowed a rich family life; patience and persistence made up the difference in the lab.

Collins: Well that’s a very interesting tale you’re telling of your own career over these decades of discovery. Who would you say, if you had to pick out, have been your own heroes? Sort of scientific examples of people that you’ve admired and then strove to sort of chart your own course in a similar way?

Rowley: Well, I had the advantage both as a medical student and then coming back on the faculty at the University, of knowing Dr. Charles Huggins fairly well.

Collins: Yes.

Rowley: He was somebody who really continued to do research up into his nineties. So he was certainly a very important role model. I have to also give my husband credit. He’s an experimental immunologist still at age seventy-five, working in the laboratory on projects and injecting mice and the rest of it. His single-minded focus on research issues and questions and how to approach them best has been a very good model for me as well.

Collins: I guess I should also ask you, Janet, if you would have things to say to young scientists, maybe particularly to young female scientists, about the challenges of trying to combine many aspects of a productive life in terms of research and some clinical medicine that you have done quite a bit of, and family, which obviously was very important to you. I know you have had a remarkable family that’s arisen from you and your husband’s dedication to that part of your lives too. Is that something you’ve found relatively easy to balance–all these responsibilities? Or did you have to make sacrifices along the way? Do you have any thoughts about that?

Rowley: That’s a very important question, especially for young people. I have two things to say on that. Firstly, as should have come out of my story, what has happened to me is totally unexpected. This is not something I strove for, was ever a goal, was ever anything I even conceived of almost anytime in my life. Not until the last fifteen years or so, when it was clear that what I had done was important; then my view of it changed. But this was never anything that I really sought.

What I think is important is that young people take a very long view of their life. Which implies that you’re going to have good health and I’m fortunate that that’s the case with me. Don’t be too impatient for things to all happen quickly. Or to think that by the time you’re thirty-five and you haven’t done very much, that you’re over the hill.

Because, again, translocations were discovered in 1972, and I was born 1925. So I was forty-seven years old before I did anything that people would really look at twice. So patience is certainly an important aspect of this. Then, I have to say that I was in an environment where people have been very supportive, very collegial, very patient, because sometimes I wasn’t as productive as I would have liked to be.

Collins: So basically, I think you’re suggesting that, yes there were a lot of pushes and pulls on your time, but you were patient enough to sort of let them find their appropriate place in your life. The role of serendipity comes across very clearly.

Rowley: Right and good luck. I have led, by and large, an extraordinarily lucky life. You just go from one thing to another. From retarded children, to chromosomes, to being in a hematology group, to banding. None of which you can predict ahead of time.

Collins: Yes. But of course others may have had that same luck and didn’t necessarily take the same advantage of it that you did.

Rowley: No.

Collins: You recognized an opportunity when it came along. And you stuck to it.

Rowley: Yes.

Collins: It’s clear. As you talked about those ten years, where you’re working away in this little lab in Chicago, pretty much by yourself, you didn’t give up. Did you ever feel discouraged? Did you ever sort of think, “Maybe I’m not really cut out to do research. Maybe I ought to just go back and be a clinical doc and give up all of this stuff?”

Rowley: No, I didn’t. But you see I think it goes back to what I said. I looked on medicine and research as a hobby. At the time that I started at the university, or a year after, I had four children. I had to be taking care of them. I only worked three days a week. So, I had a very rich, full life with my children and my husband.

The lab was a hobby. The fact that it was going slowly, well you know that didn’t bother me in the least. I never expected it to go anywhere anyhow. I shouldn’t be saying these things.

Collins: No, it’s wonderful you’re saying these things.

Rowley: Not on a web site.

Collins: Because it will be very reassuring to people who imagine that the only way you actually succeed in the competitive world of research, is becoming so single minded that you screen out every other aspect of your life. That would be a terrible message for people to receive, and you’re countering it very effectively.

Rowley: Okay.

Collins: Well Janet this has been a great pleasure for me to have the chance to lead you through this description. You’ve told me a bunch of things that I didn’t know before, about your past life and all the things you’ve done.

Again my heartiest congratulations to you for this accomplishment. You deserve it. We all celebrate it. We’re all smiling a lot after having heard this news.

Rowley: Well we can celebrate it at New York.

Collins: We will indeed.

Key Publications of Janet Rowley

Rowley, J.D. (1973) A new consistent chromosomal abnormality in chronic myelogenous leukemia. Nature 243: 290-293.

Rowley, J.D. (1975) Nonrandom chromosomal abnormalities in hematologic disorders of man. Proc. Natl acad Sci USA 72: 152-156.

Rowley, J.D., Golomb, H.M., and Vardiman, J.W. (1977) Nonrandom chromosomal abnormalities in acute nonlymphocytic leukemia in patients treated for Hodgkin’s disease and non-Hodgkin lymphomas. Blood 50: 759-770.

Thirman, M.J., Gill, H.J., Burnett, R.C., Mbankollo, D., McCabe, N.R., Kobayashi, H., Ziemin-van der Poel, S., Kaneko, Y., Morgan, R., Sandberg, A.A., Chaganti, R.S.K., Larson, R.A., LeBeau, M.M., Diaz, M.O., Rowley, J.D. (1993) Rearrangement of the MLL gene in acute lymphoblastic and acute myeloid leukemias with 11q23 chromosomal translocations. New Engl J. Med. 329: 909-914.

Rowley, J.D., Reshmi, S., Sobulo, O., Musvee, T., Anastasi, J., Raimondi, S., Schneider, N.R., Barredo, J.C., Cantu, E.S. Schlegelberger, B., Behm, F., Doggett, N.A., Borrow, J., Zeleznik-Le, N. (1997) All patients with the t(11;16)(q23;p13.3) that involves MLL and CBP have treatment-related hematologic disorders. Blood 90: 535-541.