The Baltimore Case
A Trial of Politics, Science, and Character

Read BW's Review of This Book

CHAPTER ONE

"A Beautiful Paper"

THE BALTIMORE case originated with Margot O'Toole, a postdoctoral fellow then in her early thirties whom Thereza Imanishi-Kari had hired to work in her laboratory at the Massachusetts Institute of Technology (M.I.T.) in the summer of 1985 and who eventually blew the whistle on her boss. O'Toole was asked to do experiments that would extend the work described in the contested Cell paper, and her unhappiness at not being able to get the results she sought led to the first complaint about Imanishi-Kari's data. O'Toole's dogged insistence that she was right and her supervisor was wrong lay at the heart of the affair. She became a symbol of the heroic young scientist who takes a stand against the system and prevails over powerful figures like David Baltimore.

    Margot O'Toole now does research at the Genetics Institute, a biotechnology company in Cambridge, Massachusetts, but her reputation as a scientist rests almost entirely on her conflict with Baltimore and Imanishi-Kari. She has received several awards emanating from her actions, among them the Humanist of the Year Award from the Ethical Society of Boston, and the Ethics Award of the American Institute of Chemists. O'Toole has an open Irish face and a manner that prompted a congressional investigator to say, "The first time you meet her she just reeks with integrity and credibility." I first met her one day in Cambridge in 1992, when I picked her up for lunch. She is a compelling storyteller and she held me in thrall for hours with her tales of the Baltimore case.

    Margot O'Toole, her mother once remarked, was virtually bred to confront trouble. She was raised in a strong Catholic household in Dublin, Ireland, and spent two years in a convent school. Family lore told of one grandfather, a miller, who lost his house to flames during the famine for surreptitiously diverting food from the landlords to the people; and of the other who was turned out on the street for involvement in the Land League Movement. O'Toole says that her mother relished battles, as did her father. He was an engineer for the Electricity Supply Board, and also a radio commentator and playwright. He wrote of "speaking out in the workplace, not going along," O'Toole says. One of his plays, Man Alive, which satirizes the bureaucratic complacency of a giant utility company, uncannily foreshadows key elements in the Baltimore case. O'Toole says that the Electricity Supply Board tried to block the production when it was in dress rehearsal and that, though she was merely a child when the play was produced, the fight about it was a staple of her growing up. The central character is an outspoken engineer named Tim O'Malley, who is told to keep his dissident thoughts to himself and is declared an incompetent troublemaker. He nonetheless refuses to quit, pledging at the end of the play, "As long as I stay I'll be a thorn in their backside, and every time they sit on anyone again they'll think of me."

    In 1966, when Margot was 14, the family moved to Boston, where her father eventually directed a program in science writing at Boston University. She was multiply talented, adept at swimming and French, and shared her parents' attachment to poetry. O'Toole graduated with honors in biology from Brandeis University, in 1973: Civil rights protests and demonstrations against the Vietnam War had flourished during her undergraduate years, likely encouraging her familial propensity for dissent. O'Toole spent the next academic year at Harvard, working as an assistant in the laboratory of a young immunologist named Thomas G. Wegmann. He later said that he found her very naive and her qualities of mind characteristic more of her religious upbringing than of her scientific training. He also considered her very good and productive at her technical work. Well recommended by Wegmann, she went for graduate work to Tufts University Medical School, entering in 1974 and specializing in immunology under the auspices of Henry Wortis.

    Wortis, then approaching forty and recently promoted to associate professor, was a respected immunologist who was also known among biologists of his generation for his political activism. A tall rangy man with an easygoing manner and a broad, rugged face, he wears jeans, a plaid work-shirt, and athletic shoes in the lab. He describes himself as a "red diaper baby." He is the son of politically left psychiatrists who middle-named him Havelock, after their friend, the famed sexologist Havelock Ellis; his parents spent time in the Soviet Union between the wars and were eventually called before the House Un-American Activities Committee. Wortis says that during the 1950s he was thrown out of the University of Wisconsin for his political associations, notably his chairmanship of the Marxist Labor Youth League. He militantly protested the draft and the war Vietnam during the mid-1960s, when he was a postdoctoral fellow genetics at Stanford University. O'Toole was his first doctoral student her husband, Peter Brodeur, whom she married in 1978, was his third and he admired her social conscience. He also found her bright, insightful creative, and engaging. Even now, he keeps a photograph of her tack to his bulletin board, a snapshot of an attractive young woman, her hair wet from the rain and her eyes alight with mischief.

    In 1979, O'Toole received her Ph.D., and early in 1980, she and Brodeur, both funded by fellowships from the National Institutes of Health (N.I.H.), took up postdoctoral appointments at what is now called the Fox Chase Cancer Research Center in Philadelphia. Postdoctoral fellow ships, which had long been desirable apprenticeships in the biomedical sciences, were especially choice in the 1980s. American universities were then more than doubling their annual output of biomedical doctorates The expansion both fueled the rapidly burgeoning biotechnology industry and exacerbated competition for grants and positions in academic biomedicine. Life was tense low down on the professional ladder, where fledgling scientists ambitious to stay in the game had to prove that they could fly on their own in the high-pressure atmosphere of creative research Postdoctoral fellows normally pursue the general line of investigation under way in the supervisor's laboratory. But labs can be launching pads, too, providing the freedom and facilities to open a line of independent research--one promising enough to win grant money and perhaps a university post.

    O'Toole struck scientists at Fox Chase as well versed in her field, and Donald Mosier, in whose lab she came to work, thought that she had a good conceptual grasp of scientific problems. Mosier says, however, that she made "no progress" at the bench, largely because she wanted "to a research manager without having acquired sufficient bench skills." expected O'Toole, as he did all his postdocs, to spend much of her year learning the experimental techniques necessary for the project was supposed to pursue. Mosier says that she had trouble accomplish the relatively simple task of purifying an organic chemical needed for research, even though she had help from others in the lab. He continues that she wanted to rely on technicians to perform the requisite tests procedures before she had mastered them herself and that she devised "grand, complicated experiments," tried them periodically, and failed to make them work. Mosier remembers urging that she consider a career in teaching or science writing because she commanded the concepts of science, if not manipulations at the bench.

    O'Toole, on her part, was increasingly unhappy with the situation. She told another young scientist at Fox Chase that she was uncomfortable in Mosier's laboratory because his wife, a postdoctoral fellow whom he had recently married and who remained working in the lab, received unfairly favorable treatment. Mosier calls the charge "absolutely false," explaining that he leaned over backward not to favor his wife and that, in any case, her project was remote from O'Toole's. He says that finally, after many months without significant experimental accomplishment, he asked O'Toole to leave his laboratory. In May 1982, she moved to the Fox Chase lab of the husband-and-wife team of Melvin and Gayle Bosma. By then, with Mosier's help, she had obtained new fellowship support from the National Arthritis Foundation and she settled down to a productive line of work.

    Mel Bosma found her lively and captivating, fun to be with, though he recalls that she seemed at least as absorbed with the play of personality in the laboratory as with the work at the bench. (Wortis had the impression that O'Toole was "more sensitive to the social--what I would call scientific-political--interactions around her to the point where her concern about those issues might keep her from focusing on her work.") Mosier declares that she "had an instinct for polarizing laboratory members over minor issues." By way of example, he points to her pitting his technicians against each other because one of them, who he says was better qualified than the rest, was given more authority in the lab. He says that in some respects O'Toole reminded him of "a labor organizer": She was "intense, political," ready to leap on an issue."

    People then at Fox Chase still talk about how O'Toole grappled with the issue of day care. Fox Chase was renovating several rooms in an old nearby school for the care of staff infants. Completion of the renovation was scheduled for June of 1981, but it was delayed until September. During that summer, O'Toole gave birth to a son. Other mothers made temporary alternative arrangements for their children. O'Toole brought the baby to her laboratory, in defiance of the institution's rules. The rules were intended to protect children against exposure to radioactive substances and the institution against liability. O'Toole, told to stop keeping the baby in the lab, insisted to Patricia Harsche, the administrator in charge of the renovation, that Fox Chase owed her assistance, declaring in the hall one day, "If it hadn't been for you Pat, I wouldn't be pregnant." Harsche laughed, noting, "Margot, I've been accused of many things in my life but never of having made another woman pregnant." O'Toole, Harsche remembers, did not laugh in return. Harsche, who sympathized with O'Toole as she did with other mothers with day-care problems, quickly arranged to have a small room fitted out as a nursery where O'Toole and her husband cared for the baby until their son entered the day-care facility in October.

    A year or so later, Fox Chase enlarged its child-care accommodations by renovating another old school. After this facility was occupied, the heating system gave evidence of needing replacement and inspection revealed that it was lined with asbestos. Fox Chase had the asbestos removed with the special care that the law required and engaged an independent consulting firm to evaluate whether the job had rendered the building free of asbestos. The firm said that it had, but O'Toole joined several other mothers in declaring that the consultant's report was untrustworthy because Fox Chase had paid for it. The mothers insisted that the cancer research center obtain a second evaluation by a firm acceptable to them, which it did, and the new assessment reached the same conclusion as the first.

    Late in 1984, Brodeur was offered an assistant professorship at Tufts University. Not wanting to split the family by staying in Philadelphia, O'Toole followed her husband back to Boston. The N.I.H., to which O'Toole had applied for support, told her that she might obtain funding for the project she had started with Bosma if she could devise certain experimental materials that would bolster its promise. Her thesis adviser, Henry Wortis, who says that he thought of her as a "bright, insightful scientist," helped O'Toole get temporary space in the lab of his colleague, Sidney Leskowitz. She was appointed to an assistant research professorship, a position that the department chairman could create at will for people who had their own research funds. She could work on her project until her postdoctoral money from the Arthritis Foundation ran out and continue with it if she managed to get new grant support. "That seemed great for me," she says, "because I wanted to continue it and I wanted to be in Boston. It was risky, because I had no backup. If I didn't get funded, I couldn't do it. So, I took the plunge and applied for the grants and went to Boston. But I didn't get them."

    Some time that spring, Wortis invited O'Toole to a party at his home, and there she met Thereza Imanishi-Kari. Imanishi-Kari was then forty-one, almost nine years older than O'Toole, but still a junior faculty member at M.I.T. O'Toole had certain skills that would be useful in a project that Imanishi-Kari wanted to pursue, an outgrowth of her collaboration with David Baltimore. Wortis had alerted Imanishi-Kari to O'Toole's need for grant support. On the spot, she offered O'Toole a one-year postdoctoral training fellowship supported by the N.I.H. O'Toole was to have time and facilities to strengthen her own project's eligibility for funding while she collaborated with Imanishi-Kari on extending the research that she was doing with Baltimore. O'Toole's research money was about to run out, and the offer was a godsend.

Thereza Imanishi-Kari is now a member of the pathology department at the Tufts University Medical School in Boston, where she went after leaving M.I.T. Her laboratory is a bright open room with several working credenzas laden with glassware, chemicals, and cultures, on the top floor of an old brick building with an elevator that might momentarily strand passengers between floors. Imanishi-Kari was born into an immigrant Japanese family in Brazil. She is a kind of cultural hybrid, giving the appearance of Japanese reserve but regularly shattering it with Latin expressiveness. ("I say things and face up to person," she told me.) She speaks seven languages, but her English, which is even now sometimes difficult to understand, was especially poor in the mid-1980s, when the disputed paper was published.

    Thereza Imanishi-Kari grew up in Indaiatuba, a small town near Sao Paulo, where her parents were tenant farmers, growing cotton, vegetables, and coffee. Eventually they got a mule, started transporting the neighborhood's produce to market, and soon became the owners of a small trucking business. They wanted their five children to attend school and do well, but they expected their three daughters to devote their lives to marriage and family. Thereza and her sisters fought to get an education; after her older sister left home over the battle, her parents permitted Thereza to go to high school and then a university in Sao Paolo. Her grandfather wanted her to learn about her family's culture, and in 1968 she went to Kyoto University to do graduate work in biology. Hardly any women were studying science there, so she hung out with the men, perfecting the Japanese she learned as a child into the male rather than the female version of the language.

    Kyoto University was in a state of upheaval in 1968, like most universities at the time, with fights constantly breaking out on the campus. Imanishi-Kari repaired to cafes with other students, where they studied immunology and talked about the imitative tendencies of Japanese scientists, particularly their reliance on experimental systems developed in other countries. The students considered the dependency self-defeating. Japanese students would work late hours applying a borrowed research system to a scientific problem only to find themselves preempted by the foreigners who had devised the system originally. She resolved to invent and rely on her own research system, Which is what she did while she was completing her graduate work at the University of Helsinki, in Finland. She had gone there in 1971 because she felt that the continuing disruptions at Kyoto made it impossible for her to do serious work. Using a chemical called NP ("nip," she pronounces it) that was well known to immunologists, she hit upon a method of tracking the behavior of certain immune genes in mice, the common laboratory surrogate for human beings.

    In 1974, she obtained her doctorate and married a Finnish architect, Markku Tapani Kari. She spent several postdoctoral years in the laboratory of Klaus Rajewsky in Cologne, Germany, and became known for her work on the NP system. When M.I.T. was looking for a cellular immunologist to add to its faculty in 1979, she was encouraged to apply for the position by an M.I.T. biologist named Malcolm Gefter. Imanishi-Kari was the biology department's first choice in an international search that produced about thirty candidates. When she got the offer from M.I.T., she asked Rajewsky for advice. He counseled against acceptance, observing, "M.I.T. is a very competitive place. It's like a sea full of sharks and they eat the little ones very fast." She went anyway. "That was the beginning of my nightmare," she says.

    Imanishi-Kari arrived at M.I.T. in March 1981 as an assistant professor and moved into a laboratory on the first floor at the M.I.T. Center for Cancer Research on Ames Street. She had brought with her from Germany a freezerful of valuable NP research materials. She was vivacious, competent, quick on her feet, and formidably smart. Even O'Toole says that initially she found her "very unusual and quite charming." Imanishi-Kari's group comprised several students and junior scientists on temporary billets and a technician named Chris Albanese, who had recently obtained a master's degree in biology from the State University of New York. Some--though not all--of the laboratory regulars, like Moema Reis, a visiting scientist from the Instituto Biologico in Sao Paolo, Brazil, were devoted to her. Reis, who was in her late thirties and had first gotten to know Imanishi-Kari in 1983, when she spent three months in her laboratory. She returned for a year beginning in February 1985 and became a contributor to the Cell paper. She says that the laboratory was "extremely pleasant," that Imanishi-Kari was welcoming and open, regularly sharing data and discussing experiments, and reviewing individual projects at lunch on Fridays. Reis considered her year in the laboratory "very helpful ... because I had the opportunity to work closely with somebody who considers science a vital activity," somebody of "high intelligence who has a very hard drive for working," somebody who "carries out science and research as it should be carried out."

    Imanishi-Kari broke the laboratory rules against smoking and neglected to meet M.I.T.'s requirements for getting ahead. By M.I.T.'s standards she published too few scientific papers, and part of what she published struck others as being a narrow extension of her earlier work with the NP system. However, she was enlarging her repertoire of expertise by learning to use the techniques of molecular biology. And in 1984, during her collaboration with David Baltimore on a study of the production of antibodies in a special breed of mice, her tracking system helped expose some surprising and peculiar results. The findings made no difference for her future at M.I.T. She had already been looking around for another job when in July 1985 she was told that she would not be put up for tenure. She had an offer at a biotechnology research institute in La Jolla, California, another at Mt. Sinai Medical Center in New York City, and a strong prospect at Tufts in the department where O'Toole's husband now worked and where she had several friends, including Henry Wortis, with whom she was collaborating on a research project. Her daughter wanted to remain in Boston, so she preferred Tufts. In mid-August 1985, the head of the Tufts Department of Pathology talked with Imanishi-Kari about a position, and she said that she would accept an offer when it became firm and final.

In May 1986, when Tufts was moving to make the offer final, Baltimore declared in a letter of recommendation that, although Imanishi-Kari has been "slow to get an innovative research program going," in the last few years her research had begun "to take an interesting shape." He explained referring to the work that they had just reported in Cell, that "it was the expertise in her laboratory that allowed us to understand" the odd immune response in the mice they had used in the experiment.

THE EXPERIMENT

Background: The Immune System and Rearrangement

The immune system had long interested David Baltimore and he had turned to studying it in the mid-1970s, after he won the Nobel Prize for his work in virology. In mammals, the system's principal component comprises white blood cells called lymphocytes. In 1974, at a symposium in Paris, the distinguished Danish immunologist Niels K. Jerne remarked that just twenty years earlier scientists "hardly even suspected" that "lymphocytes had anything to do with the immune system," adding that "now we know they are the immune system, or at least 98% of it." Billions of white blood cells--in fact, roughly a thousand billion of them--are present in the body, enough, taken together, to add up to a mass comparable to that of the liver or brain. They are found in the body's specialized immune organs--the thymus, bone marrow, spleen, and lymph nodes--and they circulate in the bloodstream.

    Two types of lymphocytes are central to the immune response in mammals: T cells, which are manufactured in the thymus, and B cells, which are generated in the bone marrow. Both eventually migrate to the lymph nodes and the spleen, reacting there with invading agents such as a virus or a bacterium. T cells perform several functions, one of which is to assist B cells in doing their job. B cells produce antibodies, which latch to and inactivate what the body takes to be hostile invaders. Antibodies are exquisitely specific, fitting to the invader the way a key fits to a lock.

    The variety of infectious agents that invade the body is enormous, and the most striking feature of the immune system is that it generates a comparably enormous range of antibodies. Antibodies are constructed, like an erector set, from discretely identifiable elements, but, when they are completed, most antibodies structurally resemble the letter Y (Figure 1). Each tip of the Y comprises two independent variable regions that run up opposite sides of each of the arms. Together, they provide the antibody's specificity--its ability to fit the invading agent's lock.

    Each variable region is the product of an independent set of genes which are formed from among many thousands of segments of DNA in the nucleus of the B cell. In the process of making antibodies, a small number of segments from different elements in the nucleus rearrange themselves and combine to produce genes for the two variable regions (Figure 2). The possible combinations of segments is huge, which is to say that the process generates a giant number of possibilities for each variable region. (If, say, one variable region derives from any one of a thousand genes and the other from any one of another thousand, then the number of combinations would be the product of the two, or one million from just two thousand genes.) It is this heterogeneity that allows the immune system to manufacture its mightily resourceful arsenal of antibodies. Antibody production occurs continuously throughout an animal's life, producing numerous incipient B cells. Each is committed to generating only one particular antibody, and all the B cells together account for the diversity of the animal's immune response.

    When Baltimore began immunological research, one of the great puzzles in the field was how exactly the process of genetic rearrangement is stimulated and controlled. He took up that conundrum, among others, as he increasingly oriented his laboratory to apply the powerful and rapidly developing techniques of molecular biology to problems in immunology. Perhaps the most powerful of these techniques, an invention of the mid-1970s, was recombinant DNA, which permitted scientists to snip out a gene from one organism and insert it into another--to put a human gene into a mouse, for example. Recombinant DNA provided scientists a powerful tool for studying the nature and mechanisms of gene action, and Baltimore thought he could exploit it to study the process of genetic rearrangement.

Imanishi-Kari's Method: Idiotype Tracking

Imanishi-Kari's NP tracking system provided another way of getting at the genetics of antibody formation. NP is one of a class of small organic chemicals that, when combined with a protein, will stimulate the generation of an antibody against itself. In certain strains of mice, the antibodies display a distinctive chemical feature called an idiotype. Although located in the variable regions at the tips of the antibody's arms (see Figure 1), idiotypes have nothing to do with giving antibodies the specificity of their response to a foreign agent like NP. It is more like a birthmark--an identifying signature.

    Imanishi-Kari obtained antibodies from fast-growing cell cultures called hybridomas. She formed them by joining cells from the different immune organs of her mice with rapidly multiplying myeloma cancer cells. Hybridomas are nourished in a bath of nutrients, and the antibodies they generate make their way into the fluid--a supernatant. Imanishi-Kari relied on serological methods--that is, methods used to identify the properties or contents of organic fluids such as blood sera--to characterize the antibodies, testing them with chemical or biological substances. Such substances are called reagents when they are deployed as tools in experiments, and Imanishi-Kari managed to devise reagents indicating the presence of idiotypes on antibodies to NP.

    While in Europe during the 1970s, she demonstrated that the ability of certain mice to mark such antibodies with an idiotypic birthmark is passed down from one generation to the next in accord with Mendel's laws of inheritance. Such inheritability meant that the idiotype is a product of a particular segment of DNA. Thus, the antibody's birthmark provided a serologically detectable feature that permitted tracking the behavior of the segment and its surrounding DNA in the operation of the immune system (see Figure 2). "It was a purely accidental finding," Imanishi-Kari says, "but I think I was one of the first to show that some antibodies have inheritable specificity" and to reveal "a serological marker for the variable regions."

    Imanishi-Kari's NP system exploited the marker to study antibody responses. In 1976, while working in Rajewsky's laboratory in Cologne, she and Rajewsky attended a meeting at the N.I.H. in Bethesda, Maryland, and encountered David Baltimore. Baltimore knew Klaus Rajewsky and thought highly of him, and he was intrigued to learn about Imanishi-Kari's NP system. "So, through my association with Klaus," Baltimore says, "I became associated with Thereza."

Initiating the Experiment: Mice and Molecules

In 1982, Baltimore had an idea that led eventually to the Cell paper. Imanishi-Kari had recently obtained antibodies against NP, with a distinctive idiotype, from hybridomas that she had developed from an inbred strain of mice called BALB/c. Scientists in Baltimore's laboratory, using hybridomas from Imanishi-Kari's lab, isolated and characterized the DNA responsible for the part of the variable region of these antibodies that included its idiotypic birthmark. Baltimore's idea was to use a gene engineered to contain this DNA in an experiment with a new genre of mice that had recently been introduced into the laboratory scene. These animals were just like any other laboratory mice except for one feature: A gene from another animal had been inserted into them when they were just newly fertilized eggs. When the mice were born, they contained a copy of the gene in every cell in their bodies. Scientists called the animals transgenic mice. They were highly promising instruments for biomedical research because they enabled observation and analysis of the inserted gene's impact on a living mammalian system. Baltimore expected that a suitably constructed transgenic mouse might reveal something about the action of immune genes.

    He thus proposed to insert the gene engineered from BALB/c DNA into an inbred strain of mice designated C57BL/6--black mice sometimes called "Black/6" for short. Baltimore wanted to see whether the gene would express itself--that is, contribute to the formation of antibodies--in the recipient mice as though it was native to them. Beyond that, he hoped that the presence of the foreign gene in the Black/6 mice would reveal something about how the process of rearrangement is controlled. Biologists thought that once the DNA segments in a cell were rearranged to produce a specific antibody, a kind of negative feedback process prevented any rearrangement from occurring in parallel parts of the cell's DNA. They believed that this process prevented any one B cell from producing more than one kind of antibody. The engineered gene would comprise DNA segments that had already been rearranged. The question was whether it would, as theory suggested, inhibit rearrangement of the truly native DNA segments in the mice so that they would not produce antibodies of their own against NP.

    In 1983, Baltimore sent a suitably engineered gene to a biologist at Columbia University named Frank Constantini, who knew how to make transgenic mice. The gene was designated "17.2.25," after the hybridoma from which it had been obtained, and would serve as what scientists call the transgene--that is, the foreign gene that would be inserted into the mouse. Constantini inserted 17.2.25 into the newly fertilized eggs of normal Black/6 mice. The eggs were introduced into the wombs of surrogate mother mice, and some of the altered eggs developed into transgenic mice--that is, mice that carried the foreign gene for the antibody to NP (Figure 3).

    Baltimore next established a small colony of the transgenic mice from Constantini in his laboratory on the fifth floor of the M.I.T. cancer center. He enlisted two of his postdoctoral fellows--David Weaver and Rudolf Grosschedl--to work with them, but he remembers remarking one day, "You know, we really ought to look at the immune response in these animals in a more serious way, and I don't understand how to do that." Baltimore wanted to learn about the kind of antibodies that were circulating in the blood of the mice, and such research required serology. He had neither the skills nor the tools of serology in his laboratory, but Imanishi-Kari commanded both. Since her arrival at M.I.T., Imanishi-Kari had permitted herself little involvement with Baltimore's research. Baltimore says that the fact that they had previously published together had "sort of chilled our ability to collaborate because, as a young person, she was very afraid of being seen as the adjunct of the senior person .... I tried to keep a distance." Once he had the transgenic mice, Baltimore recalls, "I went to Thereza and I said, `Look. I know you've been worried about the relationship, but here's an opportunity where your expertise will make a big difference. And I think you can only benefit both of us and science if you'll put some effort into it.' And she agreed."

    In late 1983, Grosschedl and Weaver went downstairs to Imanishi-Kari's laboratory for instruction in the rudiments of serology and hybridomas. Their job was ultimately to analyze the antibodies produced by the hybridomas using the techniques of molecular biology. They first checked the antibodies that were circulating in the blood of the living transgenic mice. Baltimore's group found elevated levels of NP-sensitive antibodies that had been produced by the 17.2.25 gene. However, they also detected normal levels of antibodies native to the mice, which suggested that the introduction of the transgene had not inhibited rearrangement of the DNA that generated them. In mid-1984, intrigued by the result, Baltimore's group started to investigate the phenomenon in the region of the mice where the antibodies were actually produced--that is, at the level of individual B cells.

    About this time, Grosschedl left for a post on the West Coast, so Weaver became the linchpin between the laboratories of Baltimore and Imanishi-Kari. With a Ph.D. in biochemistry from Harvard Medical School, he was particularly interested in DNA rearrangement, which occurs in cancer as well as in immune responses. In a sense, the experiment that led to the disputed Cell paper was his; he would be the senior author on the publication. It was commonly referred to as "Weaver et al.," and Weaver himself, a leanly built man who is precise in speech and quietly resolute in manner, says that it gave him an early but uncomfortable kind of notoriety.

    Imanishi-Kari instructed Weaver in how to grow hybridomas. In late 1984, after a number of failed attempts, he succeeded with her help in obtaining growth in four batches of hybridomas that she had given him--two of lymph cells and two of spleen cells from the transgenic Black/6 mice. The experiment now proceeded along two complementary lines: serological analysis by Imanishi-Kari of the antibodies produced in the hybridomas and molecular analysis of them by Weaver. (The fact that the experiment depended on these two types of research, which are technically quite different from each other, helped to make the effort difficult to understand for non-specialists, including some federal investigators.) David Baltimore recalls that he and his collaborators had expected the experiment "to be fairly straightforward" but that it wasn't. In fact, the results that came from both Imanishi-Kari's and Weaver's investigations were unexpected, right from the beginning.

Imanishi-Kari's Contribution: Serology

To characterize the antibodies, Imanishi-Kari decanted the supernatants into small cuplike wells--they measure about a quarter inch across the top--that were indented in parallel rows on a rectangular plate of clear plastic measuring several inches on a side. (Most of the plates used in the experiment contained 96 wells, in eight rows of twelve.) The sides of the wells were coated with either a reagent that would grab antibodies against NP or a reagent that would capture antibodies with the telltale idiotype. Imanishi-Kari then washed the supernatant out of the wells and determined the characteristics of whatever antibodies had stuck to the sides (Figure 4). She was first interested in identifying their isotypes--a characteristic of antibodies that would help reveal whether they had been produced by the transgene or genes native to the mice.

    Antibodies are also called immunoglobulins--"Ig" (scientists say "eye-gee"), for short. In vertebrates--mice, for example--they are divided into five general classes. One of them, labeled IgM, is the first class of antibody produced by a developing B cell. Another, labeled IgG, is the most common immunoglobulin in the blood. All antibodies that belong to one class of immunoglobulins are said to have the same "isotype" (not to be confused with "idiotype"). Isotype is determined by the composition of a long chain of amino acids--called the heavy chain--that runs from the upper tip of each arm of the Y down to the bottom of its trunk. The variable region of the chain forms its upper stretch. The rest of the chain is called the constant region. The chemical composition of the heavy-chain constant region defines the antibody's isotype; it is essentially the same in all the antibodies that comprise any one class. In IgM antibodies, the isotype is termed mu. In the IgG variety, it is termed gamma.

    When Imanishi-Kari checked the isotypes of the antibodies generated by the first transgenic hybridomas analyzed in the experiment, she discovered that many of them were gamma. Here was one of the first surprises. The transgene could only produce antibodies that were mu. Since the antibodies were gamma rather than mu, Imanishi-Kari was forced to conclude that they could not have come from the transgene. She says she wondered about the result, noting, "If you have something very unusual, you think you maybe did something wrong .... Because with serology, you don't know whether what you're seeing is real or whether something's wrong with the reagent you're using. It's always ambiguous."

    To learn more about what was happening, Imanishi-Kari sought to characterize more specifically the antibodies against NP that the hybridomas were generating. Like the transgene, native genes in the Black/6 mice could also produce mu antibodies. But mu constant regions can vary slightly from each other by a tiny bit of chemistry, producing differences in what scientists call their allotypes. The allotype of the transgenic antibody was designated mu-a; that of the native antibody was designated mu-b. To determine the origin of the antibodies that the hybridomas were producing, Imanishi-Kari needed reagents that would distinguish between the two. Some time after August 1984, she obtained a newly available reagent from Henry Wortis that would react only with mu-b antibodies--that is, only with antibodies emanating from a native gene. It was called AF6-78.25. Early in 1985, she got hold of a reagent for detecting transgenic antibodies. It had been devised in the laboratory of a scientist at the N.I.H. named William Paul, and it was called Bet-l. Under suitable conditions, it was far more likely to latch on to mu-a, the transgenic antibody, than to mu-b, the native one.

    Imanishi-Kari tested for mu-a antibodies using a radioimmune assay. For this procedure, the wells in the plastic plates were coated with a reagent sensitive to the idiotype and then filled with supernatant from the hybridomas. Antibodies with the idiotypic birthmark would stick to the coat. She then washed the wells. She made Bet-l radioactive by combining it chemically with radioactive iodine, which gives off gamma rays (a form of high-energy radiation; not to be confused with gamma isotypes; which the rays have nothing to do with): Bet-l was then applied to the wells and would attach itself to any mu-a antibodies that might have adhered to the coated sides. The wells were then cut out of the plate by using a hot wire to slice horizontally through the plastic, a process that generated a lot of smoke. Each well was tweezed into its own small test tube. A rack held the tubes side by side in the order in which the wells in them had occurred on the plastic plate (see Figure 4).

    The rack was designed for insertion in a gamma radiation counter that was available in a small room near Imanishi-Kari's laboratory. She shared the counter, along with additional equipment, with other scientists on the first floor of the M.I.T. cancer center. The machine processed each tube successively, measuring the gamma radiation emanating from each well for up to a minute. The detection of gamma radiation meant that Bet-l and, hence, transgenic antibody was present (because it was to that antibody that radioactive Bet-l preferentially attached itself). The intensity of the radiation, quantified as gamma counts per minute, indicated how much of the antibody was there. The radiation counter was hooked up to a teletypewriter that printed a record of the counts from each well in a column on a roll of paper (Figure 5). A similar procedure carried out with the AF6-78.25 reagent tested whether the supernatants contained mu-b antibodies, that is, antibodies originating from a native gene.

    The serological analysis demanded huge amounts of care and labor. It meant preparing and checking reagents like Bet-l. It required the management of hundreds of wells at any one time--keeping track of the reagents used on the hybridomas, of the concentrations of antibodies in the supernatants, and of which wells went into which radiation-counter tubes. "Sometimes I made mistakes," Imanishi-Kari says. Her collaborator Moema Reis recalls that she would start an experiment, run the tubes with their cutout wells into a radiation counter, and find the data emerging from the counter when she was already absorbed with a second experiment. She would set the counter printouts aside for later collection and collation, Imanishi-Kari might leave hers in file folders or drawers or on window sills for months.

By the summer of 1985, Imanishi-Kari and Reis had publishable data on transgenic hybridomas in 340 wells. The antibodies produced in 172 of them displayed idiotypic birthmarks that were not the same as those from the transgene but were closely related to them. However, only 42, fewer than a quarter, of these wells contained mu-a antibodies, which meant that the antibodies in fewer than a quarter of the wells derived from the transgene. Another 11 wells contained antibodies that were mu-b, which meant that they had been produced by native genes. The remaining 119 wells generated antibodies that were neither mu-a nor mu-b; they all had to come from native genes, too. Thus, the hybridomas in the vast majority of the wells--119 plus 11, making a total of 130 in all--produced antibodies to NP with idiotypic birthmarks similar to the transgene's, but these antibodies had been generated by genes native to the normal Black/ 6 mice.

Weaver and Baltimore's Contribution: Molecular Biology

All the while, Weaver had investigated the behavior of the hybridomas at a molecular level. He first looked for expression of the transgene's DNA. The machinery of gene expression produces a molecule called RNA (ribonucleic acid) that complements the coding regions in the gene's DNA. Weaver thus determined the characteristics of the RNA that the hybridomas were producing. He used a standard technique of joining the RNA in question with a radioactive molecule likely to resemble it, stimulating the ensemble to migrate through a gel several inches long under the force of an electric field, then taking a picture--a radiograph--of the gel (Figure 6). The radiograph revealed the distance the radioactive ensemble had moved through the gel. He could then compare that distance with the distance traveled by RNA whose originating DNA was known--in this case the transgene DNA. The comparison would indicate whether the transgene or some other gene had expressed itself.

    Weaver's early radiographic probes indicated--here was another of the surprises--that while the transgene was expressed in some of the hybridomas, it was not expressed at all in many others. Its DNA was definitely present, a point that Weaver troubled to make sure of, but it was quiescent. Weaver's analysis also revealed the presence of RNA that traveled the distance characteristic of the DNA for a gamma rather than a mu isotype--a result consistent with Imanishi-Kari's serological isotyping.

    Weaver then sought to identify by molecular means just which genes were generating antibodies in a selection of the transgenic hybridomas he had grown with Imanishi-Kari. He analyzed thirty-one such hybridomas and Albanese examined another three, making a total of thirty-four. Weaver determined that in many of these hybridomas, the gene that was being expressed belonged to the native repertoire of the Black/6 mice, which also complemented Imanishi-Kari's serological findings. Albanese's molecular data reinforced Weaver's. "No data was thrown out because it didn't fit a story," Albanese later told federal investigators, adding that "in many cases, what you find that's surprising is very interesting."

    Weaver did his work mainly in the new Whitehead Institute, of which Baltimore was the director and to which he had moved his laboratory and the two dozen or so people then working in it in the summer of 1984. Affiliated with M.I.T., the Whitehead Institute was only a couple of blocks from the cancer center, and Weaver visited Imanishi-Kari regularly while Baltimore discussed the developing data with her in telephone calls and meetings in his laboratory and hers. Baltimore later told several N.I.H. staff investigators that he took special care to figure out what Imanishi-Kari was telling him, explaining, English is "about her third or fourth [language], and by the time things got translated from Portuguese through Japanese into English, with an occasional foray into Finnish and German ... some things are not perfectly clear. So I had, all through the time I have dealt with her, times when I was a little uncertain about exactly what was being said, and then we would just sit down and I would go over it until I felt comfortable."

The Collaborators' Conclusions

The collaborators wanted to be sure that normal Black/6 mice did not generate the kind of NP-sensitive antibodies with the idiotypic birthmark that showed up in their transgenic siblings. Imanishi-Kari's characterization of antibodies in the blood of the normal mice had already indicated that they did not. Now, during the course of the experiment, she and Moema Reis each separately constructed and tested hybridomas with lymph cells and spleen cells taken from normal mice. Between them, they found that only one among 143 of these hybridomas produced antibodies with the distinctive idiotype.

    Late in the summer of 1985, Weaver drafted the paper about their results. Taken together, its findings were remarkable. Many of the hybridomas produced antibodies that had the idiotypic birthmark resembling the transgene's but that derived from genes native to the transgenic mice. It seemed that the introduction of the transgene did not inhibit rearrangement in the Black/6 mice's B cells. On the contrary, it seemed that somehow the transgene stimulated the abundant production of antibodies that their B cells would have otherwise produced infrequently or not at all. O'Toole, who by then had been in the laboratory for several months, was asked to read the draft critically. She gave it "a very careful review," she says. She supplied a number of editorial suggestions, including a rewording of the title into the one that was actually used. "It was a beautiful paper, beautiful data, dramatic findings," she remembered her reaction.

    During the fall of 1985, it was passed back and forth between the two laboratories, undergoing several revisions and rewriting by Baltimore. The collaborators, thinking that their experimental results raised important questions about how immune genes were rearranged to produce antibodies, submitted their paper to Cell on December 13, 1985. They subsequently dealt with the comments of referees--the reports were enthusiastic, praising the data as "very convincing and properly detailed" and calling the research "an important study"--and submitted their revised article on February 10, 1986. It appeared in the journal's issue for April 25.

    While the collaborators agreed that the foreign gene stimulated the abundant production of antibodies in the Black/6 mice that they did not ordinarily produce, they disagreed about why the phenomenon occurred Imanishi-Kari was inclined to think that the responsible mechanism was a process that Niels Jerne had proposed in 1974. Called idiotypic mimicry it figured in a kind of network of antibody responses that mobilized the immune system. She says that when she left Germany she doubted that any idiotype network was significant in the actual functioning of the immune system but that the data from the transgenic mice compelled her to reconsider.

    Baltimore, skeptical of idiotypic mimicry, thought that some kind of molecular mechanism within the transgenic mouse cells accounted for the unexpected antibody production. But whatever his interpretive disagreement with Imanishi-Kari, he was convinced that the high incidence of native antibodies to NP reported in the Cell paper was genuine. Two independent lines of analysis led to the same conclusion. Baltimore later told a congressional subcommittee, "For [Imanishi-Kari] to elaborate fraudulent data would have been most unlikely because the redundancy in the study would so likely have shown it up."