Robert A. Weinberg
Robert A. Weinberg; drawing by David Levine

Vivian Bearing, the main character in Margaret Edson’s Pulitzer Prize-winning play, Wit, suffers from advanced metastatic ovarian cancer. Having undergone eight full-dose treatments of an experimental chemotherapy, she muses that her doctors will no doubt write an article for a journal about her.

But I flatter myself. The article will not be about me, it will be about my ovaries. It will be about my peritoneal cavity, which, despite their best intentions, is now crawling with cancer.

What we have come to think of as me is, in fact, just the specimen jar, the dust jacket, just the white piece of paper that bears the little black marks.1

Bearing’s bitter reflections on the connection between her “me” and her disease call to mind long-running disputes about the sources of cancer. Does it originate in the self’s genes, a sentence of fate to the DNA “me”? Does it come from the behavior of the autonomous self, the “me”‘s styles of life and consumption? Or does it originate outside the self, in constituents of the environment that invade the organs of the “me”? The answers to these questions bear on how cancer might be treated or, better yet, prevented, and who should bear the costs of dealing with the disease.

At the turn of this century, physicians knew that some cancers ran in families—an indication that the disease came from some inheritable essence of the self. Yet they were also aware that the large majority of cancers occurred sporadically, independent of any prior family history. The latter evidence suggested that cancer originated in agents lurking in the environment. As early as 1775, a British physician named Percival Pott had linked scrotal cancer in former London chimney sweeps to their exposure to soot. By the early twentieth century a growing body of observations suggested linkages between cancer and the materials of modern industrial life.

German physicians and scientists, following their country’s strong tradition of social medicine, took a vigorous interest in the causes of cancer external to the natural self. After 1933, the Nazis mounted a broad-gauged “war on cancer,” to borrow from the title of Robert Proctor’s arresting and important exploration of the Hitler regime’s concern with the environmental sources of the disease. Proctor, a highly regarded historian of science whose works include the standard account in English of Nazi eugenics,2 is by no means an apologist for Nazi medical science. His purpose is partly to point out that it was complex, that along with sadism it displayed “fertile, creative faces.” He contends, indeed, that in cancer research good science did not flourish despite Nazism but because of it.

That science of high quality can arise in brutal, antidemocratic regimes should not by now be surprising, even for science under the Nazis. Proctor’s point is nevertheless worth repeating, since it remains a commonplace that science flourishes only as a ward and ally of democracy. But the value of his unblinking book lies in its revelations about why the Nazis were absorbed with the problem of cancer, what they learned about the sources of the disease, and the actions they took to prevent it.

During the 1930s, stomach cancer was the most prevalent form of the malady, which prompted Nazi scientists, like their predecessors, to put forward theories about the carcinogenic impact of alcohol and various foods. Lung cancer, coming up fast behind stomach cancer, was responsible for a quarter of all cancer deaths. Its incidence drew attention to the use of tobacco, which had risen fivefold since the turn of the century and which had been tentatively linked to lung cancer by earlier research. There was much speculation that the causes of cancers of the respiratory tract might be found elsewhere—for example, in the air pollution caused by automobile exhaust. But in the Nazi view, Proctor writes, tobacco was the chief culprit, “a genetic poison; a cause of infertility, cancer, and heart attacks; a drain on national resources and a threat to public health.”

Proctor argues that while the anti-cancer campaign drew on the German tradition of social medicine, it was far more deeply rooted in the Nazi ideology of improving the health of German society so that its men and women could better serve the Fatherland. Devotion to health could coexist with murderous cruelty—often in the same people—because the regime rationalized both as means of restoring Germany’s vitality. By identifying the environmental sources of cancer, the Nazi government would be able to control them and thus prevent disease and, ultimately, protect every racially worthy German.

The Nazi campaign against smoking and drinking, and in favor of a healthy diet, was a product partly of Nazi puritanism and prejudice and partly of the romantic right-wing idea that modern industrial civilization had polluted human beings, alienating them not only from nature but health. The dutifully healthy German was embodied by Hitler himself, a vegetarian who neither drank nor smoked, and refused to allow anyone, including Eva Braun and Martin Bormann, to smoke in his presence; he denounced tobacco as “one of man’s most dangerous poisons.”


Concern for pollution—whether environmental or the result of bad personal habits—was strengthened by the Nazi preoccupation with eugenics. Proctor rightly observes that, while suffused with a quackery all its own, eugenics was also seriously engaged with issues in genetics and health. In 1927 the American geneticist Hermann J. Muller had demonstrated that X-rays could produce genetic mutations in fruit flies. During the 1930s, a number of German scientists, extrapolating from Muller’s results, argued that they reinforced the notion that some cancers were heritable; they reasoned that cancer genes, once created, would be passed from one generation to the next. Such theorizing was mistaken as a general proposition, since a mutation in, say, a lung cell could not be transmitted to offspring. But it led German scientists to search for carcinogens in the environment that might damage the nation’s genetic heritage.

Proctor emphasizes that the intense Nazi interest in cancer produced important research into the disease. The regime established registry offices in selected cities and regions to record not only deaths from cancer but also every case of its occurrence. Coupled with an increased emphasis on autopsies that predated the Nazi regime, the registries produced abundant statistics on the incidence of the disease, a far more useful epidemiological measure than deaths caused by it. Proctor points out that during the 1930s the Nazis led all other nations in the number of publications on the general subject of occupational health. Searching for possible causes of carcinogenic mutations, investigators studied the relationship of cancer to radium and uranium, arsenic, chromium, asbestos, and aniline dyes, among other substances. They identified hazards not only in workplace toxins but also in alcohol, a variety of foods, and tobacco.

“Documenting the lung cancer hazard of smoking was one of the most remarkable achievements of the Nazi era,” Proctor writes. He gives particular attention to two pioneering studies—one published in 1939 by Franz H. Müller at a hospital in Cologne, the other in 1943 by Erich Schöniger and Eberhard Schairer at a new tobacco research institute in Jena. Using questionnaires, Müller made a statistical comparison of the behavior of ninety-six lung cancer patients with that of a “healthy” control group of similar size. Nonsmokers were more common in the healthy group (16 percent) than in the lung cancer group (3.5 percent), and the lung cancer victims smoked more than twice as much tobacco each day than did members of the healthy group. Müller’s demonstration of a link between smoking and lung cancer was confirmed by Schöniger and Schairer, who used his methods but, enlarging on the scope of his inquiry, also found that smoking was not a cause of stomach cancer.3

German health activists campaigned for restrictions on the use of pesticides, asbestos, and food dyes made from coal tars. Magazines concerned with the environment regularly castigated smoking, and anti-tobacco activists called for advertising bans, tobacco taxes, and prohibitions against the sale of tobacco to youth. Eventually, the Nazis installed no-smoking cars in trains and prohibited smoking in many workplaces, public buildings, hospitals, and rest homes. They also enacted laws requiring compensation for work-related illnesses such as cancers arising from asbestos or petrochemicals. Official concern with the impact of industrial pollutants on women’s reproductive functions led to regulations prohibiting the exposure of women and children to hazardous substances or processes.

The bans, however, were enforced only selectively, and the activists made little headway against powerful industrial interests, especially after Germany went to war. The claims of war production shifted attention away from ridding the industrial environment of carcinogenic hazards. Instead the Nazis purged the workforce of its sick members, many of whom were sent to the gas chambers. “The war on disease turned into a war on the diseased,” as Proctor puts it.

Environmental carcinogenesis did not get much attention in the United States during most of the middle third of the century. The demand for war production, as in Germany, and then the pressures of the cold war discouraged attention to pollution in the workplace. During the early years of the cold war, the Atomic Energy Commission, claiming that the nation could not take the risk of slowing down development of its nuclear arsenal, suppressed research into the high frequency of cancer among uranium miners and disparaged claims that radioactive fallout could cause cancer.

Besides, claims of chemical carcinogenesis were open to question. The increasing scientific evidence that smoking could lead to lung cancer focused greater attention on the causes of carcinogenesis in individual behavior, which undermined the force of theories of its workplace origins. Then, too, the incidence of cancer increased with age, which suggested it might be caused more frequently by long-term bodily processes than by environmental insults. And the disease struck only a minority of Americans, including only a minority of even those who smoked; the medical and scientific director of the American Cancer Society took these findings to mean that the most important factor responsible for cancer was “individual susceptibility” to the disease.4


However, the doctrine of environmental carcinogenesis began to take hold after Rachel Carson called attention to it in 1962 in Silent Spring.5 By the 1970s, environmental explanations of cancer had become a major force in American politics and culture. They were given authority by the environmental movement and by numerous epidemiological studies linking specific chemicals to human cancer. Environmental explanations were reinforced by tests of the cancer-producing effects of chemicals on animals and by increasing knowledge of the impact of chemicals in causing mutations. The latter had first been demonstrated during World War II, when British researchers showed that mustard gas, like X-rays, could provoke genetic changes in fruit flies. In the 1970s, Bruce Ames, a geneticist at Berkeley, devised a method for measuring the mutagenic potency of different chemicals by using bacteria growing in petri dishes. While greatly speeding and simplifying such assessments, the Ames test also showed that many of the chemicals known to cause cancer in laboratory animals such as mice incited genetic mutations in bacteria as well, thus adding weighty evidence to theories of environmental carcinogenesis.

By the early 1980s, the rapidly developing field of molecular biology had begun to increase understanding of how mutations of the basic genetic substance—since 1953 known in most organisms to be deoxyribonucleic acid (DNA)—provoked by carcinogens could turn normal cells into cancerous ones. The new understanding owed a good deal to the Rous Sarcoma Virus (RSV), which, on injection into the breast of a healthy chicken, would induce the formation of the kind of tumor called a sarcoma. By the late 1950s, several other viruses had been identified that would provoke tumorlike growth in cell cultures from other animals, and RSV itself was discovered to stimulate such growth not only in fowl but also in mammals, including mice, rats, hamsters, rabbits, and monkeys. Although some scientists speculated that human cancer might arise from tumor viruses, that theory foundered on the fact that cancer was not a contagious disease. Still, the powers of the viruses to produce tumors in animals indicated that they might well provide a key to the puzzle of how cancer occurs in humans.

A new way of thinking about that puzzle emerged in the 1960s with the breaking of the genetic code. DNA embodies information that instructs the cellular machinery to manufacture specific proteins. Some scientists reasoned that perhaps the genetic information characteristic of tumor viruses somehow got incorporated into the DNA of normal cells, where it awaited activation to pervert the cells so that they would multiply endlessly. By the 1970s, experimental techniques were available that permitted scientists to isolate particular segments of genetic material—for example, the segment in a virus that caused a tumor—and compare it with DNA from, say, a normal cell. In 1976, the biologists J. Michael Bishop and Harold Varmus, then at the University of California at San Francisco, reported that the genetic material of RSV closely complemented a gene in the cells of numerous animal species.

To Bishop and Varmus, the ubiquitousness of the gene indicated that it was probably involved in some normal and critical cellular function. It was not genetic information that was imported into a cell by a tumor virus and then awaited activation. Rather, it was a normal cellular gene. But its being a close cousin to the genetic material of RSV meant that it was a normal gene that could be turned into a cancer-causing gene—an “oncogene”—by a tumor virus or perhaps by a carcinogen. It was, scientists now said, a “proto-oncogene” that could become an oncogene. In 1978, Robert A. Weinberg, a biologist at MIT, demonstrated that the deleterious change from normal gene to cancer gene could in fact be provoked by a chemical agent and that when the perverted gene was introduced into normal mouse cells, it transformed them into cancerous ones.

These results, obtained from research with tumor viruses and animal cells, strongly implied that human cancers were also generated by the transformation of proto-oncogenes into oncogenes. Still, the relationship between the presumed human oncogenes and the increasing number of different oncogenes found in animal tumors was uncertain. Tumor viruses appeared to be responsible for most of the animal variety, but they were not found in most human cancers. However, in mid-1982, several laboratories independently reported that an oncogene isolated from human bladder cancer cells had been identified as a member of the family of animal oncogenes called ras (so named because they were first identified in a rat sarcoma). The discovery suggested that the same set of proto-oncogenes existed in the cells of all vertebrates and that while tumor viruses transformed them into oncogenes in animals, the trick was accomplished in human beings by chemical carcinogens. Later that year, three groups of researchers discovered that the ras oncogene differed from its normal cellular counterpart by a single point mutation—that is, by the alteration of just a single chemical rung in the double helix of its DNA.

In a book published several years ago, Robert Weinberg provided an extensive account of this revolution in understanding, bringing the story up to the mid-1980s.6 Now, in One Renegade Cell, he combines a compressed version of that account with a discussion of major subsequent developments, addressing what has been learned about the interplay between oncogenes and cellular behavior. It is difficult to know what to make of the two books as history. Weinberg draws on his own experience, conversations with many contributors to the revolution, and no doubt the numerous scientific papers that undergird it. However, he provides neither references for particular events nor specific sources for the major parts of his story. Nor does he offer a balanced overview of events; he acknowledges that, in order to keep his story clear and accessible to lay readers, he omits some scientific episodes that are more consequential than those he explores.

Yet if his accounts are less than satisfactory as history, they are valuable as memoirs, commentaries, and expositions of science. Weinberg describes the scientists involved in the genetics of cancer, including himself, with refreshing frankness, reporting false starts, mistaken expectations, failed experiments, and intense competition. He explains technical issues with engaging lucidity. One Renegade Cell is, all in all, an absorbing and instructive book, an effective primer on the state of knowledge of how cancer arises, including a sketch of what that knowledge implies for prevention, treatment, and cure.

The living body contains roughly ten trillion cells. Ordinarily, “these cells are well behaved and public-spirited,” Weinberg writes, multiplying enough to fulfil their vital functions but not so much as to oppress their neighbors. But at times, a “renegade cell” will embark on a growth program of “unlimited expansion,” heedless of the well-being of the cells in surrounding tissue. After many years, a tumor mass will appear. By the time the mass is detectable, it usually comprises a billion or more cells, each one of them a lineal descendant of the original renegade. Cancer thus originates in the transformation of a single civic-minded cell into a self-multiplying outlaw.

Every cell possesses a program for self-restrained growth that is coded into its genes, including its proto-oncogenes. Initial expectations that the program could be thrown out of kilter by the creation of a single oncogene proved unfounded, both from epidemiological considerations and from laboratory results. Since the incidence of cancer rises with age, it seemed that the transformation of a cell into the seedling of cancer must require multiple events that cause mutations—a sequence of carcinogenic “hits” in different genes stretched over time. Weinberg’s lab showed that while introduction of a single oncogene into embryonic mouse cells did not make them cancerous, the insinuation of two oncogenes did. Weinberg writes that the multi-hit theory is consistent with observations of growths in the colon. They range from an excess of normal-looking cells, to clumps of quasi-cancerous ones, to large cellular masses known as polyps, and, finally, to malignant growths. The typological difference indicates that cells progressing from normal to malignant pass through phases, evidence consistent with the multistep theory of oncogenesis. A family’s predisposition to cancer could mean that its genes had been hit by mutation once and that anyone in the family awaited only a second hit to develop a malignancy.

The multi-hit theory was seemingly confounded in the early 1980s, when searches for multiple oncogenes in human tumor cells came up emptyhanded. Many of the cells carried one oncogene but few carried two or more. However, in the mid-1980s research by a scientist at Oxford University indicated that normal cells contain genes that act as a brake on runaway growth. Isolated and identified some years later, they are called tumor suppressor genes. A mutational hit on this kind of gene, so that it no longer retards growth, figures as importantly in making a cell a renegade as does the creation of an oncogene, which promotes growth. More than a dozen such tumor suppressor genes are now known. Consistent with the multi-hit theory of oncogenesis, human cancers can arise when deleterious changes strike both proto-oncogenes and tumor suppressor genes—or, as Weinberg puts it, when the cell machinery undergoes “the simultaneous flooring of the accelerator and loss of brakes.”

Weinberg observes that not all carcinogens possess mutagenic powers. Alcohol and estrogen, for example, score weakly on the Ames test, yet they are linked with the formation of cancers. A combination of epidemiological analysis and molecular speculation traces their carcinogenic impact to the copying—or, rather, miscopying—of DNA during normal cell division. The copying process can misfire, introducing a mistake into the daughter DNA that amounts to a mutation. If the mistake occurs in a proto-oncogene or a tumor suppressor gene, the miscopying can be the equivalent of a carcinogenic hit. The likelihood of such a miscopying hit is thus increased by whatever promotes cell division, which includes exposure to estrogen or heavy drinking of alcohol. Alcohol kills cells in the lining of the mouth and throat, triggering their replacement by the division of surviving cells.

During the last decade, considerable light has been thrown on the puzzle of why the creation of an oncogene or the loss of a tumor supressor gene can incite a cell to become a renegade. Weinberg notes that the new understanding has not come primarily from cancer research but from investigations of what governs the normal processes of cell growth and division in simple organisms, notably baker’s yeast, fruit flies, and earthworms. Cells behave as good citizens by growing and dividing mainly in response to signals of encouragement from the community of cells that surround them. The signals are embodied in growth-factor proteins which are sent out by one cell and attach to receptor molecules tailored to receive them that lie on the surface of another cell. The receptor molecule, its structure now altered, sends biochemical growth signals to the cell’s interior via proteins that operate as signal “transducers”: they receive the signals, process them, and pass them along to the relevant cellular entities. A number of transducer proteins are coded for by proto-oncogenes—which is to say that these proteins are deeply embedded in the major signal-processing pathways of the cell. Weinberg explains that when the protein is produced by an oncogene, it can flood the cell with growth-stimulating signals. It might otherwise be restrained by a tumor suppressor gene; but if that gene is deactivated, the transducer protein will behave with unrestrained recklessness.

Still, Weinberg points out, the cellular system is remarkably stable; it is marked by several features that resist the appearance or proliferation of a renegade member. The cell’s signaling circuitry is sufficiently redundant that, to be perverted, more than one gene involved in it must undergo a carcinogenic hit. If DNA is miscopied in the course of cell division, a process of DNA repair will usually fix the error. Cells likely die naturally after forty to fifty doublings; a would-be renegade cell must neutralize the mechanism that normally causes the cells to die. Our bodies can also induce deleterious cells to commit suicide, a phenomenon called “apoptosis.” If a renegade cell survives these impediments and begins multiplying, the resulting clump of malignant cells will run out of oxygen and nutrients when it reaches approximately the size of one millimeter. Unless the clump connects to the body’s blood supply, it may remain in a static state for years. Weinberg writes, “Cancer is usually held at bay because it depends on a convergence of rare events that are unlikely to occur in an average human life span.”

Weinberg, like many others in his field, holds that the growing knowledge of the cellular processes that produce cancer “should take us far in conquering this disease.” He predicts that within ten to fifteen years biomedical scientists will know the elements of cellular wiring in exhaustive detail; they will have a catalog of tumor suppressor genes, and be able to predict quickly an individual’s susceptibility to a spectrum of cancers. He expects that basic molecular understanding of cancer will likely lead to “a new generation of antitumor drugs” designed to attack specific culprit proteins. For example, pharmaceutical companies are already trying to develop drugs that will interfere with the cell’s ability to manufacture the growth-promoting ras protein. Although not much is yet known about apoptosis, Weinberg believes that the greatest innovation in cancer chemotherapy may come from finding new ways of inducing cancer cells to commit suicide. “The prospects for the development of totally novel anticancer therapeutics are bright!” he writes.

But Weinberg also writes that “the big decreases in cancer deaths will… come from preventing disease rather than discovering new cures.” By this he means dealing with the “ultimate causes” of cancer—those that “really begin far outside the individual cell, in our environment, in the food we ingest, and the smoke we inhale.” In 1930, the annual rate of cancer mortality in the United States was 143 per hundred thousand; in 1990, adjusted for the rising age of the population, it was 190 per hundred thousand. Weinberg reports that almost all the increase derives from the use of tobacco. If we omit lung cancer, the overall adjusted cancer death rate would have dropped 14 percent between 1950 and 1990. The decline in smoking in recent years appears to have reversed the upward trend of male deaths from lung cancer. A similar result might be achieved if large numbers of people change to a low-fat, low-meat diet, Weinberg says, estimating that “diet plays a critical role in perhaps half of all human cancers.”

Some analysts contend that cancer could be further reduced in incidence by purging the environment of carcinogens. It has been authoritatively estimated that at least 6 percent of cancers arise from pollution of the workplace and the natural environment. Proctor, who is not alone in his views, holds that the fraction of cancers arising from environmental insults is likely higher, and that we know less than we should about environmental carcinogenisis because we devote relatively few resources to research in the subject. A number of scientists point to numerous “linkages” and “associations” found between different cancers and a variety of chemicals. The environmentalists also point to “cancer clusters”—concentrations of cancer victims in particular neighborhoods or regions—which make local air, soil, or water suspect. Proctor writes that the incidence of cancers—for example, non-Hodgkins lymphoma—that have little if anything to do with smoking or diet has not been flat in recent decades but has been rising. So, it appears, have the incidence of lung cancer among nonsmokers and the frequency of childhood cancers.7

Weinberg, for his part, writes that a great deal of epidemiological evidence runs counter to claims that the “industrialized West is being inundated with a cancer epidemic, and that most of this inundation is traceable to chemical pollutants in the air and in the food chain.” Hundreds of cancer clusters have been investigated, some in connection with lawsuits like the one narrated in the recent book and film A Civil Action. They have usually been found to lack statistical significance and a convincing local carcinogenic cause.8 The seemingly dominant view among epidemiologists is that the rates of most kind of cancers have held steady over the last fifty years, a period of rapidly increasing environmental pollution. Apart from tobacco-related cancers, the one significant exception is breast cancer, which in the United States has shown a slow but steady rise in incidence of roughly one percent annually. While a number of observers say the rise is caused by chemical pollution, Weinberg writes that a view “rapidly gaining ground” in the biomedical community attributes most of the increase to modern nutrition and reproductive habits, which have increased women’s lifetime exposure to estrogen.9

Still, in the twentieth century we have learned, even from Nazi science, that cancer is uniquely the product of neither genes, behavior, nor the environment. It can arise from genes alone or from interactions between our genes and what we ingest, inhale, or encounter. The corollary of that understanding is that the disease is best contested simultaneously on multiple fronts—by environmental research and regulation, by discouraging self-destructive behavior, and by encouraging the kinds of basic inquiry into cell growth and multiplication that have already produced such richly informative results.

Recent results from Weinberg’s lab have demonstrated once again that research in nonhuman organisms can point to processes of tumor formation in human beings. His lab had shown in 1983 that the introduction of two oncogenes into rodent cells could transform them into cancerous types. Although a number of labs had since then tried to achieve the same transformations with human cells, none had succeeded. However, this summer, in an announcement that attracted worldwide attention, Weinberg reported that his lab had accomplished the feat with human cells by inserting into them three genes: the oncogene ras, which promotes cell growth; a gene whose product neutralizes the function of tumor suppressor proteins; and a gene for the enzyme telomerase, whose action prevents the death that normally befalls a cell line after it undergoes forty to fifty doublings.10

The human genome contains its own gene for telomerase. The gene is active during embryonic development but then turns off in most human cells. In cells that become cancerous, it is somehow turned on again, permitting the cells to double indefinitely, forming tumors. Weinberg points out that telomerase is thus a distinctive feature of cancer cells and as such may represent their “Achilles heel.” He writes that anti-cancer drugs might be developed that “attack and inhibit telomerase while leaving the thousands of other enzymes in cells untouched,” adding, “Such a highly targeted drug may stop cancer cells in their tracks while having little if any effect on normal cells.”11

Like many scientists on the cutting edge of a field, Weinberg may be overly optimistic about the speed with which further understanding of carcinogenesis will develop and about the prospects of devising more effective therapies and cures.12 But there is no exaggerating the need for them. Mutagenic radiation strikes us from the heavens; carcinogens are contained in even the freshest of vegetables; and the unrepaired miscopying of DNA is an inevitable consequence of the trillions of cell divisions that take place within each of our bodies during a lifetime. Of every ten Americans virtuous enough neither to smoke nor to eat much meat and animal fat, two of them will still come down with cancer and need treatment.

This Issue

September 23, 1999