From Fisher’s Constitutional Hypothesis to Behavioral Genetics
The Legacy documents reveal that the tobacco industry undertook two separate lines of research to bolster the credibility of the constitutional hypothesis: that one or more genes affect a person’s risk for lung cancer, and that at least some of the same genes influence whether people become smokers. The latter component of the Fisher hypothesis – that genes influence smoking behavior – is a question that now falls under the rubric of behavioral genetics.
In 1962, during a lawsuit against the American Tobacco Company [17
], a new interpretation of the association between smoking and lung cancer was presented to company lawyers [18
]. This internal paper, written by American Tobacco Company President Robert Karl Heimann and a consultant, analyzed previously published data [19
] on the smoking habits of American Tobacco Company’s Richmond, Virginia employees. American Tobacco employees had been reported to smoke at a rate twice that of the general United States population, yet in more than fourteen years the mortality rate of over 11,000 employees from all causes (including cancer) was 30% lower than the general population [22
]. This result contrasted with numerous other surveys of smoking and lung cancer rates, and one goal of this internal paper was to reconcile these findings.
While the 1964 Surgeon General’s report explained the discrepancy in findings based on a healthy worker bias (i.e., since some categories of sick or debilitated persons in the population are not employed, it is inaccurate to compare a disease prevalence in a workforce to the general population) [14
, at 182], President Heimann and his co-author provided a different interpretation, that: “excessive smoking may be diagnostic of other factors predisposing to higher mortality rates” [22
]. They hypothesized that heavy smokers are “excess-prone” (see Appendix S3
online for full description). The tobacco industry representatives argued that their analysis of the published data showed that it was not cigarettes that caused lung cancer, but a third element (namely, genetic makeup) that caused people to both “take risks” (excess smoking, eating, etc.) and develop lung cancer.
However, this paper was not published because the American Tobacco Company lawyers feared that the opposition could use the results of this analysis against them [18
]. While this new study postulated an explanation for lung and other cancers that does not blame smoking, it did not prove a causal relationship and could turn court battles into a question of whose theory was superior [16
, at 441; 23]. Industry lawyers also thought it unwise for the manufacturers to support a paper that argued that smoking excessively indicates poor health, even if it does not cause it. This paper was withheld from publication, though Edward S. Harlow, Managing Director of Research and Development, expressed regret at the decision to shelve a promising alternative explanation [24
]. However, the data on American Tobacco Company employees continued to be used to cast doubt on a causal link between smoking and cancer by providing a sample of people who smoked heavily but had lower rates of cancer than the general population.
The “excess-proneness” model of smoking behavior was revisited in 1972 when Fred R. Panzer, Vice President of Public Relations for the Tobacco Institute, wrote a memo (“The Roper Proposal”) [25
] in response to a suggestion by Burns (Bud) W. Roper, an expert on public opinion polls, who recommended researching people’s “excessivity” [25
]. Panzer’s memo is an often-cited document on the tobacco industry’s strategy detailing the plan of “creating doubt of the health charge without actually denying it” [25
]. According to Roper’s hypothesis, if someone is driven psychologically and/or physiologically to over-smoking, over-drinking, etc., then this lifestyle of excess may contribute to an early death. If validated, Roper contended, “The focus would change from one product to a type of person” [emphasis in original] [26
]. Panzer saw a “credible alternative” to the causal theory in the connection between a theory of excess and the constitutional hypothesis. The public, Panzer wrote, “must perceive, understand, and believe in evidence to sustain their opinions that smoking may not be the causal factor” [25
]. Roper’s idea, grounded in the constitutional hypothesis, gave Panzer confidence: “It is my strong belief that we now have an opportunity to take the initiative in the cigarette controversy, and start to turn it around.”
Roper’s data were therefore used to support a constitutional hypothesis. For example, a correlation between divorce and smoking was explained in an internal Philip Morris memo [27
]: “Our position, of course, is neither that divorce causes the smoking nor that the smoking causes the divorce, but that both tendencies proceed from common constitutional factors relating to the difficulty of making a smooth adaptation to the life situation.” Behaviors relating to smoking and divorce were seen as the result of one’s genetic makeup [27
Panzer’s sole criticism of the Roper Proposal was that it was not “strictly scientific” [25
]. Whether searching for “excess-proneness” or “hereditary tendencies,” a scientific strategy to “take the initiative in the cigarette controversy” would not present itself to the tobacco industry for several more years.
Tobacco Industry Research on the Genetics of Nicotine Addiction
From the 1950s into the 1970s, the tobacco industry’s evolving defense against the causal theory of smoking leading to lung cancer or poor health increasingly made use of scientific research. Both the developing field of behavioral genetics and the idea of excess proneness became an area of special interest to the tobacco industry. In 1954 the industry published a “Frank Statement” to the American public and created the Tobacco Industry Research Committee (TIRC), which was nominally intended to provide research addressing the “smoking and health controversy.” The TIRC, subsequently renamed the Council for Tobacco Research (CTR), served as part of a public relations strategy to promote controversy and refute the unhealthy impact of tobacco [4
]. The initial critique by Fisher, epidemiological work by Heimann, and a later survey by The Roper Organization were a part of this effort to fund research to counter the growing body of evidence in support of a causal link between smoking and a host of diseases. The idea that smokers and nonsmokers are constitutionally different was a cornerstone of the industry’s argument, but a rigorous scientific means to advance this position was unavailable.
This changed with the advent of behavioral genetics. Edwin Jacob, a tobacco industry lawyer who managed secret “special projects” that funded scientists to produce results that supported the industry’s positions, comments in a 1974 internal report that “it has become increasingly apparent that constitutional hypotheses merit massive investigation,” and, “At present, happily, newly developed research knowledge and techniques – especially in genetics – provide the possibility of much more extensive and promising exploration of the constitutional hypothesis than has heretofore even been conceivable.” [28
] The tobacco industry capitalized on the advances being made in genetics with the intent of bringing new hope to their research and legal defense strategies.
In the mid-1970s, a group of researchers approached CTR with a proposal to study the genetics of “smoking behavior.” A 1974 draft of the study, titled “Genetic and Environmental Basis of Tobacco-related Behavior,” describes its research questions:
Specific research questions will include: (1) Is it possible to demonstrate unequivocal genetic involvement in smoking behavior? (2) What is the relative magnitude and nature of the genetic control? (3) Is it possible to determine the genetically influenced neurophysiological and pharmacological correlates of tobacco-related behavior? (4) Is it possible to discriminate between smokers and non-smokers without reference to smoking behavior? [29
The fourth question has the same research goal as previous work on excess-proneness: what separates the smoker and the non-smoker, apart from smoking? This 1974 proposal explicitly states that the purpose is to test the Fisher hypothesis: “Data obtained from investigations into these questions should contribute substantially to understanding one aspect of the ‘Fisher hypothesis’ – the hereditary influence on smoking behavior. In addition, we shall eventually be able to address the total hypothesis directly” [30
]. The plan was to first determine a hereditary influence on smoking, and then attempt to show a linked hereditary influence on disease.
Three related projects were proposed: 1) A human study using populations in Hawaii to be conducted by Dr. Geoffrey C. Ashton; 2) Animal studies using mice to be conducted at the Institute of Behavioral Genetics (IBG) of the University of Colorado, with Dr. Gerald E. McClearn as principal investigator, and; 3) Studies using twins from the Swedish Twin registry under the guidance of Dr. David W. Crumpacker, also of the University of Colorado.
The proposed coordinator of the three components was McClearn, who in 1967 established the IBG in Colorado and was its first director. McClearn and the IBG were at the forefront of behavioral genetics [31
], and McClearn was the President-elect of the Behavioral Genetics Association when the first proposal was submitted to the CTR in 1974 [32
]. The CTR Scientific Advisory Board recognized that McClearn’s alcohol studies “seem[ed] successful in leading to important and new concepts and biological interpretations” [33
], and that McClearn’s animal study was a chance to test the Fisher hypothesis: “It is a remarkable opportunity to find out the degree of genetic influence upon the smoking habit or nicotine… [I]f some part of the smoking public inherits this tendency, and their medical data are collected, the facts would likely revolutionize the general scientific attitude in this field” [34
]. Sommers also focused on the ability of this work to “provide a base of knowledge to build on,” and set the foundation for further advances on the genetics of smoking behavior [34
]. Proposed projects were given separate grades for “merit” and “relevance” by reviewers, and Sommers gave this work an “A” for both.
CTR leaders acknowledged that this study conveyed both benefits and risks to the tobacco industry. The deficiency of the animal study was that, “This proposal is essentially a study of the biological and behavioral responses of mice to nicotine. The assumption is that nicotine may be habituating or addictive, or that genetic differences in tolerance may exist and may be identified or found” [33
]. At this time the tobacco industry maintained publicly that nicotine was not addictive, though internal memos indicate the industry had been aware of the addictive nature of nicotine for over fifteen years [4
]. While this avenue of research may validate the Fisher hypothesis, investigating a genetic propensity to nicotine addiction could be seen as contradicting the industry’s stance that nicotine is not addictive.
The potential conflict over nicotine may have been one reason this study was proposed as a CTR “Special Project” [35
]. Using these Special Projects, industry lawyers supported research that could refute the relationship between smoking and disease or continue to encourage controversy [16
]. In a confidential memo, industry lawyers wrote, “CTR Special Projects instead were primarily a means of attracting researchers to areas of scientific inquiry not being addressed and, in particular, were focused on satisfying the needs of lawyers involved in defending the tobacco industry in litigation” [36
]. Proposing research on the genetics of smoking behavior as a CTR Special Project suggests industry lawyers thought the work would be helpful in litigation defense.
Edwin Jacob’s 1974 report on Special Projects for the tobacco industry expressed support for the IBG studies “with the conviction that, for the first time, the Tobacco Industry has available, within realistic compass, the scientific means to carry out a major investigation of the constitutional hypothesis” [28
]. In March 1975, at a meeting on industry-sponsored research, the work of McClearn was specifically addressed [37
]. Two industry lawyers were present, including David R. Hardy of the law firm Shook, Hardy, and Bacon [38
]. Commenting on the Fisher hypothesis, Hardy said, “to prove this is the industry’s best hope.” Drs. Gardner and Sommers, who attended the meeting, agreed. Both the animal study and the Swedish twin study were approved [39
While both McClearn’s animal study and Crumpacker’s research using the Swedish Twin Registry were funded by the CTR, Ashton’s Hawaii population study was not. In 1993, the Wall Street Journal
reported, “Dr. Ashton says the lawyer told him ‘the presidents of the tobacco companies had turned down the proposal because they didn’t think the outcome would be useful to them’” [40
]. Not denying this statement, industry lawyers responded in an internal memo that stated: “A response to this ‘allegation’ is probably unnecessary. The claim however, could be used to support our arguments about the purposes of the CTR Special Projects and to rebut the argument that CTR Special Projects were how the industry funded ‘adverse’ research it might want to ‘hide’” [36
]. In contrast to Aston’s (unfunded, potentially “adverse”) research study, the IBG studies at the University of Colorado were seen as potentially beneficial to the industry’s position by investigating the constitutional or Fisher hypothesis, and represented Special Projects that would not be construed as “hidden” or “adverse” research.
Papers from the IBG were mentioned in CTR Annual Reports from 1976 until 1983. These Annual Reports listed abstracts for published papers completed with CTR support, and were sent to physicians, researchers, tobacco industry leaders and others throughout the world. McClearn received more than $530,000 in support of his mouse studies between 1976 and 1979 [41
]. During the same period, Crumpacker received over $550,000 in funding for work with the Swedish Twin Registry [44
] originally developed in part to study the genetics of smoking behaviors and smoking-related disease [47
Dr. Allan C. Collins also emerged as a major researcher of the genetics of nicotine addiction at the IBG. Collins joined the faculty at the University of Colorado in 1972, and developed mouse models for studying the genetics of smoking behavior. Successful site visits prompted the CTR to praise Collins and support funding for his work. From 1978 to 1981, Collins received over $150,000 from the CTR for research on the genetics of smoking behavior in mice [48
By the 1980s the idea that smoking behaviors may have a genetic component had begun to enter the scientific mainstream. Tobacco industry funding of researchers like Crumpacker, McClearn, and Collins contributed to the development of this field. In 1983 Collins wrote a letter thanking CTR “for providing the funding that allowed us to get a solid start in this area,” and suggested that he would be able to obtain funding from the National Institute on Drug Abuse (NIDA) to continue the research [50
]. These researchers, differently motivated than industry leaders, sought elucidation and testing of hypotheses in behavioral genetics. However, any NIDA-funded research by Collins and others could not be done with planning, direction, and oversight by industry personnel. To this end, an R. J. Reynolds internal memo included recommendations that this work be re-funded directly by Reynolds [51
]. Between 1986 and 1994, and apparently independent of CTR, R. J. Reynolds’ extramural research funding to Collins was in excess of $700,000 to continue basic research on the genetics and neurobiology of nicotine response and addiction [52
In 1995, at the first annual meeting of the Society for Research on Nicotine and Tobacco, there was a symposium titled “Genetic Influences on Nicotine Dependence.” Both Collins and McClearn were featured in this symposium and have been honored for their groundbreaking research [57
]. An article published in Addiction
describes the meeting, including the presentation of Collins’ research on mouse strains and McClearn’s work from quantitative genetic theory, concluding that, “Evidence from several lines of research suggests that the use and effects of nicotine or tobacco smoking is influenced in part by genetic factors” [63
]. A legal representative for Brown & Williamson attended the conference and reported that the genetic work was “the most significant scientific session of the meeting” [64
Regardless of the scientific motivations or intent of the scientists involved, the research was funded by an industry that felt it had much to gain from work upholding a genetic basis of tobacco dependence. Though one could argue that all research stems from and is funded by “interested” parties, we suggest that tobacco industry funding is of particular concern because of the large body of research documenting the tobacco industry’s activities: publicly denying but privately accepting the addictive nature of nicotine [4
]; holding “judicial seminars” designed to influence court rulings [65
]; using scientific knowledge to create products that were increasingly addictive [66
]; and paying scientists to speak publicly in an attempt to keep controversy alive regarding passive smoking (secondhand smoke) [67
]. That tobacco industry lawyers have both directed much of the industry-sponsored research and stifled “unfavorable” studies speaks to the extreme nature of the bias in this body of research.
Our 21st century perspective is, of course, informed by these revelations; but we must use extreme caution in evaluating the actions of behavioral genetics researchers who accepted industry funding in the 1970s. It would be both unfair and inappropriate to judge their actions by today’s standards. It was only in the mid-1980s that suspicions about tobacco industry activities began to raise concerns among university-based researchers, who asked: Is accepting industry funding ethical [69
]? By 1995, condemnation of tobacco industry funding of research was widely voiced, including recommendations published in the Journal of the American Medical Association
]. Our historical account is not meant to suggest that pioneering behavioral genetics researchers were violating ethical norms when they sought funding from the tobacco industry. Explicit standards condemning such funding have only been established with the passage of time, and in an era of declining funding, whether to accept support from the tobacco industry remains a matter of considerable dispute.