Bat's Research Projects, 1960s–1980s

Project Janus

BAT did not rely solely on the UK industry-wide cooperative lab at Harrogate for its long-term toxicology program. In 1965 it set up a contract operation of its own, called Project Janus. Project Janus was winding down in 1977 {1164.23}, but continued at least into 1978. BAT contracted with Battelle-Institut Frankfurt am Main to conduct long-term mouse skin-painting experiments as part of BAT's product development program. This was another branch of the same laboratory that had conducted the nicotine projects Hippo I and Hippo II (discussed in chapter 3) a few years earlier. Battelle hired a staff of between twenty and

twenty-seven people for Project Janus and built a special building to house it {1138.03}. Battelle undertook a wide variety of studies under Project Janus. The core activity was mouse skin painting, and different ideas were tested to examine the toxicity of specific cigarette components and of different proposed modifications. In addition, Battelle developed a variety of short-term biological tests under this project.

The selection of projects initially seems to have been decided by the priorities of the BAT staff in the United Kingdom. Early on, before the project had even started, Dr. R. B. Griffith of B&W expressed concern that the initial testing was entirely related to British-type 100 percent flue-cured cigarettes and ignored the blended cigarette typical of the US market {1105.01}. B&W did, however, contribute materials for a variety of Janus experiments. The preparation of samples for testing at Harrogate and at Frankfurt consumed a good bit of time at the Southampton lab in the late 1960s {1138.01, p. 3}, but B&W was also involved in this work {1143.01}.

A glimpse at how Southampton contributed to Janus comes from an undated, two-page fragment of an annual report from about the late 1960s to the early 1970s {1138.01, pp. 3–4}. The lab's work is categorized into "(a) the reduction in biological activity, and (b) the understanding of consumer acceptance." The report indicates,

Additive-treated tobacco and sheet materials show considerable merit biologically, indicating that tobacco can be altered advantageously. It would appear that the tobacco type and the processing it receives may be the major factors. ... The formulation of non-tobacco materials has also been undertaken.

The investigation of additives to reduce aromatic polycyclic hydrocarbons has continued. Several have been found to be effective, and two have sufficient commercial potential to warrant detailed study regarding levels and methods of addition [emphasis added]. {1138.01, p. 3}

The preparation of samples for Project Janus testing at the Battelle facility in Frankfurt and at the Tobacco Research Committee (TRC) lab at Harrogate (TRC was the successor to TMSC) consumed a substantial amount of effort.

The results of a study of the toxicity of inhaled tobacco smoke at the Battelle laboratory are discussed in a 1972 report {1152.01}. Groups of male mice were exposed to the smoke of one of two cigarette types; their fertility then was compared to that of a group of mice that had not been exposed to smoke and to each other. Sexual activity was less in the experimental groups than in the controls, and the controls had a higher fertility rate than the smoke-exposed groups.

In 1967 Janus testing had established that the additive potassium carbonate, even though its use resulted in lower benzpyrene levels, was associated with higher rates of tumor formation in mice {1109.01, p. 2}.

Part of the Project Janus research plan in 1968 was to test the effects of different levels of cigarette ventilation on condensate carcinogenic activity and to test processed cigarette leaf (PCL) as a substrate {1112.01}.

In an overview of Janus mouse skin–painting experiments prepared in April 1971 {1138.04; 1163.13}, E. B. Wilkes summarized results from six different experimental test preparations. The rates of tumor production at the lowest of the three doses tested for each preparation are presented here in order of the size of the effect {1138.04, pp. 5–6}:

By late 1971, however, BAT was expanding its options for toxicology testing. A progress report on Project Janus for 1971, written by a BAT scientist, notes that BAT expects to set up its own laboratory facilities within a year {1163.10}. Moreover, this new lab would concentrate on inhalation studies. The new inhalation work would feature precise control and measurement of the retained dose from different cigarettes. Nevertheless, the author of this report continues to expect that a role for mouse skin–painting experiments will remain.

[I]t is anticipated that mouse-skin painting experiments will be continued at Battelle and[,] although the demands are reducing[,] the next long-term test has already been planned. Future requirements are somewhat less certain but it is likely that such tests will be required for the examination of new technical developments in the reconstitution process and the incorporation of different materials or additives in the process. {1163.10, p. 10}

A Project Janus experiment on the effect of different puff volumes on tumorigenicity was reported by the Battelle lab to BAT in 1973 {1138.02}. At issue was whether cigarette smoke condensates taken at

puff volumes of 10, 25, and 50 ml were equally potent when the same amounts (50 mg) were applied to mice in the standard manner. While the eventual rate of tumor formation was similar in all three groups, there was a clear dose-response relationship in the time of appearance of tumors. The 10-ml puff volume was associated with the most rapid appearance of tumors, the 25-ml puff volume was intermediate, and the 50-ml puff volume was associated with the most gradual onset of the appearance of tumors. In this experiment, the carcinogens in the condensates seemed to be in higher concentration in the tar from the smaller puffs than the larger puffs. Combustion may have been more complete in the later stages of the larger puffs, resulting in a dilution of carcinogens from the first part of the puff. The report describes the results as showing a dose-response relationship. The presence of a dose-response relationship is very strong scientific evidence that the material being tested, in fact, causes cancer.

Janus, which continued for more than a dozen years, made major contributions to BAT's understanding of safe cigarette strategies. It utilized long-term and short-term assays to help the R&D staff make decisions about product design. Even though the work led at best to only marginal improvements in cigarettes, the goal of the work was always clear: to help make cigarettes less toxic. To achieve this goal, BAT was willing to support a large contract research operation in Frankfurt from 1965 until at least 1978.

Project Rio

A review of BAT's biological testing program was held at Southampton in May 1983 {1164.26}. The notes on this meeting contain the first reference in the documents to Project Rio, a project designed to organize the company's research on cigarettes having reduced biological activity. Additional information about Project Rio is found in the report of the Rio de Janeiro research conference held in August 1983 {1180.07}. (Chapter 7 discusses the concern that B&W lawyers had about the possible discovery of Project Rio by lawyers for plaintiffs in products liability lawsuits.)

Three distinct components were recognized in Project Rio:

The work was to be done through the laboratories in Hamburg, Montreal, and Southampton. The first two parts of the project sound very similar to the things that the R&D group had been involved with over the past generation. The third, though, was something of a departure: the use of additives to reduce biological activity. The document mentions one such possibility: Vitamin A.

The current status of Vitamin A as an anti-cancer agent should be reviewed in the context of the possible addition of Vitamin A (or some derivative) to tobacco. {1164.26, p. 2}

Such a proposal would make sense only if one believed that tobacco causes cancer. Compounds related to Vitamin A found in food had been related to reduced cancer risk, and this information was coming to public attention around this time. However, the use of such cancer-reducing materials as additives "might pose problems similar to those faced by the pharmaceutical industry" {1164.26, p. 1}. The reference here appears to be to FDA regulation.

Toward the end of the 1980s, a B&W marketing executive, Douglas Keeney, left the company to start a new company, CA Blockers. CA Blockers was set up to market a cigarette that used an additive, N-Bloctin, which promised to reduce the consumer's exposure to nitrosamines. The product, Spectra, was regulated as a drug by the FDA because of the implied health claim (2). The FDA regarded N-Bloctin as a drug, since it was intended to reduce the absorption of carcinogens from tobacco smoke, and the agency declared that any product containing N-Bloctin was also a drug. Therefore, it exerted jurisdiction.

By August 1983 the R&D scientists at BAT had decided against actively pursuing Vitamin A as an additive. The decision was based on the results of a literature review, which led to the conclusion "that Vitamin A (and closely related compounds) does not present an opportunity for the cigarette industry directly to influence human response to smoke" {1180.07, p. 4}

A fear expressed during the discussions about Project Rio in the initial 1983 concept paper was that competitors might publish a ranking of cigarette brands according to the results of mutagenicity tests such as the Ames test. (The Ames test uses bacteria that because of a defective gene are unable to make a particular nutrient that is lacking in a specially prepared growth medium. A mutation at this gene locus can permit the bacteria to grow in the deficient medium. The relative mutagenicity of a test material—such as cigarette tar—can be estimated by the rate at which the material induces mutations that permit the bacteria to grow.)

The possibility of competitors producing a biological ranking of brands e.g. based on Ames test data is real and we should be in a position to respond to such a situation. {1164.26, p. 2}

It is interesting that this threat was seen as coming from other tobacco companies and not from a source such as the Reader's Digest , which had previously caused such problems with its publication of tar and nicotine yields. Clearly, in a country such as the United States, the only competing company that would publish such a list would be one whose product had the least Ames mutagenic activity. In other words, the competitor would have to be a start-up company such as CA Blockers. Nonetheless, the concern that someone else might develop a brand-specific table of mutagenic activity led to a plan to evaluate the Ames test mutagenicity of company brands and those of competitors in selected countries from around the world {1180.07, pp. 3–4}. This work would actually have enabled BAT to publish a biological index table. The work was to be completed on this initial testing by mid-1984.

Project Rio was to be the company's major biological research activity for the 1984–86 period {1180.07, p. 18}. While the work would concentrate on the Ames test, the minutes reflect an attitude of caution about "over-dependence" on this single measure; "more direct tests" were preferred if possible.

Preliminary results from Project Rio are summarized in the minutes of the biological conference held at Southampton in April 1984 {1181.06}.

The Ames is the main screening assay and from the results to date it is clear that:

Clearly we need tests in addition to the Ames test and Southampton is obtaining encouraging results with an enzyme induction assay. There was general agreement that the work should be actively pursued. Similarly it was agreed that the "yeast system" showed promise as a useful assay and should be followed up.

The eventual need for a long-term bioassay, preferably based on inhalation, was considered. The significance of such a test to the tobacco industry was discussed, particularly if the animals used were pre-treated with a known initiator such as radon. No firm conclusions were reached but it was agreed that Southampton should explore the feasibility of a long-term assay [italic emphasis added]. {1181.06, pp. 1–2}

BAT scientists had been able to stratify cigarette brands according to Ames activity, but not with mouse skin painting. The constituents of cigarette smoke responsible for the Ames test activity seemed to be different from those associated most strongly with carcinogenesis in the mouse skin–painting model. This result suggests that these were additional constituents of tobacco smoke with negative health effects.

A joint R&D/marketing conference held in Montreal in July 1984 included a report on Project Rio from the Montreal lab {1226.01, p. 85}. The report presents data showing that changes in smoker behavior (compensation effects) could alter Ames test results.

Traditionally, mouse skin painting has been used as an indicator of biological activity. Recently, short term tests such as the Ames mutagenicity test have

also been used to determine biological activity. With the increase [sic ] use of short term tests, there has been, within the BAT group, discussions focussing on the possibility of legislative bodies using the Ames or other short term tests to assess the "tar quality" or arriving at biological league tables. Project RIO is an example of BAT's response to a potential need for a reduced biologically active product.

...

In all cases, the biological activity of the human generated smoke condensate was compared to standard machine smoking condensate.

...

These results appear to indicate the following:

Further studies will revisit this area and investigate in more detail:

The study compared human and machine smoking of two Imperial Tobacco Ltd (Canada) brands, Matinée Extra Mild and duMaurier Light King Size. Matinée had a greater level of "specific biologic activity under human smoking conditions"; that is, Matinée smoke had a higher level of activity per microgram of smoke. In contast, deMaurier had a higher level of activity than Matinée when the data were examined on a "total" (per cigarette) basis. Human smoking profiles had greater "specific" and "total" biological activity associated with them than did the machine standard {1126.01, from overheads used in the presentation}. These data seemed to confirm the study hypothesis that human smoking behavior differs from machine-based smoking enough to be of practical importance.

All in all, Project Rio seemed to leave unresolved problems that had plagued the search for a safe cigarette from the beginning. The Ames test offered a credible short-term assay, but it measured different things from the mouse skin test that had been the standard for decades. The focus at this late stage, three decades on, was still on developing a good assay, when the technological difficulties involved in reducing toxins to safe levels while still delivering a satisfying smoke remained formidable. The whole matter was further complicated by the fact that people do not smoke like machines.

The bleak prospects for making meaningful progress on this front are reflected in the five themes for R&D found in the minutes of a Group R&D meeting at Wallingford, England, in September 1985 {1182.01}.

Progress toward a safe cigarette had become the third objective, was regarded as a long-term objective, and had been diluted to the point that a reasonable goal was seen as merely offering management a "recognised step" in this direction.

Short-Term Biological Tests

In order to develop a cigarette that was less prone to cause cancer, BAT scientists needed an assay procedure to test out various possible product modifications. Mouse skin painting was BAT's standard test for carcinogens, but it required up to two years to get the results of a single experiment. This built-in delay and expense made rapid progress impossible. Accordingly, Sir Charles Ellis suggested that BAT develop a battery of short-term tests for use in-house. In January 1964, coinciding with the publication of the US Surgeon General's report, Sir Charles wrote Richard P. Dobson at BAT headquarters, recommending that BAT develop a set of short-term toxicological tests for the use of B&W {1103.02}.

Sir Charles suggested that Battelle, the contract research organization that did other work for BAT on both nicotine pharmacology (Project Hippo) and carcinogenicity of tobacco smoke (Project Janus), or some other organization be engaged to set up the ciliastasis test. He wanted to use it in assessing cigarette brands selected by B&W. In addition, he wanted to develop two other tests from among the five suggested, or to develop other sorts of short-term bioassays. The proposed testing program seems to have been mainly for the benefit of B&W. It is unclear why the work was to be done at Battelle rather than in-house, except that Battelle was shortly to do the company's mouse skin–painting work to

evaluate carcinogenic potency of cigarette smoke under the Janus project. Having the toxicology work done in an overseas, outside laboratory also provided B&W a buffer against discovery in legal proceedings (see also the section headed "Specific Strategies to Avoid Discovery" in chapter 7).

In his letter Sir Charles speculates that having such a battery of tests available would be of similar importance to conventional quality control:

If this project were to prove feasible it might lay the basis for a continuing health monitoring service on our cigarettes analogous to the Quality Control practised for physical characteristics. The work would be quite distinct from the researches into the relation between smoking and lung cancer which is carried out on a co-operative basis by T.R.C. [at Harrogate]. {1103.02, p. 3}

A week later, in Louisville, Tom Wade of R&D at B&W wrote an analysis of this memorandum for Ed Finch, B&W's president {1103.01}. A copy was hand-carried to Mr. Dobson of BAT when he was in New York on the following day. Wade endorses the concept of rapid tests without commenting on the technical feasibility of the specific proposals made by Sir Charles. Wade also blurs the distinction between tests to look for carcinogenic potential and tests for immediate toxicity. He concludes, "Naturally the whole purpose back of this is to get a reasonably rapid method to determine differences" {1103.01}.

The documents do not include further information on this attempt to establish short-term biological testing within BAT's research establishment. Nearly three years later, though, in October 1966, an internal progress report on a research project at the Southampton lab demonstrated that there was by then active work on the development of shortterm assays. In this case the goal was the development of a short-term test that would predict carcinogenic activity {1107.01}.

By November 1968, as a memo from Dr. R. A. Sanford, technical manager of research at B&W, to J. W. Burgard (with copies to the company president, general counsel, and the director of research and development) indicates, the contract lab operated by Battelle in Frankfurt, Germany, had achieved some success with a short-term test that measured hyperplasia in mouse skin as a predictor of malignant transformation {1111.01}. The test required only eight days to perform. It measured thickening of the skin on the backs of mice exposed to the test materials. This test had shown agreement with the mouse skin–painting test in most of the following situations:

The addition of certain PCL's, Celanese synthetic materials, or certain additives to the blend reduces biological activity rating.

Filters containing charcoal, cellulose acetate, PEI (polyehtyleneimine), or paper do not affect tumorigenic results.

With the exception of one result, increasing puff volume progressively from 10 to 50 ml reduces activity.

U. S. cigarettes are less active [i.e., less capable of inducing cancerous changes in the mice] than English varieties. The addition of either burely or up to 50% CRS [cut, rolled stems] to flue-cured tobacco reduces activity. {1111.01, p. 1}

Dr. Sanford suggested that some experimental versions of Viceroy be submitted for testing in this system and that it be used in connection with an upcoming project called Project Hilton {1111.01}. Project Hilton was to be an inhalation study under the management of Battelle, with the goal of looking at the short-term toxicity of tobacco smoke in animals.

By the middle of December, Dr. Sanford had received clearance to send samples to Europe for the hyperplasia test {1112.04}. He sent current versions of Kool and Viceroy, the same brands with 28 percent WTS (water-treated system), and Life filters and Life filters without PEI (polyethyleneimine) but with a Viceroy tobacco rod. The mention of Life filters without PEI suggests that the normal filter for Life contained this additive, an ingredient that was known to increase the proportion of "extractable nicotine" in cigarette smoke {1205.03}. It had not, however, reduced tumorigenic activity of smoke condensate {1111.01}. By 1968, then, B&W had at its disposal a short-term bioassay that correlated reasonably well with the standard test for carcinogenesis, mouse skin painting.

Notes on the 1974 Duck Key, Florida, research conference refer tersely to a short-term bioassay test that the BAT lab in Germany had developed.

BAT (Germany) NMFI test is proving of significant value in rapid prediction of mouse skin activity. We propose not to make this test available to competitors at this time since it might be of considerable commercial advantage [emphasis in original]. {1125.01, p. 2}

This assay was the subject of a research report from the BAT affiliate in Canada three years later. A 1977 research report from Imperial Tobacco Limited (Canada) details the possible use of a test called "the nitromethane fraction index (NMFI)" as an indicator of "biological activity" {1129.01}. The NMFI test was designed to be an inexpensive and quick test that predicted carcinogenicity in mouse skin. The test involved the extraction of cigarette smoke in a way known to concentrate polycyclic aromatic hydrocarbons; this smoke fraction was then mixed with egg albumin, and the resulting degree of binding by fluorescence

was determined. The degree of binding in turn showed a correlation with the biological activity of the smoke.

The test had been developed at the BAT lab in Germany and also was in use at Southampton. The test was found to be reproducible and to correlate well with the results of mouse skin–painting carcinogenicity tests conducted under the Janus program. Despite some reservations, the report recommends the routine inclusion of the NFMI test among the tests done "in future projects where there is a possibility of a change in the biological activity of the smoke" {1129.01, p. 4}. The report emphasizes that the NMFI test is for internal use only; it is not to be shared with other companies. In this context, the term "biological activity" is used as a euphemism for "carcinogenic activity." The development of this test is one more example of how the company knew its products caused cancer, and tried to do something about it, while publicly denying this fact.