JOURNAL TRANSCRIPT
Regulatory Forum Toxicologic Pathology, 39: 716-744, 2011 Copyright # 2011 by The Author(s) ISSN: 0192-6233 print / 1533-1601 online DOI: 10.1177/0192623311406935
An Analysis of Pharmaceutical Experience with Decades of Rat Carcinogenicity Testing: Support for a Proposal to Modify Current Regulatory Guidelines FRANK D. SISTARE1, DANIEL MORTON2, CARL ALDEN3, JOEL CHRISTENSEN1, DOUGLAS KELLER4, SANDRA DE JONGHE5, RICHARD D. STORER1, M. VIJAYARAJ REDDY1, ANDREW KRAYNAK1, BRUCE TRELA6, JEAN-GUY BIENVENU7, SIVERT BJURSTRO¨M8, VANESSA BOSMANS5, DAVID BREWSTER9, KARYN COLMAN10, MARK DOMINICK11, JOHN EVANS8, JAMES R. HAILEY12, LEWIS KINTER8*, MATT LIU1, CHARLES MAHRT13, DIRK MARIEN5, JAMES MYER12, RICHARD PERRY2, DANIEL POTENTA10, ARTHUR ROTH2, PHILIP SHERRATT1, THOMAS SINGER9þ, RABIH SLIM9, KEITH SOPER1, RONNY FRANSSON-STEEN8, JAMES STOLTZ14, OLIVER TURNER10, SUSAN TURNQUIST2, MARJOLEIN VAN HEERDEN5, JOCHEN WOICKE11, AND JOSEPH J. DEGEORGE1 1
Merck and Co., Inc., West Point, Pennsylvania, USA 2 Pfizer, Groton, Connecticut, USA 3 Millenium, Cambridge, Massachusetts, USA 4 Sanofi Aventis, Malvern, Pennsylvania, USA 5 Janssen Pharmaceutica NV, Beerse, Belgium 6 Abbott Laboratories, Abbott Park, Illinois, USA 7 Lab Research, Laval, Que´bec, Canada 8 AstraZeneca R & D, So¨derta¨lje, Sweden, or *Wilmington, Delaware, USA 9 Hoffman-La Roche, Nutley, New Jersey, USA, or þBasel, Switzerland 10 Novartis Pharmaceutical Corporation, East Hanover, New Jersey, USA 11 Bristol-Myers Squibb, Mt. Vernon, Indiana, USA 12 GlaxoSmithKline, Research Triangle Park, North Carolina, USA 13 Eli Lilly & Co., Inc., Indianapolis, Indiana, USA 14 Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut, USA ABSTRACT Data collected from 182 marketed and nonmarketed pharmaceuticals demonstrate that there is little value gained in conducting a rat two-year carcinogenicity study for compounds that lack: (1) histopathologic risk factors for rat neoplasia in chronic toxicology studies, (2) evidence of hormonal perturbation, and (3) positive genetic toxicology results. Using a single positive result among these three criteria as a test for outcome in the two-year study, fifty-two of sixty-six rat tumorigens were correctly identified, yielding 79% test sensitivity. When all three criteria were negative, sixty-two of seventy-six pharmaceuticals (82%) were correctly predicted to be rat noncarcinogens. The fourteen rat false negatives had two-year study findings of questionable human relevance. Applying these criteria to eighty-six additional chemicals identified by the International Agency for Research on Cancer as likely human carcinogens and to drugs withdrawn from the market for carcinogenicity concerns confirmed their sensitivity for predicting rat carcinogenicity outcome. These analyses support a proposal to refine regulatory criteria for conducting a two-year rat study to be based on assessment of histopathologic findings from a rat six-month study, evidence of hormonal perturbation, genetic toxicology results, and the findings of a six-month transgenic mouse carcinogenicity study. This proposed decision paradigm has the potential to eliminate over 40% of rat twoyear testing on new pharmaceuticals without compromise to patient safety. Keywords:
carcinogenicity testing; carcinogenicity prediction criteria; preclinical research and development; pharmaceutical database; rat neoplasia; hormonal perturbation; genetic toxicology testing; regulatory guidelines.
Address correspondence to: Frank D. Sistare, Safety Assessment, Merck and Co., Inc., WP 45-205, 770 Sumneytown Pike, West Point, PA 19486–0004, USA; e-mail:
[email protected]. The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The authors received no financial support for the research, authorship, and/or publication of this article. This is an opinion article submitted to the Regulatory Forum and does not constitute an official position of the Society of Toxicologic Pathology or the journal Toxicologic Pathology. The Regulatory Forum is designed to stimulate broad discussion of topics relevant to regulatory issues in toxicologic pathology. Readers of Toxicologic Pathology are encouraged to send their thoughts on these articles or ideas for new topics to
[email protected]. 716
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
INTRODUCTION The two-year rat carcinogenicity study is longer, more costly, and uses more animals than any other standard toxicology study required to support pharmaceutical development. Specific regulatory requirements for carcinogenicity assessments of new pharmaceuticals are described in International Conference on Harmonisation (ICH) guidance documents (ICH Guidelines M3, S1A, S1B, S1C, S2, and S6). These ICH agreements clarify, in general, that a two-year rat study plus either a two-year mouse or an alternative six-month transgenic mouse carcinogenicity study of small molecules are required for human pharmaceuticals that would be dosed for six months or longer, or in a frequent and intermittent manner. We reasoned that the absence of specific histopathologic findings that can be considered risk factors for rat neoplasia from six-month rat chronic toxicology studies together with negative findings in certain other, shorter-term studies may provide sufficient information to predict rat carcinogenicity outcome and exempt these compounds from the need for two-year rat carcinogenicity studies. Recently, Reddy et al. (2010) evaluated the predictivity of histopathologic findings considered risk factors for rat neoplasia (hypertrophy, hyperplasia, and foci of cellular alteration) observed microscopically in six- and twelve-month rat chronic toxicology studies, with tumor outcomes in rat twoyear carcinogenicity studies for eighty pharmaceuticals. In agreement with previous investigations (Allen et al. 2004; Jacobs 2005; Melnick et al. 1996), they concluded that the presence of such histopathologic risk factors was not a reliable predictor of tumor outcome in the corresponding tissue. However, Reddy et al. (2010) found that the absence of evidence of these histopathologic rat neoplastic risk factors in any tissue was highly predictive (88%) of a compound’s lack of carcinogenic potential in the rat. The authors proposed that the compounds with no histopathologic risk factors for rat neoplasia in any tissue, thus taking into consideration the whole-animal response in chronic rat studies, be considered rat noncarcinogens and be exempt from the requirement for testing in rat two-year carcinogenicity studies. Only the compounds with histopathologic neoplastic risk factors in any tissue would need to be tested in the two-year rat carcinogenicity study, and all tissues in these studies would be examined. Fifty percent of the two-year rat carcinogenicity studies in the sample of eighty compounds would have been eliminated based on this proposal, whereas five rat carcinogens of questionable human relevance did not produce histopathologic risk factors of rat neoplasia in chronic rat studies. Encouraged by these promising findings, we combined available data from thirteen major pharmaceutical companies
717
accumulated over three decades of carcinogenicity testing to examine whether it is warranted to propose modifications to carcinogenicity testing principles formulated in the 1960s and 1970s. It is well known that genotoxicity, assessed by the Ames assay and in vivo micronucleus tests, is a risk for tumorigenicity in both rodents and humans. It is well known that chronic hormonal perturbation is a risk for tumorigenesis in rats that may or may not always translate to humans. In this manuscript, the ‘‘whole-animal negative predictivity’’ hypothesis proposed by Reddy et al. (2010) was further evaluated with additional pharmaceutical compounds using an expanded collection now including those that have never been approved for marketing. The notion that histopathologic or hormonal perturbation evidence from shorter-term animal studies may predict tumor induction in longer-term studies is not a novel or contemporary concept (Muehlbock 1959; Neumann 1991; Ward 1980). Here we comprehensively and systematically assessed and confirmed the link between negative carcinogenicity outcome across 182 rat two-year carcinogenicity studies and the absence of genotoxicity, the absence of any knowledge or significant evidence of hormonal perturbation activity, and the absence of evidence of histopathologic risk factors of rat neoplasia in all tissues examined in the corresponding chronic rat toxicology study conducted at similarly matching doses. Based on these data, we now propose the implementation of practical changes to current drug development regulatory guidelines. Although we also recognize that immunosuppression represents a known risk factor for human carcinogenesis, rat carcinogenicity testing results with such agents are known not to reliably reflect this human risk (Bugelski et al. 2010), so we did not incorporate this category as a criterion to warrant rat twoyear testing. Furthermore, we evaluated how this decision paradigm performs when applied to all International Agency for Research on Cancer (IARC) known and probable human chemical carcinogens, as well as the ten pharmaceuticals that have been withdrawn from the market as a result of carcinogenicity concerns. We propose from these evaluations that compounds that are negative for genotoxicity, negative for hormonal perturbation activity, lacking specified histopathologic risk factors for rat neoplasia in chronic rat toxicology studies, and negative in an alternative six-month transgenic mouse carcinogenicity study need not require a two-year rat carcinogenicity study in the course of pharmaceutical development. The analyses support the contention that additional rat carcinogenicity testing for two years with such compounds would represent a needless use of animals and a waste of valuable time and resources, with no additional value brought to the assessment of human safety.
Abbreviations: CCK, cholecystokinin; CYP, cytochrome P450; EMA, European Medicines Agency; FDA, United States Food and Drug Administration; FN, false negative; FP, false positive; GI, gastrointestinal; GLP, Good Laboratory Practice; IARC, International Agency for Research on Cancer; ICH, International Conference on Harmonisation; JPMA, Japan Pharmaceutical Manufacturers Association; LH, luteinizing hormone; LHRH, luteinizing hormone releasing hormone; NEG CARC Rat, negative for endocrine, genotoxicity, and chronic study-associated histopathologic risk factors for carcinogenicity in the rat; NTP, National Toxicology Program; PDR, Physicians’ Desk Reference; PhRMA, Pharmaceutical Research and Manufacturers of America; PPAR, peroxisome proliferator-activated receptor; T3, triiodothyronine; T4, tetraiodothyronine; TN, true negative; ToxNet, National Library of Medicine Tox Net database; TP, true positive; TSH, thyroid-stimulating hormone.
718
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
TABLE 1.—Rules for defining valid pairs of rat chronic and two-year carcinogenicity studies based on comparable dose levels.a Study outcomes Rule no.
Chronic study dose levels versus carcinogenicity study dose levels
1
Chronic
Carco
Action
Justification or assumption
Chronic > carco
þ/
Accept
2
Chronic > carco
þb
þ/
Reject
3
Chronic > carco
þc
þ/
Accept
4
Chronic < carco
þ
þ/
Accept
5
Chronic < carco
þd /
Reject
6
Chronic < carco
þe
Accept
Negative histopathology at higher dose levels is expected to be negative at lower dose levels as well Positive histopathology seen only at higher dose levels may not be positive at lower, untested dose levels Accept matching dose data based on carco high dose levels being adequately tested in chronic Positive histopathology at lower dose levels is expected to be positive at higher dose levels as well Negative histopathology at lower carco matching dose levels may not be negative at higher dose levels, which introduces uncertainty Overlapping chronic dose levels considered to adequately test the positive matching carco dose levels
Abbreviations: carco, rat two-year carcinogenicity study; chronic, chronic rat toxicology study of six- or twelve-month duration; , negative; þ, positive. a All data with matching dose levels (+25%) accepted for evaluation. b Positive only at nonmatching dose(s). c Positive at lower overlapping dose(s). d Positive at nonmatching dose(s) only. e Positive at lower matching dose(s).
METHODS Pharmaceutical Company Data Sources The rat carcinogenicity and chronic toxicology data used for evaluation were deposited by each of thirteen participating companies onto a Web site database maintained by the Pharmaceutical Research and Manufacturers of America (PhRMA) (Appendix I). Individual companies were blinded to the identity of the compounds and the company sources of all but their own data. Only the PhRMA database administrator was unblinded to the sources of the data. This evaluation focused on the rat carcinogenicity studies and not on the mouse carcinogenicity studies, as chronic toxicity studies for pharmaceuticals are routinely conducted in rats, but not in mice. The judgment of whether there was a significant genetic toxicology test result, whether there were treatment-related histopathologic findings defined by this group as risk factors for rat neoplasia, or whether there was evidence for hormonal perturbation was ultimately dependent upon the sponsors’ reviews of their own archived data. The carcinogenicity studies in this database were initiated between 1972 and 2007, with 4% initiated before Good Laboratory Practices (GLP) study standards were mandated in 1978, and 66% were initiated after 1991, the year that the FDA initiated the Carcinogenicity Assessment Committee. The number of rats per sex per dose level for the six-month and twelvemonth studies ranged from ten to twenty-five. The number of rats per sex per dose level was generally fifty or sixty for the two-year carcinogenicity studies, with a range of fifty to seventy noted, often with two control groups per sex and either three or
four treatment dose levels. All tissues examined histologically in chronic rodent toxicology studies are examined in two-year carcinogenicity studies, and these extensive tissue lists have been widely used for decades across the pharmaceutical industry (Bregman et al. 2003). Dose-Matching Rules for Data Inclusion from Internal Pharmaceutical Company Study Archives As summarized in Table 1, six rules were applied for inclusion of studies, following the approach of Reddy et al. (2010). Studies were included in the evaluation when relevant dose levels in the chronic study and the carcinogenicity study did not differ by more than 25%. The value of +25% was chosen to allow for and accommodate slight differences in dose level selection, lot and batch number of test articles, and dose formulation differences between chronic toxicology and carcinogenicity studies. Compounds were also included for analysis where: (1) the chronic study was negative at a dose higher than those tested in the carcinogenicity study (rule 1); (2) chronic study doses were higher than doses used in carcinogenicity studies, but at least one dose in the chronic study was positive and also matched the high dose used in the carcinogenicity study (rule 3); (3) the chronic study was positive for the histopathologic signal, even if the relevant doses were less than 75% of the doses used in carcinogenicity studies (rule 4); and (4) the chronic study was negative and the top dose was less than 75% of the top dose used in carcinogenicity studies, but lower doses in the carcinogenicity study matched those in the chronic study and there were positive tumor findings (rule 6).
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
Compounds were excluded where: (1) the chronic study was positive for the histopathology signal only at doses more than 25% higher than the highest dose used in the carcinogenicity study (rule 2); or (2) the highest dose in the chronic study was less than 75% of the lowest dose in the carcinogenicity study and was negative for histopathologic risk factors for neoplasia (rule 5). Overall, these rules provided the most balanced and unbiased data set for analysis of the predictive value of the chronic studies. Compounds were screened by sponsors as nonmatching according to these rules and were either not submitted or were submitted for confirmation (twelve compounds) and subsequently adjudicated by panel members to be nonmatching. The critical importance of matching doses between chronic and carcinogenicity studies to avoid bias has been described previously (Reddy et al. 2010). The twelve compounds submitted with nonmatching doses appear in Appendix I as compounds 183–194. Rules for Compound Classifications Criteria for Positive Histopathologic Risk Factors for Rat Neoplasia: The compounds were scored by originating sponsors as positive for histopathologic risk factors for rat neoplasia when hyperplasia, cellular hypertrophy, or atypical cellular foci were reported as treatment-related findings in the original rat chronic study report. Treatment-related neoplasia in chronic rat studies was also considered positive. The compounds were scored as negative for histopathologic risk factors for rat neoplasia when these changes were absent or not considered increased by treatment. The specific diagnostic terms found in reports and tables also included ‘‘multinucleated cells,’’ ‘‘basophilia,’’ ‘‘basophilic foci,’’ ‘‘cellular enlargement,’’ ‘‘cytomegaly,’’ ‘‘cellular swelling,’’ ‘‘cellular alteration,’’ ‘‘dysplasia,’’ ‘‘eosinophilic foci,’’ ‘‘karyomegaly,’’ or ‘‘tumor.’’ It is stressed here that the terms ‘‘hyperplasia’’ and ‘‘hypertrophy’’ were operationally considered as histopathologic risk factors for rat neoplasia, because progression of either change to neoplasia cannot be presumed and must not be misconstrued to be diagnostic of preneoplasia. The specific diagnostic terms were context dependent with respect to their association with risk of rat neoplasia. Histologic alterations, for example, noted and not considered as histopathologic risk factors for rat neoplasia were vaginal metaplasia and myocardial hypertrophy. These alterations were excluded because vaginal metaplasia can be a cyclical change seen during estrus, and cardiac hypertrophy represents a functional response known not to progress to tumor. All available reports were collected from company archives and contained both approved and nonapproved drugs. The required data were extracted from the study reports or the companies’ summary records, which contained a summary of histopathology and carcinogenicity interpretations by the study pathologists. The sponsor was directed not to change the determination of the original study pathologist. If the original study pathologist called the findings ‘‘treatment related,’’ then they were considered treatment related and considered to be positive beginning at the dose originally considered the low effect dose
719
for that lesion. Individual sponsors reviewed the context of the specific diagnostic terms used in their original reports and determined whether the findings were described in such a manner to be considered hyperplasia, hypertrophy, atypical foci, or neoplasia. The collection of the histopathology data from original reports of the chronic and carcinogenicity studies was done by each company in parallel. The data were then tabulated. Then evaluations were made as to whether each case represented a rat true positive (TP), rat true negative (TN), rat false positive (FP), or rat false negative (FN) and explained in accordance with the established rules. For this analysis, the ‘‘wholeanimal approach’’ was used, as described previously (Reddy et al. 2010), wherein any microscopic histopathologic risk factor of rat neoplasia in any tissue was deemed a positive potential risk for a tumor at any site. In addition, the absence of all histopathologic risk factors of rat neoplasia from all sites was considered evidence for the lack of potential risk for a tumor at any site. Organ weight data were also collected and analyzed (data not shown), but treatment-related changes in organ weights as an additional risk factor for rat neoplasia were determined to be nonspecific and incapable of satisfactory discrimination of rat TP from rat FP, similar to that reported previously (Allen et al. 2004), and therefore this information was not included with the microscopic criteria. Criteria for a Positive Genetic Toxicology Test Result: Any clear single positive genetic toxicology result in the GLP standard battery of assays (bacterial mutation, in vitro mammalian cell mutation or clastogenicity, or in vivo clastogenicity) that was not otherwise explained as an irrelevant finding was deemed a cause for concern that would warrant carcinogenicity testing. A positive result seen only at doses associated with high levels of cytotoxicity in a chromosomal aberration study or mouse lymphoma assay, for example, would not be considered a cause for concern. Likewise, a positive chromosomal aberration test associated only with mitotic arrest ascertained from follow-up flow cytometric analysis or a positive micronucleus test resulting from clear stimulation of hematopoiesis would not be considered a signal of concern for genetic toxicity. These are hypothetical examples of positive findings for which sponsor follow-up could adequately mitigate concerns. Criteria for Positive Evidence of Hormonal Perturbation: The evaluations of hormonal perturbation were based on weight of evidence, including: (1) clear treatment-related microscopic changes observed in multiple endocrine tissues within a sex (e.g., atrophy of the uterus and ovaries); (2) treatment-related macroscopic changes in multiple endocrine organs within a sex (e.g., a clear treatment-related change in the weights of both the adrenals and pituitary gland) at doses that were not associated with overall diminished animal weights; (3) measurements of significant treatment-related changes in hormones; (4) knowledge of the pharmacologic mechanisms of action including hormonal receptor binding, alteration of hormonal levels, or alteration of the activity of endogenous hormones. It is again stressed here that a weight-of-evidence approach is taken to
720
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
TABLE 2.—Prediction of rat tumor outcome based on rat chronic toxicology study histopathology risk factors, genotoxicity, and evidence of hormonal disruption. All 182 compounds: six- or twelve-month chronic study data Chronic tox, genotox, or hormonal disruption Positive Negative Carcinogenicity Positive 52 14 Negative 54 62
% Sensitivity % Negative predictivity % Elimination of studies
79 82 42
150 compounds: six-month chronic study data Chronic tox, genotox, or hormonal disruption Positive Negative Carcinogenicity Positive 45 12 Negative 44 49
% Sensitivity % Negative predictivity % Elimination of studies
79 80 41
Eighty-five compounds: twelve-month chronic study data Chronic tox, genotox, or hormonal disruption Positive Negative Carcinogenicity Positive 23 5 Negative 22 35
% Sensitivity % Negative predictivity % Elimination of studies
82 88 47
Fifty-three compounds: six- and twelve-month chronic study data for the same compounds Six-month chronic tox, genotox, or hormonal disruption Positive Negative Carcinogenicity Positive 14 5 Negative 13 21
% Sensitivity % Negative predictivity % Elimination of studies
74 81 49
Positive Negative
Twelve-month chronic tox, genotox, or hormonal disruption Positive Negative 16 3 12 22
% Sensitivity % Negative predictivity % Elimination of studies
84 88 47
Positive Negative
Test Prediction Positive a c
% Sensitivity % Negative predictivity % Elimination of studies
a / (aþb) d / (bþd) (bþd) / (aþbþcþd)
Carcinogenicity Definitions:
True outcome
Negative b d
assess hormonal perturbation. In vitro receptor binding assay data alone for example would not be sufficient and must be weighed in combination with functional in vitro assay results, in vivo exposure data, and ultimately the results of specific and definitive in vivo tests triggered to determine whether a compound would exhibit endocrine perturbation activity under study conditions. For example, hormone levels may have been measured in the chronic rat toxicology studies or in targeted mechanistic or investigative studies. Therefore, known on-target pharmacologic effects expected to yield hormonal perturbations as well as unanticipated treatment-related perturbations were both captured using these criteria. Criteria for Classifying a Compound as Positive for Carcinogenicity Outcome: Any treatment-related neoplastic finding was deemed a positive carcinogenicity study outcome. The interpretations of treatment-related carcinogenicity findings in rat studies were based on evaluations by the study pathologists unless regulatory agencies disagreed. If there was a difference of opinion between the agency and the sponsor, as was noted for three compounds, the most conservative opinion supporting the interpretation of a positive carcinogenicity outcome was given precedence and entered into the database. The alternative position of the sponsor or the agency is captured in the notes (see Appendix 1, for compounds 6, 51, and 53).
Compounds that were positive for two-year rat carcinogenicity study tumor findings and either histopathologic risk factors for rat neoplasia in chronic rat toxicology studies, evidence of perturbation of hormonal activity, or genetic toxicology test results were considered rat TP. Compounds that were negative for two-year rat carcinogenicity study tumor findings but positive for either histopathologic risk factors for rat neoplasia in chronic rat toxicology studies, evidence of perturbation of hormonal activity, or genetic toxicology test results were considered rat FP. Compounds that were positive for two-year rat carcinogenicity study tumor findings but negative for histopathologic risk factors for rat neoplasia in chronic rat toxicology studies, evidence of perturbation of hormonal activity, and genetic toxicology test results were considered rat FN. Compounds that were negative for carcinogenicity and negative for histopathologic risk factors for rat neoplasia in chronic rat toxicology studies, evidence for perturbation of hormonal activity, and genetic toxicology test results were considered rat TN. When assessing the contribution of histopathologic risk factors of potential rat neoplasia seen in the chronic rat study for predicting tumor outcome in the matching two-year carcinogenicity study in this set of 182 compounds, if a discrepancy was seen between the six- and twelve-month chronic study, the six-month study histopathology findings were relied upon to generate the data
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
summarized in Table 2. This criterion was used because the goal is to assess current practice as defined by regulatory guidelines, which require rat chronic toxicology studies for six months rather than for twelve months. There was an exception for one compound, specifically compound 27, as the sponsor included data from a 6-month chronic and a 2year carcinogenicity study for the mouse since no rat studies were conducted for this biologic agent. These data from compound 27 have been compiled in these analyses with all other data from studies conducted in the rat. IARC-Classified Chemical Carcinogens and Pharmaceuticals Withdrawn from the Market Compounds categorized by the IARC as carcinogenic to humans (Group 1) or probably carcinogenic to humans (Group 2A) and pharmaceuticals withdrawn from the market because of concern for human tumor risk were evaluated against the proposed paradigm. Positive results in a two-year rat carcinogenicity study were predicted if a compound was either positive in at least one standard battery test for genotoxicity, had evidence of hormonal activity, or caused a histopathologic risk factor for rat neoplasia in a chronic rat toxicology study. One hundred eight chemicals, chemical mixtures, infectious agents, metals, minerals, exposure circumstances, and ionizing radiation sources are currently classified as IARC Group 1, and sixty-three are classified as IARC Group 2A. Infectious agents, metals, minerals, ionizing radiation, and in most cases, chemical mixtures and exposure circumstances were excluded from this analysis because they are not representative of pharmaceutical compounds. Ten previously marketed pharmaceuticals have been withdrawn because of heightened concerns for human carcinogenicity (United Nations 2005; Wysowski and Swartz 2005; Federal Register 1998), two of which—namely, diethylstilbestrol and phenacetin—were also evaluated by IARC and classified as Group 1 human carcinogens. Thus, thirty-five IARC Group 1 chemicals, forty-three IARC Group 2A chemicals, and eight additional pharmaceuticals withdrawn from the market were evaluated. In most cases, sufficient information was available in published IARC monographs to judge the effectiveness of the proposed criteria. In other cases, the National Toxicology Program Database Search Application (NTP), the United States National Library of Medicine ToxNet (ToxNet) database, and results of literature searches were consulted to obtain more information. Genotoxicity was assessed as clear evidence of positivity in at least one of the standard battery assays using a standard protocol (bacterial mutation, in vitro mammalian cell mutation or clastogenicity, or in vivo clastogenicity). For compounds for which multiple discordant test results were available, reduced weight was given to positive results from older studies and from studies conducted with less-frequently used cell lines or assay systems. A conservative judgment was then made as to the likelihood of a positive result in the modern test battery. For example, 4-chloro-o-toluidine is treated in this exercise as potentially negative in the standard battery. In the absence of clear evidence of genotoxicity or hormonal activity, data from chronic rat
721
studies, preferably six months in duration, were sought and evaluated according to the dose-matching and histopathologic criteria above. In one case, trichloroethylene, the only data available were from a three-month-duration study, but these were considered since the reported histologic effects were present at the end of a two-year study as well. RESULTS Data Analysis Summary of the 182 Compound PhRMA Database A collection of 182 compounds was identified by member companies with both two-year carcinogenicity and chronic studies in rats conducted according to current quality standards using doses that matched sufficiently to meet the criteria described in Table 1. Of the 182 compounds, 118 were marketed pharmaceuticals and sixty-four were not. Among the 182 compounds, 150 were tested in chronic studies of six-month duration and 85 were tested in chronic studies of twelve-month duration. For 53 compounds, both sixand twelve-month study data were available. Among the 182 compounds included in the full tabulation of six- or twelve-month chronic studies with matching carcinogenicity studies (Table 2), positive carcinogenicity findings were seen with 66 compounds, whereas 116 compounds had no treatment-related neoplastic findings. Of these 66 compounds, 44 were marketed pharmaceuticals, 16 were discontinued from development, and 6 compounds were still in development despite the positive rat carcinogenicity findings. For the same set of 182 compounds, the combination of the three test criteria of histopathologic risk factors for rat neoplasia in chronic studies, evidence of hormonal disruption, or genotoxicity were positive for 106 compounds and negative for 76 compounds. The percentage of the 182 compounds that tested negative for all three negative carcinogenicity prediction criteria was 42%. As to test sensitivity, the percentage of compounds with a two-year rat positive carcinogenicity finding predicted correctly by a positive finding in any one of the three criteria was 79%. Negative predictivity, the percentage of the compounds lacking a finding in all of the three criteria and that were also negative for carcinogenicity findings in the two-year rat study, was 82%. From a comparison of the 150 compounds supported only by six-month studies with the 85 compounds supported only by twelve-month studies (Table 2), it is evident that there is little value gained in test sensitivity by running a chronic study for twelve months (82% sensitivity) rather than six months (79% sensitivity). Furthermore, analyses of the smaller set of histopathology data for the 53 compounds that had been tested in both six- and twelve-month studies (Table 2) provide further evidence that there is little value gained in test sensitivity by running a chronic study for twelve months (84% sensitivity) rather than six months (74% sensitivity), with the difference being liver hypertrophy (compound 62) or basophilic foci (compound 54) being seen for 2 compounds at twelve months but not at six months. For those 2 compounds,
722
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
TABLE 3.—Relationships between rat carcinogenicity study outcome at two years, histopathologic risk factors for rat neoplasia in a chronic rat study, genetic toxicology test results, and evidence of hormonal disturbance. Carcinogenicity
Histopathologic risk factors
Genotoxicity
Hormonal action
No. of compounds
TP
þ þ þ þ þ þ þ
þ þ þ þ
þ þ þ n/a
þ þ þ n/a Total
4 18 19 1 2 7 1a 52
FP
þ þ þ þ þ
þ þ þ n/a
þ þ þ Total
1 8 6 22 8 8 1b 54
Abbreviations: FP, false positive; TP, true positive. a For compound 52, no data were available to evaluate genotoxicity or hormonal evidence. b For compound 104, no data were available to evaluate genotoxicity.
Table 4a.—Tissues with neoplasms in carcinogenicity studies and histopathologic risk factors for rat neoplasia at a matching site in chronic studies. For thirty-three of forty-two rat true positive compounds, histopathologic risk factors for rat neoplasia were noted in the same tissue site in the rat chronic toxicology study as a tumor site appearing in the rat two-year carcinogenicity study. Tumors with matching hypertrophy, hyperplasia, or altered foci
Other tumors with matching histopathologic risk factors
Other tumors with no matching histopathologic risk factors
Other sites with histopathologic risk factors in chronic study
Liver
Thyroid (2)b
Uterus (1), thyroid (4), testes (4), urinary bladder (1), cervix (1), vagina (1), ovaries (1), fat (1), kidneys (1)
Pituitary (1), thyroid (3)
15
Thyroid Ovaries Adrenal Mammary Testes
Liver (2) Adrenal (1) Ovaries (1), mammary (1) Adrenal (1)
Ovaries (1), kidney (1)
5 4 3 2 1
Bone Skin Adipose Urinary bladder Systemic Stomach Nasal a b
Kidney (1), liver (1), mammary (1) Liver (1)
Adrenal, cervix, heart, urinary bladder, uterus, mammary, liver, lung, lymph node, pancreas, thyroid Urinary bladder
Mammary
Kidney, liver
Ovaries
Number of compoundsa
1 1 1 1 1 1 1
Thirty-three different compounds are represented in this table. (No.) indicates the number of compounds with coincident tumors and/or histopathologic risk factors.
ten animals/sex/group were examined at six months, whereas twenty were examined at twelve months. It remains unclear whether microscopic examinations of tissues from twenty rather than ten animals treated with these 2 compounds for six months would have enabled detection of histologic risk factors, or whether a twelve-month study duration would have been required.
Fifty-two Rat TP Compounds: As shown in Table 3, among the sixty-six compounds that tested positive for tumors in the twoyear rat studies, the need for completion of such a test would have been triggered for fifty-two of the compounds based on either histopathologic risk factors for rat neoplasia (forty-two compounds), positive results in the genetic toxicology testing battery (seven compounds), or evidence of treatment-related
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
Table 4b.—Tumors in two-year rat carcinogenicity studies associated with histopathologic risk factors for rat neoplasia seen in six-month rat chronic studies at non-tumor-bearing site(s).
ID
Tumor site
Site of hypertrophy, hyperplasia, or altered foci
10 14 20 21 24 32 34 41 52
Urinary bladder Urinary bladder, adipose Thyroid Adipose Liver, thyroid Pancreas Mesovarium ligament, pituitary Kidney Liver, cervix, stomach, systemic, vagina
Liver Mammary Liver, adrenal Pituitary Stomach Liver Bone marrow Bone marrow Pituitary
hormonal perturbations (twenty-six compounds). On a whole-animal basis, microscopic findings alone again proved to be a significant contributor to the prediction of neoplasia development with an overall sensitivity of 64% (forty-two of sixty-six). As reported previously (Reddy et al. 2010), the sensitivity of microscopic findings for predicting rat neoplasia on an organby-organ basis was lower than on a whole-animal basis. For the forty-two compounds classified as rat TP based on microscopic histopathologic risk factors for rat neoplasia in the chronic study, the relationship between the site of histopathologic risk factors and the tumor site evident in the two-year rat study was investigated. For thirty-three of the forty-two (79%) compounds, a tumor site appeared in the same tissue in which a histopathologic risk factor had been observed in the chronic study and often also at other sites that did not present with histopathologic risk factors in the chronic studies (Table 4a). For nine of the forty-two compounds, the tumor site seen at two years did not match any of the tissues presenting with histopathologic risk factors in the chronic study (Table 4b). Figure 1 shows that eleven tissues served as sentinels in six- and twelve-month studies for almost 90% of the tumors from these forty-two rat TP compounds, specifically the liver, thyroid, adrenals, ovaries, mammary gland, bone, pituitary, urinary bladder, kidney, skin, and stomach. Among the tumorbearing tissues, nine tissues accounted for over 80% of the treatment-related tumors seen, specifically the liver, thyroid, ovaries, testes, urinary bladder, skin, mammary gland, kidneys, and adrenals. These findings are similar but not identical to results from previously published analyses (Contrera et al 1997). Table 3 shows that twenty compounds are rat TP because they are positive for histopathologic risk factors only, and not because of the contributions of genetic toxicology test findings or evidence of hormonal effects. Among those twenty compounds, six compounds, specifically compounds 30, 31, 35, 37, 38, 39 (Table 5), would be missed as rat TP if cellular hypertrophy in liver and thyroid were ignored as risk factors for rat neoplasia, and the overall test sensitivity would drop to 69% (forty-six of sixty-six). Furthermore, for those six compounds, tumors were seen in the liver or thyroid after two years, when cellular hypertrophy was the only call at six months at its matching tissue site, thereby
723
justifying inclusion of cellular hypertrophy as a risk factor of potential future neoplasia in the rat. Among the rat TP, there were three compounds that tested positive for genotoxicity, compounds 42, 43, and 44 (Table 5 and Appendix 1), that did not also induce histopathologic risk factors in the chronic rat studies. All are listed in Appendix 1 as antiviral agents. Risk-benefit considerations likely played a role in decisions to continue their development. Compound 43 was negative in Ames and in vivo mouse micronucleus assays but positive in three tests for in vitro clastogenicity, specifically the CHO cell chromosomal aberration test, the human lymphocyte assay, and the mouse lymphoma assay. Compound 44 was positive in the mouse lymphoma, human lymphocyte, and in vivo mouse micronucleus assays. Compound 42 was positive for chromosomal aberrations in CHO cells and also showed evidence of hormonal perturbation. Eight of the fifty-two rat TP presented with evidence of hormonal disruption without also presenting with the specified histopathologic risk factors for tissue neoplasia (Tables 3 and 5). In six cases, knowledge of the pharmacologic target was sufficient itself, or led the sponsor to measure hormones (compounds 42, 46, 47, 49, 50, 51), and in two cases, unknown off-target mechanisms presented a constellation of phenotypic changes in chronic rat studies that was diagnostic of endocrine hormone disruption. Compound 45 induced significant treatment-related decreases in the weights of ovaries and uteri, and compound 48 produced microscopically evident atrophy of the uterus, vagina, cervix, or seminiferous tubules. For the fifty-two rat TP compounds, the impact of the positive rat carcinogenicity findings on marketing status was investigated (Table 5). Thirty-four were marketed despite the positive findings in the rat study. Ten compounds were terminated and not marketed for other reasons that did not relate to the positive rat carcinogenicity findings. Four compounds were still being developed despite treatment-related neoplastic findings. Only four of these fifty-two compounds were discontinued because of the positive findings in the rat carcinogenicity study, specifically compounds 7 (respiratory indication), 10 (PPAR agonist), 14 (diabetes indication), and 52 (antiviral), as described in Tables 5 and 6 and Appendix I. For three of these four compounds, the two-year mouse carcinogenicity study data were available and two of the three also tested positive in mice. Although no compounds discontinued from development for two-year rat carcinogenicity study findings may have been genotoxic, all yielded tumors in multiple tissue sites, were positive in both male and female rats, and showed significant increases in tumors at two or more tested dose levels. Each of these four compounds had clear evidence of histopathologic risk factors for rat neoplasia in the chronic rat studies, and three of the four had evidence of hormonal disruption activity, and they were therefore positive for two criteria that are proposed here to warrant twoyear rat carcinogenicity testing. Fifty-four Rat FP Compounds: For the 116 compounds that tested negative for tumors in the two-year rat study, the need for completion of such a test would have been triggered for
724
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
FIGURE 1.—Concordance between sites of histopathologic risk factors for rat neoplasia (x-axis) observed in chronic six- or twelve-month rat studies and the tumor sites (y-axis) seen in a two-year rat carcinogenicity study with the same rat TP compound. The numbers on each axis refer to the total number of unique compounds with either histopathology in the chronic study (x-axis) or tumor in the two-year study (y-axis) in the listed tissue sites. Data are from the forty-two rat TP compounds presenting with histopathologic risk factors of rat neoplasia in chronic studies. The numbers within the grids refer to the number of unique compounds presenting with the specified tissue site combinations of two-year study tumor and chronic study histopathology.
54 compounds based on either chronic study microscopic histopathologic findings (thirty-eight compounds), positive results in the genetic toxicology testing battery (seventeen compounds), or evidence of hormonal perturbation activity (fifteen compounds). This poor positive predictive value or high FP rate of 49% with 54 of 106 compounds falsely predicted to be rat carcinogens must be appreciated and the limitation of the approach for positive prediction, therefore, well understood. Among the 38 histopathology positive rat FP, 23 compounds are positive solely because of histopathologic risk factors seen in chronic studies (Table 3). Among those 23 compounds, 9 are positive solely because of cellular hypertrophy findings (Table 5), specifically compounds 82, 83, 89, 95, 96, 100, 101, 102, and 104. Cellular hyperplasia was noted either alone or in the same study with cellular hypertrophy but at different tissue sites among nine additional compounds,
specifically compounds 84, 85, 86, 87, 90, 92, 97, 99, and 103. As expected, cellular hypertrophy and hyperplasia seen in six- or twelve-month chronic studies clearly do not always result in tumors. Furthermore, tumors were noted for one rat FP compound, compound 88, in chronic studies, whereas no significant treatment-related tumor response was seen after two years. It is clear that the presence of the defined histopathologic risk factors of rat neoplasia observed sometimes at low incidence among rats in chronic studies will contribute to the high FP prediction of tumor outcome in a subsequent two-year study. For the thirty-eight compounds classified as FP based on histopathology in the chronic study, the site of histopathologic evidence of risk for rat neoplasia was in one of the following five tissues for 74% of the histopathologic findings, specifically liver, kidney, adrenal, thyroid, and gastrointestinal and
725
Result including genotox, hormonal
TP
TP
TP TP
TP TP
TP
TP TP
TP
TP TP
TP TP
TP
TP
TP TP TP TP
TP TP
TP TP TP TP TP
TP
ID
1
2
3 4
5 6
7
8 9
10
11 12
13 14
15
16
17 18 19 20
21 22
23 24 25 26 27e
28
Liver a, c; Thyroid a Adrenals c, p
Nasal t; Ovaries t
Adipose lipoma Adrenals, a; Liver, a, c; Ovaries, gtt
U.B. tc pap, c; Multi tissues sarcomas Mammary a, fa; Ovarian g, gtt Bone osteoma, osteoblastoma, osteosarcoma Liver a Adipose lipomas, ls; U.B. c, pap Adrenals a, c; Mammary fibroadenoma Mammary fa
þ þ
þ
þ þ
þ
Adipose lipomas, ls Testes it
Thyroid fc a Thyroid fc a, c; Liver a, c Liver a, c; Thyroid fc a, c U.B. pap Mammary a; Systemic lymphoma Thyroid fc a, c; Testes it; Kidney a, c; Liver t
þ þ
þ þ þ þ þ
þ
Liver a; Uterus t Kidney a, c; Ovaries g Thyroid fc, a Thyroid fc, a
þ þ þ þ
þ
þ
þ þ
þ þ
þ þ
þ
Cervix t; Liver t; Testes it; Vagina t Liver a, c, hs, ch, chc; U.B. tc, c Liver a Liver a, chc
Tissuea/ tumorb
þ
Carco result
NM
TP TP TP TP TP
TP NM
TP nd TP nd
nd
TP
TP TP
TP TP
TP
TP TP
TP
TP TP
TP TP
nd
TP
Thyroid hp Stomach hp Liver ht, f; Thyroid hp, ht U.B. hp Kidney ht; Systemic lymphoma na
Pituitary hp na
Liver ht; Pituitary ht nd Liver ht; Thyroid, hp nd
nd
Adrenals ht; Mammary hp
Liver hp, ht; Thyroid ht Mammary hp
Ovary hp Bone ht
U.B. hp; Liver ht
Adipose hp Ovaries granulosa theca cell neoplasms
Liver ht Kidneys f; Ovaries ht, hp; Adrenals ht Nasal hp
Liver ht, Thyroid ht Liver ht
nd
Liver ht
Six-month resultc Tissue/histo changed
TP
TP TP TP TP nd
nd TP
nd TP TP TP
TP
nd
nd nd
nd nd
nd
nd TP
nd
nd nd
nd TP
TP
nd
Liver t, hp, ht, f; Kidney hp
nd Signals seen at multiple sites including testes Thyroid hp Stomach hp Liver ht; Thyroid hp U.B. hp nd
nd Ovaries hp, t Liver ht; Thyroid ht, hp Adrenals ht; Liver ht
Mammary hp
nd
nd nd
nd nd
nd Ovaries granulosa theca cell neoplasms nd
nd
nd nd
nd Liver ht, f
Liver ht, f
nd
TwelveTissue/ month resultc histo changed
–
– – – – –
– –
– – – –
– –
– –
–
– –
þ þ
þ
þ
Genotox results
TABLE 5.—Summary table of all 182 compounds in the PhRMA database.
–
– – – – –
þ þ
þ þ þ þ
þ
þ
þ þ
þ þ
þ
þ þ
þ
þ þ
– –
–
–
Hormonal evidence
Not marketed NOT because of carco
Marketed Marketed Marketed Marketed Marketed
Not marketed NOT because of carco Marketed In progress Marketed Not marketed NOT because of carco In progress Marketed
Marketed Not marketed because of positive carco Marketed
Not marketed because of positive carco In progress Marketed
Not marketed because of positive carco Marketed Marketed
Marketed Not marketed NOT because of carco Marketed Marketed
Marketed
Marketed
Marketing status
(continued)
þ
– þ þ – þ
þ þ
– þ - Tg –
þ
þ
þ þ
þ nd
þ
– –
–
þ –
þ –
þ
þ
Mouse carco results
726
Result including genotox, hormonal
TP
TP TP
TP TP TP
TP
TP TP
TP TP
TP
TP
TP
TP
TP TP TP TP
TP TP TP
TP TP
FN FN FN FN FN
ID
29
30 31
32 33 34
35
36 37
38 39
40
41
42
43
44 45 46 47
48 49 50
51 52
53 54 55 56 57
Liver a Liver t; Ovaries t; Testes it
Liver a
Kidney a, c
Stomach ne a, c
Zymbal’s c; Pancreas a, c; Kidney t; Liver a, c; Brain glioma; Skin fibroma; Uterus hs Vagina sq c Mammary t Testes it Mammary fa, c
þ þ
þ
þ
þ
þ
þ þ þ þ þ
þ þ
þ þ þ
þ þ þ þ
þ þ
þ
Ovaries g; Testes it Adrenals p Mammary t; Pituitary a; Pancreas a Pancreas a Liver c; Cervix sq c; Stomach sq pap, c; Systemic hs; Vagina sq c Thyroid pfc a Liver c Lymph node hemangioma Mesovarium leomyoma Liver a; Uterus ac
Pancreas a, c Stomach carcinoid Mesovarian ligament leiomyoma; Pituitary a Testes a; Liver a, c; Thyroid fc a, c Ovaries t Fat fs; Liver a; Thyroid a, c
þ þ þ
þ þ
Skin lipoma, fl, ls; U.B. tc pap, c Thyroid fc a, c Liver a, c; Thyroid fc a, c
Tissuea/ tumorb
þ
Carco result
FN FN FN FN FN
FN TP
FN FN FN
FN FN FN FN
FN
FN
TP
TP
TP TP
TP TP
TP
nd TP TP
TP TP
TP
none none none none none
none Pituitary hp, ht
none none none
none none none none
none
Thyroid hp, ht; Liver karyomegaly Spleen hp; Bone marrow hp none
Liver ht Liver ht
Ovaries hp Liver ht
Liver ht
nd Stomach hp Marrow hp
Skin hp, fibroplasias; U.B. hp; Pituitary hp Thyroid ht Liver ht; Thyroid ht
Six-month resultc Tissue/histo changed
FN TP nd nd nd
nd nd
nd nd TP
FN FN nd nd
nd
nd
nd
nd
nd nd
TP nd
nd
TP TP nd
TP TP
nd
none Liver f nd nd nd
nd nd Mammary hyperplastic acini nd nd
none none nd nd
nd
nd
nd
nd
nd nd
Ovaries hp nd
nd
Thyroid hp, t Liver ht; Thyroid ht; U.B. hp Liver cytomegaly Stomach hp nd
nd
TwelveTissue/ month resultc histo changed
TABLE 5.—(continued)
– – – – –
– N/A
– – –
þ – – –
þ
þ
–
–
– –
– –
–
– – –
– –
–
Genotox results
– – – – –
þ N/A
þ þ þ
– þ þ þ
–
þ
–
–
– –
– –
–
– – –
– –
–
Hormonal evidence
Marketed Marketed Marketed Marketed In progress
Marketed Not marketed because of positive carco
Marketed Marketed Marketed Not marketed NOT because of carco Marketed Marketed Marketed
Not marketed NOT because of carco Marketed
In progress
Marketed Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed
Marketed
Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed Marketed Marketed
Marketing status
(continued)
þ þ þ þ nd
– N/A
– þ –
þ þ – –
þ
þ
–
þ
þ þ
– nd
þ
– – þ
– þ
–
Mouse carco results
727
Result including genotox, hormonal
FN FN FN FN FN FN FN
FN
FN FP
FP FP FP FP FP FP FP FP
FP
FP FP FP
FP FP
FP FP FP FP FP FP
FP FP
FP FP
ID
58 59 60 61 62 63 64
65
66 67
68 69 70 71 72 73 74 75
76
77 78 79
80 81
82 83 84 85 86 87
88 89
90 91
U.B. pap, c none
þ –
– –
– –
– – – – – –
– –
– – –
–
none none
none none
none none none none none none
none none
none none none
none
none none none none none none none none
Pancreas a, c
þ
– – – – – – – –
Soft tissue sarcoma Pancreas a Liver t Mammary ac Liver a; Thyroid pfc a, fc a Adipose hibernoma Testes it; LGL leukemia
Tissuea/ tumorb
þ þ þ þ þ þ þ
Carco result
FP FP
FP FP
FP FP FP FP FP FP
FP nd
NM FP FP
FP
nd FP FP FP nd FP nd FP
NM FP
FN
FN FN nd FN FN FN FN
Kidney hp Cecum hp; Kidney f
Liver ht Liver ht Liver ht; Lymph nodes hp U.B. hp Stomach hp Liver bile duct proliferation, hp Mammary hp, t Adrenals ht
Liver ht; Ovaries proliferation na Liver hp, f; Thyroid ht Adrenal ht; Liver ht; Pancreas hp; Thyroid ht Kidney hp nd
nd Spleen hp Liver ht Marrow hp nd Cecum hp nd Kidney hp, ht
na Adrenal ht
none
none none nd none none none none
Six-month resultc Tissue/histo changed
FP nd
nd FP
FP nd nd TN FP nd
nd FP
FP nd nd
nd
FP nd nd nd FP nd FP FP
FN nd
nd
FN nd FN FN TP nd nd
Kidney hp nd
nd Adrenals ht
nd Signals seen at multiple sites Liver ht nd nd none Liver f nd
Testes hp nd nd
Liver f nd nd nd Kidney hp, f nd Kidney ht Liver hp; Kidney hp, ht nd
none nd
nd
none nd none none Liver ht nd nd
TwelveTissue/ month resultc histo changed
TABLE 5.—(continued)
– –
– –
– – – – – –
– –
– – –
– –
– –
– – – – – –
þ þ
þ þ þ
þ
– – – – – – – –
þ þ þ þ þ þ þ þ –
N/A þ
–
– – – – – – –
Hormonal evidence
– þ
–
– – – – – – –
Genotox results
In progress Not marketed NOT because of carco Marketed Marketed
Marketed Marketed Marketed Marketed Marketed In progress
Marketed Marketed
Marketed In progress Marketed Marketed Marketed Marketed Not marketed NOT because of carco Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed In progress Marketed Marketed Marketed Marketed Marketed Not marketed NOT because of carco Not marketed NOT because of carco Marketed Marketed In progress
Marketing status
(continued)
– –
nd –
– þ þ þ þ –
– –
þ nd þ
þ
– – þ - Tg – – – –
– þ
–
– – – – – – –
Mouse carco results
728
FP FP FP TN TN TN TN
118 119 120 121 122 123 124
102 103 104
FP FP FP FP
FP FP FP
98 99 100 101
114 115 116 117
FP FP FP FP
96 97
FP FP FP FP
FP FP
95
110 111 112 113
FP
94
FP FP FP
FP
93
107 108 109
FP
92
FP FP
FP
ID
105 106
Result including genotox, hormonal
– – – – – – –
– – – –
– – – –
– – –
– –
– – –
– – – –
– –
–
–
–
–
Carco result
none none none none none none none
none none none none
none none none none
none none none
none none
none none none
none none none none
none none
none
none
none
none
Tissuea/ tumorb
TN nd TN TN TN TN TN
TN TN TN TN
TN nd TN nd
TN TN nd
TN TN
FP FP FP
FP FP FP FP
FP nd
FP
FP
FP
FP
none nd none none none none none
none none none none
none nd none nd
none none nd
none none
Liver f Adipose hp; Liver ht Adrenal ht Adrenal ht; Liver ht; Thyroid ht Liver ht Kidney hp; Thyroid ht Liver ht
Adrenal ht; Liver ht; Thyroid ht Liver ht nd
Liver hp, f
Liver f
Liver biliary hp
Six-month resultc Tissue/histo changed
TN TN TN nd nd TN nd
TN nd TN nd
nd TN TN TN
TN nd TN
nd nd
FP nd nd
nd nd nd nd
nd FP
nd
nd
nd
nd
none none none nd nd none nd
none nd none nd
nd none none none
none nd none
nd nd
Liver ht nd nd
nd Cecum hp; Colon hp nd nd nd nd
nd
nd
nd
nd
TwelveTissue/ month resultc histo changed
TABLE 5.—(continued)
– – – þ
þ þ þ –
– – – – – – –
þ þ þ – – – –
þ þ þ þ
– – –
þ þ þ
– – – –
– –
– – –
– – – –
– –
–
–
–
–
Hormonal evidence
þ þ
– – N/A
– – – –
– –
–
–
–
–
Genotox results
Marketed Marketed In progress Not marketed NOT because of carco In progress Marketed Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed Marketed Not marketed NOT because of carco Marketed Marketed Marketed Not marketed NOT because of carco Marketed Marketed Marketed Not marketed NOT because of carco Marketed In progress In progress In progress Marketed Marketed Marketed
Marketed Marketed
Not marketed NOT because of carco Not marketed NOT because of carco Not marketed NOT because of carco Marketed
Marketing status
(continued)
– – – þ þ þ þ
þ þ þ –
nd – – –
– – þ
– –
– þ N/A
– – – þ
þ - Tg
–
þ
–
þ
Mouse carco results
729
Result including genotox, hormonal
TN
TN
TN TN TN TN TN TN
TN
TN TN TN TN TN
TN TN TN TN TN
TN TN TN
TN TN
TN TN TN
TN TN
TN TN TN
TN TN
ID
125
126
127 128 129 130 131 132
133
134 135 136 137 138
139 140 141 142 143
144 145 146
147 148
149 150 151
152 153
154 155 156
157 158
– –
– – –
– –
– – –
– –
– – –
– – – – –
– – – – –
–
– – – – – –
–
–
Carco result
none none
none none none
none none
none none none
none none
none none none
none none none none none
none none none none none
none
none none none none none none
none
none
Tissuea/ tumorb
TN nd
TN TN TN
TN TN
TN nd TN
TN TN
TN nd NM
TN nd TN NM nd
nd TN TN TN TN
TN
TN TN TN TN TN TN
TN
TN
none nd
none none none
none none
none nd none
none none
none nd na
none nd none na nd
nd none none none none
none
none none none none none none
none
none
Six-month resultc Tissue/histo changed
nd TN
nd TN nd
TN nd
TN TN nd
TN TN
TN TN TN
TN TN TN TN TN
TN TN TN TN TN
TN
nd nd nd TN nd nd
nd
nd
nd none
nd none nd
none nd
none none nd
none none
none none none
none none none none none
none none none none none
none
nd nd nd none nd nd
nd
nd
TwelveTissue/ month resultc histo changed
TABLE 5.—(continued)
– –
– – –
– –
– – –
– –
– – –
– – – – –
– – – – –
–
– – – – – –
–
–
Genotox results
– –
– – –
– –
– – –
– –
– – –
– – – – –
– – – – –
–
– – – – – –
–
–
Hormonal evidence Not marketed NOT because of carco Not marketed NOT because of carco In progress Marketed Marketed Marketed Marketed Not marketed NOT because of carco Not marketed NOT because of carco Marketed Marketed Marketed In progress Not marketed NOT because of carco Marketed Marketed Marketed Marketed Not marketed NOT because of carco Marketed Marketed Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed Marketed Not marketed NOT because of carco Marketed Not marketed NOT because of carco Marketed Marketed Not marketed NOT because of carco Marketed Marketed
Marketing status
(continued)
– –
– – –
– –
– – –
– –
– – –
– – – – N/A
– – – – –
–
þ þ þ þ – –
þ
þ
Mouse carco results
730
– – – –
– – – – – –
– –
–
– – – –
–
– – – – – –
Carco result
none none none none
none none none none none none
none none
none
none none none none
none
none none none none none none
Tissuea/ tumorb
TN TN TN TN
TN TN TN nd nd nd
TN nd
TN
TN TN TN TN
TN
TN nd TN TN TN TN
none none none none
none none none nd nd nd
none nd
none
none none none none
none
none nd none none none none
Six-month resultc Tissue/histo changed
TN TN TN NM
nd nd nd TN TN TN
nd TN
nd
nd TN nd nd
TN
nd TN nd nd TN nd
none none none na
nd nd nd none none none
nd none
nd
nd none nd nd
none
nd none nd nd none nd
TwelveTissue/ month resultc histo changed
– – – N/A
– – – – – –
– –
–
– – – –
–
– – – – – –
Genotox results
– – – –
– – – – – –
– –
–
– – – –
–
– – – – – –
Hormonal evidence
Marketed Marketed Marketed Marketed Marketed Not marketed NOT because of carco Not marketed NOT because of carco Marketed Marketed Marketed Not marketed NOT because of carco Not marketed NOT because of carco In progress Not marketed NOT because of carco Marketed Marketed Marketed Marketed Marketed Not marketed NOT because of carco Marketed Marketed Marketed Marketed
Marketing status
– – – –
– – – – – –
– –
–
– – – –
–
– – – – – –
Mouse carco results
Abbreviation: Tg, six-month transgenic mouse study conducted; N/A, not available; U.B., urinary bladder. a Abbreviations for this column: fc, follicular cell; ne, neuroendocrine cell; pfc, parafollicular cell; sq, squamous cell; tc, transitional cell. b Abbreviations for this column: a, adenoma; ac, adenocarcinoma; c, carcinoma; ch, cholangioma; chc, cholangiocarcinoma; fa, fibroadenoma; fl, fibrolipoma; fs, fibrosarcoma; g, granulosa tumors; gtt, granulosa thecal tumors; hs, hemangiosarcoma; it, interstitial cell tumors; Lct, Leydig cell tumor; ls,liposarcoma; pap, papillomas; p, pheochromocytomas; t, tumors. c Abbreviations for this column: FN, false negative; FP, false positive; nd, not done; NM, non-matching based on Table 2 rules; TN, true negative; TP, true positive. d Abbreviations for this column: f, foci of cellular alteration; hp, hyperplasia; ht, hypertrophy; t, tumor; na, not applicable because of non-matching. e These data represent the results from a six-month chronic and two-year carcinogenicity study conducted in the mouse.
TN TN TN TN
170
179 180 181 182
TN
166 167 168 169
TN TN TN TN TN TN
TN TN TN TN
165
173 174 175 176 177 178
TN
159 160 161 162 163 164
TN TN
TN TN TN TN TN TN
ID
171 172
Result including genotox, hormonal
TABLE 5.—(continued)
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
731
TABLE 6.—A comparison of the fourteen rat FN compounds with the four rat true positive compounds (7, 10, 14, and 52) whose marketing status was negatively influenced by positive rat two-year carcinogenicity study findings. Cmpd:
55
53
54
56
62
64
57
59
65
60
66
63
58
61
10
14
Positive Prediction Evidence
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Rat FN
Hormonal evidence and histology
– – þ þ
– – þ þ
– – þ –
– – þ –
– – – þ
– – – þ
– – – –
–
–
–
–
–
–
–
Multiple Species Multiple Sites Both Genders Seen at 2 or more doses Marketing Affected
þ – þ –
þ – – –
þ – – –
þ – – –
– þ þ þ
– þ – –
NA þ – –
–
–
–
–
–
–
–
b
52
7
Hormonal evidence and histology
Histologya
Hormonal evidence and histology
þ þ þ þ
þ þ þ þ
NAb þ þ þ
– þ þ þ
þ
þ
þ
þ
Abbreviations: FN, false negative; NA, not available; þ, yes; , no. Insufficient data available to inform a conclusion regarding hormonal perturbation. b No mouse carcinogenicity study data available. a
TABLE 7.—Ten compounds from the 182-compound PhRMA database with rat carcinogenicity findings without genetic toxicity, hormonal activity, mouse tumor findings, or histopathologic risk factors for rat neoplasia in six- or twelve-month rat studies. Compound no. and chronic study length Neoplastic finding 64, six months
62, six months, twelve months
63, six months 59, six months 60, twelve months
61, six months, twelve months 65, six months
66, six months, twelve months
57, six months 58, six months, twelve months
M: Large granular lymphocytic leukemia M: Testicular interstitial cell neoplasms M: Hepatocellular adenoma F: Hepatocellular adenoma M: Parafollicular cell adenoma F: Parafollicular cell adenoma M: Follicular cell adenoma F: Follicular cell adenoma M: Hibernoma M: Pancreatic acinar adenoma F: Pancreatic acinar adenoma M: Liver adenoma M: Liver carcinoma F: Liver adenoma F:Liver carcinoma F: Mammary adenocarcinoma M: Pancreatic acinar adenoma þ carcinoma F: Pancreatic acinar adenoma þ carcinoma M: Bladder papilloma F: Bladder papilloma M: Bladder carcinoma F: Bladder carcinoma F: Hepatocellular adenoma F: Uterine adenocarcarcinoma M: Soft tissue sarcoma
Incidence of neoplasia in rats C C L M1 M2 H 23/60
22/60 27/60
41/60
47/60
53/60 53/60
57/60
0/100 1/50 3/50 4/50 6/100 2/50 2/50 6/50 7/100 6/50 6/50 4/50 16/100 5/50 8/50 6/50 4/100 2/50 3/50 6/50 0/100 0/50 1/50 2/50 0/130 0/65 1/65 0/120 0/60 8/60 0/120 0/60 0/60 3/100 3/50 0/50 0/100 0/50 0/50 0/100 0/50 0/50 2/100 0/50 0/50 Data not available—tumor incidence increased.
6/50 7/50 6/50 1/50 5/50 1/50 2/65 4/60 1/60 5/50 3/50 3/50 0/50
0/100
15/50 24/50 32/50 29/50
0/100 1/100 1/100 0/100 0/100 1/100 10/100 1/50
1/50 3/50 0/50 0/50 0/50 0/50 0/50 0/50 0/50 0/50 1/50 2/50 5/50 6/50 1/50 3/50
5/50
Mouse tumor findings
Comments
Negative
Not marketed for reasons unrelated to cancer risk. CNS indication.
Negative
Marketed anti-inflammatory / COX-2 inhibitor.
Negative Negative Negative
Marketed for CNS indication. Still in development for CNS indication. Marketed antifungal agent
Negative
Marketed antiviral agent.
Negative
Not marketed for reasons unrelated to cancer risk. Cardiovascular indication.
4/50 8/50 Negative Marketed for respiratory 8/50 indication. 1/50 1/50 4/50 Not completed Still in development for cardio21/50 vascular indication. 6/50 Negative Marketed antifungal agent. Believed secondary to chronic inflammation of connective tissue, including histiocytosis of mesenteric lymph node.
Note: Compound number is a randomly assigned database identifier. Sponsors and compound identities are masked. Abbreviations: C, control group; CNS, central nervous system; H, high-dose group; L, low-dose group; M1, mid-dose group 1; M2, mid-dose group 2 (if there are four treated dose groups).
732
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
TABLE 8.—Four compounds from the 182 compound PhRMA database that were tumorigenic in rats and mice but negative for histopathologic risk factors in chronic rat studies, negative for evidence of hormonal perturbation activity, and negative in genetic toxicology tests. Compound no. and chronic study length Neoplastic finding
Incidence of neoplasia in rats C C L M1 M2 H
54, six months, twelve months
M: Hepatocellular carcinoma
0/100
0/50
53, six months, twelve months
M: Thyroid C-cell adenoma
7/130
7/75
55, six months
M: Hemangioma F: Hemangioma
6/69 2/68
56, six months
F: Leiomyoma of mesovarium 0/100
2/50
Mouse tumor findings
Comments
6/50
4/69 2/68 2/70 3/69
7/68 4/70
M: Hepatocellular carcinoma Marketed for metabolic disease F: Hepatocellular carcinoma indication. F: Pulmonary adenoma 13/75 F: Harderian gland adenoma Marketed for bone indication. The sponsor considered the rat carcinogenicity study negative, but the label in the Physician’s Desk Reference indicates positive carcinogenicity findings. 28/70 Hepatocellular adenoma Marketed for cardiovascular 6/70 indication.
0/50
0/50
1/50
F: Uterine leiomyoma
Marketed inhaled respiratory product.
Note: Compound number is a randomly assigned database identifier. Sponsors and compound identities are masked. Abbreviations: C, control group; H, high-dose group; L, low-dose group; M1, mid-dose group 1; M2, mid-dose group 2 (if there are four treated dose groups).
therefore similar to the tissues presenting with microscopic risk factors for the rat TP compounds (Table 5, Appendix 1). Sixty-two Rat True-Negative Compounds: For the 116 compounds that tested negative for tumors in the two-year rat study, the need for completion of such a test would not have been triggered for 62 compounds based on negative microscopic evidence for rat neoplastic risk factors in chronic studies, negative results in the genetic toxicology testing battery, and negative evidence of hormonal perturbations. Among these 62 compounds, 41 were marketed, 17 compounds were terminated from development, and 4 compounds were still in development. Thus, the percentage of compounds marketed among the rat TN and rat TP were notably similar, 66% and 65%, respectively. Fourteen Rat FN Compounds: For the sixty-six compounds that tested positive for neoplasms in the two-year rat study, fourteen compounds lacked histopathologic risk factors for rat neoplasia, were negative in genetic toxicology tests, and lacked evidence of hormonal perturbation activity. Tables 7 and 8 present the dose and specific tumor response summaries for these fourteen rat FN compounds, whereas Table 6 summarizes parameters of interest and compares them with the four rat TP compounds whose marketing was influenced. Although thirteen of these fourteen rat FN compounds were also tested in the two-year mouse carcinogenicity study (Appendix 1), only four tested positive. All four of these compounds produced neoplasms only at the highest dose level tested in the two-year rat study, and only at a single tumor site (specifically, compounds 53, 54, 55, and 56), and furthermore, for compounds 53, 54, and 56 the tumors were seen only in a single rat sex. Overall, nine of the fourteen rat FN
compounds yielded neoplasms only at the highest dose level tested. Only three among the fourteen compounds yielded neoplasms at more than one tissue site (specifically, compounds 62, 64, and 57), and for compounds 64 and 57, neoplasms were seen only at the highest dose level tested in the rat two-year carcinogenicity study. Only the single-species rat tumorigenic compound 62, among the fourteen rat FNs yielded tumors in both sexes at multiple sites and at two of the four doses tested, 30 mg/kg/day and 100 mg/kg/day. This compound is a rat TP at twelve months where a microscopic diagnosis of centrilobular hepatocellular hypertrophy was made at twelve months but not at six months, and only at the highest dose tested in the chronic study, 30 mg/kg/day. The liver and thyroid tumor pattern seen for compound 62 along with the noted observation of increased liver weights seen at 30 mg/kg/day (Appendix I) at both six and twelve months suggests induced thyroid hormone metabolism, a mechanism considered to be rat specific and human irrelevant (see below). Of these fourteen rat FN compounds, ten were marketed, including compound 62, despite the positive tumor findings in the two-year rat study. Two compounds were terminated and not marketed for reasons unrelated to the positive rat carcinogenicity findings. Two compounds were still under development despite the positive findings. Data Analyses Summary Applying the Proposed Paradigm to IARC Group 1 and 2A Chemicals and to Pharmaceuticals Withdrawn from the Market because of Cancer Concerns IARC Group 1 Compounds: Of thirty-five IARC Group 1 agents (carcinogenic to humans) evaluated, twenty-eight had clear positive findings in one or more standard genetic
733
þb þ
þb –
nt þb
Mesothelioma, mesothelial hyperplasia, hyperplasia in urinary bladder at thirteen and twenty-six weeks.
Proliferative responses in olfactory mucosa and urinary bladder Hepatocellular adenomas, carcinomas as early as six months. Hepatocellular hypertrophy, low level multinucleated hepatocytes at fourteen weeks, increased at thirtyone weeks.
Some changes in lymphoid tissues in three to eight months Some indication of early changes and/or early tumors Some indication of early changes and/or early tumors Steatosis, fibrosis in liver
Chronic-dosed studies: histomorphology
Abbreviations: (f only), tested only in females; IARC, International Agency for Research on Cancer; LP, limited positive; nt, not tested; , sufficient evidence for genotoxicity in standard battery tests, hormonal activity, or immunosuppression; ?, results unclear; þ or -, carcinogenicity result from National Toxicology Program database, unless otherwise noted. a Crude preparation of aflatoxin(s) positive in male mice (L. S. Gold 2008, Carcinogenic Potency Database); aflatoxin B1 is negative in mice. b From L. S. Gold 2008, Carcinogenic Potency Database. c From IARC Monographs. d Genotoxicity summarized in IARC Supplement 7, 1987 as generally equivocal or negative, but details not given. e This listing constitutes a broad category of compounds, some of which have been shown to be genotoxic. f A large number of genotoxicity test results are reported in IARC Volume 99, 2010. The results of most but not all tests using standard battery assays are negative.
nt þb
Treosulfan [299–75–2] Vinyl chloride [75–01–4]
þ þ
þ
þ
þ þ
þ
þ
þ þ þ þ þb þb nt þc þc þc þb nt nt þ
?a þ þ þb þ þb þb LPc þb þ nt þb þc
Mouse carcinogenicity
þ þ þ b (f only) þ þb þb LPc þb þb b (f only) þb þc
Rat carcinogenicity
þ þ þ nt þb þb þ þb þc nt þb þb þb þ
Immuno-suppression
Hormonal activity
?f
?d e
Positive in standard battery
Thiotepa [52–24–4] ortho-Toluidine [95–53–4]
2,3,7,8-Tetrachlorodibenzo-para-dioxin [1746–01–6]
Tamoxifen [10540–29–1]
Diethylstilboestrol [56–53–1] Estrogen-progestogen, steroidal and nonsteroidal Ethanol [64–17–5] (in alcoholic beverages) Ethylene oxide [75–21–8] Formaldehyde [50–00–0] Melphalan [148–82–3] 8-Methoxypsoralen (Methoxsalen) [298–81–7] plus ultraviolet A radiation Methylenebis(chloroaniline) (MOCA) [101–14–4] MOPP and other combined chemotherapy including alkylating agents Mustard gas (Sulfur mustard) [505–60–2] 2-Naphthylamine [91–59–8] N’-Nitrosonornicotine (NNN) [16543–55–8] 4-(N-Nitrosomethylamino)-1-(3-pyridyl)-1-butanone (NNK) [64091–91–4] Phenacetin [62–44–2]
Aflatoxins 4-Aminobiphenyl [92–67–1] Aristolochic acid Azathioprine [446–86–6] Benzene [71–43–2] Benzidine [92–87–5] and dyes metabolized to benzidine Benzo[a]pyrene [50–32–8] N,N-Bis(2-chloroethyl)-2-naphthylamine (Chlornaphazine) [494–03–1] Bis(chloromethyl)ether [542–88–1] and chloromethyl methyl ether [107–30–2] 1,3-Butadiene [106–99–0] 1,4-Butanediol dimethanesulfonate (Busulphan; Myleran) [55–98–1] Chlorambucil [305–03–3] 1-(2-Chloroethyl)-3-(4-methylcyclohexyl)-1-nitrosourea (Methyl-CCNU; Semustine) [13909–09–6] Cyclophosphamide [50–18–0] [6055–19–2] Cyclosporine [79217–60–0]
Chemical [CASRN]
TABLE 9.—IARC Group 1 ‘‘carcinogenic to humans’’ chemicals.
734
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
TABLE 10.—Pharmaceuticals withdrawn from the market because of cancer concerns.
Chemical [CASRN]
Positive in standard battery
Hormonal activity
Immunosuppression
Chlormadinone Chloroform [67–66–3]
Rat carcinogenicity
Mouse carcinogenicity
nt þ
þ þ
þ þ
þ þ
Danthron Diethylstilbestrol [56–53–1]a
?b
Iodinated glycerol [5634–39–9] Methapyrilene
?d
þc þ
þc nt
Nitrofurazone Phenacetin [62–44–2]a
þ (benign) þ
þ þ
Phenolphthalein [77–09–8] Potassium arsenite
þc nt
þc nt
Chronic-dosed studies: histomorphology Mechanism likely involves sustained toxicity with compensatory regeneration in mice and rats Some indication of early changes and/or early changes Hepatocellular carcinomas in rats at 26 weeks Proliferative responses in olfactory mucosa and urinary bladder
Note: Vinyl chloride (IARC group 1) and ethyl carbamate (urethane, IARC group 2a) are named in some lists of withdrawn pharmaceuticals. However, they were used only as inactive ingredients. Abbreviations: IARC, International Agency for Research on Cancer; LP, limited positive; nt, not tested; , sufficient evidence for genotoxicity in standard battery tests, hormonal activity, or immunosuppression; ?, results unclear; þ or -, carcinogenicity result from National Toxicology Program database, unless otherwise noted. a Diethylstilbestrol and phenacetin also appear in Table 7: IARC group 1 ‘‘carcinogenic to humans.’’ They are listed here for clarity but are evaluated as part of the IARC group. b Genotoxicity summarized in IARC Supplement 7, 1987 as generally equivocal or negative, but details not given. c From L. S. Gold 2008, Carcinogenic Potency Database. d Discordant results, see Lee et al. (1994), Snyder (1998).
negative compounds and the former pharmaceutical phenacetin are discussed in Appendix II.
toxicology battery assays (Table 9). Of the seven compounds in this group that did not have positive genetic toxicology findings, three (diethylstilbestrol [also listed as a withdrawn pharmaceutical, Table 10], estrogen-progestogen combined therapy, and tamoxifen) are hormonally active based on known pharmacologic effects and also had histopathologic risk factors for rat neoplasia in six-month studies. One compound, cyclosporine, is an immunosuppressant and thus would be expected to be negative in a two-year rat study but tumorigenic in humans. Two (2,3,7,8-tetrachlorodibenzo-para-dioxin and ortho-toluidine) had histopathologic risk factors for rat neoplasia in sixmonth chronic studies. The dose levels used in the chronic study of o-toluidine were technically nonmatching according to Table 1 rules because histopathologic risk factors were observed at 5,000 ppm in the feed (the only dose tested, NTP Technical Report 44, 1996), whereas tumors were observed in the two-year study at 3,000 and 6,000 ppm in the feed. However, o-toluidine is considered a rat TP given the limited test data available. Finally, among the thirty-five IARC Group 1 agents evaluated here, only ethanol may be considered a failure of the paradigm to predict the need to conduct a rat carcinogenicity study for a known human carcinogen. However, ethanol is considered carcinogenic to humans through chronic toxicity and other indirect mechanisms (Boffetta and Hashibe 2006; IARC Vol. 96, 2010). Since rodents do not tolerate the toxic levels required over two years to develop a carcinogenic response, it is negative for rat carcinogenicity (Appendix II) and therefore, the rat study outcome would be accurately predicted to be negative using the three negative carcinogenicity prediction criteria. Several of these genotoxicity battery-
IARC Group 2A Compounds: Of forty-three IARC Group 2A agents (probably carcinogenic to humans) evaluated, thirtynine had clear positive findings in one or more assays of the standard genetic toxicology battery (Table 11) and one was an androgenic steroid that would be expected to be hormonally active and thus be triggered for rat two-year carcinogenicity testing. Of the remaining three compounds, the genotoxicity of 4-chloro-ortho-toluidine in standard battery tests is unclear (Go¨ggelman et al. 1996), and it was negative for carcinogenicity in two strains of rats (NCI TR No. 165, 1979; Weisburger et al. 1978); trichloroethylene had no reported chronic oral gavage studies of six or twelve months, but a three-month study demonstrated histopathologic risk factors for rat neoplasia (NTP TR 243, 1990); and of five studies across three strains of rats conducted with tetrachloroethylene, only one was positive for neoplasia (Ishmael and Dugard 2006), indicating that the finding is not reproducible in rats. Previously Marketed Pharmaceuticals Withdrawn for Cancer Concern: Ten pharmaceuticals have been withdrawn from the market over the past several decades because of known or suspected risk of carcinogenicity in humans (Table 10): chlormadinone, chloroform, danthron, iodinated glycerol, methapyrilene, nitrofurazone, phenolphthalein, and potassium arsenite (Fowler’s Solution), and diethylstilbestrol and phenacetin, both of which were evaluated above as IARC 734
735
þ þ þa þa
þ þ nt þa
þ
þ
No triggering pathology in chronic studies; Of four strains tested, only F344 rats are positive for carcinogenicity in a two-year study. Very slight/minimal cytomegaly, karyomegaly in rat kidney (a tumor site) at three months, more prevalent at two years.
No three- to twelve-month rat studies reported
In rat liver at three months: stanozolol increased mitosis, binucleation, variability in cell size, xantogranuloma
Chronic-dosed studies: histomorphology
Abbreviations: (f only), tested only in females; IARC, International Agency for Research on Cancer; LP, limited positive; nt, not tested; , sufficient evidence for genotoxicity in standard battery tests, hormonal activity, or immunosuppression; ?, results unclear; þ or -, carcinogenicity result from National Toxicology Program database, unless otherwise noted. a From L. S. Gold 2008, Carcinogenic Potency Database. b From IARC Monographs. c From OECD Environment, Health and Safety Publications Series on Testing and Assessment: Number 103, 2009. d Discordant results, as reviewed in Go¨ggelmann et al. (1996). e Assumed to be carcinogenic on the basis of class effect (nitrosamines).
1,2,3-Trichloropropane [96–18–4] Tris(2,3-dibromopropyl) phosphate [126–72–7] Vinyl bromide [593–60–2] considered to act similarly to the human carcinogen vinyl chloride Vinyl fluoride [75–02–5] considered to act similarly to the human carcinogen vinyl chloride
Trichloroethylene [79–01–6]
e
nt nt þa þ þa nt þ
e
þb þ nt þc þb þb þb nt þ þ nt (f only)a þ þ nt þb þ þa nt þa þa þa
þb þb þc nt þb þb þb nt nt þb þa nt þ þa þb þ þa nt nt þa þa
?d þa þa þa þ þa nt þ/
þ ?b þa nt nt, þa, þa,nt
nt nt þ
Mouse carcinogenicity
þb þa (f only)a nt, nt, -a,nt
þ þb þ
a
Rat carcinogenicity
Immuno-suppression
Azacitidine [320–67–2] Bischloroethyl nitrosourea (BCNU) [154–93–8] Captafol [2425–06–1] Chloramphenicol [56–75–7] a-Chlorinated toluenes (benzal chloride [98–87–3], benzotrichloride [98–07–7], benzyl chloride [100–44–7]) and benzoyl chloride [98–88–4] (combined exposures) 1-(2-Chloroethyl)-3-cyclohexyl-1-nitrosourea (CCNU) [13010–47–4] 4-Chloro-ortho-toluidine [95–69–2] Chlorozotocin [54749–90–5] Cisplatin [15663–27–1] Cyclopenta[cd]pyrene [27208–37–3] Dibenz[a,h]anthracene [53–70–3] Dibenzo[a,1]pyrene [191–30–0] Diethyl sulfate [64–67–5 and 98503–29–8] Dimethylcarbamoyl chloride [79–44–7] 1,2-Dimethylhydrazine [540–73–8; 306–37–6 (-HCl)] Dimethyl sulfate [77–78–1] Epichlorohydrin [106–89–8] Ethyl carbamate (urethane) [51–79–6] Ethylene dibromide [106–93–4] N-Ethyl-N-nitrosourea [759–73–9] Etoposide [33419–42–0] Glycidol [556–52–5] IQ (2-Amino-3-methylimidazo[4,5-f]quinoline) [76180–96–6] 5-Methoxypsoralen [484–20–8] Methyl methanesulfonate [66–27–3] N-Methyl-N0 -nitro-N-nitrosoguanidine(MNNG) [70–25–7] N-Methyl-N-nitrosourea [684–93–5] Nitrate or nitrite (ingested)/endogenous nitrosation Nitrogen mustard [51–75–2] N-Nitrosodiethylamine [55–18–5] N-Nitrosodimethylamine [62–75–9] Procarbazine hydrochloride [366–70–1] Styrene-7,8-oxide [96–09–3] Teniposide [29767–20–2] Tetrachloroethylene [127–18–4]
Hormonal activity
Positive in standard battery
Acrylamide [79–06–1] Adriamycin [23214–92–8] Androgenic (anabolic) steroids
Chemical [CASRN]
TABLE 11.—IARC Group 2A ‘‘probably carcinogenic to humans’’ chemicals.
736
SISTARE ET AL.
Group 1 agents. Of the first eight, five are positive in the standard genotoxicity test battery (danthron, iodinated glycerol, nitrofurazone, phenolphthalein, and potassium arsenite). Chlormadinone is a progestogen and thus triggered for carcinogenicity testing due to known hormonal activity. Chloroform is non-genotoxic and appears to induce kidney tumors in rats through chronic cellular toxicity and sustained regenerative cellular proliferation. Kidney samples collected at six, twelve, and twenty-four months revealed nuclear crowding and basophilia in the mid- to deep cortex at all time points for the two dose groups which were associated with significant increases in renal adenomas and carcinomas (Hard et al. 2000). Lastly, methapyrilene is negative in most reports of genotoxicity battery tests, but rapidly induces liver tumors in rats. In F344/Ncr rats fed methapyrilene, hepatocellular eosinophilic foci, basophilic foci, adenomas, and carcinomas developed over the period of ten to forty weeks (Ohshima et al. 1984). DISCUSSION Based on our analyses of data from 182 pharmaceutical compounds, from seventy-eight IARC Group 1 and 2A chemical carcinogens, and from eight additional previously marketed pharmaceuticals withdrawn for carcinogenicity concerns, we conclude that negative histopathologic risk factors for rat neoplasia in chronic rat studies, negative genetic toxicology test results, and negative evidence for hormonal perturbation activity predict negative findings in rat carcinogenicity studies with sufficient accuracy that two-year carcinogenicity studies of such compounds could have been eliminated without significantly altering the human carcinogenicity risk assessment. The thorough investigation of the nature of the fourteen rat FN indicates that there is no problem with the performance of the proposed new approach, with 80% test sensitivity for predicting the outcome of a rodent test that has a high FP rate for human carcinogenicity prediction. This approach would be considered fit-for-purpose if none of the rat FN represents a flagrant ‘‘miss.’’ In fact, the data clearly demonstrate that each of the rat FN are not of human carcinogenicity concern. Indeed, all compounds that would clearly be considered to be of human health concern and that can be identified by two-year rodent tests are identified, including the four candidates terminated from development because of rodent carcinogenicity findings, the ten pharmaceuticals withdrawn from the market because of carcinogenicity concerns, and the seventy-six additional IARC human carcinogens. This finding indicates 100% test sensitivity for all rat tumorigens among these ninety chemicals considered to have human health relevance. The tumor signals among the fourteen rat FN are marginal, inconsistent across sexes, inconsistent across species, and tend to be seen only at high doses. Human health relevance for these fourteen is therefore questionable. It is proposed, therefore, that rat carcinogenicity studies should not be required by regulatory authorities for registration of future pharmaceutical candidates demonstrated to lack all
TOXICOLOGIC PATHOLOGY
three criteria of risk for rat carcinogenicity. Additionally, an alternative six-month transgenic mouse carcinogenicity study is expected to be conducted to assess carcinogenic potential in vivo. If a small organic pharmaceutical molecule intended for chronic use in humans is assessed as positive for any of these criteria for risk of rat neoplasia, it may be reasonable to conduct a two-year rat carcinogenicity study prior to registration to help further understand that potential risk. Furthermore, if a six-month alternative transgenic mouse model is positive for carcinogenic findings and the three criteria of risk for rat neoplasia are all negative, conduct of a two-year rat study may also be indicated. An Assessment of the Fourteen Rat FN Compounds Fourteen compounds producing positive findings in rat carcinogenicity studies from the 182 compound PhRMA database were negative for genetic toxicology findings, hormonal activity, and histopathologic risk factors for rat neoplasia in sixmonth or twelve-month rat studies (Tables 7 and 8) and were therefore classified as rat FN compounds. However, the positive two-year rat study findings did not result in the registration failure or discontinuation of development for any compound among these fourteen compounds. The compound class information provides limited insight to the question of whether risk-benefit considerations had an impact on marketing decisions. For three compounds, specifically 58, 60, and 61, an antifungal, antifungal, and antiviral, respectively, the impetus to fill an unmet medical need for a serious indication may have been a factor in successful registration, but for the other eleven compounds, there is no clear evidence for this being a factor. The effects of positive rat carcinogenicity findings on the drug labels are not known because of the confidentiality of compound identities. However, single-organ, single-sex, highdose only, and single-species tumorigens in rodents generally raise less concern for human carcinogenic risk than chemicals that produce neoplasms in multiple organs, both sexes, or more than one species and at multiple and near clinical exposure levels. Eleven of these fourteen rodent FN compounds produced neoplasia in a single organ and tissue type, whereas the remaining three compounds produced neoplasia in two tissues, and none of those three was carcinogenic in mice. Several hormonal or receptor-mediated mechanisms have been demonstrated to induce genital tract tumors through rodentspecific mechanisms. In the case of b2-adrenergic receptor stimulants (e.g., bronchodilators), mesovarian leiomyomas are the result of prolonged stimulation of the b2-adrenergic receptors that mediate muscle relaxation (Apperley et al. 1978) with subsequent proliferation of the mesovarian smooth muscle (Jack et al. 1983). Among the fourteen rat FN, compound 56 (Table 8) appears to be such an example and is not considered a relevant human risk. Similarly, as mentioned previously, compound 62 appears to be an example of a drug inducing hepatic enzymes, which accelerates breakdown of thyroid hormones, resulting in subnormal levels of T4 and T3, a compensatory increased secretion of pituitary thyroid-
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
stimulating hormone, and thyroid tumors (Capen 1997; Capen 1999; McClain 1989) that is considered to be a rat-specific process. As discussed below, pancreatic acinar cell tumors can be readily induced in rats and are noted for the rat FN compounds 59 and 65, whereas these tumors are rare in humans. The fourteen rat FN compounds are collectively considered to represent molecules posing low risk for human cancer. An Evaluation of IARC Classified Human Chemical Carcinogens and Pharmaceuticals Withdrawn from the Market Because of Carcinogenicity Concerns A review of IARC Group 1 and 2A agents presents a unique opportunity to evaluate the utility of the proposed paradigm for compounds adjudicated by IARC to have human carcinogenic potential. For these compounds, human carcinogenicity is either known or suspected as probable owing to multiple lines of experimental and epidemiologic evidence. Of the seventyeight known and probable human carcinogens identified by IARC and evaluated here, and eight additional formerly marketed pharmaceuticals withdrawn for cancer concerns, most (seventy-two of the composite eighty-six) are positive in the genetic toxicology standard testing battery. Notably, sixtytwo of these seventy-two are positive in the bacterial mutation assay and, of the ten that are not, nine are positive in an in vivo micronucleus assay. The tenth compound, an Ames and in vivo micronucleus assay negative chemical, chloramphenicol, is positive in the mouse lymphoma assay. Of the fourteen genotoxicity-negative compounds, ten would be targeted for testing in a two-year rat study by virtue of their known hormonal activity or by virtue of the appearance in chronic rat studies of histopathologic risk factors for rat neoplasia. The four remaining compounds needing further consideration are ethanol, 4-chloroortho-toluidine, tetrachloroethylene, and trichloroethylene. As noted above and in Appendix II, the available data for ethanol, 4-chloro-ortho-toluidine, and tetrachloroethylene indicate that these compounds are not rat carcinogens at doses that can be tolerated in two-year rat studies. For trichloroethylene, sufficient chronic test study data are unavailable, but subchronic toxicity studies reveal a histopathologic risk factor for rat neoplasia. Thus, these data support the validity of using the three negative rat carcinogenicity prediction criteria to assess the need to conduct a two-year rat carcinogenicity study. Among these eightysix known or suspected human chemical carcinogens, there are no discrepancies to the conclusion that negative findings in all three criteria accurately predict negative outcome in a twoyear rat carcinogenicity study. An Evaluation of the Fifty-two Rat TP Compounds and Evidence for Indirect Mechanisms of Tumorigenesis In this survey of 182 pharmaceuticals, eleven tissues evaluated in chronic rat toxicology studies served as histologic sentinels of carcinogenesis for almost 90% of the positive responses observed in two-year rat carcinogenicity studies, specifically the liver, thyroid, adrenals, ovaries, mammary gland, bone, pituitary, urinary bladder, kidney, skin, and
737
stomach. Among the tumor-bearing tissues, nine sites accounted for over 80% of the induced tumors, specifically liver, thyroid, ovaries, testes, urinary bladder, skin, mammary gland, kidneys, and adrenals (Figure 1, Tables 3 and 5). These patterns suggest that tissues with the highest expected drug or metabolite exposures after dosing (stomach, liver, kidney, urinary bladder) or with a high sensitivity to hormonal perturbations (thyroid, mammary gland, ovaries, adrenals, pituitary) would be most predictive of potential tumor risk to the rat. The rat liver was the single most frequent site for occurrence of histopathologic risk factors for rat neoplasia, and the most frequent site for occurrence of treatment-related neoplasia. As a site presenting in chronic studies with histopathologic risk factors of rat neoplasia, more often than any other tissue site, the liver was also associated with distal sites of compound-induced rat tumorigenesis. Many of the rat TP compounds are hormonal agents or belong to pharmacologic classes known to cause hormonal changes in rats. Among these are compounds 5, 11, 12, 16, 18, 19, 20, 46, 48, and 51, all of which were intended to alter hormonal activity. These compounds were associated with ovarian granulosa cell, bone, mammary, testicular, pancreatic, and/or thyroid tumors, and all had earlier documented effects on hormones or hormonally regulated tissues in the rat that are related to the tumors seen after two years. Endocrine tumors are now well known to arise in rat carcinogenicity studies at tissue sites distal to the tissue that the compound may be directly affecting. These neoplasms result from initially poorly understood, but now well understood, indirect and often rodent-specific mechanisms, specifically associated with chronic trophic hormonal stimulation at the target site for tumorigenesis. An excellent review of some imbalances of hormone levels associated with tumorigenesis has been published (Alison et al. 1994), and much additional information on nongenotoxic indirect tumorigenic mechanisms has been gathered over the past fifteen years (Greaves 2007). To exemplify, several rat TP compounds in this database (such as compounds 5, 19, 20, 23, 24, 25, 28, 30, 31, 35, and 37) caused thyroid follicular cell tumors, often with histopathologic findings in the liver and often with demonstrated hepatic enzyme induction mediating the thyroid tumorigenesis. The mechanism is generally recognized to lead to disruption of the synthesis and secretion or breakdown of thyroid hormones, resulting in decreased levels of T4 and T3 followed by a compensatory increased secretion of pituitary thyroid-stimulating hormone (Capen 1997; Capen 1999; McClain 1989), a chronic trophic hormonal stimulus, which over the lifetime of the rat induces thyroid tumors and is considered to be rat specific. Pancreatic tumors can also arise by a number of distally mediated mechanisms that are distinct for the exocrine and endocrine pancreas (Wilson and Longnecker 1999). Proliferative changes of the pancreatic islets leading to tumors have been observed with several neuroleptics or other drugs that induce hyperprolactinemia (Iatropoulos 1983), an initially unexpected tumor location now considered an anticipated phenomenon. Hyperplastic islets, b cell adenomas, and carcinomas
738
SISTARE ET AL.
have been reported following chronic dosing with glucocorticoids and are believed to occur secondarily to chronic demand for insulin production owing to sustained increases in blood glucose, and peripheral insulin resistance (Dillberger 1994; Zwicker and Eyster 1993). Most commonly pancreatic acinar tumors are caused by indirect mechanisms in which the compound increases levels of cholecystokinin (CCK), or CCK-A receptors, secondarily stimulating mitogenic activity in acinar cells in rodents that is considered to be irrelevant to humans (Ross and Barrowman 1987; Woutersen et al. 1991). The rat pancreas can also be a target for compounds that cause peroxisome proliferation. In particular, PPAR-a agonists cause three tumor types in rats: pancreatic acinar cell, hepatocellular, and testicular Leydig cell tumors (Klaunig et al. 2003). The mechanisms involved in tumor induction have been shown to be associated with alterations in bile acid synthesis and flow, leading to elevated CCK-mediated stimulation of pancreatic exocrine tissue as well as enzyme induction and peroxisome proliferation in the liver. The testicular tumors are related to increased 17b-estradiol secondary to aromatase induction in the liver and/or decreased testosterone synthesis by Leydig cells, leading to elevated luteinizing hormone (LH) levels as the chronic trophic hormonal stimulus. These modes of action are generally not considered relevant for humans because: (1) there are key species differences in CCK actions and control mechanisms, (2) the level of PPAR-a expression and peroxisome proliferation in liver in humans is minimal, and (3) human Leydig cells are far less sensitive to LH stimulation than rat Leydig cells (Klaunig et al. 2003). Rats develop ovarian granulosa cell and stromal hyperplasia and tumors when estrogen levels are reduced, which results in indirect sustained increases in LH trophic stimulation. Such LH increases have been shown to occur, reversibly, following treatment with selective estrogen receptor modulators such as raloxifene (Long et al. 2001). Similarly, the LHRH analog naferelin produced benign ovarian tumors in rats. The rodent uterus is also similarly susceptible to chronic hormonal trophic tumorigenesis induced indirectly by mechanisms that modulate estrogen and progesterone levels (Nagaoka et al. 1990). For example, decreased prolactin from administration of dopamine agonists causes an estrogen dominance, which leads to endometrial stimulation. Prolonged estrogen/progesterone imbalance can lead to uterine tumor formation in rodents, but the analogous situation has not been observed in humans treated with agents such as bromocriptine (Alison et al. 1994). Testicular Leydig cell tumors can be generated indirectly by induction of a sustained elevation of circulating pituitary LH levels. Increased secretion of LH can be the result of interference with the hypothalamic–pituitary–testis axis at various points (e.g., androgen receptor antagonists, testosterone biosynthesis inhibitors, 5a-reductase inhibitors, estrogen agonists, LHRH analogues, and dopamine agonists; Cook et al. 1999) or of enhanced hepatic testosterone metabolism and clearance by induction of CYP forms involved in testosterone catabolism (Coulson et al. 2003).
TOXICOLOGIC PATHOLOGY
It is concluded that Leydig cell tumors in rats are generally not predictive or relevant for man because rats are much more sensitive, as rat Leydig cells have fourteen-fold more LH receptors than human Leydig cells (Cook et al. 1999; Prentice and Meikle 1995). In the rat, prolonged hyperprolactinemia, as induced by anti-dopaminergic compounds, can result in the development of pituitary and mammary gland tumors. Because of major differences in hormonal and reproductive physiology between rodents and humans, this is considered a rodent-specific mechanism (Gopinath 1995; Harvey 2005). A prolactin-mediated effect was concluded to be the mechanism of tumorigenesis for rat TP compounds 15, 16, 45, 47, and 50. As discussed above, indirect rodent-specific hormonal mechanisms leading to pancreatic, testicular, ovarian, cervical, vaginal, or uterine tumors may underlie the effects seen for the rat TP compounds 9, 11, 17, 18, 22, 28, 32, 36, 39, 46, 50, 51, and 52. Since proprietary information relating to specific drug targets was not shared among participants, the linkage between specific compounds and likely mechanisms is uncertain in some cases. For 79% of the forty-two rat TP compounds that presented with histopathologic risk factors for rat neoplasia, in chronic studies, at least one of the same tissue sites presented with neoplasia after two years of dosing (Table 4a). However, for nine rat TP compounds (21%) with histopathologic risk factors for rat neoplasia, tumors developed at other sites and not at the site of hypertrophy, hyperplasia, or altered foci seen in the chronic study. Three of these nine (specifically, compounds 10, 20, and 32 [Table 4b]) had histopathologic risk factors for rat neoplasia in liver in the chronic study, whereas tumors developed after two years in thyroid, pancreas, or urinary bladder, but not in liver. Such indirect mechanisms likely involving the liver directly and leading to distal tumors in the thyroid and pancreas are discussed above. One other compound, number 52, induced a histopathologic risk factor for rat neoplasia in the pituitary in the chronic study, but tumors occurred in cervix and vagina after two years, suggesting a hormonally linked mechanism. Another, compound 24, presented with microscopic findings in the stomach, a site of consistently higher initial exposure after dosing, whereas tumors appeared after two years in liver and thyroid. Table 4b shows that among the remaining four rat TP compounds in this set, a potential cause-andeffect relationship between chronic study findings and neoplasia after two years was unclear, specifically for compounds 14, 21, 34, and 41. Although Tables 4a and 4b illustrate that cause-and-effect relationships between chronic study findings and neoplasia after two years in the rat may be unclear for a small minority of the compounds, in most instances, the site of alteration in the chronic study serves as a reasonably logical indicator of the neoplasia that will be seen after two years. Regardless of the relevance of these indirect mechanisms for human cancer risk, these data clearly reveal a rationale for discordant relationships between the site of occurrence of the histopathologic risk factors seen in chronic rat studies and the site of eventual
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
tumorigenesis in two-year rat studies. As discussed above, defining the underlying mechanisms of neoplasia seen at tissue sites distal to the sites of direct compound activity has proved useful in addressing concern for humans but may take many years of research (Hernandez et al. 2009; Holsapple et al. 2006). New explanations for distally mediated indirect mechanisms that could contribute to rodent tumorigenesis continue to accrue. For example, it has recently been reported (Ilnytsky et al. 2009; Koturbash et al. 2006) that a focused tissue x-irradiation in animals leads to epigenetic alterations in DNA methylation in distant bystander tissues. The ‘‘whole-animal’’ paradigm proposed previously (Reddy et al. 2010) and further corroborated here relies on the concept that even if the histologic risk factors for rat neoplasia may not directly identify the target issue for ultimate tumorigenesis, it nonetheless accurately predicts for tumorigenesis at some location in the rat even though the causeand-effect relationship may occasionally not be understood. Tissues exposed to repeated higher local concentrations of chemicals (e.g., injection sites, the gastrointestinal tract and liver for orally administered drugs, the kidney and urinary bladder during elimination) may be prone to serve as early sensitive sentinels of histopathologic risk factors for neoplasia, although tumors may appear later only at alternate tissue sites. Indirect processes may be triggered to result in cellular alterations of concern within six months that yield tumors later only at alternate tissue sites. In time, additional mechanistic and cellular proliferation assessments, such as those described by Cohen (2004, 2010), may further improve scientific understanding and human carcinogenicity risk assessment. The mode of action framework approach described by Cohen could complement this proposal and potentially further reduce the number of rodent carcinogenicity studies that are required. When indirect and human irrelevant mechanisms may be concluded from recognized tissue patterns of histopathologic changes in chronic rat studies together with knowledge of pharmacology, analyses of molecular and biochemical tissue changes from targeted investigative studies, and measurements of alterations in circulating hormone levels, there may in fact be little value in conducting a two-year rat carcinogenicity study only to tabulate a final number of the expected tumor types. However, pursuit of such mode of action framework approaches is unwarranted when all tumorigenic risk factor criteria for the rat are negative, so routine deployment would add no benefit. Limitations of Human Tumor Prediction with Immunosuppressants in the Rat We classified the compounds in this manuscript as ‘‘rat’’ TP, ‘‘rat’’ FN, and so on to emphasize that positive neoplastic findings in rodents may or may not be relevant for human risk assessment and that this analysis is focused on prediction of ‘‘rat’’ carcinogenicity outcome. We have described above documented examples from the literature in which positive rat carcinogenicity findings have been
739
demonstrated to be irrelevant for human health. In a companion activity (Alden et al. 2011), rodent carcinogenicity findings were compared to label descriptions for carcinogenicity risk using the approved drug labels in the electronic Physicians’ Desk Reference. The conclusion might be drawn from such analyses that the two-year bioassay in and of itself is not a predictor of human carcinogenicity and may be losing credibility. It raises the issue of the questionable value of pursuing a two-year carcinogenicity study under any circumstances, but this issue is beyond the scope of the current proposal. Alden et al. (2011) note that it is widely recognized that a high percentage of marketed compounds with labeled carcinogenicity test results were positive for rodent neoplasia (Oosterhout et al. 1997), citing prototypical over-the-counter rodent carcinogenic and human noncarcinogenic drugs in this category such as omeprazole, acetaminophen, sodium fluoride, benzoyl peroxide, cetirizine, sodium ascorbate, doxylamine, minoxidil, and lansoprazole. On the other hand, immunosuppressive or immunomodulatory drugs are among the drugs with labels indicating high association with human cancer risk despite the lack of positive evidence from rodent testing. Immunosuppression is a risk for tumorigenesis in humans that is generally recognized as not reliably detected in two-year rat studies (Bugelski et al. 2010), even though intended immunosuppressive activity can be determined in subchronic toxicology studies. The relative insensitivity of rats as compared to humans to immunosuppressants has long been well appreciated, as immunosuppressive compounds are known to facilitate neoplasias induced by infectious agents, and rodent colonies are bred and maintained for their entire lifetimes under carefully controlled and closely monitored study conditions to purposely eliminate or minimize exposures to infectious agents (Cohen et al. 1991). Since the two-year rat carcinogenicity study adds little value for assessing the risk from immunosuppressants as human tumorigens, evidence of immunosuppression is not included in the proposed decision tree of criteria for rat carcinogenicity testing. Therefore, two-year rat carcinogenicity testing is not expected to add value and is not recommended for assessing human cancer risk associated with immunosuppressants unless such compounds are also associated with off-target effects noted in genetic toxicology tests, show hormonal perturbation evidence, or present with the described additional microscopic histopathologic risk factors of rat neoplasia at relevant doses in conventional chronic rat studies. There are likely significant differences between broad-based immunosuppressants and selective immune modulatory compounds that would be important to understand in helping to provide perspective for human risk assessment. Alternative proposals to the standard two-year rodent assays have been described to improve human cancer risk associated with immunomodulatory drugs (Bugelski et al 2010). Value of Transgenic Mouse Models Pharmaceuticals are being tested with increasing frequency in genetically modified mice, specifically the TgrasH2 and
740
SISTARE ET AL.
TOXICOLOGIC PATHOLOGY
FIGURE 2.—Proposed Negative for Endocrine, Genotoxicity, and Chronic study Associated histopathologic Risk factors for Carcinogenicity in the Rat decision tree. If any of the three negative prediction criteria is positive, the decision will be to conduct a two-year rat carcinogenicity study. If all three criteria are negative, the opportunity exists to complete the carcinogenicity assessment of the pharmaceutical candidate with the conduct of a mouse carcinogenicity study. If the mouse study is negative, then no relevant cancer risk for humans is projected. The dotted lines convey that if the mouse study is positive when all three criteria for conducting a two-year rat carcinogenicity study are clearly negative, the decision to conduct a two-year rat carcinogenicity study may need considerable discussion depending, for example, on the strength and nature of the neoplastic signal(s) in the mouse study, or the exposure margins associated with a tumor response.
p53þ/ models with study durations of six months (JacobsonKram et al. 2004). Genetically modified mice with insertion or deletion of genes relevant in human cancer were approved as alternatives to the lifetime mouse bioassay through ICH S1B to improve the predictive utility of the cancer hazard
identification process for human risk associated with pharmaceutical agents, and their practical value in support of pharmaceutical development has been reviewed recently (Storer et al. 2010). The improved concordance with human experience in testing with these transgenic mouse models has been
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
741
FIGURE 3.—Gantt chart to drug development. In the current state, the start of the two-year rat carcinogenicity study is generally delayed until after sufficient efficacy and safety information are gained from Phase 2 clinical trials because of a combination of factors, including the significant development failure rates for compounds at this stage together with the high resource burdens associated with the two-year study. Following completion of the conduct of the two-year rat study, the need to process and analyze all tissues from 500 to 600 animals adds additional delays to New Drug Application (NDA) filing dates. In the future, the option exists to conduct a six-month transgenic mouse study using fewer than 200 animals to complete the carcinogenicity assessment on approximately 40% of development programs in which histopathologic risk factors are absent in chronic rat studies and the compounds have no evidence of genotoxicity or hormonal perturbation activity. For such transgenic mouse studies, the tissue processing, microscopic analyses, and report completion dates are substantially shorter, enabling NDA filing sooner after study completion. The six-month transgenic mouse study, which demands far less in the way of resources, opens more opportunities for conducting studies earlier before completion of Phase 2 studies, and for those clinical programs in which proof of efficacy may not require years of treatment, NDA filing dates can be shortened by two to three years. For programs in which the decision may be to conduct the transgenic mouse study later following completion of Phase 2 trials, NDA filing timelines may still be accelerated by more than a year.
chronicled by workers in the National Toxicology Program (NTP). The transgenic alternatives (specifically, the combined p53þ/ for genotoxic molecules and the TgrasH2 for nongenotoxic molecules) provide comparable predictivity to the rat and mouse lifetime bioassays, alone or in combination, for human responses to known IARC Group 1 and probable IARC Group 2A human carcinogens. The NTP workers report that twelve Group 2A molecules sufficiently tested in their survey all tested positive using a combined p53þ/ and TgrasH2 test system. Similarly, among the fourteen Group 1 carcinogens considered, eleven yielded positive responses in this combined TgrasH2 and p53þ/ test system, one yielded a negative response, and two were untested. In contrast, applying the conventional two–species lifetime rodent test system to this same list of fourteen Group 1 carcinogens, five positive tests and one
negative test were reported, with eight untested (Pritchard et al. 2003). Furthermore, these two alternative genetically modified mouse model systems reduced the high false positive rate associated with the two-year rodent bioassays from 35% (twentythree of sixty-six human noncarcinogens) to 15% (eight of fifty-three human noncarcinogens), but, as described above, they did not compromise the ability to identify known and probable human carcinogens. A Proposed Decision Tree for Conducting Two-year Rat Carcinogenicity Studies From the data presented here and the collective preponderance of published information, it is apparent that two-year carcinogenicity studies in two species are not always necessary to
742
SISTARE ET AL.
identify pharmaceuticals that pose minimal carcinogenic risk to humans. In the absence of histopathologic risk factors of rat neoplasia in a chronic rat study, with negative genotoxicity test results, and with no evidence of hormonal perturbation activity, there is a low likelihood of a carcinogenic response in the twoyear rat study, and thus a human cancer risk assessment can be performed upon completion of the six-month rat study and a six-month transgenic mouse carcinogenicity study in the absence of a rat carcinogenicity study result. If global regulatory authorities would agree that a two-year rat carcinogenicity study is not required in the absence of genotoxicity test findings, evidence of hormonal perturbation activity, or histopathologic risk factors of rat neoplasia from a chronic rat study, what could a future carcinogenicity testing strategy look like? A decision tree termed the Negative for Endocrine, Genotoxicity, and Chronic study Associated histopathologic Risk factors for Carcinogenicity in the Rat (NEG CARC Rat) testing paradigm, which outlines a potential approach, is provided in Figure 2. Despite its simplicity, several critical factors must be addressed to allow practical implementation. A first critical factor is to reach global agreement on criteria used to determine that a rat carcinogenicity study is not required. Seeking prior agreement with all regulatory authorities on a case-bycase basis is impractical in a world of increasingly global drug development. Thus, clear international alignment and guidance on what constitutes the requirements for rat carcinogenicity studies will be needed to enable implementation of this new NEG CARC Rat testing paradigm proposal. For the NEG CARC Rat testing paradigm to work practically, regulatory authorities must also agree on the adequacy of the rat chronic study design, (e.g., dose levels, exposure margins, extent of toxicity observed). As noted above, such prior agreement is routinely achieved with the U.S. Food and Drug Administration in selecting study design aspects for rodent carcinogenicity studies, but to expect routine regulatory agreement on design of each individual six-month rat study prior to initiation would be a substantial resource burden for regulatory authorities and could significantly delay the nonclinical drug development process. As a final point, consideration needs to be given to the consequences that would emerge if the mouse study is ‘‘positive’’ and a rat two-year carcinogenicity study has not been conducted. If only a short-term alternative transgenic mouse carcinogenicity study had been conducted and tested positive, would a two-year rat study then be needed to supplement the interpretation of the mouse result? Alternatively, could a risk-benefit analysis be conducted based on six-month rat chronic study results in the absence of a two-year rat study, considering the nature of the tumors observed in the mouse? Although not essential to initiate the proposed NEG CARC Rat testing paradigm, it is useful to have a general understanding of such potential consequences that could arise from practical application of the approach. In summary, all sources of data available to the authors indicate that human cancer risk assessment would not be adversely affected by the elimination of rat carcinogenicity
TOXICOLOGIC PATHOLOGY
studies for novel small organic pharmaceuticals that lack histopathologic risk factors for rat neoplasia in chronic rat studies, positive genotoxicity findings, and evidence of hormonal perturbation. Furthermore, since sharing this proposal at an earlier stage with scientists at both the Japan Pharmaceutical Manufacturers Association and the U.S. Food and Drug Adminstration, the suggestion has been favorably received (Jacobson-Kram 2010a), and similar data collection and analyses efforts have been launched independently by these groups, each accessing distinct datasets. The analyses being generated from studies of over 100 additional compounds to further assess the validity of the NEG CARC Rat testing proposal strategy are yielding similar conclusions (JacobsonKram 2010b; Dr. Shigeru Hisada, personal communication). Although the avenues and outcomes of regulatory engagement in addressing this proposed NEG CARC Rat testing paradigm are unclear at this time, the potential benefits for drug development timelines and resource commitments following such an approach are relatively easy to foresee. Approximately 600 fewer animals would be required to be used for each twoyear rat carcinogenicity study avoided. A further 400 fewer animals would be used if the approach and timeline considerations encourage expanded use of an alternative mouse model of carcinogenicity in place of the conventional two-year mouse study. The timeline for completion of nonclinical studies supporting registration potentially could be shortened by two to three years, and registration timelines could be accelerated depending on the clinical program timeline (Figure 3). Both industry and regulatory resources applied in support of product development would be reduced, including up to an estimated $3.75 M reduction in costs for all efforts associated with the completion and evaluation of a two-year rat carcinogenicity study. The potential to eliminate much of the uncertainty around carcinogenic risk earlier in development, even if for only 40% of compounds, would be significant and could improve portfolio management. All of these goals can be achieved by first using the three negative carcinogenicity prediction criteria to determine the need to conduct a rat two-year study without compromising cancer risk assessment for human pharmaceuticals. A conservative estimate is that if the proposed NEG CARC Rat decision tree had been used to define the test paradigm to support development of the 182 compounds in the PhRMA database, approximately 86,000 fewer rodents would have been used and $290 M could have been directed to better uses with no compromise to human cancer risk recognition. AUTHORS NOTE Appendix I and II are available online at http://tpx.sagepub .com/supplemental. ACKNOWLEDGMENTS The authors acknowledge the excellent project management provided by PhRMA staff Vail Fucci, Justin Sill, and Rosemary Cook to construct and maintain the database and manage and coordinate project support.
Vol. 39, No. 4, 2011
A NEW CARCINOGENICITY TESTING PROPOSAL
REFERENCES Alden, C. L., Lynn A., Bourdeau A., Morton D., Sistare F. D., Kadambi V. J., and Silverman L. (2011). A critical review of the effectiveness of rodent pharmaceutical carcinogenesis testing in predicting for human risk. Vet Pathol. Epub ahead of print. March 7, 2011; doi: 10.1177/ 0300985811400445. Alison, R. H., Capen, C. C., and Prentice, D. E. (1994). Neoplastic lesions of questionable significance to humans. Toxicol Pathol 22, 179–86. Allen, D. G., Pearse, G., Haseman, J. K., and Maronpot R. R. (2004). Prediction of rodent carcinogenesis: An evaluation of prechronic liver lesions as forecasters of liver tumors in NTP carcinogenicity studies. Toxicol Pathol 32, 393–401. Apperley, G. H., Brittain, R. T., Coleman, R. A., Kennedy, I. and Levy, G. P. (1978). Characterization of the b-adrenoceptors in the mesovarium of the rat. Br J Pharmacol 63, 345–98. Boffetta, P., and Hashibe, M. (2006). Alcohol and cancer. Lancet Oncol 7, 149–56. Bregman, C. L., Adler, R. R., Morton, D. G., Regan, K. S., and Yano, B. L. (2003) Recommended tissue list for histopathologic examination in repeat-dose toxicity and carcinogenicity studies: A proposal of the Society of Toxicologic Pathology (STP). Toxicol Pathol 31, 252–53. Bugelski, P. J., Volk, A., Walker, M. R., Krayer, J. H., Martin, P., and Descotes, J. (2010). Critical review of preclinical approaches to evaluate the potential of immunosuppressive drugs to influence human neoplasia. Int J Toxicol 29, 435–66. Capen, C. C. (1997). Mechanistic data and risk assessment of selected toxic end points of the thyroid gland. Toxicol Pathol 25, 39–48. Capen, C. C. (1999). Thyroid and parathyroid toxicology, mechanisms of toxicity: Thyroid follicular cells. In Endocrine and Hormonal Toxicology (P. W. Harvey, K. C. Rush, and A. Cockburn, eds.), pp. 42–47. John Wiley & Sons, New York. Capen, C. C. (2004). Mechanisms of hormone-mediated carcinogenesis of the ovary. Toxicol Pathol 32 (Suppl. 2), 1–5. Cohen, S. M. (2004). Human carcinogenic risk evaluation: An alternative approach to the two-year rodent bioassay. Toxicol Sci 80, 225–29. Cohen, S. M. (2010). Evaluation of possible carcinogenic risk to humans based on liver tumors in rodent assays: The two-year bioassay is no longer necessary. Toxicol Pathol 38, 487–501. Cohen, S. M., Purtilo, D. T., and Ellwein, L. B. (1991). Ideas in pathology. Pivotal role of increased cell proliferation in human carcinogenesis. Mod Pathol 4, 371–82. Contrera, J. F., Jacobs, A. C., and DeGeorge, J. J. (1997). Carcinogenicity testing and the evaluation of regulatory requirements for pharmaceuticals. Regul Toxicol Pharmacol 25, 130–45. Cook, J. C., Klinefelter, G. R., Hardisty, J. F., Sharpe, R. M., and Foster, P. M. (1999). Rodent Leydig cell tumorigenesis: A review of the physiology, pathology, mechanisms and relevance to humans. Crit Rev Toxicol 29, 169–261. Coulson, M., Gibson, G. G., Plant, N., Hammond, T., and Graham, M. (2003). Lansoprazole increases testosterone metabolism and clearance in male Sprague-Dawley rats: Implications for leydig cell carcinogenesis. Toxicol Appl Pharmacol 192, 154–63. Department of Health and Human Services, National Toxicology Program Database Search Application Web site: http://ntp-apps.niehs.nih.gov/ ntp_tox/index.cfm. Dillberger, J. E. (1994). Age-related pancreatic islet changes in SpragueDawley rats. Toxicol Pathol 22, 48–55. Federal Register (1998). List of Drug Products That Have Been Withdrawn or Removed From the Market for Reasons of Safety or Effectiveness. Food and Drug Administration. 21 CFR Part 216 63, 54082–89. Go¨ggelmann, W., Bauchinger, M., Kulka, U., and Schmid, E. (1996). Genotoxicity of 4-chloro-o-toluidine in Salmonella typhimurium, human lymphocytes and V79 cells. Mutat Res 370, 39–47. Gold, L. S. (2008). Carcinogenic Potency Database. University of California– Berkeley Web site: http://potency.berkeley.edu/index.html. Gopinath, C. (1995). The predictive value of pathological findings in animal toxicity studies. J Toxicol Pathol 8, 89–100.
743
Greaves, P. (2007) Endocrine glands. In Histopathology of Preclinical Toxicity Studies, pp. 818–19. Elsevier Ltd., New York. Hard, G. C., Boorman, G. A., and Wolf, D. C. (2000). Re-evaluation of the 2-year chloroform drinking water carcinogenicity bioassay in OsborneMendel rats supports chronic renal tubule injury as the mode of action underlying the renal tumor response. Toxicol Sci 53, 237–44. Harvey P. W. (2005). Human relevance of rodent prolactin-induced nongenotoxic mammary carcinogenesis: Prolactin involvement in human breast cancer and significance for toxicology risk assessments. J Appl Toxicol 25, 179–83. Hernandez, L. G., van Steeg H., Luijten, M., and van Benthem, J. (2009). Mechanisms of non-genotoxic carcinogens and importance of a weight of evidence approach. Mutat Res 682, 94–109. Holsapple, M. P., Pitot, H. C., Cohen, S. H., Boobis, A. R., Klaunig, J. E., Pastoor, T., Dellarco, V. L. and Dragan, Y. P. (2006). Mode of action in relevance of rodent liver tumors to human cancer risk. Toxicol Sci 89, 51–56. Iatropoulos, M. J. (1983). Pathologist’s responsibility in the diagnosis of oncogenesis. Toxicol Pathol 11, 132–42. Ilnytsky, Y., Koturbash, I., Kovalchuk, O. (2009). Radiation-induced bystander effects in vivo are epigenetically regulated in a tissue-specific manner. Environ Mol Mutagen 50, 105–13. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. http://monographs.iarc.fr/ ENG/Monographs/PDFs/index.php. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. IARC Supplement 7: Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42 (1987). http://monographs.iarc.fr/ENG/Monographs/ PDFs/index.php. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. IARC Volume 50: Pharmaceutical Drugs (1990). http://monographs.iarc.fr/ENG/Monographs/PDFs/ index.php. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. IARC Volume 96: Ethanol Consumption and Ethyl Carbamate (2010). http://monographs.iarc.fr/ ENG/Monographs/PDFs/index.php. International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. IARC Volume 99: Some Aromatic Amines, Organic Dyes, and Related Exposures (2010). http:// monographs.iarc.fr/ENG/Monographs/PDFs/index.php. International Conference on Harmonisation. ICH Guideline M3 (R2): Guidance on Nonclinical Safety Studies for the Conduct of Human Clinical Trials and Marketing Authorization for Pharmaceuticals. June 2009. International Conference on Harmonisation. ICH Guideline S1A: Need for Carcinogenicity Studies of Pharmaceuticals. November 1995. International Conference on Harmonisation. ICH Guideline S1B: Testing for Carcinogenicity of Pharmaceuticals. July 1997. International Conference on Harmonisation. ICH Guideline S1C: Dose Selection for Carcinogenicity Studies. March 2008. International Conference on Harmonisation. ICH Guideline S2: Guidance on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use. March 2008. International Conference on Harmonisation. ICH Guideline S6: Preclinical Safety Evaluation of Biotechnology-Derived Pharmaceuticals. July 1997. Ishmael, J. and Dugard, P. H. (2006). A review of perchloroethylene and rat mononuclear cell leukemia. Regul Toxicol Pharmacol 145, 178–84. Jack, D., Poynter, D., and Spurling, N. W. (1983). Beta-adrenoceptor stimulants and mesovarian leiomyomas in the rat. Toxicology 27, 315–20. Jacobs, A. (2005). Prediction for 2-year carcinogenicity study results for pharmaceutical products: How are we doing? Toxicol Sci 88, 18–23. Jacobson-Kram, D. (2010a). Cancer risk assessment approaches at the FDA/ CDER: is the era of the 2-year bioassay drawing to a close? Toxicol Pathol 38, 169–170. Jacobson-Kram, D (2010b). A proposed vision on the future of carcinogenicity testing. Toxicologist 114, 138.
744
SISTARE ET AL.
Jacobson-Kram, D., Sistare, F. D., and Jacobs, A. C. (2004) Use of transgenic mice in carcinogenicity hazard assessment, Toxicol Pathol 32(S1), 49–52. Klaunig, J. E., Babich, M. A., Baetcke, K. P., Cook, J. C., Corton, J. C., David, R. M., DeLuca, J. G., Lai D. Y., McKee R. H., Peters J. M., Roberts R. A., and Fenner-Crisp, P. A. (2003). PPARa agonist-induced rodent tumors: Modes of action and human relevance, Crit Rev Toxicol 33, 655–780. Koturbash, I., Rugo, R. E., Hendricks, C. A., Loree, J., Thibault, B., Kutanzi, K., Pogribny, I., Yanch, J. C., Engelward, B. P., and Kovalchuk, O. (2006). Irradiation induces DNA damage and modulates epigenetic effectors in distant bystander tissue in vivo. Oncogene 25, 4267–75. Lee, H. K., Kim, Y.-H., and Roh, J. K. (1994). Clastogenicity of methapyrilene hydrochloride in cultured Chinese hamster cells. Mutat Res 341, 77–82. Long, G. G., Cohen I. R., Gries, C. L., Young, J. K., Francis, P. C. and Capen, C. C. (2001). Proliferative lesions of ovarian granulosa cells and reversible hormonal changes induced in rats by a selective estrogen receptor modulator. Toxicol Pathol 29, 719–26. McClain, R. M. (1989). The significance of hepatic microsomal enzyme induction and altered thyroid function in rats: Implications for thyroid gland neoplasia. Toxicol Pathol 17, 294–306. Melnick, R. L., Kohn, M. C., and Poirier, C. J. (1996). Implications for risk assessment of suggested nongenotoxic mechanisms of chemical carcinogenesis. Environ Health Perspect 104 (Suppl 1), 123–34. Muehlbock, O. (1959). Hormones in the genesis of cancer. Acta Un Int Cancr 15, 62–66. Nagaoka, T., Onodera, H., Matsushima, Y., Todate, A., Shihutani. M., Ogasawara, H., and Maekawa, A. (1990). Spontaneous uterine adenocarcinomas in aged rats and their relation to endocrine imbalance. J Cancer Res Clin Oncol 116, 623–28. National Cancer Institute (1979). Bioassay of 4-chloro-o-toluidine hydrochloride for possible carcinogenicity. http://ntp.niehs.nih.gov/ntp/ htdocs/LT_rpts/tr165.pdf. Technical Rpt Series No.165, DHEW Pub No. (NIH) 79–1721. National Toxicology Program (1986). Toxicology and carcinogenesis studies of tetrachloroethylene (perchloroethylene) in F344/N rats and B6C3F1 mice (inhalation studies). U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, NC. NTP TR No 311, NIH publication no. 86–2567. National Toxicology Program (1990). Carcinogenesis studies of trichloroethylene (without epichlorohydrin) in F344/N rats and B6C3F1 mice (gavage studies). NTP, Research Triangle Park, NC. NTP Toxicity Report Number 243, NIH Publication 96–1779. National Toxicology Program (1996). NTP Technical Report on Comparative Toxicity and Carcinogenicity Studies of o-Nitrotoluene/o-Toluidine (CAS Nos. 88–72–2 and 636–21–5) Administered in Feed to Male F344/N Rats. NIEHS, Research Triangle Park, NC. NTP Toxicity Report Number 44, NIH Publication 96–3936. Neumann, F. (1991). Early indicators for carcinogenesis in sex-hormonesensitive organs. Mutat Res 248, 341–56. Ohshima, M., Ward, J. M., Brennan, L. M. and Creasia, D. A. (1984). A sequential study of methapyrilene hydrochloride-induced liver carcinogenesis in male F344 rats. J Natl Canc Inst 72,759–68.
TOXICOLOGIC PATHOLOGY
Oosterhout, J. Van Der Laan, J., Waal, E.,Olejniczak, K. Hilgenfeld, M., Schmidt, V., and Bass, R. (1997). The utility of two rodent species in carcinogenic risk assessment of pharmaceuticals in Europe. Regul Toxicol Pharmacol 25, 6–17. Organisation for Economic Co-operation and Development (OECD) Environment, Health and Safety Publications (2009). Series on Testing and Assessment, No. 103: Detailed Review Paper on Transgenic Rodent Mutation Assays. OECD, Paris, France. Prentice, D. E., and Meikle, A. W. (1995). A review of drug-induced Leydig cell hyperplasia and neoplasia in the rat and some comparisons with man. Hum Exper Toxicol 14, 562–72. Pritchard, J. B., French, J. E., Davis, B. J., and Haseman J. K. (2003). The role of transgenic mouse models in carcinogen identification. Environ Health Perspect 111, 444–54. Reddy, M. V., Sistare, F. D., Christensen, J. S., DeLuca, J. G., Wollenberg, G. K., and DeGeorge, J. J. (2010). An evaluation of chronic six- and twelve-month rat toxicology studies as predictors of two-year tumor outcome. Vet Pathol 47, 614–29. Ross, J., and Barrowman, J. A. (1987). Effect of experimental pancreatic growth on the content of xenobiotic-metabolizing enzymes in the pancreas. Gut 28 (Suppl), 197–201. Snyder, R. D. (1998). A review and investigation into the mechanistic basis of the genotoxicity of antihistamines. Mutat Res 411, 235–48. Storer, R. D., Sistare, F. D., Reddy, M. V., and DeGeorge, J. J. (2010) An industry perspective on the utility of short-term carcinogenicity testing in transgenic mice in pharmaceutical development. Toxicol Pathol 38, 51–61. United Nations (2005). Consolidated List of Products Whose Consumption and/or Sale have been Banned, Withdrawn, Severely Restricted, or Not Approved by Governments–Pharmaceuticals. 12th issue. United Nations, New York, NY. Ward, J. M. (1980). Unique toxic lesions induced by chemical carcinogens in rodents: indications of a short-term bioassay for carcinogenesis testing of environmental chemicals. Med Hypotheses 6, 421–25. Weisburger, E. K., Russfield, A. B., Homburger, F., Weisburger, J. H., Boger, E., and Van Dongen, C. G. (1978). Testing of twenty-one environmental aromatic amines or derivatives for long-term toxicity or carcinogenicity. J Environ Path Toxicol 2, 325–56. Wilson, G. L., and Longnecker, D. S. (1999). Pancreatic toxicology. In Endocrine and Hormonal Toxicology, pp. 125–53. John Wiley & Sons, New York. World Health Organization, International Agency for Research on Cancer Web site: http://monographs.iarc.fr/. Woutersen, R. A., van Garderen-Hoetmer, A., Lamers, C. B. H. W., and Scherer, E. (1991). Early indicators of exocrine pancreas carcinogenesis produced by non-genotoxic agents. Mutat Res 248, 291–302. Wysowski, D. K., and Swartz, L. (2005) Adverse drug event surveillance and drug withdrawals in the United States, 1969–2002. Arch Intern Med 165, 1363–69. Zwicker, G. M., and Eyster, R. C. (1993). Chronic effect of corticosteroid oral treatment in rat on blood glucose and serum insulin levels, pancreatic islet morphology and immunostaining characteristics. Toxicol Pathol 21, 502–8.
For reprints and permissions queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.