Carcinogenicity testing and the evaluation of regulatory requirements for pharmaceuticals

Quantification of Hepatocellular Mitoses in a Toxicological Study in Rats Using a Convolutional Neural Network

Convolutional neural networks (CNNs) have been recognized as valuable tools for rapid quantitative analysis of morphological changes in toxicologic histopathology. We have assessed the performance of CNN-based (Halo-AI) mitotic figure detection in hepatocytes in comparison with detection by pathologists. In addition, we compared with Ki-67 and 5-bromodesoxyuridin (BrdU) immunohistochemistry labeling indices (LIs) obtained by image analysis. Tissues were from an exploratory toxicity study with a glycogen synthase kinase-3 (GSK-3) inhibitor. Our investigations revealed that (1) the CNN achieved similarly accurate but faster results than pathologists, (2) results of mitotic figure detection were comparable to Ki-67 and BrdU LIs, and (3) data from different methods were only moderately correlated. The latter is likely related to differences in the cell cycle component captured by each method. This highlights the importance of considering the differences of the available methods upon selection. Also, the pharmacology of our test item acting as a GSK-3 inhibitor potentially reduced the correlation. We conclude that hepatocyte cell proliferation assessment by CNNs can have several advantages when compared with the current gold standard: it relieves the pathologist of tedious routine tasks and contributes to standardization of results; the CNN algorithm can be shared and iteratively improved; it can be performed on routine histological slides; it does not require an additional animal experiment and in this way can contribute to animal welfare according to the 3R principles.

Survey of tumorigenic sensitivity in 6-month rasH2-Tg mice studies compared with 2-year rodent assays

The pharmacokinetic endpoint of a 25-fold increase in human exposure is one of the specified criteria for high-dose selection for 2-year carcinogenicity studies in rodents according to ICH S1C(R2). However, this criterion is not universally accepted for 6-month carcinogenicity tests in rasH2-Tg mice. To evaluate an appropriate multiple for rasH2-Tg mice, we evaluated data for 53 compounds across five categories of rasH2-Tg mouse-positive [(1) genotoxic and (2) non-genotoxic] carcinogens and rasH2-Tg mouse-negative [(3) non-genotoxic carcinogens with clear or uncertain human relevance; (4) non-genotoxic rodent-specific carcinogens; and (5) non-carcinogens], and surveyed their tumorigenic activities and high doses in rasH2-Tg mice and 2-year rodent models. Our survey indicated that area under the curve (AUC) margins (AMs) or body surface area-adjusted dose ratios (DRs) of tumorigenesis in rasH2-Tg mice to the maximum recommended human dose (MRHD) were 0.05- to 5.2-fold in 6 category (1) compounds with small differences between models and 0.2- to 47-fold in 7 category (2) including three 2-year rat study-negative compounds. Among all 53 compounds, including 40 compounds of the rasH2-Tg mouse-negative category (3), (4), and (5), no histopathologic risk factors for rodent neoplasia were induced only at doses above 50-fold AM or DR in rasH2-Tg mice except for two compounds, which induced hyperplasia and had no relationship with the tumors observed in the rasH2-Tg mouse or 2-year rodent studies. From the results of these surveys, we confirmed that exceeding a high dose level of 50-fold AM in rasH2-Tg mouse carcinogenicity studies does not appear to be of value.

A rat subchronic study transcriptional point of departure estimates a carcinogenicity study apical point of departure

Considerations of human relevance and animal use are driving research to identify new approaches to inform risk assessment of chemicals and replace guideline-based rodent carcinogenicity tests. Here, the hypothesis was tested across four agrochemicals that 1) a rat 90-day transcriptome-based BEPOD is protective of a rat carcinogenicity study and 2) a subchronic liver or kidney BEPOD would approximate a cancer bioassay apical POD derived from other organs and a rat subchronic BEPOD would approximate a mouse cancer bioassay apical POD. Using RNA sequencing and BMDExpress software, liver and/or kidney BEPOD values were generated in male rats exposed for 90 days to either Triclopyr Acid, Pronamide, Sulfoxaflor, or Fenpicoxamid. BEPOD values were compared to benchmark dose-derived apical POD values generated from rat 90-day and rodent carcinogenicity studies. Across all four agrochemicals, findings showed that a rat 90-day study BEPOD approximated the most sensitive apical POD (within 10-fold) generated from the 90-day rat study and long-term rodent carcinogenicity studies. This study supports the conclusion that a subchronic transcriptome-based BEPOD could be utilized to estimate an apical POD within a risk-based approach of chronic toxicity and carcinogenicity agrochemical assessment, abrogating the need for time- and resource-intensive rodent carcinogenicity studies and minimizing animal testing.

Moving forward in carcinogenicity assessment: Report of an EURL ECVAM/ESTIV workshop

There is an increased need to develop novel alternative approaches to the two-year rodent bioassay for the carcinogenicity assessment of substances where the rodent bioassay is still a basic requirement, as well as for those substances where animal use is banned or limited or where information gaps are identified within legislation. The current progress in this area was addressed in a EURL ECVAM- ESTIV workshop held in October 2016, in Juan les Pins. A number of initiatives were presented and discussed, including data-driven, technology-driven and pathway-driven approaches. Despite a seemingly diverse range of strategic developments, commonalities are emerging. For example, providing insight into carcinogenicity mechanisms is becoming an increasingly appreciated aspect of hazard assessment and is suggested to be the best strategy to drive new developments. Thus, now more than ever, there is a need to combine and focus efforts towards the integration of available information between sectors. Such cross-sectorial harmonisation will aid in building confidence in new approach methods leading to increased implementation and thus a decreased necessity for the two-year rodent bioassay.

Carcinogenicity Evaluation: Comparison of Tumor Data from Dual Control Groups in the Sprague–Dawley Rat

Following recent clarification in Europe that a single control group is now acceptable for rodent carcinogenicity studies, the use of dual controls may be reduced or disappear. To date, virtually nothing has been published on whether this latter situation has improved the identification of tumorigenic risk potential in these studies. In this paper, the results of 13 rat carcinogenicity studies, performed between 1991 and 2002, with 2 control groups, are presented. Although no major differences in tumor incidences between these dual control groups were found, some interstudy variation occurred. In cases where a notable difference was seen, the use of 2 control groups, as well as robust, contemporary background data, allowed an easier interpretation of findings in drug-treated groups. Thus, the continued use of dual control groups has a vital role in the assessment of tumoriogenic risk. The paper also presents an update on survival, on the range and extent of background spontaneous neoplasms, and comments on genetic drift in this commonly used rat strain.

Carcinogenicity Evaluation: Comparison of Tumor Data from Dual Control Groups in the CD–1 Mouse

Current regulatory thinking allows for the use of single control groups for rodent carcinogenicity testing although there has been a trend until recently to use dual control groups. To date, virtually nothing has been published on whether a shift from dual to single control groups will affect the identification of tumorigenic risk potential in these studies. A recent evaluation of dual control carcinogenicity data in the rat (Baldrick, Toxicol Pathol 2005, 33: 283–291) showed that although no major differences in tumor incidences between the control groups were found, some interstudy variation occurred and in cases were a notable difference was seen, the use of 2 control groups, as well as robust, contemporary background data, allowed an easier interpretation of findings in drug-treated groups. In this paper, the results of 10 mouse carcinogenicity studies, performed between 1991 and 2004, with 2 control groups, are presented. As in the rat, interstudy variation was seen and in some cases, the use of dual control groups assisted in the tumor risk assessment. Thus, the continued use of 2 control groups can have a vital role in mouse carcinogenicity studies. The paper also presents an update on survival, on the range and extent of background spontaneous neoplasms and comments on genetic drift in this commonly used mouse strain.

The use of genetically modified mice in cancer risk assessment: challenges and limitations

The use of genetically modified (GM) mice to assess carcinogenicity is playing an increasingly important role in the safety evaluation of chemicals. While progress has been made in developing and evaluating mouse models such as the Trp53⁺/⁻, Tg.AC and the rasH2, the suitability of these models as replacements for the conventional rodent cancer bioassay and for assessing human health risks remains uncertain. The objective of this research was to evaluate the use of accelerated cancer bioassays with GM mice for assessing the potential health risks associated with exposure to carcinogenic agents. We compared the published results from the GM bioassays to those obtained in the National Toxicology Program's conventional chronic mouse bioassay for their potential use in risk assessment. Our analysis indicates that the GM models are less efficient in detecting carcinogenic agents but more consistent in identifying non-carcinogenic agents. We identified several issues of concern related to the design of the accelerated bioassays (e.g., sample size, study duration, genetic stability and reproducibility) as well as pathway-dependency of effects, and different carcinogenic mechanisms operable in GM and non-GM mice. The use of the GM models for dose-response assessment is particularly problematic as these models are, at times, much more or less sensitive than the conventional rodent cancer bioassays. Thus, the existing GM mouse models may be useful for hazard identification, but will be of limited use for dose-response assessment. Hence, caution should be exercised when using GM mouse models to assess the carcinogenic risks of chemicals.

Reassessing the two-year rodent carcinogenicity bioassay: a review of the applicability to human risk and current perspectives

The 2-year rodent carcinogenicity test has been the regulatory standard for the prediction of human outcomes for exposure to industrial and agro-chemicals, food additives, pharmaceuticals and environmental pollutants for over 50 years. The extensive experience and data accumulated over that time has spurred a vigorous debate and assessment, particularly over the last 10 years, of the usefulness of this test in terms of cost and time for the information obtained. With renewed interest in the United States and globally, plus new regulations in the European Union, to reduce, refine and replace sentinel animals, this review offers the recommendation that reliance on information obtained from detailed shorter-term, 6 months rodent studies, combined with genotoxicity and chemical mode of action can realize effective prediction of human carcinogenicity instead of the classical two year rodent bioassay. The aim of carcinogenicity studies should not be on the length of time, and by obligation, number of animals expended but on the combined systemic pathophysiologic influence of a suspected chemical in determining disease. This perspective is in coordination with progressive regulatory standards and goals globally to utilize effectively resources of animal usage, time and cost for the goal of human disease predictability.

Science, politics, and health in the brave new world of pharmaceutical carcinogenic risk assessment: technical progress or cycle of regulatory capture?

The carcinogenicity (cancer-inducing potential) of pharmaceuticals is an important risk factor for health when considering whether thousands of patients on drug trials or millions/billions of consumers in the marketplace should be exposed to a new drug. Drawing on fieldwork involving over 50 interviews and documentary research spanning 2002–2010 in Europe and the US, and on regulatory capture theory, this article investigates how the techno-regulatory standards for carcinogenicity testing of pharmaceuticals have altered since 1998. It focuses on the replacement of long-term carcinogenicity tests in rodents (especially mice) with shorter-term tests involving genetically-engineered mice (GEM). Based on evidence regarding financial/organizational control, methodological design, and interpretation of the validation and application of these new GEM tests, it is argued that regulatory agencies permitted the drug industry to shape such validation and application in ways that prioritized commercial interests over the need to protect public health. Boundary-work enabling industry scientists to define some standards of public-health policy facilitated such capture. However, as the scientific credibility of GEM tests as tools to protect public health by screening out carcinogens became inescapably problematic, a regulatory resurgence, impelled by reputational concerns, exercised more control over industry’s construction and use of the tests, The extensive problems with GEM tests as public-health protective regulatory science raises the spectre that alterations to pharmaceutical carcinogenicity-testing standards since the 1990s may have been boundary-work in which the political project of decreasing the chance that companies’ products are defined as carcinogenic has masqueraded as techno-science.

Testing strategies for embryo-fetal toxicity of human pharmaceuticals. Animal models vs. in vitro approaches: a workshop report

Reproductive toxicity testing is characterized by high animal use. For registration of pharmaceutical compounds, developmental toxicity studies are usually conducted in both rat and rabbits. Efforts have been underway for a long time to design alternatives to animal use. Implementation has lagged, partly because of uncertainties about the applicability domain of the alternatives. The reproductive cycle is complex and not all mechanisms of development can be mimicked in vitro. Therefore, efforts are underway to characterize the available alternative tests with regard to the mechanism of action they include. One alternative test is the mouse embryonic stem cell test (EST), which has been studied since the late 1990s. It is a genuine 3R "alternative" assay as it is essentially animal-free. A meeting was held to review the state-of-the-art of various in vitro models for prediction of developmental toxicity. Although the predictivity of individual assays is improving, a battery of several assays is likely to have even higher predictivity, which is necessary for regulatory acceptance. The workshop concluded that an important first step is a thorough survey of the existing rat and rabbit studies, to fully characterize the frequency of responses and the types of effects seen. At the same time, it is important to continue the optimization of in vitro assays. As more experience accumulates, the optimal conditions, assay structure, and applicability of the alternative assays are expected to emerge.

Power, expertise and the limits of representative democracy: genetics as scientific progress or political legitimation in carcinogenic risk assessment of pharmaceuticals?

In modern 'representative' democratic states, the legitimacy of governments' actions rests on their publicly declared commitment to protect the interests of their citizens. Regarding the pharmaceutical sector in most democracies, new drug products are developed and marketed by a capitalist industry, whose member firms, via shareholders, have commercial interests in expanding product sales. In those democracies, states have established government agencies to regulate the pharmaceutical industry on behalf of citizens. State legislatures, such as the US Congress and European Parliaments, have charged government drug regulatory agencies with the legal responsibility to protect public health. Yet, this paper argues that government drug regulatory agencies in the EU, Japan, and USA have permitted the pharmaceutical industry to reshape the regulatory guidance for carcinogenic risk assessment of pharmaceuticals in ways that are not techno-scientifically defensible as bases for improved, or even equivalent, protection of public health, compared with the previous techno-regulatory standards. By adopting the industry's agenda of streamlining carcinogenicity testing in order to accelerate drug development and regulatory review, it is contended that these regulatory agencies have allowed the techno-regulatory standards for carcinogenic risk assessment to be loosened in ways that are presented as scientific progress resulting from new genetics, but for which there is little evidence of progress in public health protection.

An analysis of pharmaceutical experience with decades of rat carcinogenicity testing: support for a proposal to modify current regulatory guidelines

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 two-year testing on new pharmaceuticals without compromise to patient safety.

Structural diversity of biologically interesting datasets: a scaffold analysis approach

Background

The recent public availability of the human metabolome and natural product datasets has revitalized "metabolite-likeness" and "natural product-likeness" as a drug design concept to design lead libraries targeting specific pathways. Many reports have analyzed the physicochemical property space of biologically important datasets, with only a few comprehensively characterizing the scaffold diversity in public datasets of biological interest. With large collections of high quality public data currently available, we carried out a comparative analysis of current day leads with other biologically relevant datasets.

Results

In this study, we note a two-fold enrichment of metabolite scaffolds in drug dataset (42%) as compared to currently used lead libraries (23%). We also note that only a small percentage (5%) of natural product scaffolds space is shared by the lead dataset. We have identified specific scaffolds that are present in metabolites and natural products, with close counterparts in the drugs, but are missing in the lead dataset. To determine the distribution of compounds in physicochemical property space we analyzed the molecular polar surface area, the molecular solubility, the number of rings and the number of rotatable bonds in addition to four well-known Lipinski properties. Here, we note that, with only few exceptions, most of the drugs follow Lipinski's rule. The average values of the molecular polar surface area and the molecular solubility in metabolites is the highest while the number of rings is the lowest. In addition, we note that natural products contain the maximum number of rings and the rotatable bonds than any other dataset under consideration.

Conclusions

Currently used lead libraries make little use of the metabolites and natural products scaffold space. We believe that metabolites and natural products are recognized by at least one protein in the biosphere therefore, sampling the fragment and scaffold space of these compounds, along with the knowledge of distribution in physicochemical property space, can result in better lead libraries. Hence, we recommend the greater use of metabolites and natural products while designing lead libraries. Nevertheless, metabolites have a limited distribution in chemical space that limits the usage of metabolites in library design.

Evaluation of carcinogenicity studies of medicinal products for human use authorised via the European centralised procedure (1995-2009)

Carcinogenicity data of medicinal products for human use that have been authorised via the European centralised procedure (CP) between 1995 and 2009 were evaluated. Carcinogenicity data, either from long-term rodent carcinogenicity studies, transgenic mouse studies or repeat-dose toxicity studies were available for 144 active substances contained in 159 medicinal products. Out of these compounds, 94 (65%) were positive in at least one long-term carcinogenicity study or in repeat-dose toxicity studies. Fifty compounds (35%) showed no evidence of a carcinogenic potential. Out of the 94 compounds with positive findings in either carcinogenicity or repeat-dose toxicity studies, 33 were positive in both mice and rats, 40 were positive in rats only, and 21 were positive exclusively in mice. Long-term carcinogenicity studies in two rodent species were available for 116 compounds. Data from one long-term carcinogenicity study in rats and a transgenic mouse model were available for eight compounds. For 13 compounds, carcinogenicity data were generated in only one rodent species. One compound was exclusively tested in a transgenic mouse model. Six compounds were tumourigenic in repeat-dose toxicity studies in rats. The majority of tumour findings observed in rodent carcinogenicity studies were considered not to be relevant for humans, either due to a rodent-specific mechanism of carcinogenicity, a high safety margin between exposures at the NOAEL (No Observed Adverse Effect Level) in rodents and recommended therapeutic doses in humans, or based on historical control data, a small effect size and lack of dose-response relationship and tumours typically observed in rodent strains used, or were considered not to be relevant for humans based on literature and clinical data or likely differences in metabolism/local concentrations between rodents and humans. Due to the high number of rodent tumour findings with unlikely relevance for humans, the value of the currently used testing strategy for carcinogenicity appears questionable. A revision of the carcinogenicity testing paradigm is warranted.

Marine collagen peptides prepared from chum salmon (Oncorhynchus keta) skin extend the life span and inhibit spontaneous tumor incidence in Sprague-Dawley Rats

To observe the effects of marine collagen peptides (MCPs) prepared from chum salmon (Oncorhynchus keta) skin on life span and spontaneous tumor incidence, Sprague-Dawley rats were fed diets supplemented with MCP at concentrations of 0%, 2.25%, 4.5%, and 9% (wt/wt) from the age of 4 weeks until natural death. There were 40 rats in each group (male:female ratio = 1:1). The results showed that the MCP did not significantly influence body weight or food consumption of rats of either sex throughout the life span; it did dose-dependently inhibit the age-related decrease in the activities of antioxidant enzymes and the age-related increase in the levels of lipid peroxidation product in both sexes. MCP notably increased the mean life span, the life span of the last 30% of the survivors, and the maximal life span; it decreased overall spontaneous tumor incidence of both sexes with significance in the 4.5% and 9% MCP-treated male groups and 9% MCP-treated female group. Compared to the control group, the incidence of death from tumors was decreased in MCP groups in comparison with the control group of both sexes. Therefore, we concluded that MCPs dose-dependently increase life span and decrease spontaneous tumor incidence in Sprague-Dawley rats. Moreover, the antioxidative property of MCPs may be responsible for the increased life span and protection against tumor development.

An Evaluation of Chronic 6- and 12-Month Rat Toxicology Studies as Predictors of 2-Year Tumor Outcome

Chronic 6- and 12-month rat toxicology studies were evaluated for their ability to predict tumor outcome in 2-year rat carcinogenicity studies for 80 pharmaceuticals from commercial and Merck databases. The data consisted of 62 (6-month) and 54 (12-month) studies, which included 30 rat carcinogens and 50 noncarcinogens in 2-year studies. The histopathology findings considered as evidence of potential preneoplasia in the chronic studies were hyperplasia, cellular hypertrophy, and atypical cellular foci. The authors hypothesized that a whole animal–based approach should be taken, wherein (1) evidence of potential preneoplasia in any tissue may serve as a sensitive predictor of tumor outcome in any tissue in the whole animal and not necessarily the same tissue and (2) the absence of evidence for potential preneoplasia in all tissues may serve as a strong negative predictor of tumor outcome in any tissue. Based on this whole animal approach, 25 of 30 rat carcinogens showed histopathologic signals in chronic toxicology studies, yielding a test sensitivity of 83%. The negative predictivity of the absence of histopathology findings in chronic toxicology studies of 50 nontumorigenic compounds was 88%. The value of the extra 6-month treatment was not apparent. The 5 false negatives (negative chronic studies but positive 2-year studies) are for marketed drugs approved for non-life-threatening conditions and associated with rat-specific mechanisms. The absence of preneoplasia in the whole animal is a reliable predictor of negative tumor outcome in 2-year rat studies, and approximately 50% rat carcinogenicity studies could be eliminated for the 80 pharmaceuticals examined, with no risk to humans and with a substantial reduction in animal usage and drug development time.

Predictive models for carcinogenicity and mutagenicity: frameworks, state-of-the-art, and perspectives

Mutagenicity and carcinogenicity are endpoints of major environmental and regulatory concern. These endpoints are also important targets for development of alternative methods for screening and prediction due to the large number of chemicals of potential concern and the tremendous cost (in time, money, animals) of rodent carcinogenicity bioassays. Both mutagenicity and carcinogenicity involve complex, cellular processes that are only partially understood. Advances in technologies and generation of new data will permit a much deeper understanding. In silico methods for predicting mutagenicity and rodent carcinogenicity based on chemical structural features, along with current mutagenicity and carcinogenicity data sets, have performed well for local prediction (i.e., within specific chemical classes), but are less successful for global prediction (i.e., for a broad range of chemicals). The predictivity of in silico methods can be improved by improving the quality of the data base and endpoints used for modelling. In particular, in vitro assays for clastogenicity need to be improved to reduce false positives (relative to rodent carcinogenicity) and to detect compounds that do not interact directly with DNA or have epigenetic activities. New assays emerging to complement or replace some of the standard assays include Vitotox, GreenScreenGC, and RadarScreen. The needs of industry and regulators to assess thousands of compounds necessitate the development of high-throughput assays combined with innovative data-mining and in silico methods. Various initiatives in this regard have begun, including CAESAR, OSIRIS, CHEMOMENTUM, CHEMPREDICT, OpenTox, EPAA, and ToxCast. In silico methods can be used for priority setting, mechanistic studies, and to estimate potency. Ultimately, such efforts should lead to improvements in application of in silico methods for predicting carcinogenicity to assist industry and regulators and to enhance protection of public health.

Prediction of rodent carcinogenic potential of naturally occurring chemicals in the human diet using high-throughput QSAR predictive modeling

Consistent with the U.S. Food and Drug Administration (FDA) Critical Path Initiative, predictive toxicology software programs employing quantitative structure-activity relationship (QSAR) models are currently under evaluation for regulatory risk assessment and scientific decision support for highly sensitive endpoints such as carcinogenicity, mutagenicity and reproductive toxicity. At the FDA's Center for Food Safety and Applied Nutrition's Office of Food Additive Safety and the Center for Drug Evaluation and Research's Informatics and Computational Safety Analysis Staff (ICSAS), the use of computational SAR tools for both qualitative and quantitative risk assessment applications are being developed and evaluated. One tool of current interest is MDL-QSAR predictive discriminant analysis modeling of rodent carcinogenicity, which has been previously evaluated for pharmaceutical applications by the FDA ICSAS. The study described in this paper aims to evaluate the utility of this software to estimate the carcinogenic potential of small, organic, naturally occurring chemicals found in the human diet. In addition, a group of 19 known synthetic dietary constituents that were positive in rodent carcinogenicity studies served as a control group. In the test group of naturally occurring chemicals, 101 were found to be suitable for predictive modeling using this software's discriminant analysis modeling approach. Predictions performed on these compounds were compared to published experimental evidence of each compound's carcinogenic potential. Experimental evidence included relevant toxicological studies such as rodent cancer bioassays, rodent anti-carcinogenicity studies, genotoxic studies, and the presence of chemical structural alerts. Statistical indices of predictive performance were calculated to assess the utility of the predictive modeling method. Results revealed good predictive performance using this software's rodent carcinogenicity module of over 1200 chemicals, comprised primarily of pharmaceutical, industrial and some natural products developed under an FDA-MDL cooperative research and development agreement (CRADA). The predictive performance for this group of dietary natural products and the control group was 97% sensitivity and 80% concordance. Specificity was marginal at 53%. This study finds that the in silico QSAR analysis employing this software's rodent carcinogenicity database is capable of identifying the rodent carcinogenic potential of naturally occurring organic molecules found in the human diet with a high degree of sensitivity. It is the first study to demonstrate successful QSAR predictive modeling of naturally occurring carcinogens found in the human diet using an external validation test. Further test validation of this software and expansion of the training data set for dietary chemicals will help to support the future use of such QSAR methods for screening and prioritizing the risk of dietary chemicals when actual animal data are inadequate, equivocal, or absent.

In silicoapproaches to explore toxicity end points: issues and concerns for estimating human health effects

The European Chemicals Bureau and the Organisation for Economic Cooperation and Development are currently compiling a sanctioned list of quantitative structure-activity relationship (QSAR) risk assessment models and data sets to predict the physiological properties, environmental fate, ecological effects and human health effects of new and existing chemicals in commerce in the European Union. This action implements the technical requirements of the European Commission's Registration, Evaluation and Authorisation of Chemicals legislation. The goal is to identify a battery of QSARs that can furnish rapid, reliable and cost-effective decision support information for regulatory decisions that can substitute for results from animal studies. This report discusses issues and concerns that need to be addressed when selecting QSARs to predict human health effect end points.

Comparative overview of current international strategies and guidelines for genetic toxicology testing for regulatory purposes

National and international regulatory agencies historically have used genotoxicity information as part of a weight-of-evidence approach to evaluate potential human carcinogenicity. Additionally, some agencies consider heritable mutation a regulatory endpoint. Furthermore, genotoxicity has the potential to contribute to other adverse health conditions. This article provides a comparative overview of the testing strategies used by regulatory agencies throughout the world. Despite minor variations in details, the genotoxicity test schemes for most regulatory entities generally comprise three tests: a bacterial gene mutation assay, an in vitro mammalian cell assay for gene mutation and/or chromosome aberrations, and often an in vivo assay for chromosomal effects. In some cases, fewer than these three tests are required. In other cases, when exposure data, structure-activity considerations, or other factors warrant, even chemicals negative in the three baseline tests may be subject to additional testing. If genotoxicity is identified by the baseline screening tests, assessment of the ability of the chemical to interact with DNA in the gonad may be required. This may apply regardless of whether or not a cancer bioassay has been triggered. Mutagens positive in second stage gonadal assay(s) may be tested in third stage in vivo rodent tests to provide data for a quantitative risk assessment. In all testing, theutilization of internationally-recognized protocols, where they exist, is advisable, although not in all instances required. When testing for regulatory purposes, it is advisable to verify the testing program with the specific regulatory body or bodies responsible forregulatory oversight before beginning testing.

Preclinical Development of New Drugs that Enhance Thyroid Hormone Metabolism and Clearance: Inadequacy of Using Rats as an Animal Model for Predicting Human Risks in an IND and NDA

New drugs that enhance metabolism or clearance of thyroid hormones in rats often trigger a sequence of toxicity events during chronic administration: reduction of thyroxine, elevation of thyroid-stimulating hormone (TSH) levels, and thyroid gland hyperfunction/growth. Hepatocellular hypertrophy and thyroid follicular hyperplasia are often observed with increased liver and thyroid organ weights. This unique toxicity profile seems to be species-specific because the thyroxine in rodents is metabolized rapidly, without thyroid hormone-binding globulin that serves as a reserve, as in humans. Thus, elevations of TSH were not reported in humans for drugs such as delavirdine, fluvastatin, nicardipine, phenobarbital, simvastatin, and spironolactone, all of which produce thyroid hyperplasia or tumors in rats. Further, the human thyroid is less sensitive to prolonged TSH stimulation than that of the rat (eg, endemic goiter patients with high TSH due to iodine deficiency do not develop thyroid cancer). In view of the species difference in sensitivity of the thyroid between rodents and humans, using the rat as an animal model to explore target organs of toxicity for a new drug that significantly enhances thyroid hormone metabolism/clearance and increases TSH levels would not be adequate. In this case, a compromised and dysfunctional hypothalamo-pituitary-thyroid system would confound the toxicity profile explored in preclinical toxicity testing and render the model an inadequate risk predictor for the new drug in humans. Under such conditions, IND and NDA sponsors of drugs exhibiting this activity profile should be encouraged to use alternative animal species for toxicity exploration to provide a more meaningful human risk prediction.

An analysis of genetic toxicity, reproductive and developmental toxicity, and carcinogenicity data: I. Identification of carcinogens using surrogate endpoints

A retrospective analysis of standard genetic toxicity (genetox) tests, reproductive and developmental toxicity (reprotox) studies, and rodent carcinogenicity bioassays (rcbioassay) was performed to identify the genetox and reprotox endpoints whose results best correlate with rcbioassay observations. A database of 7205 chemicals with genetox (n = 4961), reprotox (n = 2173), and rcbioassay (n = 1442) toxicity data was constructed; 1112 of the chemicals have both genetox and rcbioassay data and 721 chemicals have both reprotox and rcbioassay data. This study differed from previous studies by using conservative weight of evidence criteria to classify chemical carcinogens, data from 63 genetox and reprotox toxicological endpoints, and a new statistical parameter of correlation indicator (CI, the average of specificity and positive predictivity) to identify good surrogate endpoints for predicting carcinogenicity. Among 63 endpoints, results revealed that carcinogenicity was well correlated with certain tests for gene mutation (n = 8), in vivo clastogenicity (n = 2), unscheduled DNA synthesis assay (n = 1), and reprotox (n = 3). The current FDA regulatory battery of four genetox tests used to predict carcinogenicity includes two tests with good correlation (gene mutation in Salmonella and in vivo micronucleus) and two tests with poor correlation (mouse lymphoma gene mutation and in vitro chromosome aberrations) by our criteria.

Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens I. Sensitivity, specificity and relative predictivity

The performance of a battery of three of the most commonly used in vitro genotoxicity tests--Ames+mouse lymphoma assay (MLA)+in vitro micronucleus (MN) or chromosomal aberrations (CA) test--has been evaluated for its ability to discriminate rodent carcinogens and non-carcinogens, from a large database of over 700 chemicals compiled from the CPDB ("Gold"), NTP, IARC and other publications. We re-evaluated many (113 MLA and 30 CA) previously published genotoxicity results in order to categorise the performance of these assays using the response categories we established. The sensitivity of the three-test battery was high. Of the 553 carcinogens for which there were valid genotoxicity data, 93% of the rodent carcinogens evaluated in at least one assay gave positive results in at least one of the three tests. Combinations of two and three test systems had greater sensitivity than individual tests resulting in sensitivities of around 90% or more, depending on test combination. Only 19 carcinogens (out of 206 tested in all three tests, considering CA and MN as alternatives) gave consistently negative results in a full three-test battery. Most were either carcinogenic via a non-genotoxic mechanism (liver enzyme inducers, peroxisome proliferators, hormonal carcinogens) considered not necessarily relevant for humans, or were extremely weak (presumed) genotoxic carcinogens (e.g. N-nitrosodiphenylamine). Two carcinogens (5-chloro-o-toluidine, 1,1,2,2-tetrachloroethane) may have a genotoxic element to their carcinogenicity and may have been expected to produce positive results somewhere in the battery. We identified 183 chemicals that were non-carcinogenic after testing in both male and female rats and mice. There were genotoxicity data on 177 of these. The specificity of the Ames test was reasonable (73.9%), but all mammalian cell tests had very low specificity (i.e. below 45%), and this declined to extremely low levels in combinations of two and three test systems. When all three tests were performed, 75-95% of non-carcinogens gave positive (i.e. false positive) results in at least one test in the battery. The extremely low specificity highlights the importance of understanding the mechanism by which genotoxicity may be induced (whether it is relevant for the whole animal or human) and using weight of evidence approaches to assess the carcinogenic risk from a positive genotoxicity signal. It also highlights deficiencies in the current prediction from and understanding of such in vitro results for the in vivo situation. It may even signal the need for either a reassessment of the conditions and criteria for positive results (cytotoxicity, solubility, etc.) or the development and use of a completely new set of in vitro tests (e.g. mutation in transgenic cell lines, systems with inherent metabolic activity avoiding the use of S9, measurement of genetic changes in more cancer-relevant genes or hotspots of genes, etc.). It was very difficult to assess the performance of the in vitro MN test, particularly in combination with other assays, because the published database for this assay is relatively small at this time. The specificity values for the in vitro MN assay may improve if data from a larger proportion of the known non-carcinogens becomes available, and a larger published database of results with the MN assay is urgently needed if this test is to be appreciated for regulatory use. However, specificity levels of <50% will still be unacceptable. Despite these issues, by adopting a relative predictivity (RP) measure (ratio of real:false results), it was possible to establish that positive results in all three tests indicate the chemical is greater than three times more likely to be a rodent carcinogen than a non-carcinogen. Likewise, negative results in all three tests indicate the chemical is greater than two times more likely to be a rodent non-carcinogen than a carcinogen. This RP measure is considered a useful tool for industry to assess the likelihood of a chemical possessing carcinogenic potential from batteries of positive or negative results.

Estimating the safe starting dose in phase I clinical trials and no observed effect level based on QSAR modeling of the human maximum recommended daily dose

Estimating the maximum recommended starting dose (MRSD) of a pharmaceutical for phase I human clinical trials and the no observed effect level (NOEL) for non-pharmaceuticals is currently based exclusively on an extrapolation of the results of animal toxicity studies. This process is inexact and requires the results of toxicity studies in multiple species (rat, dog, and monkey) to identify the no observed adverse effect level (NOAEL) and most sensitive test species. Multiple uncertainty (safety) factors are also necessary to compensate for incompatibility and uncertainty underlying the extrapolation of animal toxicity to humans. The maximum recommended daily dose for pharmaceuticals (MRDD) is empirically derived from human clinical trials. The MRDD is an estimated upper dose limit beyond which a drug's efficacy is not increased and/or undesirable adverse effects begin to outweigh beneficial effects. The MRDD is essentially equivalent to the NOAEL in humans, a dose beyond which adverse (toxicological) or undesirable pharmacological effects are observed. The NOAEL in test animals is currently used to estimate the safe starting dose in human clinical trials. MDL QSAR predictive modeling of the human MRDD may provide a better, simpler and more relevant estimation of the MRSD for pharmaceuticals and the toxic dose threshold of chemicals in humans than current animal extrapolation based risk assessment models and may be a useful addition to current methods. A database of the MRDD for over 1300 pharmaceuticals was compiled and modeled using MDL QSAR software and E-state and connectivity topological descriptors. MDL QSAR MRDD models were found to have good predictive performance with 74-78% of predicted MRDD values for 120 internal and 160 external validation compounds falling within a range of +/-10-fold the actual MRDD value. The predicted MRDD can be used to estimate the MRSD for pharmaceuticals in phase I clinical trials with the addition of a 10-fold safety factor. For non-pharmaceutical chemicals any compound-related effect can be considered an undesirable and adverse toxicological effect and the predicted MRDD can be used to estimate the NOEL with the addition of an appropriate safety factor.

Current in vivo assays for cutaneous toxicity: local and systemic toxicity testing

This review provides an update on current (June 2004) in vivo assays for evaluations of local and systemic cutaneous toxicity. As the science of toxicology and safety assessment grows and evolves, an increasing number of in vitro assays and screening procedures designed to minimize the use of animals in cutaneous toxicity evaluations have been validated or are currently in development. However, in some cases, it remains necessary to evaluate the effects of new drugs or chemicals in an animal model in order to protect human health. Current regulatory requirements and study designs for local and systemic in vivo cutaneous toxicity evaluations, as well as discussions of procedures unique to cutaneous toxicity evaluation are presented in this review.

Building an organ-specific carcinogenic database for SAR analyses

FDA reviewers need a means to rapidly predict organ-specific carcinogenicity to aid in evaluating new chemicals submitted for approval. This research addressed the building of a database to use in developing a predictive model for such an application based on structure-activity relationships (SAR). The Internet availability of the Carcinogenic Potency Database (CPDB) provided a solid foundation on which to base such a model. The addition of molecular structures to the CPDB provided the extra ingredient necessary for SAR analyses. However, the CPDB had to be compressed from a multirecord to a single record per chemical database; multiple records representing each gender, species, route of administration, and organ-specific toxicity had to be summarized into a single record for each study. Multiple studies on a single chemical had to be further reduced based on a hierarchical scheme. Structural cleanup involved removal of all chemicals that would impede the accurate generation of SAR type descriptors from commercial software programs; that is, inorganic chemicals, mixtures, and organometallics were removed. Counterions such as Na, K, sulfates, hydrates, and salts were also removed for structural consistency. Structural modification sometimes resulted in duplicate records that also had to be reduced to a single record based on the hierarchical scheme. The modified database containing 999 chemicals was evaluated for liver-specific carcinogenicity using a variety of analysis techniques. These preliminary analyses all yielded approximately the same results with an overall predictability of about 63%, which was comprised of a sensitivity of about 30% and a specificity of about 77%.

Prediction of Chemical Carcinogenicity from Molecular Structure

Carcinogens represent a serious threat to human health. In vivo determination of carcinogenicity is time-consuming and expensive, thus in silico models to predict chemical carcinogenicity are highly desirable for virtual screening of compound libraries of both pharmaceutically and other commercially interesting molecules. In the present study, a PLS-DA (partial least squares discriminant analysis) model was developed to predict carcinogenicities in each of four rodent models: male mouse (MM), female mouse (FM), male rat (MR), and female rat (FR). The data set that was used contained over 520 compounds from both the NTP and the FDA databases. All the models were built from the same molecular descriptor system, which is based on atom typing [Sun, H. J. Chem. Inf. Comput. Sci. 2004, 44, 748-757], enabling the comparison of atomic contributions to carcinogenicity with respect to species and gender. Using four components, the models were able to achieve excellent fitting and prediction, with r(2) = 0.987 and q(2) = 0.944 for MM, r(2) = 0.985 and q(2) = 0.950 for FM, r(2) = 0.989 and q(2) = 0.962 for MR, and r(2) = 0.990 and q(2) = 0.965 for FR. The models were further validated by response permutation testing and external validation, and the results indicated that the models were both statistically significant and predictive. Variable influence on projection (VIP) analysis identified the key atom types and fragments that contributed to carcinogenicities and response differences across species and gender.

Predicting the carcinogenic potential of pharmaceuticals in rodents using molecular structural similarity and E-state indices

MDL QSAR (formerly SciVision QSAR IS) software is one of the several software systems under evaluation by the Informatics and Computational Safety Analysis Staff (ICSAS) of the FDA Center for Drug Evaluation and Research for regulatory and scientific decision support applications. MDL QSAR software contains an integrated set of tools for similarity searching, compound clustering, and modeling molecular structure related parameters that includes 240 electrotopological E-state, connectivity, and other descriptors. These molecular descriptors can be statistically correlated with toxicological or biological endpoints. The goal of this research was to evaluate the feasibility of using MDL QSAR software to develop structure-activity relationship (SAR) models that can be used to predict the carcinogenic potential of pharmaceuticals and organic chemicals. A validation study of 108 compounds that include 86 pharmaceuticals and 22 chemicals that were not present in a control rodent carcinogenicity data set of 1275 compounds demonstrated that MDL QSAR models had excellent coverage (93%) and good sensitivity (72%) and specificity (72%) for rodent carcinogenicity. The software correctly predicted 72% of non-carcinogenic compounds and compounds with carcinogenic findings. E-state descriptors contributed to more than half of the SAR models used to predict carcinogenic activity. We believe that electrotopological E-state descriptors and QSAR IS (MDL QSAR) software are promising new in silico approaches for modeling and predicting rodent carcinogenicity and may have application for other toxicological endpoints.

Mechanism of clobazam-induced thyroidal oncogenesis in male rats

In order to elucidate the mechanisms by which long-term treatment with clobazam (CLB), 1,5-benzodiazepine, induces thyroid follicular cell tumors in male rats, male Sprague-Dawley (SD) rats were treated orally with 400 mg/kg of CLB for up to 4 weeks, and the contribution of feedback through elevated thyroid stimulating hormone (TSH) was investigated. Measurements taken after 1, 2, and 4 weeks of treatment revealed that thyroxine (T4)-UDP-glucuronosyltransferase (T4-UDPGT) activity was higher than that of untreated animals. This change was accompanied by increase in liver weights and centrilobular hepatocyte hypertrophy. In addition, plasma total triiodothyronine (T3) and T4 levels were lower than in the untreated rats when measured after 1 week of treatment. However, a high plasma TSH level was sustained throughout the 4-week treatment. Thyroid follicular cell hypertrophy began after 1 week of treatment, followed by increased thyroid weight after 2 weeks. Clearance of exogenous [125I] T4 from the blood of treated rats, determined after 4 weeks of treatment, was significantly faster than that in untreated rats, whereas iodine uptake and organification in the thyroid glands were not affected. These results suggest that CLB increases hepatic T4-UDPGT activity leading to acceleration of T4-clearance, which results in decreased plasma thyroidal hormones followed by compensatory increase of TSH biosynthesis and secretion. Chronic high levels of TSH would exert a continuous growth pressure on the thyroid, under which hypertrophic follicular cells can ultimately progress to frank neoplasms.

Selection of drugs to test the specificity of the Tg.AC assay by screening for induction of the gadd153 promoter in vitro

Short-term assays for carcinogenicity testing of chemicals that use transgenic mice designed to have altered expression of genes mechanistically relevant to carcinogenesis are attractive alternatives to two-year dosing studies in rodents. The models that have been the received the greatest level of performance evaluation include p53(+/-), rasH2, Xpa/p53(+/-), and Tg.AC mice. For use of these models in a regulatory setting to evaluate the carcinogenic potential of pharmaceuticals, it is important to establish an assurance of assay specificity and positive predictivity based on studies using drugs with a wide spectrum of pharmacologic activity. For this purpose, 99 noncarcinogenic drugs were prioritized based on their activity in an in vitro induction assay correlative with a positive response in the Tg.AC assay (induction of the gadd153 promoter in HepG2 cells). Activities in two assays less predictive of Tg.AC activity (induction of c-fos and zeta-globin gene promoters) were also measured. Nine percent of the screened drugs induced the gadd153 promoter by at least fourfold. Several criteria were used to select candidates for subsequent in vivo testing in the Tg.AC assay: (1) sufficient drug solubility in appropriate skin paint vehicles to elicit systemic toxicity, (2) the level of induction of the gadd153 promoter by the drug, (3) the in vitro potency of the drug, and (4) the cost of the drug required for a 6-month study. Based on these criteria, amiloride, dipyridamole, and pyrimethamine were selected from 99 rodent noncarcinogens in a drug database for testing the specificity of the Tg.AC assay.

Reshaping the carcinogenic risk assessment of medicines: international harmonisation for drug safety, industry/regulator efficiency or both?

The most significant institutional entity involved in the harmonisation of drug testing standards worldwide is the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH), which comprises the three pharmaceutical industry associations and regulatory agencies of the EU, US and Japan. It is often claimed that such harmonisation will both accelerate the development and approval of new drugs and preserve safety standards, if not strengthen safety regimes. Drawing on extensive documentary research and interviews, this paper systematically examines whether the efforts by the ICH to improve industrial and regulatory efficiency by harmonising drug testing requirements is likely to raise, maintain or compromise safety standards in carcinogenic risk assessment of pharmaceuticals. The evidence suggests that, in the field of carcinogenicity testing, the ICH management of international harmonisation of medicines regulation is not achieving simultaneous improvements in safety standards and acceleration of drug development. Rather, the latter is being achieved at the expense of the former. Indeed, the ICH may be converting permissive regulatory practices of the past into new scientific standards for the future. These findings are significant as many expert scientific advisers to drug regulatory agencies seem to have accepted uncritically the conclusions reached by the ICH, which may affect a potential patient population of half a billion and tens of thousands of clinical trials.

The science and politics of medicines control

Drug development and regulation are often presented as purely matters of technical science. In this paper it is argued that, in principle, toxicology, clinical pharmacology and pharmacovigilance in drug testing and regulation are necessarily a combination of science and politics. This has important implications for how one attempts to make progress in drug regulation, such as in interpreting technical evidence and in the setting of regulatory standards with which evidence should be evaluated. In practice, drug testing and regulation are shown to be hybrids of science and politics. Moreover, drawing on existing empirical evidence, it is suggested that this mixture currently, and for some time, has had the wrong ingredients for optimal drug safety and public health outcomes. For example, too often the balance of the scientific doubts about drug safety are weighed to the interests of manufacturers rather than to those of patients and public health, while some scientific standards with which drug safety is to be interpreted are being reshaped in ways that give insufficient priority to the protection of public health. Finally, it is proposed that: drug regulation should include comparative efficacy testing; regulatory agencies should conduct some key tests, charging the costs to industry and without duplication; and the regulatory system should be less secretive and more accountable to public scrutiny. Greater efforts should be made to eliminate experts' conflicts of interest within the regulatory process.

Tumorigenicity of chloral hydrate, trichloroacetic acid, trichloroethanol, malondialdehyde, 4-hydroxy-2-nonenal, crotonaldehyde, and acrolein in the B6C3F(1) neonatal mouse

The tumorigenicity of chloral hydrate (CH), trichloroacetic acid (TCA), trichloroethanol (TCE), malondialdehyde (MDA), crotonaldehyde, acrolein, and 4-hydroxy-2-nonenal (HNE) was tested in the B6C3F(1) neonatal mouse. Mice were administered i.p. injections of CH (1000, 2000, 2500, and 5000 nmol per animal), TCA (1000 and 2000 nmol), TCE (1000 and 2000 nmol), MDA (1500 and 3000 nmol), crotonaldehyde (1500 and 3000 nmol), acrolein (75 and 150 nmol), and HNE (750 and 1500 nmol) at 8 and 15 days of age. At 12 months, only male mice treated with the positive control chemicals, 4-aminobiphenyl (500 and 1000 nmol) and benzo[a]pyrene (150 and 300 nmol), had incidences of tumors in the liver significantly higher than the solvent control. Additional male mice were dosed as described above and their livers were excised at 24, 48 h, and 7 days after the final dose. Liver DNA was isolated and analyzed by 32P-postlabeling/high-performance liquid chromatography (HPLC) and HPLC/electrochemical detection for MDA-derived adduct (M(1)G) and 8-oxo-2'-deoxyguanosine (8-OHdG) formation, respectively. At 24 and 48 h after the final dose, CH- and TCA-treated mice exhibited significantly higher M(1)G levels than the controls. 8-OHdG formation was also induced by CH, TCA, and MDA. These results suggest that under these experimental conditions the B6C3F(1) neonatal mouse is not sensitive to carcinogens that induce an increase in endogenous DNA adduct formation through lipid peroxidation or oxidative stress.

Toxicity and carcinogenicity studies of chlorpromazine hydrochloride and p-cresidine in the p53 heterozygous mouse model

The carcinogenic potential of chlorpromazine hydrochloride, a psychotropic agent, was assessed in the p53 heterozygous mouse assay. In a 4-week dose range finding study in p53 wild-type mice, doses of 20,40, 60, and 80 mg/kg were poorly tolerated because of mortality secondary to the severe sedative and hypotensive effects of chlorpromazine. Based on 40% mortality at a dose of 20 mg/kg in the dose-range finding study, a high dose of 10 mg/kg was chosen for the 26-week carcinogenicity study in p53 heterozygous mice. Doses of 2.5, 5, and 10 mg/kg chlorpromazine hydrochloride were well tolerated in the 26-week study. The administration of chlorpromazine hydrochloride at dose levels up to and including 10 mg/kg to p53 heterozygous and wild-type mice did not result in a dose-related increase in tumor incidence or in the type of tumors seen in comparison to controls. Findings related to the administration of chlorpromazine in the 26-week study were limited to minimal uterine and ovarian atrophy in p53 wild-type mice dosed with 10 mg/kg chlorpromazine hydrochloride. However, p53 heterozygous mice administered 400 mg/kg p-cresidine, a genotoxic carcinogen commonly used as a positive control for this model, developed urinary bladder tumors. Administration of p-cresidine also resulted in a regenerative anemia, splenic and hepatic hemosiderosis, renal findings, and ovarian and uterine atrophy. This study demonstrated that chlorpromazine hydrochloride, at the doses tolerated, was not carcinogenic in the p53 heterozygous mouse assay.

Neonatal mouse model: review of methods and results

The neonatal mouse model, in various forms, has been used experimentally since 1959 and a large number of chemicals have been tested. The neonatal model is known to be very sensitive for the detection of carcinogens that operate via a genotoxic mode of action. In contrast, it is known not to respond to chemicals that act via epigenetic mechanisms, commonly observed in the two-year carcinogenicity studies. As such, the model has a high sensitivity and specificity in its response. Dose selection for the neonatal model is based on the maximum tolerated or feasible dose. Traditionally, compounds have been tested via the IP route of administration in this model. In some cases, this has limited the amount of material that can be administered because of the low dosing volumes (10 to 20 microL) that can be administered IP. For the ILSI project, the neonatal model was adapted for oral administration, which has the advantages of being the same route for which most pharmaceuticals are administered. In addition, a 10-fold increase in the volume of administration (100 to 200 microL) and the ability to dose drugs in suspension, permits much higher doses to be used as compared to the IP route of administration. The spontaneous tumors in the neonatal model occurred mainly in the liver of male mice and lung of male and female mice with a few tumors observed in the Harderian gland. The positive control, DEN produced a robust, uniform, and reproducible tumor response with the target organs essentially limited to liver and lung. A total of 13 compounds out of the 21 ILSI ACT compounds were evaluated in the neonatal model involving 18 studies with duplicate studies for some compounds. The genotoxic carcinogens including those used as positive controls were clearly positive (cyclophosphamide, diethylnitrosamine, 6-nitrochrysene). The non-genotoxic rodent carcinogens were clearly negative (chlorpromazine, sulfisoxazole, sulfamethoxazole, clofibrate, DEHP, haloperidol, metaproteranol, and phenobarbital). The non-genotoxic human carcinogen (cyclosporin) was clearly negative. The two other human carcinogens phenacetin and DES were negative and interestingly estradiol was negative in one of the two oral studies, but was clearly positive in the other. Considering the mode of action for three of the human carcinogens (DES, cyclosporin and phenacetin), which were negative in this model, the mode of action in humans is likely to be epigenetic. Overall, for the 3 clearly genotoxic chemicals, all were positive. For the 9 clearly non-genotoxic chemicals, all 9 were negative. The two human carcinogens for which genotoxicity may or may not play a role (DES and phenacetin) were negative and estradiol was positive in I of the two oral studies. Overall, the extensive database for compounds tested in the neonatal mouse model would support its use as an alternative model for the assessment of the carcinogenic potential of a chemical. The model responds to chemicals that act via a genotoxic mode of action that represent a greater concern for human cancer risk.

A perspective on current and future uses of alternative models for carcinogenicity testing

This perspective is based upon the data presented at the International Life Sciences Institute (ILSI), Health and Environmental Sciences Institute Workshop on the Evaluation of Alternative Methods for Carcinogenicity Testing (ILSI Workshop). It is important to understand that all models discussed at the Workshop have limitations and that they are not designed to be employed as stand-alone assays. Although they may have other, appropriate applications. I do not recommend use of the SHE cell assay and the Tg.AC model for the regulatory purposes of a safety assessment. In my view, the neonatal mouse, p53+/-, XPA-/-, XPA-/- and p53+/-, and the rasH2 models can, as a component of an overall assessment, provide information on potential carcinogenicity of a chemical that is appropriate for consideration in a regulatory context. Generally, these models exhibit the ability to detect genotoxic compounds. In most cases these compounds would be detected in a standard battery of genotoxicity tests and, therefore, quite often the use of an alternative is not necessary. Actually, I believe that a bioassay in rats will suffice most of the time, that is, in my view, a routine bioassay in mice is not necessary. Specific circumstances where data obtained from one of the "recommended" alternative models might be helpful are discussed. With regard to lessons for the future, there is a particular need for models that are responsive to chemicals that exhibit a nongenotoxic mode of action. Additionally, new models will continue to be developed and their half-life will likely be substantially shorter than the time required for traditional validation. The development of enhanced paradigms for validation should be a priority so that improved safety assessment decisions can be made more quickly. However, while evaluating and validating such models, it is important to consider the fundamental issues, for example, rational dose selection, evaluation of mode of action in the context of dose-response relationships including the existence of thresholds and secondary mechanisms, and species-to-species extrapolation. The alternatives to carcinogenicity testing project was a very major undertaking. In addition to the valuable information provided, it serves to illustrate the value of cooperation between academia, government, and industry. Furthermore, the involvement of the International Life Sciences Institute as the overall organizing, facilitating umbrella was crucial for the success of the project.

Alternative models for carcinogenicity testing

The International Conference on Harmonisation Expert Working Group on Safety suggested that under certain circumstances, data from alternative assays could be used in safety evaluation in place of a second bioassay. Several alternatives were discussed. Six of these models were evaluated in a collaborative effort under the auspices of the Health and Environmental Sciences Institute (HESI) branch of the International Life Sciences Institute (ILSI). Standard protocols, pathology review, and statistical evaluations were developed. Twenty-one chemicals were evaluated, including genotoxic, nongenotoxic, carcinogenic, and noncarcinogenic chemicals. The models that were evaluated included the p53(+/-) heterozygous knockout mouse, the rasH2 transgenic mouse, the TgAC transgenic mouse (dermal and oral administration), the homozygous XPA knockout and the XPA/p53 knockout mouse models. Also evaluated were the neonatal mouse models and the Syrian Hamster Embryo (SHE) transformation assay. The results of this comprehensive study suggest that some of these models might be useful in hazard identification if used in conjunction with information from other sources in a weight of evidence, integrated analysis approach to risk assessment.