The NIEHS Predictive-Toxicology Evaluation Project.

The Predictive-Toxicology Evaluation (PTE) project conducts collaborative experiments that subject the performance of predictive-toxicology (PT) methods to rigorous, objective evaluation in a uniquely informative manner. Sponsored by the National Institute of Environmental Health Sciences, it takes advantage of the ongoing testing conducted by the U.S. National Toxicology Program (NTP) to estimate the true error of models that have been applied to make prospective predictions on previously untested, noncongeneric-chemical substances. The PTE project first identifies a group of standardized NTP chemical bioassays either scheduled to be conducted or are ongoing, but not yet complete. The project then announces and advertises the evaluation experiment, disseminates information about the chemical bioassays, and encourages researchers from a wide variety of disciplines to publish their predictions in peer-reviewed journals, using whatever approaches and methods they feel are best. A collection of such papers is published in this Environmental Health Perspectives Supplement, providing readers the opportunity to compare and contrast PT approaches and models, within the context of their prospective application to an actual-use situation. This introduction to this collection of papers on predictive toxicology summarizes the predictions made and the final results obtained for the 44 chemical carcinogenesis bioassays of the first PTE experiment (PTE-1) and presents information that identifies the 30 chemical carcinogenesis bioassays of PTE-2, along with a table of prediction sets that have been published to date. It also provides background about the origin and goals of the PTE project, outlines the special challenge associated with estimating the true error of models that aspire to predict open-system behavior, and summarizes what has been learned to date.

Summary and recommendations for session B: activity classification and structure-activity relationship modeling for human health risk assessment of toxic substances.

The major theme of Session B explored and assessed the current status of activity-classification (AC)1 and structure-activity-relationship (SAR) methods developed to model adverse health effects that can result when biological systems are exposed to various chemical substances. The output from such models is intended to be used as information that supports risk assessments performed on toxic substances. Speakers gave special attention to the requirements and applications of hazard identification models. Specific aspects of the broad subject matter were augmented and explicated by audience and panel discussions during the 1.5 days available. This format stimulated the exchange of a surprisingly broad range of information and stimulating ideas. In order to gather the diverse aspects of Session B in one place, the Rapporteurs agreed that this summary would aim at providing a comprehensive overview, while Dr. Feldman's would amplify selected points of general interest.

Chemical analysis of human blood for assessment of environmental exposure to semivolatile organochlorine chemical contaminants.

A chemical method for the quantitative analysis of organochlorine pesticide residues present in human blood was scaled-up to provide increased sensitivity and extended to include organochlorine industrial chemicals. Whole blood samples were extracted with hexane, concentrated, and analyzed without further cleanup by gas chromatography with electron capture detection. The methodology used was validated by conducting recovery studies at 1 and 10 ng/g (ppb) levels. Screening and confirmational analyses were performed by gas chromatography/mass spectrometry on samples collected from potentially exposed residents of the Love Canal area of Niagara Falls, New York and from volunteers in the Research Triangle Park area of North Carolina for 25 specific semivolatile organochlorine contaminants including chlorobenzene and chlorotoluene congeners, hexachloro-1,3-butadiene, pesticides, and polychlorinated biphenyls as Aroclor 1260. Dichlorobenzene, hexachlorobenzene, and beta-hexachlorocyclohexane residues fell in the range of 0.1 to 26 ppb in a high percentage of both the field and volunteer blood samples analyzed. Levels of other organochlorine compounds were either non-detectable or present in sub-ppb ranges.

NTP 3-month toxicity studies of estragole (CAS No. 140-67-0) administered by gavage to F344/N rats and B6C3F1 mice.

Estragole is a natural organic compound that is used as an additive, flavoring agent, or fragrance in a variety of food, cleaning, and cosmetic products; as an herbal medicine; as an antimicrobial agent against acid-tolerant food microflora; and to produce synthetic anise oil. Estragole was nominated for toxicity testing by the National Institute of Environmental Health Sciences to characterize its toxicity when administered by gavage to F344/N rats and B6C3F1 mice and to determine how similar its effects might be to those of the structurally related compound, methyleugenol. Male and female F344/N rats and B6C3F1 mice were given estragole (greater than 99% pure) in corn oil by gavage for 3 months. Genetic toxicology studies were conducted in Salmonella typhimurium and mouse peripheral blood erythrocytes. Core and special study (rats only) groups of 10 male and 10 female rats and mice were administered 37.5, 75, 150, 300, or 600 mg estragole/kg body weight in corn oil by gavage, 5 days per week. The core study groups were given estragole for 3 months and the special study groups for 30 days. All core study rats survived the 3-month exposure period. Mean body weights of the 300 and 600 mg/kg groups were 73% to 92%, respectively, of those of the vehicle control groups. A staining pattern on the ventral surface anterior to the genitalia beginning at week 9 in the 300 and 600 mg/kg groups was attributed to residue of estragole or metabolites in the urine. Alterations in the erythron related to estragole administration occurred in male and female rats; male rats demonstrated a stronger response. The changes in the erythron were characterized as a microcytic, normochromic, nonresponsive anemia. There were decreases in serum iron concentration in the 300 mg/kg females and 600 mg/kg males and females. The average percent saturation of total iron binding capacity was decreased in the 600 mg/kg males and females. Dose-related increases in platelet counts occurred in most of the dosed groups of rats; the effect appeared to be stronger in males. The increase could be consistent with a reactive thrombocytosis. Increases in the serum alanine aminotransferase and sorbitol dehydrogenase activities suggested a hepatocellular effect (increased leakage) and were consistent with the morphological liver changes observed. There were dose-related increases in serum bile salt concentration in most treated male rats at all time points; females were less affected. Absolute and relative liver weights were significantly increased in 300 and 600 mg/kg males and in 75 mg/kg or greater females. Relative kidney weights were significantly increased in all dosed groups of male rats and in female rats given 75 mg/kg or greater. Absolute and relative testis weights of 300 and 600 mg/kg males were significantly decreased. Two 600 mg/kg male rats had multiple cholangiocarcinomas in the liver and a third had an hepatocellular adenoma. All 600 mg/kg males exhibited cholangiofibrosis. All 75 mg/kg or greater males and all 150 mg/kg or greater females had hepatocellular hypertrophy. Incidences of bile duct hyperplasia, oval cell hyperplasia, and chronic periportal inflammation were significantly increased in all dosed groups. Incidences of basophilic and mixed cell foci were significantly increased in 150 mg/kg or greater males and females. Incidences of eosinophilic focus were significantly increased in 300 and 600 mg/kg males and 600 mg/kg females. Incidences of cellular infiltration of the periportal region by histiocytes increased significantly in all dosed groups of males and in 150 mg/kg or greater females. Incidences of bone marrow hyperplasia were significantly increased in 75, 300, and 600 mg/kg male rats. Incidences of renal tubule papillary mineralization were significantly increased in 300 mg/kg males and females and 600 mg/kg males. Incidences of cortical renal tubule pigmentation were significantly increased in 150 mg/kg or greater males, and the incidence of renal tubule regeneration was significantly increased in 600 mg/kg females. Incidences of degeneration of the olfactory epithelium in the nose were significantly increased in 300 and 600 mg/kg rats. Incidences of hypertrophied chromophobe cells in the pars distalis of the pituitary gland were significantly increased in 300 and 600 mg/kg males. Cytoplasmic alteration of the submandibular salivary gland occurred in all 75 mg/kg or greater rats. Incidences of atrophy of the gastric glands in the stomach were significantly increased in 150 mg/kg or greater rats. Bilateral degeneration of the germinal epithelium in the testes and bilateral hypospermia of the epididymis occurred in all 300 and 600 mg/kg males. In the special study, serum gastrin concentration and stomach pH were significantly increased in rats exposed to 600 mg/kg for 30 days. Gastric gland atrophy was significantly increased in the stomach of 300 and 600 mg/kg rats. Hepatic 7-pentoxyresorufin-O-deethylase activity was significantly increased in all exposed groups except 37.5 mg/kg females, and the increases were generally dose related. In the mouse core study, a 600 mg/kg male died during week 9, and all 600 mg/kg female mice died during week 1; the female deaths were attributed to liver necrosis caused by estragole exposure. Mean body weights of 300 and 600 mg/kg males and 75 mg/kg or greater females were 79% to 89% those of the vehicle control groups. Liver weights were generally increased in 75 mg/kg or greater males and in 300 mg/kg females. Relative thymus weights were significantly increased in all dosed groups of female mice. The incidences of hepatocellular hypertrophy and hepatocellular degeneration were significantly increased in 300 and 600 mg/kg male mice and 150 and 300 mg/kg female mice. Incidences of oval cell hyperplasia were significantly increased in 300 and 600 mg/kg males and in 75 mg/kg or greater females. Liver necrosis occurred in all 600 mg/kg female mice, along with a significant increase in the incidence of diffuse fatty change. In addition, 600 mg/kg females exhibited significant increases in the incidences of degeneration of the gastric glands of the glandular stomach, as well as squamous hyperplasia, mineralization, and ulcer in the forestomach. Degeneration of the olfactory epithelium in the nose occurred in all 300 and 600 mg/kg mice. Estragole was not mutagenic in Salmonella typhimurium strains TA98, TA100, TA1535, or TA1537 when tested in the presence or absence of exogenous metabolic activation enzymes. No increases in the frequencies of micronucleated normochromatic erythrocytes were observed in peripheral blood samples from male and female mice in the 3-month study. Under the conditions of these 3-month studies, estragole showed carcinogenic activity based on the occurrence of two cholangiocarcinomas and one hepatocellular adenoma in the liver of three of 10 male F344/N rats in the high dose group. Because rats and mice were exposed for only 3 months, these studies do not access the full carcinogenic potential of estragole. Nonneoplastic effects were observed in the liver, glandular stomach, nose, kidney, and salivary gland of male and female rats and in the testes, epididymides, and pituitary gland of male rats. Nonneoplastic effects were also observed in the liver and nose of male and female mice and in the stomach of female mice.

The induction of rules for predicting chemical carcinogenesis in rodents.

This paper presents results from an ongoing effort in applying a variety of induction-based methods to the problem of predicting the biological activity of noncongeneric (structurally dissimilar) chemicals. It describes initial experiments, the long-term goal of which is to assist toxicologists, cancer researchers, regulators, and others to predict the toxic effects of chemical compounds. We describe a series of experiments in tree and rule induction from a set of example chemicals whose carcinogenicity has been determined from long-term animal studies, and compare the resulting classification accuracy with eight published human and computer predictions for a common set of 44 test chemicals. The accuracy of our system is comparable to the most accurate human expert prediction yet published, and exceeds that of any of the computer-based predictions in the literature. The induced rules provide confirmation of current expert heuristic knowledge in this domain. These early results show that an inductive approach has excellent potential in predictive toxicology.

Symbolic, neural, and Bayesian machine learning models for predicting carcinogenicity of chemical compounds.

Experimental programs have been underway for several years to determine the environmental effects of chemical compounds, mixtures, and the like. Among these programs is the National Toxicology Program (NTP) on rodent carcinogenicity. Because these experiments are costly and time-consuming, the rate at which test articles (i.e., chemicals) can be tested is limited. The ability to predict the outcome of the analysis at various points in the process would facilitate informed decisions about the allocation of testing resources. To assist human experts in organizing an empirical testing regime, and to try to shed light on mechanisms of toxicity, we constructed toxicity models using various machine learning and data mining methods, both existing and those of our own devising. These models took the form of decision trees, rule sets, neural networks, rules extracted from trained neural networks, and Bayesian classifiers. As a training set, we used recent results from rodent carcinogenicity bioassays conducted by the NTP on 226 test articles. We performed 10-way cross-validation on each of our models to approximate their expected error rates on unseen data. The data set consists of physical-chemical parameters of test articles, alerting chemical substructures, salmonella mutagenicity assay results, subchronic histopathology data, and information on route, strain, and sex/species for 744 individual experiments. These results contribute to the ongoing process of evaluating and interpreting the data collected from chemical toxicity studies.

Studies on the short-term toxicity of theophylline in rats and mice.

The purpose of these studies was to evaluate the short-term toxicity of theophylline, a compound present in tea and used in a variety of clinical applications. Fourteen-day repeated-dose toxicity studies were conducted in B6C3F1 mice and F344 rats of both sexes. Theophylline was administered in feed (0, 500, 1000, 2000, 4000, and 8000 ppm) or by gavage in corn oil (12.5-twice daily, 25, 50, 50-twice daily, 100, 200, 200-twice daily, and 400 mg/kg). Dosed-feed exposure to theophylline at concentrations up to 8000 ppm induced no significant toxicity except for dose-related uterine hypoplasia in rats. Palatability problems at that level precluded administration of higher concentrations. In the gavage study, 400 mg/kg was acutely toxic for both species, but mice and rats differed in that this same daily dose administered as two separate doses of 200 mg/kg was acutely toxic in rats but not in mice. No dose-related weight gain depression was evident in mice; weight gain was depressed in the majority of dose levels in rats and was pronounced at the higher levels. Clinical signs in mice were squinting and distended testes in males, and in rats, rapid respiration (all doses), squinting, and hunching. Gross necropsies, organ weights, clinical pathology, and pathology identified no target organs in mice, while histopathologic observations in rats suggested heart and stomach as possible target organs. Histopathologic effects in a number of other tissues, including lung, thymus, bone marrow, spleen, and uterus, were considered to reflect agonal changes in treated rats, possibly related to inanition. The results suggest that both species and sex differences exist with respect to sensitivity to theophylline toxicity, with F344 rats being more sensitive than B6C3F1 mice and male rats being more sensitive than female rats.