key: cord-311260-eyvaazfj
authors: Rao, Ghanta N.; Huff, James
title: Refinement of long-term toxicity and carcinogenesis studies()
date: 2004-09-27
journal: Fundam Appl Toxicol
DOI: 10.1016/0272-0590(90)90160-l
sha: 
doc_id: 311260
cord_uid: eyvaazfj

The chance that alternatives will completely replace animals for toxicology research in the foreseeable future is nil. Continual refinement of animal toxicity and carcinogenesis studies, however, can be an effective means of reducing the numbers of animals used and conserving time and resources without compromising scientific quality. We must continue to strive to find species and strains that can metabolize chemicals similar to humans, are small enough to be housed in large numbers, and have low prevalence of spontaneous lesions with sufficient life span to express the toxic and carcinogenic potential of chemicals. Adequate care of animals with control of variables such as light, temperature, diet, bedding, diseases, and genetic characters of laboratory animals will decrease the variability. Humane considerations and euthanasia of animals with large masses and other conditions interfering with eating and drinking, major injuries and ulcers related to husbandry and treatment, and diseases indicating pain and suffering will help not only to alleviate further pain and distress but also to facilitate collection of tissues without secondary complications for detection of chemical treatment-related lesions. Limiting the duration of studies to decrease the variability due to ageassociated changes will also refine long-term studies. Other considerations for refinement of carcinogenesis studies include selection of the most sensitive sex of one or more species for evaluation of selected chemicals in a class where toxic and carcinogenic potential of other representative chemicals are known. Genetically engineered animal models with known oncogenes may reduce the duration and increase the sensitivity of carcinogenesis studies with a reduction in the use of animals.

of Long-Term Toxicity and Carcinogenesis Studies. RAO, G. N., AND HUFF, J. (1990) . Fundam.

Appl. Toxicol. 15, 33-43. The chance that alternatives will completely replace animals for toxicology research in the foreseeable future is nil. Continual refinement of animal toxicity and carcinogenesis studies, however, can be an effective means of reducing the numbers of animals used and conserving time and resources without compromising scientific quality.

We must continue to strive to find species and strains that can metabolize chemicals similar to humans, are small enough to be housed in large numbers, and have low prevalence of spontaneous lesions with sufficient life span to express the toxic and carcinogenic potential of chemicals.

Adequate care of animals with control of variables such as light. temperature. diet, bedding, diseases, and genetic characters of laboratory animals will decrease the variability. Humane considerations and euthanasia of animals with large masses and other conditions interfering with eating and drinking, major injuries and ulcers related to husbandry and treatment, and diseases indicating pain and suffering will help not only to alleviate further pain and distress but also to facilitate collection of tissues without secondary complications for detection of chemical treatment-related lesions. Limiting the duration of studies to decrease the variability due to ageassociated changes will also refine long-term studies. Other considerations for refinement of carcinogenesis studies include selection of the most sensitive sex of one or more species for evaluation of selected chemicals in a class where toxic and carcinogenic potential of other representative chemicals are known. Genetically engineered animal models with known oncogenes may reduce the duration and increase the sensitivity of carcinogenesis studies with a reduction in the use of animals.

6 1990 Society ofToxicology.

The search for alternatives to the use of animals in research and testing remains a valid goal of researchers. but the chance that alternatives will completely replace animals in the foreseeable future is nil. (National Research Council, 1988) .

We must use alternatives to whole animals such as bacterial systems and cell culture methods only when they are reliable predictors of toxicity and carcinogencity. However, we must evaluate potential alternative tests objectively, free of subjective attachment. In the field of toxicology alone there are hundreds (Goldberg, 1985; Tennant et al., 1987; Mehlman, 1989) of in vitro assays and other alternative methods for evaluation of toxic and carcinogenic potential of drugs and chemicals. Many of these alternative tests with cell systems and subcellular components will aid in understanding effects and mechanisms at cellular and subcellular levels. But many of these tests do not predict the re-studies on these chemicals were considered to sponse of the whole animal (Tennant et al., be incomplete or inadequate. Continual re-1987) . There are a number of reasons for this finement and optimization of animal toxicity difference. In a living animal during the ex-and carcinogenesis studies can be an effective pression of physiologic, pharmacologic, and means of reducing the numbers of animals toxic responses to an agent, in addition to used and of conserving time and resources subcellular component interactions, there are without compromising scientific quality. Esother interactions. They include (a) cell to cell sential components for refinement of longcommunication between cells of the same or term studies include selection of proper spedifferent types in organs such as liver, and (b) cies and strains, proper care of the animals, organ to organ communication, interaction, humane considerations during the course of and compensation between organs of a living a study, and adequate duration of the studies. animal such as liver and kidney, thymus and Each of these essential components is dislymphoid tissue, pituitary and other endo-cussed below. crine organs, which we cannot predict by alternative assays. In addition the living animal as a whole by its complex feedback mecha-SELECTION OF PROPER SPECIES, nisms between organs and tissues adjusts to STRAINS, AND NUMBER varying chemical and environmental insults, OF ANIMALS compensates, and survives or overreacts and dies, which we cannot predict by any single

We should continue to strive to find species or battery of in vitro and alternative tests. and strains of experimental animals (a) that Notwithstanding these drawbacks, we must metabolize or dispose a chemical in a manner continue to strive for reliable in vitro or alter-similar to humans; (b) with (median) life native methods to the use of whole animals. It spans long enough to adequately assess the is the most humane, ethical, and economical carcinogenic potential of chemicals (mouse direction for us to follow. But diseases like strains such as A and AKR with relatively cancer and acquired immunodeficiency syn-short life spans may be useful for assessing the drome (AIDS) humble us with regard to limi-carcinogenic potential of potent carcinogens tations in our understanding of complex bio-but will not provide adequate information on logic systems such as mammals. Alternatives weaker carcinogens); (c) with low spontaneare useful for mechanistic research and a lim-ous lesions and tumors, especially in the tarited number of alternative tests currently get organs of the chemical, to enhance the have some utility for assessing the toxic and specificity and sensitivity and to facilitate incarcinogenic potential of selected classes of terpretation; and (d) small enough that they chemicals (e.g., polycyclic aromatic hydro-can be housed conveniently in large numcarbons). However, the probability that alter-bers, thus allowing the use of a greater numnatives will completely replace animals for ber of animals to increase the sensitivity (statoxicology research in the foreseeable future tistical power) for detecting toxic responses, is nil.

especially carcinogenic effects. The number If a study must be done using animals, then of animals per group should be adequate to we are obligated to do it right thejirst time. In assess the toxic and carcinogenic potential. the long run this will save money and time, However, increasing the number of animals will decrease the number of animals used, beyond 50 to 60 per group does not result in and is one of the best refinements we can do. a proportional increase in sensitivity. As In recent years a number of chemicals were shown in Table 1 , for animals having a backselected by the National Toxicology Program ground tumor rate of 5% if the number per (NTP) for reevaluation, because the previous group is increased by 40 animals from 20 to 60, there will be more than a doubling in sensitivity (after compensating for background rate); whereas the same increase of 40 animals from 60 to loo/group will result in an increase of sensitivity by only a third. Approximately 50 animals per group (Portier and Hoel, 1983: NTP, 1984 ) achieves a reasonable balance between the cost, animal use, and study sensitivity.

Much of the variation between animals and between studies is a product of genetic and environmental factors. Contribution of genetic variability may be substantial when outbred stock (e.g., Sprague-Dawley rats, ICR-Swiss mice) are used. However, with genetically defined inbred and hybrid strains, most of the variability is due to environmental factors. The environmental variables could be physical or biological factors and some of these are listed in Table 2 . The effects of these factors were discussed in detail elsewhere (Rao, 1986: Haseman et al., 1989, and  the references cited therein). Some recent findings on the effects of light, diet, and viral infections are discussed below.

Light. The duration, intensity, and quality of light will influence many physiologic responses and reactions. Light may cause eye lesions and influence tumor incidences (Greenman et al., 1984; Wiskemann et al., 1986) . High light intensity may cause eye lesions in albino rodents, (Bellhorn, 1980; Greenman et al., 1982) . In some NTP studies, for example, Fischer 344 rats housed in the top row and side columns of racks where they were exposed to more light than the rats in other cages have been observed to develop opacity of the eyes and this appears to be due to inflammation of various structures of the eye. This lesion later progressed to generalized inflammation of the eye with retinal degeneration and cataract formation. Prevalence of eye lesions by cage position and light intensity in male F344 rats is illustrated by an example shown in Table 3 . If the chemical under investigation has the potential to cause eye lesions, then the light intensity related eye lesions may complicate the interpretation of chemical effects. As shown in Table 4 (NTP, 1989) high doses of the chemical appear to have increased the incidence of eye lesions in male F344 rats. However, the contribution of light intensity for development of eye lesions in the treated groups is not known. The prevalence of light intensity-associated eye lesions in rodents could be markedly decreased by (a) reducing the light intensity in the cages to < 15 foot-candles or the animal room light intensity to ~50 foot-candles at 5 ft from the Rao, 1986. floor and (b) rotating the cages of each column of a rack from top to bottom when cages or racks are changed (NTP, unpublished data).

Diet. A casein-based purified diet (Newberne et al., 1973) considered to be adequate for a multigeneration reproduction study in rats was found to be deficient in manganese, warranting invalidation of a 2-year in utero exposure chemical carcinogenesis study (Rao, 1988a) . Optimal diets for rodents in long-term toxicity and carcinogenesis studies should be nutritionally adequate for growth and maintenance without excesses of high energy and growth enhancing nutrients. There should be established formulation with standards for ingredients, nutrient concentra-tions, and contaminant limits. Contaminant concentrations should be as low as practical. Each batch/lot of diet should be analyzed for macronutrients and selected micronutrients with complete micronutrient analyses on selected batches/lots (Rao, 1988a) . Viral infections. Sendai virus (SV), pneumonia virus of mice (PVM). and mouse hepatitis virus (MHV) are the most common viral infections of mice and SV, PVM, and rat corona virus/sialodacryoadenitis virus (RCV/SDAV) infections are most common in rats (Boorman et al., 1986) . Viral infections may complicate toxicology research (Collins, 1986) . In a systematic comparison of control and chemically treated groups of B6C3Fl mice, viral infections did not cause consistent adverse effects on survival and tumor prevalences. However, SV infection in mice was associated with significantly (r, < 0.05) higher survival and survival-associated increase in liver tumors (Rao et al.. 1989a ). Viral infections did not cause consistent adverse effects on survival or tumor prevalences in control groups of F344 rats (Rao et al., 1989b) . However, viral infections were associated with nonneoplastic lesions in lungs, nasal cavity, liver, and other organs of rats and mice and may complicate the identification and interpretation of toxic effects of chemicals (NTP, unpublished data).

Animals used in toxicology or carcinogenesis studies should be monitored closely by experienced professional scientists, especially if the chemical is toxic or if there are increases in tumor-bearing animals. There should be specific criteria supplemented with professional judgment for euthanasia of moribund animals during the course of these long-term studies. The objective of these criteria is not only to relieve excessive pain and distress to animals but also to allow collection of tissues for pathologic assessment that are free of secondary complications. Major reasons for eu- thanasia of animals during the course of a large mammary tumor may be injured during study include: (a) large masses and other conthe normal movement in the cage leading to ditions interfering with eating and drinking, ulceration and infection resulting in second-(b) major injuries and ulcers related to husary lesions in the tumor and in the animal. bandry, fighting, or chemical exposure, and Major injuries such as accidental fracture of (c) diseases and conditions indicating pain a limb or wounds resulting from fighting of and suffering as judged by an experienced group-caged animals may lead to unrelieved laboratory animal specialist. For example, a pain and nonhealing lesions. For humane reasons and to prevent complications of chemical-induced lesions by secondary changes, it is prudent to euthanatize and necropsy the animal. In studies such as dermal toxicity and carcinogenesis where treatment causes persistent or enlarging ulcers and fast growing tumors which are injured by the animal or components of the cage, it is appropriate to euthanatize and necropsy the animal not only to relieve pain and distress but also to prevent infections. Infections and inflammation may lead to secondary lesions which can complicate the interpretation of toxic effects and may decrease the sensitivity of a toxicology study. Other considerations for euthanasia of rodents in chronic studies are listed in Table 5 ; most of these conditions will change the normal physiology and may affect morbidity and mortality. Any one of these conditions may not lead to mortality in a short period, but a combination of these conditions will compromise the physiology and health of the animals. If these animals are allowed to continue in a study, secondary affects may mask or exacerbate the chemical effects, complicate the interpretation of toxic effects, and decrease the sensitivity and quality of a long-term study.

Duration of toxicology studies with rodents may range from 3 to 18 months depending on the toxicity of the chemical (Frederick, 1986 ). In general a 3-to 6-month toxicity study may be long enough for many chemicals with considerable acute toxic effects and limited cumulative toxicity. However, with a high proportion of chemicals, due to the cumulative nature of their toxicity, 3 to 6 months duration may not be long enough and studies of 9 to 12 months duration may be necessary. Chemicals that can cause delayed toxicity, such as neurotoxicity, may have to be evaluated in studies of 12 to 18 months duration. However, when con- ducting studies longer than 12 months, one should be aware of and consider the influence of age-associated changes and complications in interpretation of the toxic responses. Duration of chemical carcinogenesis studies in rodents usually covers a major portion of the life span. Length of generally accepted chemical carcinogenesis studies in rats and mice include: (a) 21 to 24 months, (b) 30 months or 20% survival, whichever occurs first, and (c) until most or all animals in the chemically exposed groups die or are euthanatized due to moribund condition. For potent or "early" carcinogens, studies longer than 18 months may not be necessary (e.g., 1,3-butadiene, Huff et al., 1985) . For weaker carcinogens with delayed expression of carcinogenic potential, studies of longer than 21 months may be necessary. The median life span (50% survival) of rodent species and strains used for chemical carcinogenesis studies is 20 to 30 months and, therefore, studies lasting 2 1 to 27 months will cover 60-80% of the life span, assuming 90% mortality is close to the life span. In these long-term studies, especially in studies longer than 18 to 2 1 months, there are several age-associated changes that may compromise the health of the animals and complicate the expression and interpretation of carcinogenic response. Tables 6 and 7 and Figs. 1 to 3 illustrate some of the age-associated changes that may complicate long-term carcinogenesis studies.

Male F344 rats attain a maximum body weight at 80.2 * 9.6 (mean ? SD) weeks of age (Haseman et al., 1985) or some time during the 15th to 20th month of a chronic study and may lose as much as 20% of their body weight by the 24th month. The decline in body weight is accompanied by an increase in water consumption or vice versa (Fig. I) . The increase in water consumption (polydipsia) is due to nephrosis leading to polyuria with substantial changes in physiologic processes such as distribution, metabolism, and excretion of chemicals (Grice, 1984) . These changes could complicate the interpretation of a toxicology study. Even in a carcinogenesis study where the animals are steadily losing body weight, it may not be scientihcally prudent to continue the study after a substantial decline in body weight.

Female B6C3Fl mice reach maximum body weight at 98.1 t 10.2 weeks of age (Haseman et al., 1985) or during the 19th to 23rd month of a chronic study. A steep in-crease in mortality and persistent loss of body weight (Fig. 2) starts after 2 1 to 24 months of study. Even though the life span (90% mortality) of female B6C3Fl mice is about 35 months, median survival is reached at about the 28th month of a chronic study and coincides with a 15 to 20% loss of body weight. Furthermore, the prevalence of lymphoma in control mice shows a marked increase between the 2 1 st and 24th months (Fig. 2) , indicating that it may not be appropriate to con- tinue carcinogenesis studies in B6C3Fl mice beyond 24 months, especially when the lymphoreticular organs are the target tissues. Table 6 shows the prevalence of amyloidosis in ICR-Swiss (or CD-l) outbred mice (Rao et al., 1988b) . High prevalence of amyloidosis in major organs such as kidney, liver, and heart at 23-25 months of age (or 22-24 months of a chronic study) may impair the ability of these organs to metabolize and dispose the chemicals and may complicate the interpretation of neoplastic lesions.

Body weight and water consumption patterns of a strain of Syrian hamsters (Homburger et al., 1983) are shown in Fig. 3 marked increase in water consumption begins at 65-75 weeks of age together with a start of persistent decrease in body weight, indicating changes in physiologic processes. Furthermore, there is a high prevalence of amyloidosis (Newcomer et al., 1987) in major organs such as liver, kidney, and spleen in 20-to 2%month-old Syrian hamsters as shown in Table 7 . The body weight and water consumption patterns (Fig. 3 ) and the prevalence of amyloidosis (Table 7) indicate that the physiology and health of these animals may have been compromised by about 20 months of age and so carcinogenesis study of longer than 18 months may not be appropriate with Syrian hamsters.

Tables 6 and 7 and Figs. 1 to 3 identified a few of many age-associated changes observable in the whole animal. These changes are due to several pertubations at the subcellular, cellular, and organ level. Some of these changes relevant to chemical toxicology and carcinogenesis may include chemical absorption, distribution, metabolism, and excretion; protein, lipid, and carbohydrate metabolism; hepatic drug metabolism, renal excretion, and morphologic changes in liver. kidney, and other organs (Grice, 1984) .

A variety of diseases, tumors, and age-associated changes present in laboratory rodents most often increase in severity and incidence with increasing age, resulting in marked alter-ations in physiologic processes. These diseases and tumors are associated with alterations in function and morphology of many organs, making it difficult to differentiate between age-associated changes and chemically induced toxicity and lesions. Changes in organ function are associated with changes in metabolism and toxic responses to chemicals. With few exceptions, the aging animal metabolizes chemicals less effectively and is more susceptible to the toxic effects of chemicals (Grice, 1984) . Studies beyond 18 months in hamsters and 24 months in mice and rats may result in greater variability and nonneoplastic disease-associated complications in interpretation of results. Increasing the duration beyond these general limits may not be a refinement of long-term toxicity and carcinogenesis studies. However, for weak carcinogens with delayed expression of carcinogenic potential, studies of longer than 24 months in strains with long life spans and low prevalence of spontaneous lesions and tumors may be necessary.

Most of the National Cancer Institute-National Toxicology Program long-term chemical carcinogenesis studies included male and female F344 inbred rats and male and female B6C3Fl hybrid mice. Haseman and Huff (1987) reported species correlation in 266 long-term toxicology and carcinogenesis studies. Retrospective analyses of sex-species combinations for identifying potential chemical carcinogens are shown in Table 8 . Combinations of male rats and male mice or male rats and female mice detected 124 (92%) of the 135 identified carcinogens. The male rats and female mice combination "missed" 11 (8%) chemicals shown to be positive in one of the other sex-species, of which 7 were positive in male mice only, 3 in female rats only, and 1 in female rats and male mice. The male rats and male mice combination also "missed" 11 (8%) carcinogens, of which 6 were positive in female mice only, 3 in female rats only, and 2 in female rats and female mice. Prospective predictions based on these retrospective evaluations may not be appropriate for unknown chemicals. When a class of chemicals is evaluated (e.g., benzidine dyes, nitrotoluenes, glycol ethers), both sexes of rats and mice may have to be included to assess full carcinogenic potential and mechanism of carcinogenesis of a few representative chemicals of the class. However, for evaluation of other chemicals in that class, the most sensitive sex of one or more species may be adequate. This refinement may decrease the number of animals for selected chemicals and reduce costs of some long-term studies.

Chemical carcinogenesis is a multistep process and may involve the activation of one or more protooncogenes.

Genetically engineered animal models such as transgenic mice with ras, myc. net{ or other oncogenes may be useful in refinement of chemical carcinogenesis studies. The advantages of transgenie animal models could be (a) carcinogenic potential of chemicals may be detected with shorter studies, e.g., 6 to 12 months instead of 18 to 24 months: (b) sensitivity and possibly specificity may be increased with proper animal models for different types or classes of potentially carcinogenic chemicals: and (c) the number of animals per chemical may be reduced with a reduction in cost. The NTP has started studies with transgenic strains TG.SH with rus, TG.M with rrz~~, and TG.NK with neu oncogenes using known carcinogens and noncarcinogens. The purpose of these studies is to assess the usefulness and advantages of transgenic mouse models for detecting chemical carcinogens. In addition, the NTP (NIEHS) is developing an extensive program for development and evaluation of transgenic models to refine and optimize not only carcinogenesis studies but also genetic toxicology, possibly immunotoxicity, and reproductive and developmental toxicology studies.

In summary, alternative in vitro assays or other methods will not replace live animals ( 100) a Haseman and Huff (1987) . ' 13 I studies were negative in male and female rats and mice. for toxicology research. If a study must be done with animals, then we are obligated to do it right the first time. This is the best refinement one can do to decrease the use of animals. Essential components of refining toxicity and carcinogenesis studies include: (a) selection of animal models that can metabolize a chemical in a similar manner to humans, that are small enough to house in large numbers in a laboratory setting, that have a low prevalence of spontaneous lesions, and that have a long life span; (b) control of variables such as environmental conditions, diseases, and genetic characters of laboratory animals; (c) euthanasia of animals during the course of a study to alleviate pain and suffering, thus preventing development of secondary lesions and facilitating detection of chemical effects; and (d) limiting the duration of studies in rodents; which may decrease variability and complications due to nonneoplastic lesions and spontaneous tumors. All of these procedures will help to decrease the variability and increase the sensitivity--thus refining toxicity and carcinogenesis studies. Further refinements include the use of genetically engineered animal models such as transgenie mouse strains with known oncogenes to decrease the duration and increase the sensitivity of carcinogenesis studies with a decrease in the use of animals.

Lighting in the animal environment. spontaneous and induced neoplasms in mice

Age associated (geriatric) pathology: Its impact on long-term toxicity studies

Sources ofvariability in rodent carcinogenicity studies

Neoplasms observed in untreated and corn oil gavage control groups of F344/N rats and B6C3Fl mice

Species correlation in long-term carcinogenicity studies

A new firstgeneration hybrid Syrian hamster, BIO FlD Alexander for in vivo carcinogenesis bioassay, as a third species or to replace the mouse

Multiple organ carcinogenicity of 1.3-butadiene in B6C3FI mice after 60 weeks of inhalation exposure

Report cf the NTP Ad Hoc Punel on Chemical Carcinogenesis Testing and Evaluation. pp. 17 l-l 76. U.S. Department of Health and Human Services, Public Health Service. National Toxicology Program (I 989). Toxicology and Curcinogenesis Studies qf Nalidisic

Safety evaluation of fish protein concentrate over five generations of rats

Optimal design of chronic animal bioassay

Significance ofenvironmental factors on the test system

Rodent diets for carcinogenesis studies

Mouse strains for chemical carcinogenicity studies: Overview of a workshop

Influence of viral infections on body weight, survival, and tumor prevalences of B6C3Fl (C57BL/6N X C3H/HeN) mice in carcinogenicity studies. Fundam. .4pp/. Taxicd

Influence of viral infections on body weight. survival, and tumor prevalence in Fischer 344 rats of long-term studies

Prediction of chemical carcinogenicity in rodents from in vitro genetic toxicity assays

Fluorescent lighting enhances chemically induced papilloma formation and increases susceptibility to tumor challenge in mice

The authors thank Drs. Joseph Haseman and Lawrence Boone for scientific advice and Drs. Herbert Amyx and Michael Jokinen for reviewing the manuscript.