Rat Genetics and Toxicology
Michael F.W. Festing
MRC Toxicology Unit, University of Leicester, Leicester, United Kingdom
SUMMARY
There is substantial genetic variation in the response of laboratory rats to xenobiotics, and this variation has important implications for toxicologic research and screening. However, most characteristics of toxicologic interest have a polygenic mode of inheritance so the variation is not immediately apparent to most investigators who use only a single stock or strain of rats.
A survey of “rat” papers published in Toxicology and Applied Pharmacology in 1979 and 1999 showed that the use of inbred strains has increased from 7% to 31% in the 20-year period. However, given the extensive literature on the importance of using inbred (or F1 hybrid) strains and the lack of any published scientific justification for the use of outbred stocks, this slight increase suggests that toxicologists are still not giving much thought to the most appropriate choice of animals.
Most (66%) authors failed to note the strain used either in the title or abstract, and were apparently under the impression that they were studying “the rat,” even though their results gave no indication of the likely variation in response among rat strains. Only four (7%) of the 61 rat papers published in early 1999 used more than one strain, but three involved study of known genetic polymorphisms and one used two strains interchangeably. None of them would have been able to detect previously unknown genetic variation in response. Academic research workers could easily use more than one strain of rats without increasing the total numbers of animals, using factorial experimental designs. Such designs would be statistically powerful and would not present any particular statistical problems.
Given that there is substantial, and economically important, genetic variation in humans in response to xenobiotics and that rats are widely used in toxicology and pharmacology, the failure to seek genetic variation in rats, which could be used as a model of similar conditions in humans, is surprising. Unless toxicologists change their research tactics they will fail to benefit from the enormous advances currently being made in molecular genetics.
INTRODUCTION
It is now 20 years since ILAR published an excellent set of guidelines on “Laboratory Animal Management: Genetics” (ILAR 1979). These guidelines described the main types of strains and stocks then available to research workers, which included inbred strains and their derivatives such as congenic strains and recombinant inbred strains, mutants, and “stocks not genetically defined,” including outbred stocks.
In a discussion of the choice of inbred strains versus outbred stocks, the guidelines suggest, “An investigator working with species for which inbred strains are available would be well advised to use them. ” The serious limitations of genetically undefined strains were emphasized. Although such animals may be cheaper, they will be phenotypically more variable so that larger numbers are needed, they are subject to genetic drift, and colonies with the same name from different breeders may differ to a serious extent. Moreover, since that time there have been many papers published that describe the valuable properties of inbred strains and the limitations of outbred stocks (Festing 1990, 1995, 1997a,c; Festing and Wolff 1995). Yet, outbred stocks continue to be used widely, even though no scientific justification for their continued use appears to have been published in the last 20 years.
In contrast, inbred strains, which have been described as “immortal clones of genetically identical individuals,” tend to be highly uniform, they stay genetically constant for long periods, the genetic and phenotypic characteristics of most strains are well documented, and genetic quality control is relatively easy using DNA genetic markers (Festing 1997b).
FROM THE LITERATURE
Strain Differences in Response to Xenobiotics
Genetic variation in response to xenobiotics is seen most clearly as strain or stock differences, in experiments that have used more than one strain (for convenience the term strain will be used to indicate both inbred strains and outhred stocks). Examples include differences in response to DMBA among three rat strains in which strain COP was totally resistant to a dose of carcinogen that caused 100% mammary tumours in WF, with F344 being intermediate (Moore
and others 1988), large differences in response to 3,2'-dimethyl-4-aminobiphenyl among five rat strains with 48% prostate tumors being observed in F344, but none in the Wistar stock (Shirai and others 1990), and large differences between inbred ACI rats and an outbred Sprague-Dawley stock in response to diethylstilbestrol and neutron irradiation, with the Sprague-Dawley stock being completely resistant to the effects of DES, but reasonably sensitive to the neutron irradiation, with the converse being observed in strain ACI (Shellabarger and others 1978). Large strain differences in response to pharmaceutical agents are also frequently observed (Kacew and Festing 1996).
Such strain differences have important implications for toxicologic research and screening (Festing 1987, 1995, 1997a). Without some idea of the range of sensitivity seen among different strains, it clearly does not make much sense to characterize “the rat” on the basis of research done with a single strain that may be highly atypical. For example, some colonies of outbred Han:Wistar rats have a mutation in the Ah receptor, which makes them approximately 1000-fold less sensitive to the acute toxic effects of TCDD than an outbred Long-Evans stock (Pohjanvirta and others 1999; Tuomisto and others 1999). However, strain differences are of intrinsic interest because there are large differences among humans in response to xenobiotics (Evans and Relling 1999). Because rats are widely used in both the pharmacology and toxicology of drug development, knowledge of genetic variation in response would provide useful animal models of human adverse drug reactions. For example, the DA rat strain (which should not be called the Dark Agouti strain as the D stands for the D blood group and not for the word Dark) is widely used as a model to study the effects of genetic variation at the CYP2D locus (Vorhees and others 1999).
Survey of Papers Published in Toxicology and Applied Pharmacology in 1979 and 1999
To obtain more details of the way rats have been and are currently used in research, papers published in a single toxicologic journal (Toxicology and Applied Pharmacology) in 1979 and 1999 were studied. This is a well-respected journal that publishes mechanistic studies of the effects of xenobiotics on biologic systems, including laboratory rats. The results of a study of the first 45 papers published in 1979 and the first 61 papers published in 1999 using the laboratory rat are given in Table 1. The aim was to determine what progress has been made over the last 20 years in encouraging toxicologists to use genetically defined animals and to find out what proportion of papers included more than one strain so that the investigator would have become aware of genetic variation in the observed responses.
In 1979, only 7% of papers involved the use of inbred strains, but by 1999, this use had increased to 31%. This change is almost entirely accounted for by an increase in the use of F344 rats at the expense of Sprague-Dawleys, with the use
TABLE 1 Survey of the First 45 and the First 61 “Rat” Papers Published in Toxicology and Applied Pharmacology in 1979 and 1999
1979 |
1999 |
|||
No. of papers |
45 |
61 |
||
No. |
% |
No. |
% |
|
Used inbreds |
3 |
(7) |
19 |
(31) |
Used outbreds a |
42 |
(93) |
43 |
(70) |
Use strain/stock |
||||
SD |
29 |
(64) |
25 |
(41) |
Wistar |
10 |
(22) |
12 |
(20) |
F344 |
2 |
(4) |
14 |
(23) |
Other |
5 |
(11) |
9 |
(15) |
Strain in title/abstractb |
8 |
(18) |
27 |
(44) |
Used more than 1 strain |
1 |
(2) |
4 |
(7) |
Able to find new genetic variationc |
0 |
(0) |
0 |
(0) |
a One paper used both. bUsed as an indication of whether the author(s) considered that they were investigating “the rat” or a particular strain of rats. cPapers that used more than one strain comparatively when not investigating a known polymorphism. |
of Wistar and “other “ rats staying approximately constant. Thus, some slight progress has been made in encouraging the use of inbred strains, although whether this rate of progress is acceptable, given the limitations of these outbred stocks, is debatable. None of the papers gave any reasons for choosing the strain used.
In 1979, only a single paper in the sample used more than one strain, and that was a study involving a known genetic polymorphism. Thus, no papers would have detected previously unknown genetic variation. By 1999, 7% (4/61) of papers in the sample used more than one strain, but three of the papers involved studies of known genetic polymorphisms, and one used two strains interchangeably in different experiments without indicating which strain had been used when presenting the results. Thus, none of the studies was in a position to observe genetic variation that was not already known.
Toxicologists often appear to assume that the strain or stock of rats they use is representative of “the rat” in general (Festing 1990). As an indicator of this assumption, the survey also recorded what proportion of the papers failed to mention the strain of rats used either in the title or in the abstract. In 1979, 18% of papers noted the strain in this way, but this proportion had increased only to 44% by 1999. Thus, more than half the papers apparently characterized “the rat ” on the basis of work done with a single, often undefined strain of rats. Statements such as “The oral LD50 of adenine in the rat is 227 mg/kg” have very little meaning if strains can differ substantially in their response to such a xenobiotic.
Multistrain Experiments Do Not Necessarily Need to Involve the Use of More Animals
The quality of much toxicologic research could be substantially improved if toxicologists sometimes used more than one strain. Their failure to do so appears to stem from the assumption that the use of two strains would double the number of animals needed, but this assumption is wrong. In most cases, it would be entirely valid statistically to use the same number of animals, but divided among two or more strains using a factorial experimental design (Festing 1999).
There are four possible strategies with respect to genotype that could be used in studying the effects of a xenobiotic either in toxicologic research or in screening. For simplicity, it will be assumed that an experiment will involve a control and a treated group and that there will be a total of 48 rats in each group. In practice, group sizes are about this large in toxicologic screening, but there are usually three or four treated groups with different dose levels of the compound. Possible experimental designs involve the use of a single inbred strain, a single outbred stock, identical twins, or several isogenic strains but without increasing total numbers.
Single inbred strain
The first design involves the use of a single inbred strain. This design has the advantage that the treated and control groups are genetically identical at the start of the experiment, so such an experiment would tend to have high statistical power provided the strain is genetically susceptible to the compound. However, if the strain is unusually resistant, then this strategy will not be very good and, as it only uses a single strain, the experiment will not indicate whether the response is under genetic control.
Single outbred stock
The second design involves the use of a single outbred stock, as is currently most common. This design has four serious limitations: (1) If the outbred stock is genetically heterogeneous, then the treated and control group will not be genetically identical at the start of the experiment. This genetic difference will normally lead to increased phenotypic variability so that the experiment will lack statistical power. (2) The stock may, like a single inbred strain, be genetically resistant to the xenobiotic. (3) The experiment may not be repeatable elsewhere because outbred stocks with the same name often differ; and (4) Because individual genotypes or pedigrees are unknown, there will be no indication that the response is under genetic control.
Identical twins
The third design involves the use of 48 pairs of identical twins (assuming they are available). It is well established that twin experiments in humans and
cattle are extremely powerful because, on the one hand, the treated and control groups are genetically identical, and, on the other hand, the differences between twin pairs will sample a broad range of genetic variation in susceptibility. In humans, such an experiment could, in theory, include people of different racial groups to sample a wide range of human genetic polymorphism. Differences between twin pairs will give some indication of the range of genetic variation present in the human population, although a formal test is not possible unless twin pairs could be stratified, for example, by race or on the basis of some known genetic polymorphism. Notice that with twin studies, there are no particularly difficult statistical problems in analyzing the data. The group size for comparing treated with control groups is the same as if a single group had been used, although for quantitative characters, a paired rather than an unpaired t-test would be used. For qualitative characters, the total number of responders would be compared in the treated and control groups across the whole experiment. Thus, this design would, in theory, be very good, although in practice, identical twin rats are not available, and it would be inconvenient to use two rats from each of 48 strains.
Several isogenic strains, but without increasing total numbers
The fourth design is a suitable compromise between the use of a single isogenic strain and the use of twins or 48 isogenic strains. Thus, the experiment could consist of small numbers of several different strains. For example, instead of using 48 rats of a single strain, it would be possible to use, say, 12 rats of each of four isogenic strains. Strains could be chosen at random, on the basis of known susceptibility to the class of agent being studied, or to be as genetically diverse as possible. This design has the advantage that the treated and control groups are genetically identical at the start of the experiment, and the differences between the strains will sample a range of genetic variation in susceptibility. In many ways, it is very like the twin study and presents no particular problems for statistical analysis. Toxicologists sometimes mistakenly see this as four separate experiments, each of which is too small; however, given that with a twin study or when using an outbred stock it is quite permissible to average across genotypes, there is no statistical reason why the same should not be done with this design. The more strains that are used, the more statistically powerful the experiment becomes (Felton and Gaylor 1989). As the differences between inbred strains are usually quite large (which is a feature of the effects of inbreeding [Falconer 1981]), it is in some ways rather like a twin study that was able to sample different racial groups, making it quite a powerful design. This design could be used immediately by academic toxicologists who are not under the same regulatory constraints as those doing toxicologic screening for commercial purposes. In the long term, such a design could also be used in regulatory toxicology once it has been used in academic work and its useful properties have been explored in some detail.
CONCLUSIONS
Toxicologists continue to use genetically undefined outbred stocks, although the case for using inbred or F1 hybrid strains has been made repeatedly in the past, and has never been seriously criticized. Moreover, very few academic papers surveyed involved more than one strain, so toxicologists are often not aware that the responses they observe may differ to a considerable extent in a different strain.
The use of a multistrain experiment as part of a series of experiments involving the study of toxic mechanisms would alert toxicologists to the importance of genetic variation. Some investigators would then be able to start using modern tools of molecular genetics which would almost certainly lead to a better understanding of toxic mechanisms. However, until toxicologists start to use isogenic strains and begin to compare several strains as a routine part of their research, most of them will continue to be stuck in the dark ages as far as genetics is concerned.
REFERENCES
Evans, E.E., and M.V. Relling. 1999. Pharmacogenomics: translating functional genomics into rational therapeutics . Science 286:487-491.
Falconer, D.S. 1981. Introduction to quantitative genetics. 2nd edition. London: Longman.
Felton, R.P., and D.W. Gaylor. 1989. Multistrain experiments for screening toxic substances. J. Toxicol. Environ. Health 26:399-411.
Festing, M.F.W. 1987. Genetic factors in toxicology: Implications for toxicological screening . CRC Crit. Rev. Toxicol. 18:1-26.
Festing, M F.W. 1990. Contemporary issues in Toxicology: Use of genetically heterogeneous rats and mice in toxicological research: A personal perspective. Toxicol. Appl. Pharmacol. 102:197-204.
Festing, M.F.W. 1995. Use of a multi-strain assay could improve the NTP carcinogenesis bioassay program. Environ. Health Perspect. 103:44-52.
Festing, M.F.W. 1997a. Fat rats and carcinogen screening. Nature 388:321-322.
Festing, M.F.W. 1997b. Inbred strains of mice: a vital resource for biomedical research. Mouse Genome 95:845-855.
Festing, M.F.W. 1997c. Variation and its implications for the design of experiments in toxicological research. Comp. Haematol. Int. 7:202-207.
Festing, M.F.W. 1999. Warning: The use of genetically heterogeneous mice may seriously damage your research. Neurobiol. Aging 20:237-244.
Festing, M.F.W., and G.L. Wolff. 1995. Re-inbred strains of laboratory animals—Superior to outbred mice. J. Natl. Cancer Inst. 87:1715.
ILAR [Institute for Laboratory Animal Resources]. 1979. Laboratory animal management: Genetics. ILAR News 23:A1-A16.
Kacew, S., and M.F.W. Festing. 1996. Role of rat strain in the differential sensitivity to pharmaceutical agents and naturally occurring substances. J. Toxicol. Environ. Health 47:1-30.
Moore, C.J., W.A. Tricomi, and M.N. Gould. 1988. Comparison of 7,12-dimethylbenz[a]anthracene metabolism and DNA binding in mammary epithelial cells from three rat strains with differing susceptibilities to mammary carcinogenesis. Carcinogenesis 9:2099-2102.
Pohjanvirta, R., M. Viluksela, J.T. Tuomisto, M. Unkila, J. Karasinska, M.A. Franc, M. Holowenko, J.V. Giannone, P.A. Harper, J. Tuomisto, and A.B. Okey. 1999. Physicochemical differences in the AH receptors of the most TCDD-susceptible and the most TCDD-resistant rat strains. Toxicol. Appl. Pharmacol. 155:82-95.
Shellabarger, C.J., J.P. Stone, and S. Holtzman. 1978. Rat differences in mammary tumor induction with estrogen and neutron irradiation. J. Natl. Cancer Inst. 61:1505-1508.
Shirai, T., A. Nakamura, S. Fukushima, A. Yamamoto, M. Tada, and N. Ito. 1990. Different carcinogenic responses in a variety of organs, including the prostate, of five different rat strains given 3,2'-dimethyl-4-aminobiphenyl . Carcinogenesis 11:793-797.
Tuomisto, J.T., M. Viluksela, R. Pohjanvirta, and J. Tuomisto. 1999. The AH receptor and a novel gene determine acute toxic responses to TCDD: Segregation of the resistant alleles to different rat lines . Toxicol. Appl. Pharmacol. 155:71-81.
Vorhees, C.V., L.L. Morford, S.L. Inman, T.M. Reed, M.A. Schilling, G.D. Cappon, M.S. Moran, and D.W. Nebert. 1999. Genetic differences in spatial learning between Dark Agouti and Sprague-Dawley strains: Possible correlation with the CYP2D2 polymorphism in rats treated neonatally with methamphetamine. Pharmacogenetics 9:171-181.