SELECTION OF ANIMAL MODELS
The significance and validity with respect to usefulness in terms of extrapolatability of results generated in an animal model depend on the selection of a suitable animal model. A good knowledge of comparative anatomy and physiology is an obvious advantage when developing an animal model. Animal models may be found throughout the animal kingdom, and knowledge about human physiology has been achieved in species far removed from the human in terms of evolutionary development. A good example is the importance of the fruit fly for the original studies of basic genetics. Animal models are used in virtually every field of biomedical research. Virtually all medical knowledge and treatment, especially that in the last century, has involved work with laboratory animals. In the past 50 years alone, remarkable advances have been made in medicine, ranging from the development of a vaccine for polio to antibiotics for infectious diseases like pneumonia, and drugs for chronic disorders like high blood pressure, ulcers, and diabetes. Today, further progress is being made against these maladies and many others that continue to plague us, such as cancer, cystic fibrosis, AIDS, and Alzheimer disease.The rapid rise in medical knowledge after 1800 can be attributed in large part to the increased use of laboratory animals. Scientists recognized that there was little hope of coping with human illness until they first understood how living systems function when they are healthy. During the first quarter of the 19th century, animal studies were crucial for less than one-third of the major advances that occurred. Thereafter, research with animals contributed to more than half of the significant discoveries, and during some periods, animal studies accounted for more than 75% of the major advances. Two-thirds of the Nobel prizes awarded since 1901 have been for discoveries requiring the use of laboratory animals.
The increasing use of animals in research, testing, and teaching during the 19th century was followed by the emergence of the antivivisectionist (AV) movement, which became particularly prominent in Victorian England during the late 1800s. The British AV movement had separate origins and objectives from those of the animal welfare and anticruelty societies. The AVs were motivated by a fear of science and revulsion toward animal experimentation, and probably by underlying misanthropic sentiments. The current animal rights movement is derived from the Victorian AV movement. The new AVs have the same objectives and use the same arguments and tactics as did their 19th-century predecessors to persuade the public as to the correctness of their cause.
Today, breakthroughs in medicine are threatened by some well-financed animal rights/welfare organizations that seek to impede or stop the use of animals in research, testing, and teaching. Some of these groups claim that animals are no longer needed. One of the buzzwords today is "alternatives. However, it is a common misconception that "alternatives" mean complete replacements for animals. To scientists and other people concerned with animal welfare, the term also includes refinements in testing procedures to cause the animals less stress, or reductions in the number of animals required in a particular test, while maintaining the quality and validity of the scientific information obtained.
There are no real substitutes for the use of laboratory animals. Studies with bacteria, tissue culture, and computer simulation can provide useful information, but the complexity of living organisms requires research and testing on animals that are similar to humans to attain reliable and effective results. Blindness cannot be studied in bacteria or high blood pressure in tissue cultures. Surgery cannot be simulated on computers. Unforeseen side effects of inadequately tested new drugs or biomedical materials have proven the need for extensive animal tests before such drugs or materials can meet Food and Drug Administration (FDA) licensing requirements for human application.
The aim of using animal models in biomedical research is to reconcile biologic phenomena between species. I.E., we wish to examine systems existing in one species and extrapolate knowledge to another.
Researchers continually identify or develop new animal models to evaluate pathogenic mechanisms, diagnostic and therapeutic procedures, nutrition and metabolic disease, and the efficacy of novel drug development. Animal models for agricultural or food and fiber production research are also contributing data to studies of aging and disease processes of humans and animals.
The assumption that Homo sapiens are identical to other animals in his/her bodily functions has led to a number of errors in the history of medicine. Galen, the Greek physician and philosopher who lived in the 2nd century AD and was to become the leading medical authority for many centuries, is believed to be the founder of experimental physiology in the western world. His anatomical research was based almost entirely on studies of apes and pigs. Unhesitatingly, he transferred (i.e., extrapolated) his discoveries directly to humans, thus initiating many errors. The combination of Galens immense authority and the dogmatic prohibition by the Church of postmortem dissections of the human body conserved these errors well into the late 16th century. Galen was later blamed for using the wrong method. A closer look reveals that his mistake was to draw wrong conclusions from the results of these first scientific animal models because of uncritical interspecies extrapolation. Thus, the most brilliant design, the most elegant procedures, the purest reagents, along with investigator talent, public money, and animal life are all wasted if the choice of animal is incorrect.
In 1865, the famous French physiologist Claude Bernard published a book intended to give physicians rules and principles for the study of experimental medicine: Introduction to Study of Experimental Medicine. This work advocated the chemical and physical induction of disease in experimental situations, thus leading the way to the induced animal models of todays biomedical research. Bernard also emphasized the applicability of animal experiments to humans. The other authority in France who helped establish and popularize laboratory research as well as the experimental use of animals was Louis Pasteur. He and Robert Koch in Germany introduced the concept of specificity into medicine and the germ theory of disease. Kochs work on cholera and tuberculosis further promoted the use of animal models of infectious diseases and was followed by models for the screening and evaluation of new antibacterial drugs in the 20th century. Today, animal models are used in virtually every field of biomedical research.
Etymologically, the word animal derives from the Latin animal meaning soul/spirit, thus describing living organisms that are animated.
A model is an object of imitation, something that accurately resembles something else, a person or thing that is the likeness or image of another. The Holy Bible tells us that God said, Let us make man in our image, in our likeness, so God created man in his own image, in the image of God he created him; male and female he created them. God created man out of the dust of the ground and then breathed into his nostrils the breath of life to animate him. Thus, humans are animal models of God. Consequently, combining the two definitions, an animal model is an animated object of imitation, an image of Man (or other species), used to investigate a physio- or pathological circumstance in question.
The U.S. National Research Committee on Animal Models for Research on Aging attempted to define the term laboratory animal model as an animal model in which normative biology or behavior can be studied, or in which a spontaneous or induced pathological process can be investigated, and in which the phenomenon in one or more respects resembles the same phenomenon in humans or other species of animal.
What is very often meant by the term animal model is modeling humans. It is not the image of the used animal that is the focus of research but the analogy of the physiological behavior of this animal to our own (or another) species. It would, thus, be more correct to speak of human models in this context. Laboratory animal science and animal experiments are indeed much more about humans than about any other animal species.
A plethora of animal models has been used and is being used and developed for studies of biological structure and function in the human. The models may be exploratory, aiming to understand a biological mechanism, whether this is a mechanism operative in fundamentally normal biology or a mechanism associated with an abnormal biological function. Models may also be developed and applied as so-called explanatory models, aiming to understand a more or less complex biological problem. Explanatory models need not necessarily be reliant on the use of animals but may also be physical or mathematical model systems developed to unravel complex mechanisms. A third important group of animal models is employed as predictive models. These models are used with the aim of discovering and quantifying the impact of a treatment, whether this is to cure a disease or to assess toxicity of a chemical compound. The anatomy or morphology of the model structure of relevance to the studies may be of importance in all three of these model systems. The extent of resemblance of the biological structure in the animal with the corresponding structure in humans has been termed fidelity. A high-fidelity model with close resemblance to the human case may seem an obvious advantage when developing certain models. What is often more important, however, is the discriminating ability of the models, in particular the predictive models. When using models, for instance to assess the carcinogenicity of a substance, it is of the essence that at least one of the model species chosen responds in a manner that is predictive of the human response to this substance. Thus the similarity between humans and model species with respect to relevant biological mechanism is often more important than the fidelity of the model. Often the two go hand in hand, and high-fidelity models offer the best opportunity to study a particular biological function.
An animal model may be considered homologous if the symptoms shown by the animal and the course of the condition are identical to those of humans. Models fulfilling these requirements are relatively few, but an example is well-defined lesion syndromes in, for instance, neuroscience. An animal model is considered isomorphic if the animal symptoms are similar but the cause of the symptoms differs between human and model. However, most models are neither homologous nor isomorphic but may rather be termed "partial." These models do not mimic the entire human disease but may be used to study certain aspects or treatments of the human disease.
The majority of laboratory animal models are developed and used to study the cause, nature, and cure of human disorders. They may conveniently be categorized in one of the following five groups, of which the first three are the most important, as given in numerical order:
As the name implies, induced models are healthy animals in which the condition to be investigated is experimentally induced, for instance, the induction of diabetes mellitus with encephalomyocarditis virus, allergy against cow's milk through immunization with minute doses of protein, or partial hepatectomy to study liver regeneration. The induced-model group is the only category that theoretically allows a free choice of species. Although one might be tempted to presume that extrapolation from an animal species to the human is the better the closer this species resembles humans (high fidelity), phylogenetic closeness, as fulfilled by primate models, is not a guarantee for validity of extrapolation, as the unsuccessful chimpanzee models in acquired immunodeficiency syndrome (AIDS) research have demonstrated. It is just as decisive that the pathology and outcome of an induced disease or disorder in the model species resembles the respective lesions of the target species. Feline immunodeficiency virus infection in cats may therefore for many studies be a better model for human AIDS than is human immunodeficiency virus infection in simians. Although mice and rats have many biological characteristics in common, they do not necessarily serve equally well as models of human disease. For example, schistosomiasis (mansoni) infection may be studied in experimentally infected mice but not in rats, whose immune system is able to fight the infection effectively.
Most induced models are partial or isomorphic because the etiology of a disease experimentally induced in an animal is often different from that of the corresponding disease in the human. Few induced models completely mimic the etiology, course, and pathology of the target disease in the human.
These models of human disease use naturally occurring genetic variants (mutants). Many hundreds of strains and stocks with inherited disorders modeling similar conditions in humans have been characterized and conserved (see, e.g., http://www.jax.org). A famous example of a spontaneous mutant model is the nude mouse, which was a turning point in the study of heterotransplanted tumors and, for instance, enabled the first description of natural killer cells. Other famous spontaneous models include Snell's dwarf mice, without a functional pituitary, and the curly-tail mouse, in which fetuses develop a whole range of neural-tube defects. Many of the mutants are available in inbred strains, with corresponding coisogenic or congenic strains. This is very useful because the influence of just one affected gene or locus may then be studied against a reference strain with similar genetic background as the mutant.
An extensive literature is available on spontaneous models, and the majority of these are mice and rat models, although a wide range of mutants in many different species has been described.
The spontaneous models are often isomorphic, displaying phenotype similarity between the disease in the animal and the corresponding disease in the human - this is called face validity; for instance, type I diabetes in humans and insulin-requiring diabetes in the BB rat. This phenotypic similarity often extends to similar reactions to treatment in the model animal and the human patient, and spontaneous models have been important in the development of treatment regimens for human diseases.
However, if the object of a project is to study the genetic causes and etiology of a particular disease, then comparable genomic segments involved in the etiology of the disorder - construct validity - is normally a requirement. It should be remembered, however, that an impaired gene or sequence of genes very often results in activation of other genes and mobilization of compensating metabolic processes. These compensatory mechanisms may of course differ between humans and the animal model species.
The rapid developments in genetic engineering and embryo manipulation technology during the past decade have made transgenic disease models perhaps the most important category of animal disease models. A multitude of animal models for important diseases have been developed since this technology became available, and the number of models seems to be increasing quickly. Mice are by far the most important animals for transgenic research purposes, but farm animals and fish are also receiving considerable interest.
Many physiological functions are polygenic and controlled by more than one gene, and it will require considerable research activities to identify the contribution of multiple genes to normal as well as abnormal biological mechanisms. The insertion of DNA into the genome of animals or the deletion of specific genes gives rise to sometimes-unpredictable outcomes in terms of scientific results as well as in terms of animal well-being in the first generation of animals produced. Thereafter, transgenic lines can be selected and bred or cloned to avoid or select for a specific genotype. It is not an accurate science, although the methodology is constantly improving with the aim of eliminating unwanted effects. The embryo manipulation procedures themselves do not appear to affect the welfare of offspring in the mouse.
The recent completion of the maps of the genomes of mouse and human will increase the research activities in genomics and proteomics; and using high-density microarray DNA chip technology in human patients as well as in animals, it will be possible to investigate which genes are switched on or off in different diseases.
With both the human and the mouse genome maps available, this new technology is expected to rapidly increase knowledge on the genetic background and etiology of important diseases. This paves the way for a range of new homologous animal models with homology between animal and human (construct validity) for genotype as well as for phenotype. This development may result in a change in animal use from models for the identification of causative genes to models for studying the effects of changes in genetic pathways, gene-gene interactions and gene-environment interactions.
Negative model is the term used for species, strains, or breeds in which a certain disease does not develop, for instance, gonococcal infection in rabbits after an experimental treatment that induces the disease in other animals. Models of infectious diseases are often restricted to a limited number of susceptible species, and the remaining unresponsive species may be regarded as negative models for this particular human pathogenic organism. Negative models thus include animals that demonstrate a lack of reactivity to a particular stimulus. Their main application is in studies on the mechanism of resistance that seek to gain insight into its physiological basis.
Orphan model disease is the term that has been used to describe a functional disorder that occurs naturally in a non-human species but has not yet been described in humans and that is recognized when a similar human disease is later identified. Examples include Marek's disease, Papillomatosis, bovine spongiform encephalopathy, Visna virus in sheep, and feline leukemia virus. When humans are discovered to suffer from a disease similar to one that has already been described in animals, the literature already generated in veterinary medicine may be very useful.CHOOSING THE RIGHT MODEL
The selection of the laboratory species, breed, and strain to be used is one of the most important decisions to be made, and it should include consideration of non-animal methods. Over the years, many valuable non-animal models have been developed, refined, and extensively characterized. These models are useful in some types of research and testing, and they can often be used to supplement work with live animals.
Once it has been determined that the use of laboratory animals is necessary, the most appropriate species, breed, and strain must be selected. The objectives of the protocol play an important role in the selection of laboratory animal to be used. In the past, and often even today, it has been difficult to predict which organisms would yield the most useful insights and information for a given study. In biological research, model selection generally begins with a search for close homology that is judged to be good analogs. The validity of any information derived from the animal is dependent on its appropriateness. A good example of the importance of choosing the most appropriate animal for the study comes from research concerning vascular restenosis. Most of the data obtained in the pig and primate models of restenosis after angioplasty, contrary to those obtained in the smaller animal species, appear to be more closely related to the data obtained in humans. In the pig, pretreatment with an angiotensin converting enzyme (ACE) inhibitor, cilazapril, failed to attenuate carotid neointimal myoproliferation after either "deep" or "mild" balloon-mediated injury. In two other pig studies, ACE inhibition did not limit restenosis in the coronary circulation. Similarly, ACE inhibition with cilazapril in a primate model did not reduce intimal thickening in arteries injured by endarterectomy, balloon denudation, or synthetic vascular grafting. In contrast, pretreatment with ACE inhibition in the rat decreased neointima formation by 80% after balloon injury of the carotid artery. Other studies have also confirmed a similarly profound decrease in smooth muscle cell proliferation after balloon injury by pretreatment with cilazapril in the rat model.
Because new animal models are continually being identified and characterized, a thorough literature search should always be conducted to determine what models are available and which are the most relevant. There are several good general listings of available animal models for human disease. In addition, there are a number of books on models for specific areas of research.
Selection of a species should not be based solely on availability, familiarity, or cost. The readily available, familiar, or inexpensive species may not provide the genetic, physiologic or psychologic facets needed or wanted for the proposed project.
Mammals have been widely used, because of their obvious similarities in both structures and function to man. Rats, mice, guinea pigs, and hamsters came into favor because of their small size, short life span, ease of handling, and high reproductive rate. Vertebrates, especially mammals, provide essential one-to-one models for many specific human disease processes. The use of these one-to-one models has paid great dividends in understanding and controlling disease states as well as health, and is expected to continue to provide important information in biomedical research.
No single animal model can ever duplicate the original condition (i.e., the model is never the same as the prototype), and models never provide final answers, but only offer approximation. Of course, in a broad sense, experimental science itself is the study of approximations, and universal truths are rarely found.
It is virtually impossible to give specific rules for the choice of the best animal model, because the many considerations that have to be made before an experiment can take place differ with each research project and its objectives. Nevertheless some general rules can be given:
Useful web sites to assist in selecting appropriate animal models are:
1. Research Issues. Canadian Council on Animal Care. http://www.ccac.ca/en/CCAC_Programs/ETCC/Module05/toc.html
2. Animal Models & Strains Search Engine. Institute for Laboratory Animal Research. http://dels.nas.edu/ilar_n/ilarhome/search_amsst.shtml
Model selection is very much the prerogative of the individual scientist, who therefore is responsible for convincing the rest of the scientific community that her or his choice is valid. Contrary to an infamous quotation, a rat is not a pig is not a dog is not a boy!
When experimental results have been generated in an animal model, they have to be validated with respect to their applicability to the target species, which normally is the human. The term "extrapolation" is often used to describe how data obtained from animal studies reliably can be used to apply to the human. However, extrapolation is generally not performed in its mathematical sense, in which data fit a certain function that may be described graphically and the graph extends beyond the highest or lowest sets of data to describe a situation outside the window of observation. Establishing toxicity data in animals and using these to determine safe levels of exposure for people is perhaps what comes closest to mathematical extrapolation in animal studies. However, most studies of animal structure and function are never extrapolated to be applicable for describing the corresponding features in the human; this is not relevant. What laboratory animal experimentation is about is very similar to other types of experiments. Scientists aim to obtain answers to specific questions. Hypotheses are tested, and the answers are obtained, analyzed, and published. As an example of this, one might question the possible health hazards of a new synthetic steroid and ask a number of relevant questions to be answered in animal studies before deciding on the substance's potential usefulness as a human hormonal contraceptive. For instance, does it exist in the same form in humans and animals? How does it affect the estrus cycle in rodents? How does it affect endogenous hormone levels in rodents and other species? How soon after withdrawal do the animals revert to normal cyclicity? Does it interfere with pregnancy in rodents and fetal development in rodents and primates? Is the frequency of fetal malformation in mice affected? Is puberty, the ovarian cycle, and pregnancy in rodent and dog offspring of mothers treated with the substance affected? And so on. Analyzing the data from experiments of this nature would give information on the potential of the new synthetic steroid as a hormonal contraceptive in the human.
Although the predictive value of animal studies may seem high if they are conducted thoroughly and have included several species, uncritical reliance on the results of an mal tests can be dangerously misleading and has resulted in damages to human health in several cases, including those of some drugs developed by large pharmaceutical companies. What is noxious or ineffective in non-human species can be innoxious or effective in humans and vice versa; for instance, penicillin is fatal for guinea pigs but generally well tolerated by humans; and aspirin is teratogenic in cats, dogs, guinea pigs, rats, mice, and monkeys but obviously not in pregnant women, despite frequent consumption. Thalidomide, which crippled 10,000 children, does not cause birth defects in rats or many other species but does so in primates. A close phylogenetical relationship or anatomical similarity is not a guarantee of identical biochemical mechanisms and parallel physiological response, although such is the case in many instances.
The validity of extrapolation may be further complicated by the question, "To which humans?" As desirable as it often is to obtain results from a genetically defined and uniform animal model, the humans to whom the results are extrapolated are genetically highly variable, with cultural, dietary, and environmental differences. This may be of minor importance for many disease models but can become significant for pharmacological and toxicological models.It is not possible to give reliable general rules for the validity of extrapolation from one species to another. This has to be assessed individually for each experiment and can often only be verified after trials in the target species. An extensive and useful overview on the problem of predictive anthropomorphization, especially in the field of toxicology research, is Principles of Animal Extrapolation by Calabrese. The rationale behind extrapolating results to other species is based on the extensive homology and evolutionary similarity between morphological structures and physiological processes among different animal species and between animals and humans.
It is well known that the metabolic rate of small animals is much higher than that of large animals. It has also been demonstrated that capillary density in animals smaller than rabbits increases dramatically with decreasing body weight. However, considering that most animals are similar in having heart weights just above 0.5% of their body weight and a blood volume corresponding to 7% of the body weight, it becomes obvious that to supply the tissues of small animals with sufficient oxygen for their high metabolic rate, it is not sufficient to increase the stroke volume. The stroke volume is limited by the size of the heart, and heart frequency is the only parameter to increase, which results in heart rates well over 500 per minute in the smallest mammals. Other physiological variables, like respiration and food intake, are similarly affected by the high metabolic rate of small mammals.This means that scaling must be an object for some consideration when one calculates dosages of drugs and other compounds administered to animals in experiments. If the object is to achieve equal concentrations of a substance in the body fluids of animals of different body size, then the doses should be calculated in simple proportion to the animals' body weights. If the object is to achieve a given concentration in a particular organ over a certain time period, the calculation of dosage becomes more complicated, and other factors, including the physicochemical properties of the drug, become important. Drugs and toxins exert their effect on an organism not per se but because of the way that they are metabolized, the way that they and their metabolites are distributed and bound in the body tissues, and how and when they are finally excreted.
However, metabolism or detoxification and excretion of a drug are not directly correlated with body size but, more accurately, to metabolic rate of the animal.
M = 3.8 x >BW-0.25
where M is the metabolic rate (oxygen consumption in milliliters per gram of body weight per hour) and BW is body weight in grams. This equation may be used to calculate dosages for animals of different body weights if the dose for one animal (or human) is known.
Dose1/Dose 2 = BWl-0.25/BW2 -0.25
Dose1 = Dose2 x BW1-0.25/BW2 -0.25The equations should be considered as assistance for calculating dosages, but caution should be exerted with respect to too broad a generalization of their use, and the 0.50 power of body weight should be employed when dealing with animals with body weights of <100 g.> Some species react with particular sensitivity toward certain drugs, and marked variations in the reaction of animals within a species occur with respect to strain, pigmentation, nutritional state, time of day, stress level, type of bedding, ambient temperature, and so on.
Regardless of the animal model chosen, one is obliged to provide exact descriptions of the model. This includes much more than the commonly occurring statement that albino mice were used for all studies. Other identifiers that should be provided include genetic strain and sub-strain and, if applicable, special genetic features, using the correct international nomenclature. The microbial status of the animals also needs to be specified. Conventional, specific pathogen-free, and axenic mice differ in their responses to various agents and in their physiology. These descriptions should be provided, as well as the methods and results obtained when microbial or pathogen status was verified. Age, animal housing, normal maintenance, diet and husbandry also should be given; i.e., a rat maintained on autoclaved food and sterile distilled water in suspended wire cages in an isolator is a drastically different animal than a genetically identical rat maintained in an open shoe-box cage in a conventional animal room. Other information that is required in the definition of an infectious process is the strain of the organism, the methods by which the inoculum was prepared, the route by which it was inoculated, and the dose that was used. Even the decision whether to use anesthesia during inoculation may affect the model. For example, for years we thought that infection of pathogen-free F344 rats with dosages of 106 or more colony-forming units (CFU) of a single strain of Mycoplasma pulmonis delivered as an intranasal bolus was a reasonable model of naturally occurring murine respiratory mycoplasmosis (MRM). In reality, organism load in natural infection is likely to be in the range of 101 to 102 CFU/animal in barrier-maintained colonies. In addition, we now know that different genetic strains of mice and rats respond differently; different strains of organism vary in pathogenicity, and a wide variety of factors, especially ammonia, also affect the severity of the disease. Thus, our original model simulated only the most severe form of naturally occurring MRM, usually seen only in conventional rodents. However, the study of this highly artificial model led to the identification of all the factors listed above, all of which are involved in natural MRM.CONCLUSION
The choice or selection of animal model depends on a number of factors relating to the hypothesis to be tested but also on more practical aspects associated with the project and with project staff and experimental facilities. The usefulness of a laboratory animal model should be judged on how well it answers the specific questions it is being used to answer, rather than on how well it mimics the human disease. The appropriateness of any laboratory animal model will eventually be judged by its capacity to explain and predict the observed effects in the target species.REFERENCES
Calabrese, E.J.: Principles of Animal Extrapolation. Wiley, NY, 1983.
Capechhi, M.: Targeted Gene Replacement. Scientific American. March 1994, pp 52-59.
Davidson, M.K., Lindsey, J.R., and Davis, J.K. Requirements and selection of an animal model. Israel Journal of Medical Sciences 1987; 23:551-555.
Hau, J. and VanHoosier, Jr., G.L.: Handbook of Laboratory Animal Science, Second Edition. CRC Press, Boca Raton, FL, 2003.
Lewis, S.M. and Carraway, J.H.: Large Animal Models of Human Disease. Lab Animal. January 1992, pp. 22-29.
Rollin, B.E. and Kesel, M.L.: The Experimental Animal in Biomedical Research, volume I. CRC Press, Boca Raton, FL, 1990.
Thompson, F.C.: The Thompson Chain-Reference Bible: New International Version. B.B. Kirkbride Bible Co., Indianapolis, IN, 1983.
Wright, K.C.: Working with laboratory animals: general principles and practical considerations. Journal of Vascular and Interventional Radiology. May-June 1997, pp. 363-373.
This page designed by Karelia Coronado