Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Criteria for Selecting Experimental Animals SPECIES AND STOCKS Choosing a Species for Study For a scientific investigation to have the best chance of yielding useful results, all aspects of the experimental protocol should be carefully planned. If animal models will be used, an important part of the process is to con- sider whether nonanimal approaches exist. If, after careful deliberation and review of the existing literature, the investigator is satisfied that there are no suitable alternatives to the use of live animals for the study in question, the next question that should be addressed is what species would be most appropriate to use. In choosing a species for study, it is important to weigh a variety of scientific and operational factors, including the following: . In which species is the physiologic, metabolic, behavioral, or dis- ease process to be studied most similar to that of humans or other animals to which the results of the studies will be applied? · Do other species possess biologic or behavioral characteristics that make them more suitable for the planned studies (e.g., generation time and availability'? · Does a critical review of the scientific literature indicate which species has provided the best, most applicable historical data? 16
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS 17 · Do any features of a particular species or strain including ana- tomic, physiologic, immunologic, or metabolic characteristics render it in- appropriate for the proposed study? · In light of the methods to be used in the study, would any physical or behavioral characteristics of a particular species make the required physical manipulation or sampling procedures impossible, subject to unpredictable failure, or difficult to apply? . Does the proposed study require animals that are highly standard- ized either genetically or microbiologically? Those and other considerations often lead to the selection of a laboratory rodent species as the most appropriate model for a biomedical research protocol. Rodents are generally easy to obtain and relatively inexpensive to acquire and maintain. Other advantages of laboratory rodents as research models include small size, short generation time, and availability of micro- biologically and genetically defined animals, historical control data, and well-documented information on ohvsiolo~ic~ Pathologic, and metabolic ,,_^, ~ r--J~~~~~~~~7 rip processes. The order Rodentia encompasses many species. The most commonly used rodents are laboratory micel, laboratory rats (Rattus norvegicus), guinea pigs (Cavia porcellus), Syrian hamsters (Mesocricetus auratus), and gerbils (Meriones unguiculatus). All those rodents have been extensively studied in the laboratory, and information about them can be found in the peer- reviewed literature and in a number of texts (e.g., Altman and Katz, 1979a,b; Baker et al., 1979-1980; Foster et al., 1981-1983; Fox et al., 1984; Gill et al., 1989; Darkness and Wagner, 1989; Van Hoosier and McPherson, 1987; Wagner and Manning, 19761. Rodent Stocks The same factors used in selecting a species for study can be used in selecting a rodent stock. Rodents have been maintained in the laboratory environment for more than 100 years. Some, such as the mouse, have been very well characterized genetically and have undergone genetic manipula- tion to produce animals with uniformly heritable phenotypes. A hallmark of good scientific method is reproducibility, which is accomplished by minimizing and controlling extraneous variables that can alter research results. In stud- ies that are mechanistic, genetic uniformity is highly desirable. In contrast, genetic uniformity might be undesirable in studies that explore the diversity 1 laboratory mice are neither pure Mus domesticus nor pure Mus musculus; therefore, geneticists have determined that there is no appropriate scientific name (International Commit- tee on Standardized Genetic Nomenclature for Mice, 1994a).
18 RODENTS: LABORATORY ANIMAL MANAGEMENT of application of a phenomenon over a range of phenotypes, such as prod- uct-registration studies, including safety evaluation of compounds that have therapeutic potential. In many such studies, a varied genetic background might be appropriate, as long as the range of variation can be characterized and is to some degree reproducible (Gill, 19801. Genetically Defined Stocks Inbred Strains. The mating of any related animals will result in inbreed- ing, but the most common and efficacious method for establishing and main- taining an inbred strain is brother x sister (i.e., full-sib) mating in each genera- tion. Full-sib inbreeding for 20 generations will result in more than 98 percent genetic homogeneity, at which point the members of the stock are isogenic, and the stock is considered an inbred strain. Many inbred strains of mice and rats have been developed (Festing, 1989; Festing and Greenhouse, 1992), and they are widely used in biomedical research. Many of the commonly used strains have been inbred for over 200 generations. A few inbred strains of guinea pigs, Syrian hamsters, and gerbils have also been developed (Altman and Katz, 1979b; Festing, 1993; Hansen et al., 1981~. The isogeneity of the members of an inbred strain provides a powerful research tool. Although some genes might remain heterogeneous, most metabolic or physiologic processes, as well as their phenotypic expression, will be identical among individuals of an inbred strain, thereby eliminating a source of experimental variation. Isogeneity also allows exchange of tissue between individuals of an inbred strain without rejection. F1 Hybrids. F1 hybrid animals are the first filial generation (the F1 generation) of a cross between two inbred strains. They are often more hardy than animals from either of the parental strains, having what is called hybrid vigor. F1 hybrids are heterozygous at all genetic loci at which the parental strains differ; nevertheless, they are uniformly heterozygous. Be- cause of the heterogeneity, F1 hybrids will not breed true; to produce them one must always cross animals of the parental inbred strains. Reciprocal hybrids are developed by reversing the strains from which the dam and the sire are taken. Reciprocal male hybrids will have Y-chromosome differ- ences. Reciprocal female hybrids will have identical genotypes but might have differences caused by inherited maternal effects. F1 hybrids will ac- cept tissue from either parental strain, except in the case of a Y-chromo- some incompatibility (e.g., a skin graft from a male of either parental strain will be rejected by a female F1 hybrid). Special Genetic Stocks. The effects of specific genes or chromosomal regions can be studied by using various breeding or gene manipulation
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS 19 methods to create a new strain that differs from the original strain by as little as a single gene. · A segregating inbred strain is an inbred strain maintained by full- sib matings; however, male-female pairs are selected for mating so that one pair of genes will remain heterozygous from generation to generation. This method of mating permits well-controlled experiments because a single sibship contains both carriers and noncarriers of the gene of interest, and all the animals are essentially identical except for that gene. · A coisogenic strain is an inbred strain in which a single-gene muta- tion has occurred and has been preserved; it is otherwise identical with the nonmutant parental strain. If the mutation is not deleterious when homozy- gous, the strain can be maintained by simple full-sib matings. If the muta- tion adversely affects breeding performance, the coisogenic strain can be maintained by one of several special breeding systems (Green, 1981; NRC, 1989~. To avoid subline divergence between the coisogenic strain and the nonmutant parental inbred strain, periodic back-cro$sing (see next para- graph) with the parental strain is recommended. · A congenic strain is a close approximation to a coisogenic strain. It is created by mating an individual that carries a gene of interest, called the differential gene, with an individual of a standard inbred strain. An offspring that carries the differential gene is mated to another individual of the same inbred strain. This type of mating, called back-crossing, is contin- ued for at least 10 generations to produce a congenic strain. Back-crossing for 10 generations minimizes the number of introduced genes other than the differential gene and its closely linked genes. Details on developing congenic strains have been published (Bailey, 1981; Green, 1981~. Both coisogenic and congenic strains can be maintained by full-sib matings if the differen- tial gene is homozygous; however, to avoid subline divergence between the congenic strain and the standard inbred strain, periodic back-crossing with the standard strain is recommended. . A transgenic strain is similar to a coisogenic or congenic strain in that it carries a segment of genetic information not native to the strain or individual (Hogan et al., 1986; Merlino, 1991~. The introduced genetic mate rial can be from the same or another species. Transgenic animals are de scribed in more detail in Chapter 8. . Recombinant inbred (RI) strains are sets of inbred strains produced primarily to study genetic linkage. Each RI strain is derived from a cross between two standard inbred strains. Animals from the F1 generation are then bred to produce the second filial generation (the F2 generation), mem- bers of which are randomly selected and mated to produce a series of RI lines. Members of the F2 generation are used to found RI lines because, unlike the F1 generation, they are not isogenic. The mice derived from any parental pair will be genetically homogeneous when inbreeding is complete;
20 RODENTS: LABORATORY ANIMAL MANAGEMENT however, each line in a set will be homozygous for a given combination of alleles originating from the two parental inbred strains. Alleles that are linked in the parental strains will tend to remain together in the RI lines; this is the basis for their use in genetic-mapping studies. · Recombinant congenic strains are like recombinant inbred strains except that each strain of a series has been derived from a back-cross in- stead of an F2 cross (Demant, 1986~. The number of back-crosses made before full-sib inbreeding is started determines the proportion of genes from each of the parental inbred strains. Series of recombinant congenic strains are particularly useful in the genetic analysis of multiple-gene systems, such as that responsible for cancer susceptibility. Nongenetically Defined Stocks The terms noninbred, random-mated, and outbred are all used to refer to populations of animals in which, theoretically, there is no genetic uniformity between individuals. Nongenetically defined stocks make up the majority of rodents used in biomedical research and testing, and they are generally less expensive and more readily available than genetically defined stocks. Noninbred refers to a population of animals in which no purposeful inbreeding system has been established. Random-mated refers to a group of animals in which the selection of breeding animals is random. It assumes an almost infinite population with no external selection pressures. In prac- tice, such a colony probably does not exist. Outbred refers to a colony in which breeding is accomplished by a purposeful scheme that minimizes or eliminates inbreeding. Animals produced by these breeding systems have varied genotypes, and characterizing the range and distribution of pheno- types requires a large sample of the population. The degree of heterozygosity in any nongenetically defined stock is continuously varying, so two populations developed from the same parental stock will show differing degrees of heterozygosity at any loci at any time. Spontaneous mutations can occur and become fixed because no purposeful selection is imposed on the population to eliminate the mutant genes. Out- bred populations are always evolving and therefore are more variable than inbred strains. For that reason, large sample numbers are needed to account for phenotypic variation that could have an impact on the charactersitics being studied. If outbred animals are used, treatment and control groups in a study will not necessarily be identical, nor will the population of animals necessarily be identical if the study is repeated. The genetic variation in outbred stocks, which can be magnified by sampling error, can make results from different laboratories difficult to compare. Background data on stock characteristics will vary over time, so concurrent controls are needed to allow useful interpretation of data.
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS STANDARDIZED NOMENCLATURE FOR RODENTS 21 Standardized nomenclature allows scientists to communicate briefly and precisely the genetics of their research animals. The International Commit- tee on Standardized Genetic Nomenclature for Mice and the International Rat Genetic Nomenclature Committee, which are affiliated with the Interna- tional Council for Laboratory Animal Science, are responsible for maintain- ing the nomenclatures for genetically defined mice and rats, respectively, and modifying them as necessary. The sections below briefly describe the nomenclature for inbred, mutant, and outbred mice and rats. The complete rules for mice can be found in the third edition of Genetic Variants and Strains of the Laboratory Mouse (Lyon and Searle, in press). Those rules are regularly updated, and updates are published in Mouse Genome (for- merly called Mouse News Letter; Oxford University Press) and are available on-line in MOD, the Mouse Genome Database. Information on MOD can be obtained from the Mouse Genome Informatics Group, The Jackson Laboratory, Bar Harbor, ME 04609 (telephone, 207-288-3371; fax, 207-288-5079; Internet, firstname.lastname@example.org). The rules for rats have been published as an appendix to the report Definition, Nomenclature, and Conservation of Rat Strains (NRC, 1992a), and updates will be published in Rat Genome, Heinz W. Kunz, Ph.D., editor, Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261. Investigators using other labo- ratory rodents should follow the rules for mice or rats. Inbred Strains An inbred strain is designated by capital letters (e.g., mouse strains AKR and CBA and rat strains BN and LEW). The mouse rules, but not the rat rules, allow the use of a combination of letters and numbers, beginning with a letter (e.g., C3H), although this type of symbol is considered less desirable. Brief symbols (generally one to four letters) are preferred. Ex- ceptions are allowed for strains that are already widely known by designa- tions that do not conform (e.g., mouse strains I01 and 129 and rat strains F344 and DONRYU). Substrains An established strain is considered to have divided into substrains when genetic differences are known or suspected to have become established in separate branches. These differences can arise either from residual het- erozygosity at the time of branching or from new mutations. A substrain is designated by the full strain designation of the parent strain followed by a slanted line (slash) and an appropriate substrain symbol, as follows:
22 RODENTS: LABORATORYANIMAL MANAGEMENT · Mice. The substrain symbol can be a number (e.g., DBA/1 and DBA/ 2~; a laboratory code, which is defined below (e.g., C3H/lIe, where He is the laboratory code for Walter E. Heston); or, when one investigator or laboratory originates more than one substrain, a combination of a number and a laboratory code, beginning with a number (e.g., C57BL/6J and C57BL/ lOJ, where J is the laboratory code for the Jackson Laboratory, Bar Harbor, Maine). Exceptions, such as lower-case letters, are allowed for already well-known substrains (e.g., BALB/c and C57BR/cd). · Rats. The substrain symbol is always a number when genetic differ- ences have been demonstrated. The founding strain is considered the first substrain, and the use of /1 for it is optional (e.g., KGH or KGH/11. A laboratory code (e.g., Pit for the University of Pittsburgh Department of Pathology and N for the NIH Genetic Resource) is used to designate a substrain when genetic differences are probable but not demonstrated (e.g., BN/Pit and BN/N). Laboratory Codes Each laboratory or institution that breeds rodents should have a laboratory code. The registry of laboratory codes is maintained by ILAR, National Re- search Council, 2101 Constitution Avenue, Washington, DC 20418 (telephone, 202-334-2590; fax, 202-334-1687; URL:http://www2.nas.edulilarhome/~. The laboratory code, which can be used for all laboratory rodents, consists of either a single roman capital letter or an initial roman capital letter and one to three lower-case letters. . Mice. A particular colony is indicated by appending an "@" sign and the laboratory code to the end of the strain or substrain symbol (e.g., SJL@J, the colony of strain SJL mice bred at the Jackson Laboratory; C3H/ He@N, the He substrain of strain C3H bred at the NIH Genetic Resource; and CBA/Ca-se@J, the Ca substrain of strain CBA carrying the se mutation and bred at the Jackson Laboratory). If the substrain symbol and laboratory code are the same, the @ symbol and the laboratory code can be dropped for simplicity (e.g., SJL/J@J becomes SJL/J). The laboratory code is al- ways the last symbol used and is meant to indicate that the environmental conditions and previous history of a colony are unique. When a strain is transferred to a new laboratory, the laboratory code of the originating labo ratory is dropped, and the code of the recipient is appended; laboratory codes are not accumulated. . Rats. Normally, a rat strain is designated by the strain name, a slash, the substrain designation (if any), and the laboratory code (e.g., BN/ [Pit). When a strain is established in another laboratory, the new labora- tory code is appended (e.g., BN/lPitN). In general, more than two labora- tory codes are not accumulated. Intermediate codes are dropped to avoid excessively long designations.
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS 23 For both mice and rats, a strain's holder is responsible for maintaining a strain history. F1 Hybrids An F1 hybrid is designated by the full strain designation of the female parent, a multiplication sign, the full strain designation of the male parent, and F1 (e.g., the hybrid mouse C57BL/6J x DBA/2J Fl and the hybrid rat F344/NNia x BN/RijNia Fig. If there is any chance of confusion, parenthe- ses should be used to enclose the parental strain names te.g., (C57BL/6J x DBA/21)F1 and (F344/NNia x BN/RijNia)Fl]. The correct formal name should be given the first time the hybrid is mentioned in a publication; an abbreviated name can be used subsequently [e.g., C57BL/6J x DBA/2J F1 (hereafter called B6D2F1) and F344/NNia x BN/RijNia F1 (hereafter called FBNF1~. Coisogenic, Congenic, and Segregating Inbred Strains In mice, a coisogenic strain is designated by the strain symbol, the substrain symbol (if any), a hyphen, and the gene symbol in italics (e.g., CBA/H-kd). When the mutant or introduced gene is maintained in the heterozygous condition, this is indicated by including a slash and a plus sign in the symbol (e.g., CBA/H-kdl+~. A congenic strain is designated by the full or abbreviated symbol of the background strain, a period, an abbre- viated symbol of the donor strain, a hyphen, and the symbol of the differen- tial locus and allele (e.g., B10.129-H12b). Segregating inbred strains are designated like coisogenic strains; however, indication of the segregating locus is optional when it is part of the standard genotype of the strain (e.g., 129/J and 129/J-CCh/C mean the same thing, and either can be used). In rats, a coisogenic strain (except for alloantigenic systems see NRC, 1992a) is designated like a coisogenic strain in mice, except that the labora- tory code follows the substrain symbol and the gene symbol is not italicized (e.g., RCS/SidN-rdy). A con~enic rat strain (except for alloantigenic sys _ _ ~= ~ A tems) is designated l~ke a coisogenic strain (e.g., LEW/N-rnu). For segre gating inbred strains developed by inbreeding with forced heterozygosis, indication of the segregating locus is optional. Recombinant Inbred (RI) Strains The symbol of an RI strain should consist of an abbreviation of both parental-strain symbols separated by a capital X with no intervening spaces (e.g., CXB for an RI strain developed from a cross of BALB/c and C57BL mouse strains and LXB for an RI strain developed from a cross of LEW and BN rat strains). Different RI strains in a series should be distinguished by numbers (e.g., CXB1 and CXB2 in mice and LXB1 and LXB2 in rats).
24 RODENTS: LABORATORYANIMAL MANAGEMENT Genes The rules for gene nomenclature are very complicated because they apply not only to mutant genes, but also to gene complexes, biochemical variants, and other special classes of genes (e.g., transgenes). This descrip- tion will cover only a small portion of the gene nomenclature. The full rules can be found in the references given previously. The symbols for loci are brief and are chosen to convey as accurately as possible the characteristic by which the gene is usually recognized (e.g., coat color, a morphologic effect, a change in an enzyme or other protein, or resemblance to a human disease). Symbols for loci are typically two- to four-letter abbreviations of the name. For mice, the symbols are written in italics; for rats, they are not. For convenience in alphabetical lists, the initial letter of the name is usually the same as the initial letter of the symbol. Arabic numbers are included for proteins in which a number is part of the recognized name or abbreviation (e.g., in mice, C4 and C6, the fourth and sixth components of complement, respectively; in rats, C4 and C63. Except in the case of loci discovered because of a recessive mutation, the initial letter of the locus symbol is capitalized and all other letters are lower-case. Hyphens are used in gene symbols only to separate characters that together might be confusing. This rule was adopted for mice in 1993, and hyphens should be deleted from all gene symbols except where they are necessary to avoid confusion. Gene designations are appended to the desig- nation of the parental strain, and they are separated by a hyphen. Loci That Are Members of a Series A locus that is a member of a series whose members specify similar proteins or other characteristics is designated by the same letter symbol and a distinguishing number (e.g., Esl, Es2, and Es3 in mice and Esl, Es2, and Es3 in rats). For morphologic or "visible" loci with similar effects (e.g., genes that cause hairlessness', distinctive names are given because the gene actions and gene products can ultimately prove to be different (e.g., hr and nu in mice and fz and mu in rats). Alleles An allele is designated by the locus symbol with an added superscript. For mice, the superscript is written in italics; for rats, it is not. An allele superscript is typically one or two lower-case letters that convey additional information about the allele. For mutant genes, no superscript is used for the first discovered allele. When further alleles are found, the first is still designated without a superscript (e.g., nu for nude and nuStr for streaker in
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS - 25 mice and fa for fatty and faCP for corpulent in rats). If the information is too complex to be conveyed conveniently in the symbol, the allele is given a superscript (e.g., Esla and Eslb in mice and Esla and Eslb in rats), and the information is otherwise conveyed. Indistinguishable alleles of independent origin (e.g., recurrences) are designated by the gene symbol with a series symbol, consisting of an Arabic number corresponding to the serial number of the recurring allele plus the laboratory code, appended as a superscript in italics. To avoid confusing the number "1" and the lower-case letter "1," the first discovered allele is left unnumbered, and the second recurring allele is numbered 2 (e.g., bg, beige; bgJ, a recurrence of the mouse muta- tion bg at the Jackson Laboratory; and bg27, a second recurrence of the mutation bg at the Jackson Laboratory). A mutation or other variation that occurs in a known allele (except for alloantigenic systems in the rat) is designated by a superscript m and an appropriate series symbol, which consists of a number corresponding to the serial number of the mutant allele in the laboratory of origin plus the labo- ratory code. The symbol is separated from the original allele symbol by a hyphen (e.g., Mupla-m7~ for the first mutant allele of mouse Mupla found by the Jackson Laboratory). For a known deletion of all or part of an allele, the superscript m may be replaced with the superscript dl. This nomencla- ture is used for naming targeted mutations (often called "knockout" muta- tions), as well as spontaneously occurring ones. 7 _ Transgenes Nomenclature for transgenes was developed by the ILAR Committee on Transgenic Nomenclature (NRC, 1992b). A transgene symbol consists of three parts, all in reman type, as follows: TgX(YYYYYY)#####Zzz, where TgX is the mode, (YYYYYY) is the insert designation, and #####Zzz represents the laboratory-assigned number (#####) and laboratory code (Zzz). The mode designates the transgene and always consists of the letters Tg (for "transgene") and a letter designating the mode of insertion of the DNA: N for nonhomologous recombination, R for insertion via infection with a retroviral vector, and H for homologous recombination. The purpose of this designation is to identify it as a symbol for a transgene and to distinguish between the three fundamentally different organizations of the introduced sequence relative to the host genome. When a targeted mutation introduced by homologous recombination does not involve the insertion of a novel functional sequence, the new mutant allele (the knockout mutation) is des- ignated in accordance with the guidelines for gene nomenclature for each species. The gene nomenclature is also used when the process of homolo
26 RODENTS: LABORATORYANIMAL MANAGEMENT gous recombination results in integration of a novel functional sequence, if that sequence is a functional drug-resistance gene. For example, Mbpm~Dn would be used to denote the first targeted mutation of the myelin basic protein (Mbp) in the mouse made by Muriel T. Davisson (Dn). In this example, the transgenic insertion, even if it contains a functional neomycin- resistance gene, is incidental to "knocking out" or mutating the targeted locus (see also International Committee on Standardized Genetic Nomen- clature for Mice, 1994b). The insert designation is a symbol for the salient features of the transgene, as determined by the investigator. It is always in parentheses and consists of no more than eight characters: letters (capitals or capitals and lower-case letters) or a combination of letters and numbers. Italics, superscripts, sub- scripts, internal spaces, and punctuation should not be used. Short symbols (six or fewer characters) are preferred. The total number of characters in the insert designation plus the laboratory-assigned number may not exceed 11 (see below); therefore, if seven or eight characters are used, the number of digits in the laboratory-assigned number will be limited to four or three, respectively. The third part of the symbol is a number and letter combination that uniquely identifies each independently inserted sequence. It is formed of two components. The laboratory-assigned number is a unique number that is assigned by the laboratory to each stably transmitted insertion when germline transmission is confirmed. As many as five characters (numbers as high as 99,999) may be used; however, the total number of characters in the insert designation plus the laboratory-assigned number may not exceed 11. No two lines generated within one laboratory should have the same assigned number. Unique numbers should be given even to separate lines with the same insert integrated at different positions. The number can have some intralaboratory meaning or simply be a number in a series of transgenes produced by the laboratory. The second component is the laboratory code. Thus, the complete designation identifies the inserted site, provides a sym- bol for ease of communication, and supplies a unique identifier to distin- guish it from all other insertions [e.g., C57BL/6J-TgN(CD8Ge)23Jwg for the human CD8 genomic clone inserted into C57BL/6 mice from the Jack- son Laboratory (J' and the 23rd mouse screened in a series of microinjec- tions done in the laboratory of Jon W. Gordon (Jwg)~. The complete rules for naming transgenes have been published (NRC, 1992b). TBASE, a database developed at Oak Ridge National Laboratory, Oak Ridge, Tennessee, as a registry of transgenic strains, is maintained at the Johns Hopkins University, Baltimore, Maryland. Information on TBASE can be obtained from the Genome Database and Applied Research Labora- tory, The Johns Hopkins University, 2024 East Monument Street, Balti- more, MD 21205 (telephone, 410-955-1704; fax, 410-614-04341.
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS Outbred Stocks 27 An outbred-stock designation consists of a laboratory code, a colon, and a stock symbol that consists of two to four capital letters (e.g., mouse stock Crl:ICR and rat stock Hsd:LE). The 'stock symbol must not be the same as that for an inbred 'strain of the same species. As an exception, a stock derived by outbreeding a formerly inbred strain may continue to use the original symbol; in this case, the laboratory code preceding the stock symbol characterizes the stock as outbred. An outbred stock that contains a specified mutation is designated by the laboratory code, a colon, the stock symbol, a hyphen, and the gene symbol (e.g., Crl:ZUC-fa). The transfer of an outbred stock between breeders is indicated by list- ing the laboratory code of the new holder followed by the laboratory code of the holder the stock was obtained from (e.g., HsdBlu:LE for rats obtained by Harlan Sprague Dawley from Blue Spruce Farms). To avoid excessively long designations, only two laboratory codes should be used. QUALITY In selecting rodents for use in biomedical research, consideration should be given to the quality of the animals. Quality is most commonly character- ized in terms of microbiologic status and of the systems used in raising animals to ensure that a specific microbiologic status is maintained. How- ever, the genetics of an animal, as well as the genetic monitoring and breed- ing programs used to ensure genetic consistency, clearly also play an im- portant part in defining rodent quality. Microbiologic Quality Rodents can be infected with a variety of adventitious pathogenic and opportunistic organisms that under the appropriate circumstances can influ- ence research results at either the cellular or subcellular level. Some of those agents can persist in animals throughout their lives; others cause tran- sient infections and are eliminated from the animals, leaving lasting sero- logic titers as the only 'indicators that ' the organisms were present. The types of organisms that can infect rodents include bacteria, protozoa, yeasts, fungi, viruses, rickettsia, mycoplasma, and such nonmicrobial agents as helminths and arthropods. Many of the common organisms that infect laboratory rodents have been studied extensively, and some of their research interactions have been characterized (see Bhatt et al., 1986;NRC, 1991, for review). Unfortu- nately, information about the effects of many other organisms is incomplete or is not available. There is no general agreement on the importance of
28 RODENTS: LABORATORYANIMAL MANAGEMENT many organisms that latently infect rodents, especially opportunistic organ- isms that cause disease or alter research results only under narrowly defined conditions and even then usually affect only a very small proportion of the population. Any decision on the quality of rodents to be selected for a particular research project should include a realistic assessment of the or- ganisms that have a reasonable probability, as determined by documentation in the peer-reviewed literature, of producing confounding effects in the proposed study. It is commonly assumed that animals for which the most extensive health monitoring has been done and to which the most rigorous techniques for excluding microorganisms have been applied are the most appropriate for use in all studies. However, for both scientific and practical reasons, that assumption is not always valid. Rodents that are free of all microor- ganisms (axenic rodents, see definition below) or axenic rodents that have purposely been inoculated with a few kinds of nonpathogenic microorgan- isms (microbiologically associated rodents) can have altered physiologic and metabolic processes that make them inappropriate models for some studies. They can also rapidly become contaminated with common micro- organisms unless they are maintained with specialized housing and hus- bandry measures, which are expensive and can fail. The commercial avail- ability of such rodents is limited, and they are more expensive than rodents in which the microbial burden is not so restricted. For those reasons, the rodents most commonly used in research are ones that are free of a few specific rodent pathogens and some other microorganisms that are well known to have confounding effects on specific kinds of research. The quality of laboratory animals is generally related to the microbiologic exclusion methods used to breed and maintain them. There are three major types of maintenance: isolator-maintained, barrier-maintained, and no-con- tainment or conventionally maintained animals. An isolator is a sterilizable chamber that is usually constructed of metal, rigid plastic, vinyl, or polyure- thane. It usually has a sterilized air supply, a mechanism for introducing sterilized materials, and a series of built-in gloves to allow manipulation of the animals housed within. All materials moved into the isolator are sterilized, and animals raised within the isolator are generally maintained free from con- tamination by either all or specified microorganisms. Barrier-maintained animals are bred and kept in a dedicated space, called a barrier. For barrier facilities, personnel enter through a series of locks and are usually required to disrobe, shower, and use clean, disinfected clothing. All body surfaces that will potentially make contact with animals are cov- ered. All equipment, supplies, and conditioned air provided to the barrier facility are sterilized or disinfected. Barrier facilities can be of any size and can consist of one or more rooms. They are designed to exclude organisms for which rodents are the primary or preferred hosts but generally will not exclude organisms for which humans are hosts.
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS 29 Barrier maintenance can also be achieved at the cage or rack level with equipment that can be sterilized or otherwise disinfected. This type of maintenance depends heavily on providing large volumes of filtered or ster- ilized air to the animal cages. Such systems can be used successfully to maintain animals with a highly defined microbiologic status; the success of such systems depends on the techniques used and is difficult to monitor because microbiologic status might differ from cage to cage. No-containment, or conventionally maintained, animals are raised in areas that have no special impediments to the introduction of microorganisms. This method of maintaining animals cannot ensure stability of the microbiologic status, because unwanted organisms can be introduced at any time. Several classifications have been developed to define the microbiologic quality of laboratory animals, as follows (see also NRC, 1991~: · Axenic refers to animals that are derived by cesarean section or embryo transfer and reared and maintained in an isolator with aseptic tech- niques. It implies that the animals are demonstrably free of associated forms of life, including viruses, bacteria, fungi, protozoa, and other saprophytic or parasitic organisms. Animals of this quality require the most compre ~· ~: ~ . hensive and frequent monitoring of their microbiologic status and are the most difficult to obtain and maintain. . Microbiologically associated, definedflora, or gnotobiotic refers to axenic animals that have been intentionally inoculated with a well-defined mixture of microorganisms and maintained continuously in an isolator to prevent contamination by other agents. Generally, a small number (usually less than 15) of species of microorganisms are used in the inoculum, and it is implied that these organisms are nonpathogenic. · Pathogen-free implies that the animals are free of all demonstrable pathogens. It is often misused, in that there is no general agreement about which agents are pathogens, what tests should be used to demonstrate the lack of pathogens and with what frequency, and how the populations should be sampled. Use of this term should be avoided because of the lack of precision of its meaning. . Specific-pathogen-free (ape) Is applied to animals that show no evidence (usually by serology, culture, or histopathology) of the presence of particular microorganisms. In its strictest sense, the term should be related to a specific set of organisms and a specific set of tests or methods used to detect them. An animal can be classified as SPF if it is free of one or many pathogens. . Conventional is applied to animals in which the microbial burden is unknown, uncontrolled, or both. In addition, the term clean conventional is sometimes used to describe ani- mals that are maintained in a low-security barrier and are demonstrated to be free of selected pathogens. This term is even less precise than pathogen- free, and its use is discouraged (NRC, 1991~.
30 RODENTS: LABORATORY ANIMAL MANAGEMENT Commercial suppliers have coined various terms to indicate SPF status. All the terms are related to specific organisms of which the animals are stated to be free and for which they are regularly monitored. In some cases, the terms (e.g., virus-antibody-free and murine-pathogen-free) imply a quality of animals beyond the actual definitions of the terms. Virus-antibody-free animals, for example, are animals that are free of antibodies to specific rodent viruses. The term is a variation of SPF, in that it relates to specified viruses. The implied method of detection is serology. Animals might not be free of viruses other than those specified and might not be free of other . . microorganisms. Genetic Quality In spite of diligent maintenance practices that are required in any breeding colony to identify animals properly and house them securely, people can make mistakes. In addition, loose animals, including animals that escape their housing unnoticed and wild rodents, can enter cages, mate with the inhabitants, and produce genetically contaminated offspring. Good hus- bandry practices carried out by trained personnel, including keeping a pedi- gree and clearly identifying animals and cages, can help to reduce the oc- currence of such events. Nevertheless, to avoid devastating consequences of genetic contamination, a good program of genetic monitoring is war- ranted. Genetic monitoring consists of any method used to ensure that the genetic integrity of individuals of any particular strain has not been vio- lated. Several commercial sources provide genetic monitoring services for inbred mouse and rat strains. Personnel should be alert to phenotypic changes in the animals, such as unexpected coat colors or large changes in reproductive performance. In a pedigree-controlled foundation colony (see Chapter 4), it is important to monitor the breeding stock at least once every two generations so that a single erroneous mating can be detected quickly. Retired breeders or some of their progeny can be tested. In an expansion or production colony, in which it might not be cost-effective or practical to monitor so closely, sampling is recommended. The extent of such sampling can be as broad as resources and need permit. If genetic contamination occurs outside the foundation colony, contamination will eventually be purged by the infusion of breeders from the more rigorously controlled foundation colony. The extent of necessary testing depends on the number and genotypes of neighboring strains. A testing system should be capable of identifying the strain to which the individual belongs and differentiating it from other strains maintained nearby. Most strains can be identified with a small set of any genetic markers for which an assay is available. Newer DNA-typing methods that use multilocus probes, minisatellite markers, and "DNA-fin
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS 31 gerprinting" analysis are powerful tools for distinguishing strains, espe- cially strains that are closely related, but electrophoretic methods that type isoenzymes are generally more cost-effective for genetic monitoring (Hedrich, 1990; Nomura et al., 1984), in that such monitoring is most commonly done to detect mismatings. Immunologic methods are also used, and the ex- change of skin grafts between individuals of a strain is a particularly effec- tive method for screening a large number of loci in a single test. DNA from - representative breeders ot a stra~n can be stored for future use in identifying suspected genetic contaminations. Genetic monitoring is used orimarilY to verify the authenticity of a given strain; new mutations are rarely detected by this means. It is impos- sible to monitor all loci for new mutations, given the large number of unknown loci and known loci that do not produce a visible phenotype. A good breeding-management program, as described in Chapter 4, will help to reduce unwanted genetic changes caused by mutations. SELECTED ASPECTS OF EXPERIMENTAL DESIGN An experiment in which laboratory animals are used should be designed carefully, so that it produces unequivocal information about the questions that it was designed to address. The two most important requirements of proper experimental design in that connection are as follows: . Animals in different groups should vary only in the treatment that the experiment is designed to evaluate, so that the experimental outcome will not be confounded by dissimilarities in the constitution of the groups or in how they are treated or measured. · Each treatment should be given to enough animals for the experi- mental outcome to be attributed confidently to treatment difference and not merely to chance. The best way to ensure that groups of experimental animals are compa- rable is to draw them from a single homogeneous pool and to assign them randomly to treatment groups. Choosing animals of the same age, sex, and inbred strain for all treatment groups and even assigning littermates ran- domly to different treatment groups can eliminate factors that might par- tially account for group-to-group differences in experimental outcome. once animals are assigned to groups, they should be handled identi- cally, except for the treatment differences that the experiment is designed to evaluate. Food, water, bedding, and other features of animal husbandry should be the same. For long-term experiments, cages should be rotated to minimize group differences caused by cage position. For invasive experi- mental treatments, sham or placebo procedures should be performed in compar
32 RODENTS: LABORATORY ANIMAL MANAGEMENT ison groups; for example, animals given treatment by gavage should be compared with controls given the vehicle by Savage, animals treated surgi- cally should be compared with animals that undergo sham surgical opera- tions, and animals exposed to treatment by inhalation should be compared with animals placed in inhalation chambers that circulate only air. Follow- ing those precautions will ensure that differences in outcome between groups can be attributed to the experimental treatment itself and not to ancillary differences associated with the administration of the treatment. Finally, wherever possible, the outcome of interest should be measured by people who are unaware of which treatment each animal received, be- cause such knowledge can magnify or even create observed treatment dif- ferences. It is particularly important to carry out "blind" studies when the outcome is to be evaluated subjectively (e.g., by grading of disease sever- ityJ, rather than measured quantitatively (e.g., by measuring concentrations of serum constituents). The number of animals needed in each group will depend on many features of the experimental design, including the following: . . the goals of the study; · the primary outcome measure that will be compared; the number of groups that will be compared; · the expected number of technical failures or usable end points; the number and type of comparisons that will be made; · the expected animal-to-animal and measurement variability in the outcome; · the statistical design and analysis that will be used; · the magnitude of the differences between control and treatment groups that it is desirable to detect; · the projected losses; and · the maximal tolerable chance of drawing erroneous conclusions. The more variable an outcome measure is, either because outcomes in identically treated animals vary substantially or because there is a high degree of measurement variability, the more animals will be needed in each group to distinguish between group differences caused by treatment and those caused by chance. How outcome measurement variability, treatment difference to be detected, and tolerable chance of drawing an erroneous conclusion affect the required sample size depends on the measurement to be made' the type of croup comparison to be made. and the statistical , ~ .^ ~ ~ analysis to be used. Tables and formulas for comparing proportions among two or more groups have been published (Gart et al., 1986), as has useful information for other types of outcomes (Mann et al., l991J. For most experiments, it is highly desirable to collaborate with a statistician through- out, beginning with the design stage, so that appropriately defined groups of sufficient size will be available for a proper statistical analysis.
CRITERIA FOR SELECTING EXPERIMENTAL ANIMALS REFERENCES 33 Altman, P. L., and D. D. Katz, eds. 1979a. Inbred and Genetically Defined Strains of Laboratory Animals. Part I: Mouse and Rat. Bethesda, Md.: Federation of American Societies for Experimental Biology. 418 pp. Altman, P. L., and D. D. Katz, eds. 1979b. Inbred and Genetically Defined Strains of Laboratory Animals. Part II: Hamster, Guinea Pig, Rabbit, and Chicken. Bethesda, Md.: Federation of American Societies for Experimental Biology. 319 pp. Bailey, D. W. 1981. Recombinant inbred strains and bilinear congenic strains. Pp. 223-239 in The Mouse in Biomedical Research. Vol. I: History, Genetics, and Wild Mice, H. L. Foster, J. D. Small, and J. G. Fox, eds. New York: Academic Press. Baker, H. J., J. Russell Lindsey, and S. H. Wiesbroth, eds. 1979-1980. The Laboratory Rat. Vol. I, Biology and Diseases, 1979, 435 pp.; Vol. II, Research Applications, 1980, 276 pp. New York: Academic Press. Bhatt, P. N., R. O. Jacoby, H. C. Morse III, and A. E. New, eds. 1986. Viral and Mycoplas- mal Infections of Laboratory Rodents: Effects on Biomedical Research. Orlando, Fla.: Academic Press. Demant, P., A.A. Hart. 1986. Recombinant congenic strains A new tool for analyzing genetic traits determined by more than one gene. Immunogenetics 24(6):416-422. Festing, M. F. W. 1989. Inbred strains of mice. Pp. 636-648 in Genetic Variants and Strains of the Laboratory Mouse, 2d ed, M. F. Lyon and A. G. Searle, eds. Oxford: Oxford University Press. Festing, M. F. W. 1993. International Index of Laboratory Animals, 6th ed. Leicester, U.K. M. F. W. Festing. 238 pp. Available from M. F. W. Festing, PO Box 301, Leicester LE1 7RE, UK. Festing, M. F. W.. and D. D. Greenhouse. 1992. Abbreviated list of inbred strains of rats. Rat News Letter 26:10-22. Foster, H. L., J. D. Small, and J. G. Fox, eds. 1981-1983. The Mouse in Biomedical Research. Vol. I: History, Genetics, and Wild Mice, 1981, 306 pp.; Vol. II: Diseases, 1982, 449 pp.; Vol. III: Normative Biology, Immunology, and Husbandry, 1983, 447 pp.; Vol. IV: Experimental Biology and Oncology, 1982, 561 pp. New York: Academic Press. Fox, J. G., B. J. Cohen, and F. M. Lowe, eds. 1984. Laboratory Animal Medicine. Orlando, Fla.: Academic Press. 750 pp. Gart, J. J., D. Krewski, P. N. Lee, R. E. Tarone, and J. Wahrendorf. 1986. Statistical methods in cancer research. Volume III: The design and analysis of long-term animal experi- ments. Pub. No. 79. IARC Scientific Publications. Gill, T. J. 1980. The use of randomly bred and genetically defined animals in biomedical research. Am. J. Pathol. 101(3S):S21-S32. Gill, T. J., III, G. J. Smith, R. W. Wissler, and H. W. Kunz. 1989. The rat as an experimental animal. Science 245 :269-276. Green, E. L. 1981. Genetics and Probability in Animal Breeding Experiments. New York: Oxford University Press. 271 pp. Hansen, C. T., S. Potkay. W. T. Watson, and R. A. Whitney, Jr. 1981. NIH Rodents: 1980 Catalogue. NIH Pub. No. 81-606. Washington, D.C.: U.S. Department of Health and Human Services. 253 pp. Harkness, J. E., and J. E. Wagner. 1989. The Biology and Medicine of Rabbits and Rodents, 3rd ed. Philadelphia: Lea & Febiger. 230 pp. Hedrich, H. J., M. Adams, ed. 1990. Genetic Monitoring of Inbred Strains of Rats: A Manual on Colony Management, Basic Monitoring Techniques, and Genetic Variants of the Laboratory Rat. Stuttgart: Gustav Fischer Verlag. 539 pp. Hogan, B.. F. Costantini, and E. Lacy. 1986. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory. 332 pp.
34 RODENTS: LABORATORY ANIMAL MANAGEMENT International Committee on Standardized Genetic Nomenclature for Mice. 1994a. Rules for nomenclature of inbred strains. Mouse Genome 92(2):xxviii-xxxii. International Committee on Standardized Genetic Nomenclature for Mice. 1994b. Rules and guidelines for gene nomenclature. Mouse Genome 92(2):viii-xxiii. Mann, M. D., D. A. Crouse, and E. D. Prentice. 1991. Appropriate animal numbers in biomedi- cal research in light of animal welfare considerations. Lab. Animal Sci. 41(1):6-14. Merlino, G. T. 1991. Transgenic animals in biomedical research. FASEB J. 5:2996-3001. Nomura, T., K. Esaki, and T. Tomita, eds. 1984. ICLAS Manual for Genetic Monitoring of Inbred Mice. Tokyo: University of Tokyo Press. NRC (National Research Council), Institute of Laboratory Animal Resources, Committee on Immunologically Compromised Rodents. 1989. Immunodeficient Rodents: A Guide to Their Immunobiology, Husbandry, and Use. Washington, D.C.: National Academy Press. 246 pp. NRC (National Research Council), Institute of Laboratory Animal Resources, Committee on Infectious Diseases of Mice and Rats. 1991. Infectious Diseases of Mice and Rats. Washington, D.C.: National Academy Press. 397 pp. NRC (National Research Council), Institute of Laboratory Animal Resources, Committee on Rat Nomenclature. 1992a. Definition, nomenclature, and conservation of rat strains. ILAR News 34(4):S1-S26. NRC (National Research Council), Institute of Laboratory Animal Resources, Committee on Transgenic Nomenclature. 1992b. Standardized nomenclature for transgenic animals. ILAR News 34(4):45-52. Van Hoosier, G. L., Jr., and C. W. McPherson, eds. 1987. Laboratory Hamsters. Orlando, Fla.: Academic Press. 400 pp. Wagner, J. E., and P. J. Manning, eds. 1976. The Biology of the Guinea Pig. New York: Academic Press. 317 pp.