GENOMIC CONSIDERATIONS FOR THE PATHOLOGIST

Having stressed the importance of strain and substrain, it is notable that the mouse genomic community does not utilize a single strain of mouse, and when they do use a similar strain, it is often a different substrain.

Random insertion of transgenes is accomplished through pronuclear microinjection of zygotes with ectopic DNA (transgenes). This has generally been achieved using hybrid zygotes of 2 inbred parental strains, outbred Swiss mice, or from inbred Swiss FVB/N mice to take advantage of hybrid vigor to com­pensate for the trauma of microinjection and facilitate the process of microinjection by providing large pronu­clei. Transgenes become randomly integrated through­out the genome, often in tandem repeats, so that each pup within a litter arising from microinjected zygotes is hemizygous for the transgene, but is genetically distinct from its littermates. The degree of transgene expression (phenotype) varies with the location of the transgene within the genome. Each founder line of the same transgene represents a unique and nonreproducible genotype and, therefore, phenotype. Transgenes tend to be genetically unstable, and copies may be lost in subsequent generations, resulting in ephemeral phenotypes. Transgene insertions can also lead to unanticipated altered function of genes through inser- tional mutagenesis, or regulation by flanking genes within the area of insertion. Unanticipated phenotypes, such as immunodeficiency or other effects, can therefore occur. The use of hybrids or outbred mice as founders requires selective inbreeding to attain a useful model.

The discipline of mouse genomics has lent itself to incredible precision through homologous recombin­ation, with the ability to alter not only specific genes but also gene function at specific time points during development or life stage, create tissue-specific gene alter­ations, gain of function, loss of function, and targeted integration of transgenes that allow customized develop­ment of mouse models of human disease that would not ordinarily arise within the context of the indigenous mouse genome. Targeted mutant mice are often created in one of several types of 129 ES cells, and once germline transmission has been effected, the 129-type mutant mouse is usually backcrossed to a more utilitarian mouse strain, such as B6. Full backcrossing to congenic status requires 3-4 years, which is seldom fulfilled. In constructs that require cre-lox technology, mutant mice are further crossed with cre transgenic mice, which may be of another strain, substrain, or stock background. Thus, despite superb precision in altering a gene of interest, the rest of the mouse's genome can remain highly heterogeneous, which defeats the inherent value of the GEM for research, or at least limits its full potential.

ES cells, and the mutations that they carry, are most often derived from one of the 129-type mouse strains, and ES cells become mice through the generation of chimeric progeny. Insufficient backcrossing, with reten­tion of 129 characteristics, may result in erroneous assumptions about the phenotype of the targeted gene. There is considerable genetic variation among different 129 ES cell lines, which can be a potential problem for comparing phenotypes of the same gene alteration among different 129 ES cell-derived mice.

Because of the highly inbred nature of laboratory mice, experimental mutation of many genes often leads to embryonic or fetal death that precludes evaluation of phenotype in adult mice. Thus, pathologists are being increasingly called upon to familiarize themselves with fetal development and evaluate developmental defects. Fetal pathology is beyond the scope of this text, but the reader can access several excellent sources of informa­tion (see Kaufman 1995; Kaufman and Bard 1999; Rossant and Tam 2002; Ward et al. 2000). Embryonic/ fetal viability is most often influenced by abnormalities in placentation, liver function, or cardiovascular func­tion (including hematopoiesis). Particular attention should be paid to these factors. Depending upon genetic background, lethality can vary. Gene expression, and therefore circumvention of events such as embryonic lethality, can be controlled temporally and quantita­tively by tissue-specific promoters with drug-regulated transcription systems and with cre/lox deletion, in which cre recombinase can be controlled with transcrip­tion techniques. Temporal and quantitative control of transgenes poses unique challenges to pathologists when evaluating phenotype.

In addition to predicted phenotypes, GEMs often manifest unique pathology that is not present in paren­tal strains.

Consequently, the pathologist must be cognizant of general mouse pathology, strain-related patterns of spontaneous pathology, infectious disease pathology, developmental pathology, comparative pathology (to validate the model), methodology used to create the mice, predicted outcomes of the gene alteration (includ­ing effects of the promoter), potential but unexpected outcomes of the gene alteration, and Mendelian genet­ics. The pathologist must also resist temptation to over­emphasize a desired phenotype, underemphasize an undesired phenotype, or proselytize a phenotype as a model for human disease when it isn't. There is no better person to be the gatekeeper of reality in the world of functional genomics than the comparative pathologist.