removal of the lox-P-flanked selectable marker. The final step is the removal of the lox-P-flanked DNA in specific tissues of the whole animal, which is accomplished by breeding the gene-targeted mice with transgenic mice expressing Cre, as a transgene, in specific tissues.
A variation on this scheme can be used to restore normal expression of a mutated gene. In this case, a lox-P-flanked selectable marker is inserted into a gene, by gene targeting, to disrupt its function. Function can be restored in specific tissues in the resulting mutant mice by mating with transgenics expressing tissue-specific Cre in order to remove the inserted deleterious DNA, leaving behind a single lox-P site in a noncritical position.
Variations in the design of any transgene or gene-targeting construct, as well as local features of the integration or target site, will affect the outcome of transgenic experiments. Integration of a transgene is random. Thus, its expression might be affected unpredictably by other promoters or enhancers at its integration site, or the transgene might be entirely silenced by integration into a transcriptionally inactive chromosomal region. This has been termed the “neighborhood effect.” Expression might also be influenced by epigenetic phenomena, such as methylation. Another chance event that will alter the intended experimental outcome, but which can be exploited, is insertional mutagenesis, described above. If the transgene happens to integrate in a position that causes the disruption of an endogenous gene, the experiment might be more informative about the endogenous gene than the transgene.
With gene targeting, a possible complication can arise if the selectable marker, which is essentially a foreign transgene, remains in the genome. Either this gene or its promoter could potentially affect expression of the targeted gene or neighboring genes. In gene-targeting experiments that involve deletions, it is possible to unknowingly remove cryptic regulatory regions located within introns and thus, potentially, to affect expression of nearby genes.
Phenotypic effects observed following any mutational change, whether through a transgene, gene targeting, or a naturally occurring mutation, are subject to what are called “genetic background effects.” The mutant phenotype might vary in different animals depending on what other genes that animal possesses (its genetic background). For example, a mutation might show a different phenotype in two inbred strains if those strains carry different alleles in other genes that directly or indirectly modify the phenotype of the mutant gene. These background effects will be largely unpredictable, but this situation provides material to identify and isolate modifier genes and thus to increase understanding of genetic pathways. The judicious use of inbred strains, in which the mice are theoretically 98% genetically identical, allows these background effects to be studied (e.g., see Lander and Schork 1994).