Breeding strategies for tissue-specific knockout mice

Q: My lab has created a tissue-specific knockout mouse. Is it OK to maintain my colony homozygous for my loxP allele and homozygous for cre?

A: Theoretically, maintaining your colony this way may be possible. Your ability to do so depends on the particular cre that your mice express. Even if this mating scheme is possible, there are other reasons why you may not want to do so. But first, let’s deal with cre.

Many cre transgenes are produced using pronuclear microinjection. Typically, the cre-expressing transgene is integrated at a random location in the genome, and its integration site is not known. (Mapping transgene-integration sites requires multiple crosses and genomic analysis, and most investigators don’t bother.) Occasionally, transgene integrations can produce null alleles of endogenous genes that are required for either the animals’ viability or fertility, and JAX distributes several cre transgenic strains which homozygotes are either not viable or fertile (see, for example, B6.Cg-Tg(CD2-cre)4Kio/J (Stock # 008520).

Because cre transgenes generally don’t need to be homozygous to efficiently recombine loxP-flanked (floxed) alleles, many investigators never bother to breed their cre transgenic animals to homozygosity, and it is not known whether such animals will survive or breed. If this is the case for your cre mouse, then clearly you will need to discover this before you can consider maintaining your strain homozygous for cre. Also, if the integration site of your cre is not known, an allele-specific genotyping assay for distinguishing hemizygotes from homozygotes won’t be available, and you will have to rely on either Southern blots, quantitative PCR or progeny testing to identify the homozygotes.

Cre mice, despite being designed to express cre tissue specifically, also may show expression in other tissues, including the male or female germline. Although germline expression is not necessarily problematic when breeding mice that are heterozygous for a floxed allele and cre positive, it could turn into a problem if maintaining a cre-positive strain that is homozygous for a floxed allele. If you are not careful to regularly monitor the offspring in your colony for the presence of the null allele, your colony, ultimately, could start producing mice that are homozygous for the null allele globally, instead of homozygous for the conditional allele. Even only infrequent recombination in the germline could produce that null for your gene of interest over the long-term. Such events could even occur in inducible cre mice in which low levels of background recombination in the absence of inducers have been detected – see, for example Kemp et al. (Nucleic Acids Res.2004. 32911) e92) – although the overall frequency of such events is very low.

Finally, maintaining a homozygous cre, homozygous floxed colony fails to produce any experimental controls, relying instead on a second colony of mice (presumably inbred or F2 hybrids.) For reasons discussed in a previous blog post, having cre positive, floxed heterozygote and cre-positive, floxed-negative controls is often recommended to account for any phenotypes that the cre transgene, itself, might produce. Therefore, even if your mice are viable and fertile as cre and floxed double homozygotes, maintaining them this way may not be such a good idea.