Should we be more cre‐tical? A cautionary tale of recombination

Summary

The cre‐loxP system has been revolutionary in the field of immunology. The technology enables genetic deletion in mice with unprecedented precision. It is therefore now widely used to investigate gene functions in animal models of disease, and in fundamental studies of the immune response. This widespread adoption of cre‐loxP technology has allowed a thorough investigation of its strengths and weaknesses. Here we highlight an important paper which not only describes potential problems with the commonly‐used screening procedures used when identifying offspring of the correct genotype, but also describes how this screening can be improved to ensure that the right animals are produced.

Cre‐loxP technology was first described in bacteriophages.1 Within a decade the system was applied to edit fertilised murine eggs2 and couple of years after that the first mouse models started emerging.3 Since then hundreds of publications have used the system. The cre‐loxP system allows a gene of interest to be conditionally knocked out within a certain tissue, in a particular type of cell, or at a specific developmental stage. Targeting knock‐outs in this way not only provides a way to avoid the lethality that might occur when genes that also function during development are inactivated in all cells, but also enables the study of protein and gene interactions in specific cell types and/or developmental stages.

For the cre‐loxP system to work, two strains of mice are needed. One, a mouse expressing the cre recombinase under the control of a promoter for a specific gene (that defines the location/ time of the cre expression) is crossed with the other, a mouse that expresses the gene that will be deleted. The gene to be deleted is engineered to be flanked by “flox” sequences. The flox sequences are the targets for the cre recombinase. Thus, in offspring that both express cre and carry the flox sequences, the “floxed” gene is deleted. Thereafter, following crossing the heterozygous offspring of the original cross, homozygous mice can be generated that lack the gene of interest in the defined location.

The widespread use of cre‐loxP technology has enabled investigators to identify potential difficulties, including possible cre toxicity in highly proliferative cells,4 and sub‐optimal cre recombination efficiency.5 In addition, the process of recombination at the heart of cre‐loxP system can also occur spontaneously, both ectopically between sites of minimal homology, and between pseudo‐loxP sites. Because of this spontaneous recombination, careful genotyping should be carried out during the crossing of cre‐loxP mice. To mitigate these difficulties, transgenic mice that express fluorescent proteins such as GFP or YFP upon cre recombinase activity are often used as controls.6, 7

In this issue of Immunology, Wu et al8 provide a detailed analysis of the genotypes generated during generation of cre‐loxP mice. They report a cautionary tale of an unexpected outcome after using a standard cre‐loxP approach for creating a conditional gene deletion. The authors describe a preliminary attempt to generate mice in which the insulin receptor (Insr) was conditionally knocked out in T regulatory cells. For this purpose, Foxp3^YFP‐Cre mice were purchased and crossed with Insr^fl/fl mice. Surprisingly, the offspring generated did not fit the expected Mendelian ratios. The F4 generation of mice was genotyped and expected to have 50/50 ratio between homozygotes and heterozygotes as their F3 parents were a homozygote and a heterozygote. However, it was observed that a significant number of animals that appeared to be homozygotes for the wild type allele. If the genotyping had been correct in earlier generations these animals could not have been generated. This observation led the authors to investigate the accuracy of the standard genotyping used during breeding. They discovered that the PCR method used could not detect alleles that had spontaneously recombined, because the primer binding site was disrupted during the recombination. They then demonstrated that when recombination was considered during primer design, the genotypes could be accurately assessed using qPCR. Furthermore, using this improved qPCR approach the authors were able to assess the abundance of ectopic recombination across many different tissues and cell types.

This publication not only provides examples of improved controls that ensure correct genotyping when generating mice using the cre‐loxP system, it also reveals how often spontaneous and ectopic recombination can occur. While this has been reported before when using Foxp3^YFP‐Cre mice9 and concern about ectopic recombination has also been raised,10 Wu et al highlight the biological mechanisms underlying this difficulty, enabling design of better controls. This report is therefore important for any investigators generating cre‐loxP mouse lines and may help them to save both money and time.