Cas9 knock-in mice for efficient genome editing in vivo and ex vivo

Cas9 mice broaden options for genome editing

CRISPR/Cas9-mediated genomic manipulation has been increasingly adopted by researchers to generate novel mutations in a number of model systems rapidly and efficiently. While the myriad applications for the technology are exciting, applying it in vivo and ex vivo is difficult because traditional routes for delivering the endonuclease into target cells are inefficient or limited to only certain permissible cell types. In a recent Cell paper published by Dr. Feng Zhang, Assistant Professor, MIT, a novel Cre-dependent Cas9 knock-in mouse is described that enables either ubiquitous or cell type-specific activation of Cas9 and GFP marker expression (Platt et al., 2014). Using these Cas9 knock-in mice, Dr. Feng’s group demonstrate efficient editing in three different platforms: (1) modification of brain cells in vivo via adeno-associated virus (AAV)-mediated delivery of short guide RNA (sgRNA); (2) ex vivo editing of genes in difficult-to-modify primary immune cells; and (3) in vivo multiplexed editing of cancer-related genes in lung tissue via AAV particles loaded with multiple sgRNAs to induce heterogeneous adenocarcinomas.

Creation of Lox-stop-Lox-Cas9 (024857) and Cas9 (024858) mice

Delivery of Cas9 in vivo is challenging due to limitations on the amount of DNA that can be packaged into viral vectors. To circumvent this limitation and the reduced efficiency associated with viral delivery of Cas9, Platt et al. generated a Cre-dependent Lox-stop-Lox (LSL) Cas9 knock-in mouse (B6.129-Gt(ROSA)26Sortm1(CAG-xstpx-cas9,-EGFP)Fezh/J, 024857). Genome editing is enabled in these mice by breeding them to a Cre-driver strain and delivering an appropriate sgRNA. The knockin allele also contains 3xFLAG and EGFP sequences to allow for confirmation of Cas9 expression by immunoblot or fluorescence microscopy, respectively. To test the robustness of Cas9 expression following Cre expression, the researchers also generated an additional strain with constitutive Cas9 expression in most cells by crossing the conditional LSL-Cas9 mice to a beta-actin Cre-driver strain (FVB/N-Tg(ACTB-cre)2Mrt/J 003376), creating strain STOCK Gt(ROSA)26Sortm1.1(CAG-cas9,-EGFP)Fezh/J (024858). Progeny constitutively expressing Cas9 are viable and express widespread Cas9 mRNA and protein. Constitutive Cas9 mice are also fertile and produce pups in normal litter sizes with no obvious morphological abnormalities, suggesting that constitutive Cas9 expression can be maintained without toxicity or deleterious phenotypes.

In vivo brain editing
Figure 1. Cas9 mice offer a myriad of gene editing and biological model creation possibilities.

Using the conditional LSL-Cas9 mice, Platt et al. delivered AAV particles containing Cre and an sgRNA targeted to Rbfox3 (RNA binding protein, fox-1 homolog; aka NeuN) via stereotactic injection into the brain. Analysis of brain tissue and cells three weeks after injection showed generation of insertions and deletions (indels) near the predicted cleavage site in NeuN, indicating targeted activity by Cas9. Further, NeuN protein was depleted only in cells at the injection site and not in surrounding tissue or in control animals. To verify that Cas9 expression does not affect basal function of neurons, the researchers compared neuronal cell function from constitutive Cas9 mice to controls via patch-clamp assays. No significant changes in any electrophysiological parameters were found in Cas9-expressing neurons compared to controls. These experiments suggest that neurons, which are challenging and sensitive cell populations to modify, can be effectively edited in vivo without off-target alteration of their normal function by Cas9.

Genetic modification of primary cells

The Cas9 mouse models generated by Platt et al. also allow for genomic manipulation of notoriously challenging-to-modify cell types, such as dendritic cell precursors isolated from the bone marrow (BMDCs). These cells are difficult to isolate and maintain in vitro, and show only limited success when transduced with short hairpin RNAs (shRNAs) to modify gene expression. The researchers tested their ability to modify BMDCs ex vivo by isolating bone marrow cells from constitutively Cas9-expressing mice and culturing them under conditions that favor BMDC differentiation. They then delivered sgRNAs targeting Myd88 or Tnfaip3 (A20)-- positive and negative regulators of TLR4-mediated signaling, respectively-- via lentiviral transduction. One week after transduction, the cells were given lipopolysaccharide (LPS) to activate TLR4 inflammatory signaling and subsequently assayed for gene expression. Indels were observed in 70-80% of sequencing reads for both genes, and Myd88 mRNA and protein expression were reduced compared to controls. The researchers also measured the expression of 276 genes representative of the LPS-activated inflammatory response. Cells that received Myd88 sgRNAs had reduced expression of genes activated by TLR4, while targeting of A20 resulted in an increase in the expression of inflammatory response genes. These results demonstrate the utility of constitutively Cas9-expressing mice for efficient gene editing in primary cell types that are otherwise difficult to modify.

Multiplexed gene editing in the lung

A major advantage in CRISPR/Cas gene editing is that reactions can be multiplexed to introduce multiple genome modifications in a single step. To test this feature in the Cre-dependent LSL-Cas9 model, Platt et al. delivered a single AAV vector to generate loss-of-function mutations in the tumor suppressors Trp53 (p53) and Lkb1, and homology-directed repair modification of Kras to oncogenic KrasG12D simultaneously. Intra-tracheal delivery of this vector, which also carried Cre and a luciferase reporter sequence in addition to the sgRNA and KrasG12D homology repair templates, resulted in adenocarcinomas in 100% of the infected animals. Significantly, the lung tumors were heterogeneous in size, distribution and grade, reminiscent of the heterogeneous tumors observed in human lung cancer patients. These data suggest that the conditional LSL-Cas9 mice can be used to generate new models of human cancer that more closely resemble the tumor heterogeneity and mosaicism characteristic of the human disease. These two new Cas9 strains provide exciting new tools to perform targeted genomic manipulation in traditionally challenging cell types in vivo and ex vivo, and to generate novel mouse models of complex diseases, such as cancer.