Generation and comparison of CRISPR-Cas9 and Cre-mediated genetically engineered mouse models of sarcoma

Abstract

Genetically engineered mouse models that employ site-specific recombinase technology are important tools for cancer research but can be costly and time-consuming. The CRISPR-Cas9 system has been adapted to generate autochthonous tumours in mice, but how these tumours compare to tumours generated by conventional recombinase technology remains to be fully explored. Here we use CRISPR-Cas9 to generate multiple subtypes of primary sarcomas efficiently in wild type and genetically engineered mice. These data demonstrate that CRISPR-Cas9 can be used to generate multiple subtypes of soft tissue sarcomas in mice. Primary sarcomas generated with CRISPR-Cas9 and Cre recombinase technology had similar histology, growth kinetics, copy number variation and mutational load as assessed by whole exome sequencing. These results show that sarcomas generated with CRISPR-Cas9 technology are similar to sarcomas generated with conventional modelling techniques and suggest that CRISPR-Cas9 can be used to more rapidly generate genotypically and phenotypically similar cancers.

Figure 1. In vitro and in vivo validation of pX333-sgTrp53-Cre.

(a) The pX333 plasmid was modified to express Cre. A sgRNA to target Trp53 was cloned into the plasmid following a U6 promoter. (b) MEFs from Kras^LSL-G12D/+ (K), Rosa26^{LSL-Cas9-EGFP/+} (C) or Kras^LSL-G12D/+; Rosa26^{LSL-Cas9-EGFP/+} (KC) mice were harvested and cultured. MEFs were transiently transfected with pX333-sgTrp53-Cre and subsequently DNA analysis, soft agar studies and an allograft study were performed. (c) PCR shows that transient transfection of pX333-sgTrp53-Cre in MEFs resulted in recombination of Kras^LSL-G12D/+ and Rosa26^{LSL-Cas9-EGFP/+} (top two panels). Surveyor assay (bottom panel) shows that pX333-sgTrp53-Cre efficiently generates indels in the Trp53 gene. (d) MEFs transfected with pX333-sgTrp53-Cre were subjected to a soft agar assay. When KC MEFs were transfected with the plasmid, deletion of Trp53 and activation of the conditional Kras mutation resulted in anchorage-independent growth. The result is representative of at least two different experiments. Furthermore, when these transformed cells were injected into nude mice (n=4) (e), they all formed tumours that showed a spindle cell morphology. Scale bars, 100 μm.

Figure 2. Conditional expression of Cas9 to generate tumours in a primary mouse model of sarcoma.

(a) Kras^LSL-G12D/+ (K), Rosa26^{LSL-Cas9-EGFP/+} (C), or Kras^LSL-G12D/+; Rosa26^{LSL-Cas9-EGFP/+} (KC) mice received intramuscular injections of an adenovirus carrying the plasmid. (b) Adenoviral delivery of sgTrp3-Cre resulted in efficient primary sarcoma formation by activating conditional expression of the mutant Kras and Cas9 alleles and delivering a sgRNA to target Trp53. (c) PCR analysis and Surveyor assay showing recombination of Kras^LSL-G12D and Rosa26^{LSL-Cas9-EGFP}, as well as indels within the Trp53 gene, respectively. (d) Twelve of the primary tumours generated via adenoviral plasmid delivery were subject to histopathological analysis with staining for myogenic markers MyoD and myogenin. (e) The majority of tumours resembled UPS, whereas a minority were consistent with a myogenic UPS phenotype. Scale bars, 500 μm.

Figure 3. Sequencing of primary sarcomas generated using a combined CRISPR-Cas9-Cre-loxP system.

(a) Sanger sequencing of the Trp53 gene in tumours generated in KC mice after adenoviral delivery of pX333-sgTrp53-Cre. Most of the sequenced alleles demonstrated deletions but there was a single nucleotide insertion observed in mouse 2981- allele 1. (b) Targeted deep sequencing of the Trp53 gene in tumours generated in KC mice. (Blue=sgTrp53; green=PAM; red=insertion/deletion). (c) Sanger sequencing of the on-target chromosome 17 site in tumours generated in KC mice after adenoviral delivery of pX333-sgTrp53-Cre. Most of the sequenced alleles demonstrated deletions but there was a single nucleotide insertion observed in mouse 2981- allele 1 and two nucleotides insertion observed in mouse 2995 allele 2. (d) Targeted deep sequencing of the on-target chromosome 17 site in tumours generated in KC mice. (Blue=sgTrp53; green=PAM; red=insertion/deletion).

Figure 4. Generation of primary sarcomas using in vivo electroporation.

(a) K, C and KC mice received intramuscular injections of the naked pX333-sgTrp53-Cre and K, C, KC and K-loxP-C mice received intramuscular injections of the naked pX333-sgTrp53-Cre with IVE. (b) Ten of the primary tumours generated via IVE were subject to histopathological analysis with haematoxylin and eosin staining. Scale bars, 200 μm.

Figure 5. CRISPR-Cas9-mediated generation of primary MPNSTs.

(a) The px333 plasmid was modified by cloning a sgRNA targeting Nf1 into the plasmid following the first U6 promoter and a sgRNA targeting Trp53 following the second U6 promoter. (b) Wild-type 129/SvJ mice received sciatic nerve injections of an adenovirus carrying CRISPR components, and the mice were monitored for tumour formation. The formation of tumours was observed at the site of injection in these wild-type mice and (c) Surveyor analysis revealed indels in Nf1 and Trp53. (d) Haematoxylin and eosin histology and (e) positive S100 staining were consistent with MPNST. Scale bars, 100 μm.

Figure 6. Whole exome sequencing of primary sarcomas generated using Cre-loxP or a combined CRISPR-Cas9-Cre-loxP system.

(a) Ten of the primary tumours generated via Cre-loxP or a combined CRISPR-Cas9-Cre-loxP system were subject to histopathological analysis with haematoxylin and eosin staining. (b) Comparison of the days to tripling of primary sarcomas generated with Cre-loxP (n=15) and the combined CRISPR-Cas9-Cre-loxP (n=20) system. Scale bars, 200 μm. The non-synonymous mutational load (c) and CNVs (d) of primary sarcomas generated with Cre-loxP (n=7) and the combined CRISPR-Cas9-Cre-loxP (n=5) system was determined using whole exome sequencing of tumour DNA (Error bars=s.e.m.).