Here we describe a rapid and direct in vivo CRISPR/Cas9 screening methodology using ultrasound-guided in utero embryonic lentiviral injections to simultaneously assess functions of several genes in the skin and oral cavity of immunocompetent mice.
Genetically modified mouse models (GEMM) have been instrumental in assessing gene function, modeling human diseases, and serving as preclinical model to assess therapeutic avenues. However, their time-, labor- and cost-intensive nature limits their utility for systematic analysis of gene function. Recent advances in genome-editing technologies overcome those limitations and allow for the rapid generation of specific gene perturbations directly within specific mouse organs in a multiplexed and rapid manner. Here, we describe a CRISPR/Cas9-based method (Clustered Regularly Interspaced Short Palindromic Repeats) to generate thousands of gene knock-out clones within the epithelium of the skin and oral cavity of mice, and provide a protocol detailing the steps necessary to perform a direct in vivo CRISPR screen for tumor suppressor genes. This approach can be applied to other organs or other CRISPR/Cas9 technologies such as CRISPR-activation or CRISPR-inactivation to study the biological function of genes during tissue homeostasis or in various disease settings.
One of the challenges for cancer research in the post-genomic era is to mine the vast amount of genome data for causal gene mutations and to identify nodes in the gene network that can be targeted therapeutically. While bioinformatic analyses have helped immensely towards these goals, establishing efficient in vitro and in vivo models is a prerequisite to decipher the complexity of biological systems and disease states and for enabling drug development. While conventional transgenic mouse models have been used extensively for in vivo cancer genetics studies, their cost-, time- and labor-intensive nature has largely prohibited the systematic analysis of the hundreds of putative cancer genes unraveled by modern genomics. To overcome this bottleneck, we combined a previously established ultrasound-guided in utero injection methodology1,2 with a CRISPR/Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats) gene editing technology3 to simultaneously induce and study loss-of-function mutations of hundreds of genes in skin and oral cavity of a single mouse.
The methodology described here utilizes ultrasound-guided injections of engineered lentiviruses into the amniotic cavity of living mouse embryos at embryonic day E9.5. The lentiviral cargo containing CRISPR/Cas9 components transduce the single-layered surface ectoderm, which later gives rise to the epithelium of the skin and oral cavity. Skin is composed of an outer epidermis, followed by basement membrane and dermis. Epidermis is a stratified epithelium composed of an inner, basal layer, which maintains contact with the basement membrane and has proliferative and stem cell capacity. The basal layer gives rise to the differentiated layers above such as spinous, granular and stratum corneum layers2,4. Lineage tracing studies show that this direct in vivo CRISPR/Cas9 method genetically manipulates tissue-resident stem cells within the basal layer that persist throughout adulthood. As lentivirus can be titrated to transduce the E9.5 surface ectoderm at clonal density, this method can be used to generate mosaic mice harboring thousands of discrete gene knock-out clones. Next generation sequencing can then be used to analyze the effect of CRISPR/Cas9-mediated gene ablation within those clones in a multiplexed manner5.
We recently used this method to assess the function of 484 genes that show recurrent mutations in human Head and Neck Squamous Cell Carcinoma (HNSCC)5. HNSCC is a devastating cancer with a high mortality rate of 40–50% and it is the 6th most common cancer worldwide6. HNSCC arise in mucosal linings of the upper airway or oral cavity and are associated with tobacco and alcohol consumption or human papillomavirus (HPV) infection. Cutaneous Squamous Cell Carcinoma (SCC) are skin tumors and represent the second most common cancer in humans7. Cutaneous SCC and HNSCC are histologically and molecularly very similar, with a high percentage of cases exhibiting alteration in TP53, PIK3CA, NOTCH1, and HRAS8. While there are only a handful of genes mutated at high frequency, there are hundreds of genes found mutated at low frequency (< 5%), a phenomenon commonly referred to as the long-tail distribution. As the majority of the long-tail genes lack biological or clinical validation, we used this in vivo CRISPR screening technology to model loss-of-function of these genes in tumor-prone mice with sensitizing mutations in p53, Pik3ca or Hras and identified several novel tumor suppressor genes that cooperate with p53, Pik3ca or Hras to trigger tumor development5.
Here, we describe a detailed protocol to generate multiplexed sgRNA lentiviral CRISPR sgRNA libraries and perform CRISPR/Cas9 gene knock-out screens in the mouse surface ectoderm. Of note, this methodology can be adapted to incorporate other gene manipulation technologies such as CRISPR activation (CRISPRa) and CRISPR inactivation (CRISPRi) or modified to target other organ systems in mouse to study gene functions.
This protocol was approved and performed in accordance with IACUC of University of Toronto.
1. Design and cloning of pooled CRISPR libraries
2. Production of high titer lentivirus suitable for in vivo transduction
NOTE: Perform all steps in this section of the protocol in a BSL2+ facility in a Class II, Type A2 biosafety cabinet. 293FT and especially the 293NT packaging cells allow for higher virus production. Use low passage (<p20) HEK293T, 293FT or 293NT cells for transfections. Prewarm all media to 37 °C. Never allow HEK293T, 293FT or 293NT cells to become confluent while subculturing. Grow 293FT in the presence of G418 to maintain the expression of the SV40 large T-antigen.
3. Ultra-sound guided surgery and injection
NOTE: This technology was adapted from4,11. Microinjection geared towards transduction of the surface epithelium must be performed at embryonic day E9.5, when the surface ectoderm consists of a single layer and before formation of the periderm starting at E10, which would prevent transduction of this basal layer. Preferably set up mice on Friday, so that the first possible day with E9.5 embryos is the following Monday. Use Rosa26-Lox-STOP-LOX-Cas9-GFP mice (Jackson Laboratory #024857) for optimal CRISPR/Cas9 efficiency12.
4. Deep sequencing procedure
Figure 1A shows the design of the oligonucleotides for multiplexing several custom CRISPR libraries in a cost-effective manner in a single 12k or 92k oligo chip. Once the sgRNAs (blue color coded) are selected, the oligonucleotides are designed with restriction sites (orange colored BsmBI) and library specific PCR primer pairs (green color coded). Several libraries can be designed by using unique combination of primer pairs for multiplexing in a single oligo chip. When PCR amplifying the libraries using a specific PCR primer pair, always include a water only negative control and run all the reactions on an agarose gel. The negative control lane should not have any bands at 100 bp, while the library amplified lane should have a single 100 bp band. If there is no band, make sure the PCR primer pair is selected appropriately. Figure 1B shows the quality control step of the cloned library and the viral production and concentration procedure. Care should be taken to preserve the equal representation of sgRNAs from cloning till transduction. Next-generation sequencing reads of PCR amplified DNA from plasmid and lenti-viral library transduced cells should show a high correlation. Any deviation must be analyzed carefully to examine whether the representation is lost before or after the lenti-virus preparation. If the sgRNA reads from plasmid DNA do not show equal representation of sgRNAs, then the viral preparation and concentration procedure must be repeated with care. If the loss of equal sgRNA representation occurred in plasmid DNA, then the entire cloning procedure must be repeated including PCR amplification of libraries from oligo chip with reduced amplification cycles. The success of the CRISPR library screening depends critically on equal representation of the sgRNAs present in the library.
Figure 2 shows the ultrasound-guided micro-injection set up for manipulating the E9.5 embryos in pregnant mouse. The entire set-up is placed under a biosafety level class II cabinet to maintain the sterile condition of the entire procedure and to avoid any infection of the pregnant mouse due to the surgical procedures. Care should be taken to avoid pressuring the uterine horns while injecting. The needle must be very sharp, so that the wound on the uterine wall and the amniotic membrane is minimal.
Figure 3 shows the results of the ultra-sound guided injection of lenti-virus carrying Cre-recombinase in E9.5 Lox-stop-Lox (LSL) Confetti mouse pups at postnatal day (P)0. Viral titer can be adjusted to get an appropriate transduction coverage of the epidermis. While high titer of virus would result in larger coverage of the mouse skin, it would also result in transduction of multiple sgRNAs into the same cell potentially confounding the outcomes. To reduce multiple transductions of a single cell, the infection rate should be kept <20%.
Figure 4 shows next-generation sequencing reads of PCR amplified DNA from plasmid and lenti-viral library transduced cells and also reads from 4 representative tumors. sgRNA guides targeting Adam10 and Ripk4 are enriched in the tumor samples (triangles and diamonds) compared to sgRNA presentation in plasmid pool or infected cells. Adam10 and Ripk4 function as tumor suppressors5. Several hundred tumors can be multiplexed by assigning unique barcodes to each sample and deep sequenced as outlined in the protocol.
Figure 1: Cloning of targeted CRISPR library. (A) Schematic representing the design of oligonucleotides for the 12 k or 92 k custom oligo chip. BsmBI (or other compatible) restriction sites (orange color coded and underlined) were added on each side of the sgRNA. Arrow heads indicate the BsmBI cut site. For each library, a unique pair of PCR primers (green, brown and purple color coded) were added, so that it can be specifically amplified using PCR from the pooled chip. Several custom libraries can be multiplexed up to 12k or 92k oligos. (B) Graph showing sgRNA representation from a CRISPR library in plasmid DNA versus DNA from cells transduced with the same lentiviral library. Each dot represents a guide. Full representation was maintained after transduction with some correlation in abundance. Please click here to view a larger version of this figure.
Figure 2: Picture of the ultrasound-guided microinjection set-up. (A) Mouse carrying E9.5 embryos was anesthetized and placed on a heated platform with the uterus exposed in a PBS filled modified Petri dish chamber (stabilized by four modeling clay). The blue semi-round rubber supported the uterus containing embryos and the injection needle from the microinjector is positioned on the right side. The ultrasound scan head was mounted on the top relaying live image to the monitor behind with needle head visible. (B) Close-up ultra-sound image of an E9.5 embryo. Lentiviral library was injected with the needle (seen on right side) into the amniotic cavity. Please click here to view a larger version of this figure.
Figure 3: Successful transduction of surface epithelium in mouse. Fluorescent images of whole-body, oral cavity, tongue, and palate of newborn Cre-reporter LSL-Confetti mice transduced with Cre lentivirus (Inlet: corresponding white light images). Scale bars = 500 μm Please click here to view a larger version of this figure.
Figure 4: Deep sequencing of tumor samples. Graph showing sgRNA representation from a CRISPR library in plasmid DNA versus DNA from cells transduced with the same lentiviral library and in addition to reads from 4 representative tumors. The red and green circles denote the number of Adam10 and Ripk4 sgRNA reads in library and transduced cells whereas the red triangle represents the reads of Adam10 sgRNAs and the green diamond represents reads of Ripk4 sgRNAs identified in tumors from separate HNSCC mice showing a 1000x fold enrichment. Please click here to view a larger version of this figure.
PCR1 Forward Primer | GAGGGCCTATTTCCCATGATTC |
PCR1 Reverse Primer | CAAACCCAGGGCTGCCTTGGAA |
Table 1: Primers for PCR1 reaction. The forward and reverse primers used for amplification of the region surrounding sgRNA cassette from genomic DNA of cells transduced with the lenti virus.
Step | Temperature | Time | |
1 | 98 °C | 30 sec | |
2 | 98 °C | 10 sec | |
3 | 66 °C | 30 sec | |
4 | 72 °C | 15 sec | 15 cycles (step 2-4) |
5 | 72 °C | 2 min | |
6 | 4 °C | Hold |
Table 2: PCR1 cycle parameters. PCR conditions used for the amplification of the region surrounding sgRNA cassette from genomic DNA of cells transduced with the lenti virus.
501 FW | AATGATACGGCGACCACCGAGATCTACACTATAGCCTACACTCTTTCCCTACACG ACGCTCTTCCGATCTtgtggaaaggacgaaaCACCG |
||
701 RV | CAAGCAGAAGACGGCATACGAGATCGAGTAATGTGACTGGAGTTCAGACGTGTGCT CTTCCGATCTATTTTAACTTGCTATTTCTAGCTCTAAAAC |
||
* The underlined bases indicate the Illumina (D501-510 for forward and D701-712 for reverse) barcode location that were used for multiplexing. | |||
* Red colored bases indicate the sequence that bind to the target site on the lentiviral CRISPR plasmid. This region can be modified according to the lentiviral vector used. |
Table 3: Barcoding Primers for PCR2 reaction. Primers used to barcode amplify each tumor sample (either directly from genomic DNA or using the products from PCR1 as template). The underlined region indicates the unique barcode region that can be assigned for each sample for multiplexing in a deep sequencing reaction. The red colored base pairs indicate the target region in the lentiviral construct that these primers bind. This target region can be modified according to the type of lentiviral construct used.
Step | Temperature | Time | |
1 | 98 °C | 30 sec | |
2 | 98 °C | 10 sec | |
3 | 68 °C | 30 sec | |
4 | 72 °C | 15 sec | 10 cycles (step 2-4) |
5 | 72 °C | 2 min | |
6 | 4 °C | Hold |
Table 4: PCR2 cycle parameters. PCR conditions used to barcode amplify each tumor sample (either directly from genomic DNA or using the products from PCR1 as template).
CRISPR/Cas9 genome editing has been widely used in in vitro and in vivo studies to investigate gene functions and cellular processes. Most in vivo studies utilize CRISPR/Cas9 gene edited cells grafted into an animal model (allograft or xenograft). While this is a powerful tool to study cancer genetics and cellular functions, it still lacks the native tissue microenvironment and might elicit wounding and/or immune responses.
To overcome these challenges, several groups have pioneered direct in vivo CRISPR approaches generating several, multiplexed autochthonous mouse models over the last few years15,16,17,18. These projects usually rely on adeno-associated virus (AAV) to deliver sgRNAs to target organs. AAV allows for generation of very higher titer thereby facilitating production high titer virus required for in vivo manipulations. However, usage of AAVs severely limits the number of genes that can be analysed to about 40-50 genes, as AAVs, unlike lentivirus, do not stably integrate into genomic DNA, making readout of sgRNA libraries impractical. AAV based screening necessitates direct readout of mutations in target sites using e.g., targeted capture sequencing. However, multiplexed PCR-based capture sequencing restricts the number of capturable targets in a ‘screenable’ library usually to the order of dozens15,16,17,18.
Compared to AAV in vivo CRISPR screens, the methodology described here uses lentiviral sgRNAs. Given that lentiviral constructs integrate into the target cell’s DNA, sgRNA representation can be analysed in genomic DNA similar to any conventional CRISPR screens19,20. Thus, this methodology allows the simultaneous analysis of hundreds of genes and can be seamlessly scaled-up even to genome-wide in vivo screens. For example, a genome-wide screen with 78,000 sgRNAs and a coverage of 30x would require ~90 embryos as previously described for a similar shRNA screen4. As the litter sizes of most common inbred mouse strains is around 8-12 pups/litter and an experienced surgeon can easily inject all mice within a littler, only about ~10-12 dams and surgeries would be required for such a genome-wide screen.
The success of an in vivo screen hinges on high titer virus production. While lentiviral constructs that express a sgRNA of interest, Cas9 and Cre (e.g., pSECC) are a convenient and feasible all-in-one solution, they have the packaging limit for the lentiviral capsid (~10 kb in total size), leading to overall lower viral titer. To overcome this challenge, we use R26-LSL-Cas9-GFP mice and a lenti-viral construct containing only Cre-recombinase along with a sgRNA. The resulting lentiviral construct is only 7 kb in length and yields a viral titer of more than 108 PFU/mL.
Targeted and specific libraries can be quickly generated in as little as a few days and the complex network of genetic interactions can be studied in vivo within a few weeks. Cas9-mediated mutagenesis can be performed either in wild-type mice to study organ development, homeostasis or disease phenotypes (cancer, inflammatory skin diseases, etc.) or can be easily combined with mice harboring any oncogenic mutations or combination of mutations to model the genetic status found in human patients. In addition, the cloning methodology can be refined to clone CRISPRa or CRISPRi libraries by adding compatible restriction sites and sgRNA sequences in all-in-one plasmid. Of note, all the libraries that manipulate gene functions can be used not only in mouse models but also in other animal models and in organoid cultures. Together, this methodology highlights the utility and provides a boilerplate to use direct in vivo CRISPR to integrate somatic gene editing and mouse modeling to rapidly assess gene function in vivo.
The authors have nothing to disclose.
This work was supported by a project grant from the Canadian Institute of Health Research (CIHR 365252), the Krembil Foundation and the Ontario Research Fund Research Excellence Round 8 (RE08-065). Sampath Kumar Loganathan is the recipient of a Canadian Cancer Society fellowship (BC-F-16#31919).
0.45 micron filter | Sigma | S2HVU02RE | |
12k or 92k oligo chip | Customarray Inc. (Genscript) | ||
15 cm cell culture plates | Corning | ||
293FT | Invitrogen | R70007 | |
293NT | Systems Biosciences | LV900A-1 | |
Alkaline phosphatase | NEB | M0290L | |
Amplicillin | Fisher Scientific | BP1760-25 | |
ATP | NEB | 9804S | |
BsmBI | NEB | R0580L | |
Chromic gut suture | Covidien | ||
Deep sequencing (Next-Seq or Hi-Seq) | Illumina | ||
DNA-cleanup kit | Zymo Research | D4008 | |
DNAesy Blood and Tissue DNA extraction kit | Qiagen | 69506 | |
Endura electrocompetent cells | Lucigen | 60242-1 | |
Glass Capillaries | Drummond | 3-000-203-G/X | |
HEK293T cells | ATCC | CRL-3216 | |
High-Speed Centrifuge | Beckman Coulter | MLS-50 | |
LB Agar | Wisent Technologies | 800-011-LG | |
Micropipette puller | Sutter Instrument | P97 | |
Mineral oil | Sigma | M5904 | |
Mini-prep plasmid Kit | Frogga Bio | PDH300 | |
Mouse oxygen anaesthesia system | Visual Sonics | ||
Nanoject II micromanipulator | Drummond | ||
NEBuffer 3.1 (Buffer for BsmBI) | NEB | R0580L | |
Needle sharpener | Sutter Instrument | BV-10 | |
Oligo cleanup kit | Zymo research | D4060 | |
PAGE purified illumina sequencing primer | IDT DNA | ||
PEI (polyethyleneimine) | Sigma | 408727-100ML | |
Permoplast modeling clay | |||
Petridish with central opening | Visual Sonics | ||
pMD2.G | Addgene | 12259 | |
psPAX2 | Addgene | 12260 | |
Q5 Polymerase 2x Master mix | NEB | M0494L | |
Qubit Fluorometric Quantification | Invitrogen | Q33327 | |
Semicircular Silicone plug | Corning | ||
Silicone membrane | Visual Sonics | ||
T4 DNA ligase | NEB | M0202L | |
Ultra-centrifuge tubes | Beckman Coulter | 344058 | |
Vevo2000 ultrasound system | Visual Sonics |