A subscription to JoVE is required to view this content. Sign in or start your free trial.
Method Article
Here, we present a protocol for performing gene knockouts that are embryonic lethal in vivo in genetically engineered mouse model-derived tumors and then assessing the effect that the knockout has on tumor growth, proliferation, survival, migration, invasion, and the transcriptome in vitro and in vivo.
The development of new drugs that precisely target key proteins in human cancers is fundamentally altering cancer therapeutics. However, before these drugs can be used, their target proteins must be validated as therapeutic targets in specific cancer types. This validation is often performed by knocking out the gene encoding the candidate therapeutic target in a genetically engineered mouse (GEM) model of cancer and determining what effect this has on tumor growth. Unfortunately, technical issues such as embryonic lethality in conventional knockouts and mosaicism in conditional knockouts often limit this approach. To overcome these limitations, an approach to ablating a floxed embryonic lethal gene of interest in short-term cultures of malignant peripheral nerve sheath tumors (MPNSTs) generated in a GEM model was developed.
This paper describes how to establish a mouse model with the appropriate genotype, derive short-term tumor cultures from these animals, and then ablate the floxed embryonic lethal gene using an adenoviral vector that expresses Cre recombinase and enhanced green fluorescent protein (eGFP). Purification of cells transduced with adenovirus using fluorescence-activated cell sorting (FACS) and the quantification of the effects that gene ablation exerts on cellular proliferation, viability, the transcriptome, and orthotopic allograft growth is then detailed. These methodologies provide an effective and generalizable approach to identifying and validating therapeutic targets in vitro and in vivo. These approaches also provide a renewable source of low-passage tumor-derived cells with reduced in vitro growth artifacts. This allows the biological role of the targeted gene to be studied in diverse biologic processes such as migration, invasion, metastasis, and intercellular communication mediated by the secretome.
Before the last two decades, the treatment of human cancers relied heavily on radiotherapy and chemotherapeutic agents that broadly targeted rapidly proliferating cellular populations by damaging DNA or inhibiting DNA synthesis. Although these approaches did inhibit cancer cell growth, they also had deleterious side effects on normal rapidly proliferating cell types such as intestinal epithelial cells and hair follicle cells. More recently, cancer therapy has begun to utilize chemotherapeutic agents that precisely target proteins within signaling pathways that are critically important for the growth of an individual patient's neoplasm. This approach, commonly referred to as "Precision Medicine," has led to the development of an ever-expanding repertoire of monoclonal antibodies and small molecular inhibitors. These agents effectively inhibit tumor cell proliferation and survival while avoiding the deleterious side effects on normal cell types seen with conventional chemotherapeutic agents and radiotherapy. Monoclonal antibodies used for the treatment of human cancers most commonly target cell surface molecules such as growth factor receptors1 (e.g., the large family of membrane-spanning receptor tyrosine kinases) and immune response modulators2 (e.g., programmed cell death protein 1, programmed death-ligand 1). Small molecular inhibitors can inhibit either cell surface proteins or signaling proteins that are located intracellularly3. However, to effectively employ these new therapeutic agents, it must be established that a particular cancer is dependent upon the molecule that is being targeted by a candidate therapeutic agent.
Although these new therapeutic agents have more focused effects, many of them still inhibit the action of more than one protein. In addition, multiple agents with varying effectiveness and specificity are often available to target a specific protein. Consequently, during preclinical investigations, it is wise to use additional approaches such as genetic ablation to validate a candidate protein as a therapeutic target. One especially useful approach to validating a protein as a therapeutic target is to ablate the gene encoding the candidate protein in a genetically engineered animal model that develops the specific cancer type of interest. This approach can be relatively straightforward if mice with a null mutation (either a natural mutation, a genetically engineered null mutation [a "knockout"], or a null mutation introduced by a gene trap) are available, and the mice are viable into adulthood. Unfortunately, mice with a null mutation that meet these criteria are often not available, typically because the null mutation results in death embryonically or in the first days of postnatal life. In this circumstance, mice prone to develop the tumor type of interest may instead be crossed to mice in which key segments of the gene of interest are flanked by loxP sites ("floxed"), which allows the gene to be ablated by introducing a transgene expressing Cre recombinase into the tumor cells (a conditional knockout). This approach provides several advantages. First, if a Cre driver is available that is expressed in the tumor but not in the cell type that led to death in conventional knockouts, this approach can potentially validate the candidate therapeutic target. Second, ablating the gene encoding the candidate protein in tumor cells but not in other intratumoral elements such as tumor-associated fibroblasts or immune cells allows the investigator to distinguish between cell-autonomous and non-cell-autonomous effects of the therapeutic target. Finally, a tamoxifen-inducible Cre driver (CreERT2) allows the investigator to delete the gene of interest at different stages in tumor development and define the window in which the candidate therapeutic agent is most likely to be effective.
Unfortunately, there are also technical issues that can limit the use of conditional knockouts in tumors arising in GEM models. For instance, a Cre driver that is expressed in tumor cells and avoids gene deletion in normal cells essential for life may not be available. Another issue, which may be underestimated, is that Cre and CreERT2 drivers often variably ablate floxed alleles in mice, resulting in mosaicism for the null mutation in a GEM cancer. When this occurs, tumor cells in which the targeted gene has not been ablated will continue to proliferate rapidly, overgrowing the tumor cells with ablated alleles. Mosaicism in Cre driver lines can occur due to non-ubiquitous Cre expression in the lineage targeted and by failed recombination in individual cells independent of Cre expression4. This is a known phenomenon of Cre drivers that is cell-type dependent and should be considered during experimental design and data interpretation. Mosaicism can mask the effect of the knockout and lead an investigator to erroneously conclude that the gene of interest is not essential for tumor cell proliferation and/or survival and thus is not a valid therapeutic target.
Several of these problems were encountered in a previous study that attempted to determine whether the receptor tyrosine kinase erbB4 was a potential therapeutic target in MPNST cells5. In these studies, mice were used that express a transgene encoding the neuregulin-1 (NRG1) isoform glial growth factor-β3 (GGFβ3) under the control of the Schwann cell-specific myelin protein zero promoter (P0-GGFβ3 mice). P0-GGFβ3 mice develop multiple plexiform neurofibromas that progress to become MPNSTs via a process that recapitulates the processes of neurofibroma pathogenesis and plexiform neurofibroma-MPNST progression seen in patients with the autosomal dominant tumor susceptibility syndrome neurofibromatosis type 1 (NF1)6. When crossed to mice with a Trp53 null mutation, the resulting P0-GGFβ3;Trp53+/- mice develop MPNSTs de novo as is seen in cis-Nf1+/-;Trp53+/- mice.
These MPNSTs recapitulate the progression from World Health Organization (WHO) grade II to WHO grade IV lesions seen in humans7. In P0-GGFβ3 mice, MPNSTs arise within pre-existing plexiform neurofibromas in the trigeminal nerve (58%) and spinal dorsal nerve roots (68%)7; the MPNSTs arising in P0-GGFβ3;Trp53+/- mice have a highly similar distribution. In humans, MPNSTs most commonly arise in the sciatic nerve followed by the brachial plexus, spinal nerve roots, vagus, femoral, median, sacral plexus, popliteal obturator, and posterior tibial and ulnar nerves8. This tumor distribution in these GEM models is somewhat different from what is seen in humans. However, the MPNSTs that arise in P0-GGFβ3 and P0-GGFβ3;Trp53+/- mice are histologically identical to human MPNSTs, carry many of the same mutations seen in human MPNSTs, and recapitulate the process of neurofibroma-MPNST progression seen in NF1 patients. The generation of P0-GGFβ3 or P0-GGFβ3;Trp53+/- mice that were Erbb4-/- was not feasible as mice with two Erbb4 null alleles die in utero at embryonic day 10.5 secondary to cardiac defects9. Because rescuing Erbb4 expression in the heart by introducing a cardiac-specific Erbb4 transgene (α-myosin heavy chain (MHC)-Erbb4) results in viable Erbb4-/- mice10, the generation of mice with a complicated P0-GGFβ3;Trp53+/-;α-MHC-Erbb4;Erbb4-/- genotype was attempted.
However, the matings did not produce mice in the expected Mendelian ratios, indicating that the desired genotype was deleterious. Therefore, the generation of P0-GGFβ3;Trp53+/- mice with floxed Erbb4 alleles11 and a CreERT2 driver was attempted to allow the deletion of Erbb4 in the MPNSTs arising in these mice. In these animals, numerous tumor cells with intact Erbb4 alleles were still present (mosaicism). The mosaicism observed could result from inefficient tamoxifen delivery, which resulted in differences in recombination efficiency within the tissue. The possibility of spontaneous compensatory mechanisms could further contribute to mosaicism in tamoxifen-mediated recombination by bypassing the requirement for Erbb4 expression. It is feasible that the loss of Trp53 makes tumor cells susceptible to additional spontaneous "permissive" mutations that could confuse the interpretation of the data. As it seemed likely that the Erbb4-intact MPNST cells would mask the consequences of ablating Erbb4 in other tumor cells, this approach was abandoned.
These obstacles led to the development of a methodology for ablating Erbb4 in very early passage MPNST cells using an adenovirus expressing Cre recombinase and eGFP. These cells can be separated from non-infected cells using FACS, which markedly reduces mosaicism for the ablated Erbb4 gene. Below, the methods used to achieve this, together with the methods used to assess the effects of gene ablation in vitro and in vivo, are described. The following protocol is an example of how to produce tumor-bearing mice that yield tumors carrying floxed alleles of embryonic lethal genes of interest for ex vivo excision prior to in vivo allograft tumor growth assessment. This includes a description of the approaches used to analyze the effect that Erbb4 ablation exerts on tumor cell proliferation, survival, and gene expression in vitro and proliferation, survival, and angiogenesis in orthotopic allografts.
Prior to performing any procedures with mice, all procedures must be reviewed and approved by the Institutional Animal Care and Use Committee. The protocol described in this manuscript was approved by the Institutional Animal Care and Use Committee of the Medical University of South Carolina. This protocol was performed by properly trained personnel following MUSC's institutional animal care guidelines.
1. Generation of mice that develop MPNSTs homozygous for Erbb4 flox alleles
2. Ex vivo ablation of floxed Erbb4 alleles in MPNST cells
3. Proliferation and viability assays in MPNST cells with ablated Erbb4 alleles
4. RNA-Seq analyses and identification of genes whose expression is altered by Erbb4 loss
5. Orthotopic allografting of MPNST cells with ablated Erbb4 alleles and analysis of the effects of Erbb4 ablation
Figure 4 illustrates a typical result obtained when transducing P0-GGFβ3;Trp53-/+;Erbb4fl/fl MPNST cells with either the Ad5CMV-eGFP adenovirus or Ad5CMV-Cre/eGFP adenovirus (Figure 4A). Cultures are viewed with fluorescence microscopy to identify eGFP-expressing cells and by phase-contrast microscopy to determine the total number of cells present in the same field at 10x (top) and 40x (bottom). The pe...
The detailed methods presented here were developed using a GEM model of MPNSTs. However, if the tumor tissue of interest can be dispersed into individual cells, these methodologies are easily adaptable for various tumor types arising in GEMs. It is important to ensure that the floxed allele does not result in i) decreased survival that can make it difficult to obtain sufficient mice to screen for tumors, or ii) increased tumor latency that can make it difficult to obtain enough tumor-bearing mice. If the floxed allele do...
The authors have no conflicts of interest to disclose.
This work was supported by grants from the National Institute of Neurological Diseases and Stroke (R01 NS048353, R01 NS109655), the National Cancer Institute (R01 CA122804), the Department of Defense (X81XWH-09-1-0086, W81XWH-11-1-0498, W81XWH-12-1-0164, W81XWH-14-1-0073, and W81XWH-15-1-0193), and The Children's Tumor Foundation (2014-04-001 and 2015-05-007).
Name | Company | Catalog Number | Comments |
Ad5CMV-eGFP | Gene Transfer Vector Core, Univ of Iowa | VVC-U of Iowa-4 | |
Ad5CMVCre-eGFP | Gene Transfer Vector Core, Univ of Iowa | VVC-U of Iowa-1174 | |
alexa 568 secondary antibody | Thermo/Fisher | GaR A11036 | |
Bioconductor Open Source Software for Bioninformatics | Bioconductor | http://www.bioconductor.org | alternative statistical analysis tool used for step 4.4 |
CD31 | Abcam | ab28364 | |
Celigo Image Cytometer | Nexcelom Bioscience | N/A | |
Cell Stripper | Corning | 25-056-Cl | mixture of chelators |
DAB staining kit | Vector Labs | MP-7800 | |
DAVID (Database for Annotation, Visualization, and Integrated Discovery) | DAVID | https://david.ncifcrf.gov | functional enrichment analysis software used for step 4.5 |
DMEM | Corning | 15-013-Cl | |
DreamTaq and Buffer (Genotyping PCR) | Thermo/Fisher | EP0701 and K1072 | |
erbB4 antibodies | Santa Cruz | sc-284 | |
erbB4 antibodies | Abcam | ab35374 | |
erbB4 antibodies | Millipore | HFR1: 05-1133 | |
FACS Sorter | BD Biosciences | Aria II | |
Forskolin | Sigma | F6886 | |
GenomeSpace Tools and Data Sources | GenomeSpace | https://genomespace.org/support/tools/ | general resource for several types of open source bioinformatic tools for step 4.5 |
Glutamine | Corning | 25-005-Cl | |
Gorilla Gene Ontology enRIchment anaLysis and visuaLizAtion tool | Gorilla | N/A, http://cbl-gorilla.cs.technion.ac.il | functional enrichment analysis software used for step 4.5 |
GSEA Gene Set Enrichment Analysis | Broad Institute | N/A, https://www.gsea-msigdb.org/gsea/index.jsp | functional enrichment analysis software used for step 4.5 |
HSD: Athymic Nude-FOxn1nu mice | Envigo (Previously Harlan Labs) | 69 | |
Illumina HiSeq2500 (next generation DNA sequencer) | Illumina | Hi Seq 2500 | DNA sequencer used for step 4.2 |
Lasergene: ArrayStar Gene expression and variant analysis | DNAStar LaserGene software | N/A | software statistical and normalization analysis used for step 4.4 |
Lasergene: SeqMan NGen sequence alignment assembly | software alignment used for step 4.3 | N/A | software alignment used for step 4.3 |
Matrigel, low growth factor basement membrane matrix | Corning | 354230 | |
NRG1-beta | In house | Generated by SLC, also commercially available from R & D Systems(396-HB-050/CF). | |
Nuclear Stain Hoeschst 33342 | Thermo | 62249 | |
Panther Gene Ontology Classification System | Panther | http://pantherdb.org | functional enrichment analysis software used for step 4.5 |
Partek (BWA aligner and analyzer) | Partek, Ver 7 | N/A | software alignment and statistical/normalization used for step 4.3 |
Pen/Strep | Corning | 30-002-Cl | |
Primer 1: 5′-CAAATGCTCTCTCTGTTCTTTGT GTCTG- 3′ | Eurofins Genomics | Primer 1 + 2: 250 bp ErbB4 null product and a 350 bp Floxed ErbB4 product; | |
Primer 2: 5′-TTTTGCCAAGTTCTAATTCCATC AGAAGC-3′ | Eurofins Genomics | Primer 1 + 2: 250 bp ErbB4 null product and a 350 bp Floxed ErbB4 product; | |
Primer 3: 5′-TATTGTGTTCATCTATCATTGCA | Eurofins Genomics | Primer 1 + 3: 350 bp wild-type ErbB4 product. | |
Propidium Iodine | Fisher | 51-351-0 | |
Proteom Profiler Phospho-Kinase Arrays | R&D Systems | ARY003B | |
Real time glo | Promega | G9712 | bioluminescent cell viability assay |
ToppGene Suite | ToppGene | https://toppgene.cchmc.org | functional enrichment analysis software used for step 4.5 |
Trizol (acid-quanidinium-phenol and choloroform based reagent) | Invitrogen | 15596026 | |
Tyramide Signal Amplification Kit | Perkin Elmer | NEL721001EA |
Request permission to reuse the text or figures of this JoVE article
Request PermissionThis article has been published
Video Coming Soon
Copyright © 2025 MyJoVE Corporation. All rights reserved