|Home | About | Journals | Submit | Contact Us | Français|
We discuss the generation of primary soft tissue sarcomas in mice using the Cre-loxP system to activate conditional mutations in oncogenic Kras and the tumor suppressor p53 (LSL-KrasG12D/+; p53flox/flox). Sarcomas can be generated either by adenoviral delivery of Cre recombinase, activation of transgenic Cre recombinase with tamoxifen, or through transplantation of isolated satellite cells with Cre activation in vitro. Various applications of these models are discussed, including anticancer therapies, metastasis, in vivo imaging, and genetic requirements for tumorigenesis.
Soft tissue sarcomas (STS) are malignant tumors of the connective tissue that affect muscle, fibrous tissue, fat, blood vessels, and nerves . In humans, these tumors are often characterized by inactivating mutations in the p53 tumor-suppressor pathway and activating mutations in the oncogenic Kras pathway . To examine sarcoma pathogenesis, we have utilized the Cre-LoxP system to develop mouse models of soft tissue sarcoma . Cre is a site-specific recombinase that recognizes loxP sequences, resulting in excision of DNA that is flanked by two loxP sites. Sarcomas are generated by activation of Cre in mice or cells with conditional mutations in both oncogenic Kras (LSL-KrasG12D/+) and mutant p53 (p53flox/flox). In the absence of Cre recombinase, p53 is expressed at wild-type levels, whereas oncogenic Kras is not expressed. After Cre activation, the floxed p53 exons 2–10 are deleted, resulting in inactivation of both p53 alleles. At the same time, Cre activates oncogenic Kras by deleting an upstream floxed transcription/translation STOP cassette (termed “loxP-STOP-loxP,” or LSL cassette). The resulting tumor expresses KrasG12D and is p53-null.
Cre recombinase can be activated by a variety of techniques. First, Cre can be delivered directly to the animal via intramuscular injection of an adenovirus that expresses Cre (Ad-Cre) . Following intramuscular injection of Ad-Cre into the compound mutant mice (LSL-KrasG12D/+; p53flox/flox), high-grade sarcomas develop at the site of injection within 2–4 months. Additionally, approximately 40 % of these mice develop lung metastases, which are common in human soft tissue sarcoma patients. A second method for deleting the loxP alleles is to use CreER technology . This method uses a Cre recombinase that is fused to the hormone-binding domain of the estrogen receptor (ER) , generating a CreER allele that can be activated by the estrogen analog 4-hydroxy-tamoxifen (4OHT). Following 4OHT exposure, CreER translocates to the nucleus, where Cre recombinase excises DNA flanked by loxP sites. In contrast to the Ad-Cre method, this approach can restrict Cre activity to specific tissues and cell types by utilizing transgenic mice that express CreER from cell type-specific promoters. Using a Pax7-CreER allele that is specific to the Pax7+ muscle satellite cells , we recently generated rhabdomyosarcomas and undifferentiated pleomorphic sarcomas following intraperitoneal injection of tamoxifen , which is converted to 4OHT in the liver. A third option for tumor generation involves isolating tumor-initiating cells by fluorescence activated cell sorting (FACS) via expression of specific cell surface markers. These cells can be transformed in vitro by either Ad-Cre or 4-OHT (for CreER-containing cells) and then transplanted into an animal. This method allows direct isolation of putative tumor-initiating cells, facilitating the study of tumorigenesis by specific gene mutations. After isolating  and transforming satellite cells in vitro by deleting p53 and expressing oncogenic Kras, we have generated rhabdomyosarcomas following transplantation into recipient mice (Blum JM, Li Z, and Kirsch DG, unpublished data).
All of these techniques can be used to functionally test the role of novel therapies or genes that regulate sarcoma development in a genetically tractable system. These mouse models provide distinct advantages for examining sarcoma pathogenesis that are not achievable by cell transplant or xenograft studies . First, the mutant genes that initiate sarcoma growth are expressed from their endogenous promoters at physiological levels. Second, by spatially and temporally restricting tumor initiation, sarcomas develop in adult mice at a defined anatomic site. This allows close monitoring of tumor size in response to experimental treatments. Third, tumors develop within the native tissue microenvironment, surrounded by tissue that is not transformed. This facilitates studies that examine the relationship between the tumor stroma and therapeutic response. Finally, the host has an intact immune system, in contrast to xenograft model systems, which permits further examination of the tumor stroma on sarcoma biology.
These genetically engineered mouse models of STS may be used for multiple translational applications. Several specific examples are discussed below to provide the reader with examples that may stimulate their own experiments.
First, the models initiated by intramuscular injection of Adeno-Cre or 4OHT have been used to investigate various anticancer treatments. Because the tumors are generated in the leg, monitoring changes for tumor size in response to therapy is simple using calipers. We have successfully used these models to test sarcoma response to radiation therapy , molecularly targeted agents against MEK , and combination treatments, including sunitinib with radiotherapy  and PI3K inhibitors with conventional doxorubicin chemotherapy . Second, these tumor models metastasize to the lungs, as seen in sarcoma patients, and thus the model initiated by Adeno-Cre injection is adaptable for in vivo metastasis assays. Approximately 20 % of the mice will develop pulmonary metastases at the time of primary tumor development, but the primary tumors grow at such a rapid rate that the animals must be sacrificed before the true metastatic potential of each tumor can be determined. To circumvent this limitation, the tumor-bearing limb can be amputated, which extends the survival of the mice while allowing for potential lung metastases to grow. Following amputation, approximately 40 % of mice develop pulmonary metastases. We have used this amputation-based protocol to make several observations about sarcoma metastasis, including the role of hypoxia inducible factor-1alpha (HIF-1α)  and miRNAs . Third, these models can be used for in vivo imaging or surgical studies because the primary tumors develop specifically at the site of intramuscular injection. We have used these models to test several novel in vivo tumor imaging systems, including intraoperative imaging with cathepsin-activatable fluorescent imaging agents [14, 15] and dual-energy microcomputed tomography  with gold nanoparticles . Finally, the models described above can be extended to test novel gene mutations that may be relevant to distinct subtypes of sarcoma. For example, genomic analyses of patient sarcomas revealed that mutations in the neurofibromin 1 (NF1) gene are common in a number of STS. To test this observation in a genetically engineered mouse model, we crossed mice with floxed alleles for NF1 and the tumor suppressor Ink4a/Arf. Injection of Ad-Cre into either the nerve sheath or muscle of NF1flox/flox; Ink4a/Arfflox/flox mice generated sarcomas that resembled malignant peripheral nerve sheath tumors (MPNST) or high-grade myogenic sarcomas, respectively . In addition to demonstrating the utility of the model for examining different genetic mutations, this study also highlights the possibility of examining alternative anatomic sites for sarcomagenesis.
These techniques allow for many modifications. Injection of Ad-Cre into other anatomical sites can generate tumors at other sites in addition to the lower limb. Using other tissue-specific CreER drivers can potentially be used to generate sarcomas from different tumor-initiating cells following delivery of tamoxifen. When deciding between using an Ad-Cre-driven tumor model system or a tamoxifen-inducible CreER system, certain factors should be considered. Advantages of Ad-Cre delivery include generation of sarcomas even when the cell of origin is not known. In contrast, the CreER system enables the generation of sarcomas from specific cell types. Investigators must also consider the availability of mice in which CreER is expressed from cell type-specific promoters and the amount of time required to breed the animals with the CreER allele. Note that lung tumors may be generated in the LSL-Kras G12D/+; p53flox/flox mice using a similar protocol if Ad-Cre is delivered to the lungs by intranasal inhalation .
CreER is activated by exposure to tamoxifen and its analogs, the most potent of which is 4-hydroxy-tamoxifen (4OHT). Two delivery methods for 4OHT are available. For localized administration, 4OHT is injected directly into the muscle (intramuscular—IM). For systemic administration, tamoxifen is injected into the peritoneum (intraperitoneal—IP). The injected tamoxifen is then metabolized by the liver to yield 4OHT, which can recombine loxP sites throughout the body.
We thank members of the Kirsch lab for helpful discussions and for contributing to the development of these techniques. We acknowledge members of Amy Wagers lab and Christoph Lepper for developing the satellite cell isolation technique and for graciously sharing their protocols.
1Tamoxifen is poorly soluble in sterile corn oil. To aid in dissolution, a 20 mg/mL solution is heated at 50 °C for 2–5 h. 1-mL aliquots can be frozen and stored at −80 °C, though they require heating to redissolve upon thawing. Alternatively, tamoxifen can be more readily dissolved in a 5 % ethanol/95 % corn oil mixture. First, the tamoxifen is mixed in 100 % ethanol. The concentration of tamoxifen is too high for it to fully dissolve, and it generates a white suspension. Corn oil is then added to the appropriate volume. For example, we generally purchase 1 g of tamoxifen and mix with 2.5 mL of 100 % ethanol, creating a slurry. We then add 47.5 mL of corn oil to bring the mixture to the final concentration of 20 mg/mL. The solution is heated at 50 °C on a shaker for approximately 1 h to aid in dissolution of the tamoxifen. It is then frozen in 1-mL aliquots at −80 °C. In contrast to when tamoxifen is in frozen in 100 % corn oil, tamoxifen frozen in 5 % ethanol/95 % corn oil remains in solution and requires no heating prior to use.
2For Ad-Cre-generated tumors, large masses develop at the site of injection within 2–4 months in ~90 % of all animals. Once detected, tumors will grow at a rapid rate, resulting in the animal needing to be sacrificed within approximately 2 weeks. Light microscopy demonstrates that these tumors are high-grade sarcomas. Gene expression analysis shows that these tumors most closely resemble the subtype of soft tissue sarcoma referred to as undifferentiated polymorphic sarcoma (UPS) .
3Critical steps in the Ad-Cre tumor generation process include complete mixing of the viral cocktail and adherence to the 15–60-min time window for generating viral particles. Injection of the cocktail outside of this window could result in the viral precipitates being too large or too small to enter the cells. Consistency in the site of injection, depth of needle penetration, and volume injected will increase the likelihood that tumor will form at the same site on different mice. If experimental mice are scarce, then we recommend practicing Ad-Cre injections on littermates without useful genotypes to maximize reproducibility of results.
4Tumor development with the Pax7-CreER system is highly dependent on the delivery method of tamoxifen. Intraperitoneal injection of tamoxifen results in tumors at multiple sites throughout the animal by 40–60 days. Gene expression analyses show these tumors are most similar to human embryonal rhabdomyosarcoma (ERMS) or UPS . Tumors generated with intramuscular injection of 4OHT contain a mixed histology that includes rhabdomyoblasts that stain for RMS markers, including MyoD and myogenin, and other areas of UPS.
5Isolated satellite cells appear as semi-fused fibers with fibroblastic appearance by 7 days post-isolation. The tumors generated from satellite cells transformed in vitro generally appear in nude mice within 30 days (Figs. 1 and and2).2). These tumors contain rhabdomyoblasts and stain for myogenin.
6For satellite cell isolation, it is important to not completely mince the fresh muscle at the beginning of the protocol. Although it may appear that complete mincing will promote digestion of the tissue, this may kill the satellite cells you are attempting to isolate.
7Another key step in this protocol is to thoroughly homogenize the mixture with the glass pipette and rubber bulb. Skimping on the 250 required strokes will dramatically decrease the number of cells isolated.
Disclosures : None.