This protocol describes the generation of patient-derived orthotopic xenograft models by intra-vesically instilling high-grade urothelial cell carcinoma cells or intra-rectally injecting colorectal cancer cells into non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice for primary tumor growth and spontaneous metastases under the influence of lymph node stromal cells, which mimics the progression of human metastatic diseases.
Cancer patients have poor prognoses when lymph node (LN) involvement is present in both high-grade urothelial cell carcinoma (HG-UCC) of the bladder and colorectal cancer (CRC). More than 50% of patients with muscle-invasive UCC, despite curative therapy for clinically-localized disease, will develop metastases and die within 5 years, and metastatic CRC is a leading cause of cancer-related deaths in the US. Xenograft models that consistently mimic UCC and CRC metastasis seen in patients are needed. This study aims to generate patient-derived orthotopic xenograft (PDOX) models of UCC and CRC for primary tumor growth and spontaneous metastases under the influence of LN stromal cells mimicking the progression of human metastatic diseases for drug screening. Fresh UCC and CRC tumors were obtained from consented patients undergoing resection for HG-UCC and colorectal adenocarcinoma, respectively. Co-inoculated with LN stromal cell (LNSC) analog HK cells, luciferase-tagged UCC cells were intra-vesically (IB) instilled into female non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice, and CRC cells were intra-rectally (IR) injected into male NOD/SCID mice. Tumor growth and metastasis were monitored weekly using bioluminescence imaging (BLI). Upon sacrifice, primary tumors and mouse organs were harvested, weighed, and formalin-fixed for Hematoxylin and Eosin and immunohistochemistry staining. In our unique PDOX models, xenograft tumors resemble patient pre-implantation tumors. In the presence of HK cells, both models have high tumor implantation rates measured by BLI and tumor weights, 83.3% for UCC and 96.9% for CRC, and high distant organ metastasis rates (33.3% detected liver or lung metastasis for UCC and 53.1% for CRC). In addition, both models have zero mortality from the procedure. We have established unique, reproducible PDOX models for human HG-UCC and CRC, which allow for tumor formation, growth, and metastasis studies. With these models, testing of novel therapeutic drugs can be performed efficiently and in a clinically-mimetic manner.
It has been shown that lymph node (LN) metastasis is a poor prognostic indicator in many solid organ malignancies, including high-grade urothelial cell carcinoma (UCC) of the bladder and colorectal cancer (CRC)1,2. Over half of the patients with muscle-invasive UCC (MIUCC), despite curative therapy for clinically-localized disease, will develop metastases and die within 5 years. Metastatic CRC is a leading cause of cancer-related death in the US.
An estimated 81,190 new patients and 17,240 cancer specific deaths are expected to occur in 2018 in the United States due to UCC of the bladder3,4. While patients will predominantly (70%) present with non-muscle invasive disease, 30% will have MIUCC5. Despite the curative therapy (radical cystectomy [RC] with or without systemic chemotherapy) for clinically localized disease, half of the patients with MIUCC of the bladder will still develop metastases and die within 5 years3. Lymph node involvement is found in approximately 20%−25% of patients having undergone RC6,7,8. Five-year survival rate in LN positive patients is less than 35% even after RC, suggesting LN involvement as a crucial negative predictor for the prognosis in UCC patients.
Colorectal cancer is the third most common cancer diagnosed in both men and women in the United States. The patient outcomes largely depend on tumor characteristics and tumor microenvironment, such as depth of invasion, LN involvement, and distant organ metastases. Although the mortality rate in CRC decreased in the last decade due to screening and effective surgeries, it is estimated that almost 50% of CRC patients will develop metastases or recurrent disease9.
Small animal models provide an expeditious, reproducible, and modifiable platform to study tumor progression and different metastatic patterns. There are currently no described xenograft models that consistently mimic CRC and UCC metastasis seen in patients. The primary route of cancer distant metastasis is via lymphatic spread. New research suggests that the LNs provide tumors with a unique microenvironment, and are not only simply stationary targets where cancer cells transiently pass, but also play an integral role by interacting with cancer cells in the metastatic process. Indeed, our studies discovered that, in addition to educating and promoting tumor progression and metastases, the LN stromal microenvironment is also responsible for drug resistance in CRC10,11. Our lab recently confirmed the tumorigenic effects of LN stromal cells (LNSCs) on CRC and UCCs using patient-derived orthotopic xenograft (PDOX) mouse models12,13.
Developing PDOX models provides an important platform for translational cancer research14,15. By maintaining the principal histologic and genetic characteristics of their donor tumor, PDOX models remain stable across passages and make good platforms for translational cancer research12,15. PDOX models are being used for preclinical drug evaluation, biomarker identification, and preclinical evaluation of personalized medicine strategies allowing for prediction of clinical outcomes. Currently, there are no described xenograft models that consider the importance of LN involvement and are capable of consistently reproducing primary tumor and distant organ metastasis in CRC and UCC. In this study, we describe the development of PDOX models in NOD/SCID mice with reproduction of metastatic CRC and UCC diseases with LNSC involvement.
All methods described in these animal studies were conducted under the approved guidelines of the Institutional Animal Care and Use Committee of Ochsner Health System and in accordance with animal research guidelines. All patient tumors for this study were collected from consented patients undergoing cancer resection surgeries in accordance with the Ochsner Health System Investigative Review Board and the ethical standards of the Institutional Committee on Human Experimentation. Board-certified pathologists at Ochsner Health System determined the pathological diagnoses of patient specimens based on the microscopic features of tumor cells, their histological type, and grade level.
NOTE: The following protocol describes the steps for two separate xenograft models, a UCC model via the electrocauterization of the bladder wall to instill UCC cells and an intrarectal injection of CRC cells for study in a CRC model. All steps preparing for and monitoring the experiments are identical for both models, while sections 7 and 8 specifically describe the procedure for UCC instillation and CRC injection, respectively.
1. Culturing cell lines
2. Patient specimen collection
3. Expansion of patient tumor
4. Tagging and enrichment of luciferase labeled tumors
5. Select appropriate portion of tumor for enzymatic digestion
6. Enzymatic digestion of tumor
7. UCC mouse model
8. CRC mouse model
9. Bioluminescent imaging
10. Harvesting organs and tumor
11. Histological evaluation
In the UCC PDOX model, UCC patients' BlCaPt15 or BlCaPt37 cells were intra-vesically (IB) instilled in the presence of HK cells into female NOD/SCID mouse bladder (Figure 1A). Twenty-five out of thirty (83.3%) animals generated primary tumors and displayed time dependent primary tumor growth based on weekly BLI (Figure 1B,C and Table 1). Similarly, in the CRC PDOX model, 31 out of 32 (96.9%) mice grew primary tumor when intra-rectally (IR) injected with patients' CoCaPt155 or CoCaPt302 cells plus HK cells (Figure 1D-F and Table 1). Depending on the patient tumor, mouse tumor growth had a different latency period, which reflects the difference in the patient's clinical characteristics (Figure 1C,F).
In both IB and IR models, tumor cell injection not only generated orthotopic primary tumors (Figure 2A,B, blue arrows), but many mice tested also developed liver and/or lung metastases. In 10 out of 30 (33.3%) and 17 out of 32 (53.1%) mice instilled with UCC cells and CRC cells with HK cells, respectively, we detected distant organ metastasis via ex vivo BLI (Figure 2A,B and Table 1).
To confirm similar tissue morphology, H&E and IHC staining were performed comparing xenografts and primary patient tumors. Histopathology of patient bladder carcinoma was maintained in xenografts from BlCaPt15 and BlCaPt37 (Figure 3A). Results show xenograft tumor corresponding to the muscle invasive growth pattern of the patients' primary tumors. The antibody specific to human cell proliferation marker Ki67 was used in IHC. Ki67 positive nuclear staining indicates highly proliferative, fast-growing human tumor cells. The staining results from xenografts were similar to those of the original surgical biopsies. Similarly, in the IR model, H&E staining indicates the similarity of architecture between xenografts and patient tumors of both CoCaPt155 and CoCaPt302. IHC using antibody against cytokeratin 20 also showed similar tumor growth pattern in both PDOX models (Figure 3B). Thus, our PDOX model recapitulated UCC and CRC patient clinical progression.
Figure 1: Orthotopic UCC and CRC mouse models. (A-C) Intra-vesicle (IB) instillation of UCC cells into mouse bladder13.(Aa) An angiocatheter was inserted into the bladder of a female NOD/SCID mouse and an electrocautery shock was applied to the bladder wall via a guide wire. (Ab) Luciferase tagged UCC tumor cells, BlCaPt15 (2 x 104 cells), or BlCaPt37 (5 x 105 cells) with the addition of 3 x 105 LN stromal HK cells, were instilled into the NOD/SCID mouse bladder through the angiocatheter. (D-F) Intra-rectal (IR) injection of CRC cells into the submucosal tissue layer of mouse rectum17. (D) The anal canal was dilated with lubricated blunt-tipped forceps to allow access to the distal anal and rectal mucosa and a 30 G needle was inserted into the distal posterior rectal submucosa 1−2 mm above the anal canal until the bevel was covered before the injection takes place. Luciferase tagged CRC tumor cells, CoCaPt155 (5 x 105 cells), or CoCaPt302 (1 x 104 cells) with the addition of 3 x 105 HK cells were injected. Tumor burden was monitored and quantified via bioluminescent imaging (BLI; B and E).Tumor growth of luciferase tagged UCC or CRC cells was monitored kinetically via BLI and analyzed using image analysis software (C and F). Please click here to view a larger version of this figure.
Figure 2: PDOX models produce spontaneous distant organ metastases. Representative mice (top panels) from the same experiments as in Figure 1, e.g., instilled intra-vesically with luciferase tagged UCC tumor cells, BlCaPt15 or BlCaPt37 cells with HK cells (A) or intra-rectally with luciferase tagged CRC tumor cells, CoCaPt155 or CoCaPt302 cells with HK cells (B) are shown. Yellow arrows indicate mouse bladder (A). Photos taken at the time of sacrifice indicate orthotopic tumor formation (blue arrows). Liver, lung, and tumor (middle panels) collected at necropsy and their ex vivo BLI (bottom panels) showed mouse liver and lung metastasis as well as tumor with luciferase activity. Please click here to view a larger version of this figure.
Figure 3: Xenograft tumors resemble patient pre-implantation tumors. Paraffin embedded tumor tissue from patient tumor or tumors collected from mice in the same experiments as in Figure 1 were sectioned and stained by H&E (A and B) or IHC with antibodies against human Ki67 (A) or cytokeratin 20 (CK20; B). The brown color indicates positive staining. H&E staining shows tumor nests dissecting into smooth muscle bundles (A). Photographs were taken using a digital deconvoluting microscope and analyzed with an image analysis software. Scale bars: 100 µm. All images were taken in original magnification of 200×. Please click here to view a larger version of this figure.
Tumor implant (%) | Lung/liver metastasis (%) | Mortality (%) | |
UCC, n=30 | 83.3 | 33.3 | 0 |
CRC, n=32 | 96.9 | 53.1 | 0 |
Table 1: Summary of tumor formation, metastasis, and mortality in IB and IR models.
Metastatic disease is responsible for most cancer patient mortalities. In pre-clinical therapeutic tests, it is crucial to establish mouse models that most closely emulate human tumor growth with spontaneous distant organ metastases. Using murine models with implanted patient tumor derived cancer cells (xenografts) allows for a better understanding of tumor biology and predictive biomarkers as well as testing and prediction of antineoplastic effects of novel therapies18. Many models have been used to show UCC and CRC metastases in murine experiments, such as intravenous tail vein injections showing the ability to produce lung disease19 or subcutaneous implantation of tumor cells or tumor fragments into the flank for localized tumor growth20,21. One laboratory previously reported a bladder cancer murine model by using hydrochloric acid treatments to successfully promote tumor uptake22. While these methods produce reliable local growth and may demonstrate some metastatic activities, they do not specifically resemble the natural course of cancer developed in humans and do not utilize the metastatic mechanism seen in patients18,23. Other murine models were reported to mimic tumor growth by injecting tumor cells directly into organs such as the liver or mesentery, but they carried risks of tumor cell leakage and did not produce significant metastases.
We have previously demonstrated the correlation between cancer cell content in the primary tumor and LN involvement24 and the role of the cancer cell/LN stromal interaction in the course of primary tumor progression to metastatic disease10,12,17. Incorporating our previous work on the influence of the LN stromal microenvironment in metastatic progression, we established orthotopic models (especially the PDOX models) that mimic the natural course of metastatic dissemination, are technically reproducible, preserve the heterogeneity of original patient tumors, and generate consistent primary tumor and metastatic results12,13,17. Using the tumor-enhancing effects of the LN stromal microenvironment is important because it provides a similar tumor microenvironment in human UCC and CRC, develops all steps in the metastatic cascade, reduces cancer cell number required in the mouse model which minimizes number of xenograft passages, and results in a reliable model which closely mimics tumor growth and metastases in human.
We have established a unique method of IB electro-stimulation using the co-instillation of HK cells that produces a reliable model for developing MIUCC. Our model mimics the natural course of UCC progression by tumor implantation beginning in the mucosa, leading into the muscle, then metastasizing to lungs13.
Our results also show that the IR model is safe, reproducible, and successful. The orthotopic CRC mouse model features primary tumor growth and spontaneous distant metastasis12,17. The IR procedure is quick, easy to learn, technically easy to perform, and not too stressful on the animals. The IB and IR groups had zero mortality (Table 1) in the postoperative period before final BLI measurement. However, the technique requires practice. If the intrarectal injection is successful, there should be a visible "bubble" that forms as the fluid is introduced into the rectal submucosa and will result in primary tumor growth that will eventually become palpable as shown in Figure 1. If the tumor has been injected too deep into the pelvic cavity, it will be unattached to the colorectal tract and grows very large to fill the pelvis, sometimes causing obstruction. If the injection is too shallow or does not enter the rectal submucosal layer at all, it will leak out resulting in reduced or absent primary tumor burden.
We have established unique, reproducible PDOX models for human HG-UCC and CRC. These models allow for tumor formation and metastasis studies. We can now use these models as the primary method to continue to study the LN stromal microenvironment and its interaction with patient primary tumors. These models will also allow us to investigate therapies that interfere with the pro-tumorigenic effects of the LNSC on primary tumor. With these models, testing of novel therapeutic drugs can be performed efficiently and in clinically-mimetic manners.
The authors have nothing to disclose.
The authors thank Brian Reuter, Danielle Bertoni, Peter Miller, and Shannon McChesney who helped to initiate these studies for their excellent technical support. The authors also thank Heather Green Matrana, Margaret Variano, Sunil Talwar, and Maria Latsis for assistance in consenting patients and providing tumor specimens.
Avidin-biotin-peroxidase | Vector Labs Inc | PK-6100 | |
Biotinylated secondary antibody | Vector Labs Inc | BA-1000 | |
Collagenase IV (1.5 mg/mL) | Worthington Biochemical Corporation | LS004189 | |
Deoxyribonuclease I (0.1 mg/mL) | Sigma | D4263 | |
D-Luciferin (150 mg/kg) | Perkin Elmer | 122796 | |
Formalin (10% neutral buffered) | Leica | 46129 | |
glutamine (2 nM) | Fisher Scientific | 35050061 | |
Hair Removal Cream | Church & Dwight Co., Inc | 1 (800) 248-8820 | |
Hanks Balanced Salt Solution (HBSS) | Fisher Scientific | SH30016.02 | |
Hyaluronidase (20 mg/mL) | Sigma | H3884 | |
Isoflurane | Henry Schein Animal Health | 108333 | |
Luc/RFP-lentivirus | From our collaborators. See reference 13: Gills, J. et al. A patient-derived orthotopic xenograft model enabling human high-grade urothelial cell carcinoma of the bladder tumor implantation, growth, angiogenesis, and metastasis. Oncotarget. 9, 32718-32729, doi:10.18632/oncotarget.26024 (2018). | ||
McCoy’s medium | Life Technologies | 110862 | |
penicillin/streptomycin 100 mL (100 U/mL) | Fisher Scientific | 15140-122 | |
RPMI-1640 Medium | American Type Culture Collection | 110636 | |
Trypan Blue | Sigma | T6146 | |
Trypsin/EDTA | Life Technologies | 15400-054 | |
Name | Company | Catalog Number | Comments |
Gas | |||
100% Oxygen | Airgas Inc | OX USP200 | |
100% CO2 | Airgas Inc | CD USPE | |
Name | Company | Catalog Number | Comments |
Mice | |||
6-8 week old NOD/SCID Mice (male) | Jackson Lab | 001303 | |
6-8 week old NOD/SCID Mice (female) | Jackson Lab | 001303 | |
Name | Company | Catalog Number | Comments |
Immunohistochemistry | |||
Hematoxylin | Sigma | GHS232 | |
Ki-67 Rabbit Monoclonal Antibody | Thermo Scientific | RM-9106-S | |
Name | Company | Catalog Number | Comments |
Tools | |||
40 µm cell strainer | Fisher Scientific | 08-771-1 | |
100 µm cell strainer | Fisher Scientific | 08-771-19 | |
15 mL Conical Tube | Sarstedt | 11799 | |
50 mL Conical tube | Sarstedt | 15762 | |
150 mm Tissue Culture Dish | USA Scientific Inc | CC7682-3614 | |
96 Well plate | USA Scientific Inc | CC7682-7596 | |
Forceps | Symmetry Surgical Inc | 06-0011 | |
Surgical scissors | Symmetry Surgical Inc | 02-2011 | |
Name | Company | Catalog Number | Comments |
Equipment | |||
5% CO2 humidified incubator | Thermo Scientific | 3110 | |
Bioluminescent (BLI) Imaging Machine | Perkin Elmer | CLS136334 | |
BLI Imaging Machine Software | Perkin Elmer | CLS136334 | |
Centrifuge | Beckman | 366830 | |
Deconvoluting Microscope | Intelligent Imaging Innovations | Marianas | |
Deconvoluting Microscope Imaging Software | Intelligent Imaging Innovations | +1 (303) 607-9429 x1 | |
Digital caliper | Fowler Tools and Instruments | 54-115-330 | |
Dissecting microscope | Precision Instruments LLC | (504) 228-0076 | |
Electrosurgical generator | ValleyLab | FORCE1C20 | |
Isoflurane Induction Chamber | Perkin Elmer | 119038 | |
Microtome | American Optical Corporation | 829 | |
Pipet Aid | Fisher Healthcare | 13-681-15E | |
Serological pipet (10 mL) | Sarstedt | 86.1254.001 |