This protocol describes the process of the generation and characterization of mouse urothelial organoids harboring deletions in genes of interest. The methods include harvesting mouse urothelial cells, ex vivo transduction with adenovirus driving Cre expression with a CMV promoter, and in vitro as well as in vivo characterization.
Bladder cancer is an understudied area, particularly in genetically engineered mouse models (GEMMs). Inbred GEMMs with tissue-specific Cre and loxP sites have been the gold standards for conditional or inducible gene targeting. To provide faster and more efficient experimental models, an ex vivo organoid culture system is developed using adenovirus Cre and normal urothelial cells carrying multiple loxP alleles of the tumor suppressors Trp53, Pten, and Rb1. Normal urothelial cells are enzymatically disassociated from four bladders of triple floxed mice (Trp53f/f: Ptenf/f: Rb1f/f). The urothelial cells are transduced ex vivo with adenovirus-Cre driven by a CMV promoter (Ad5CMVCre). The transduced bladder organoids are cultured, propagated, and characterized in vitro and in vivo. PCR is used to confirm gene deletions in Trp53, Pten, and Rb1. Immunofluorescence (IF) staining of organoids demonstrates positive expression of urothelial lineage markers (CK5 and p63). The organoids are injected subcutaneously into host mice for tumor expansion and serial passages. The immunohistochemistry (IHC) of xenografts exhibits positive expression of CK7, CK5, and p63 and negative expression of CK8 and Uroplakin 3. In summary, adenovirus-mediated gene deletion from mouse urothelial cells engineered with loxP sites is an efficient method to rapidly test the tumorigenic potential of defined genetic alterations.
Bladder cancer is the fourth most common cancer in men and affects more than 80,000 people annually in the United States1. Platinum-based chemotherapy has been the standard of care for patients with advanced bladder cancer for more than three decades. The landscape of bladder cancer treatment has been revolutionized by the recent Food and Drug Administration (FDA) approval of immunotherapy (anti-PD-1 and anti-PD-L1 immune checkpoint inhibitors), erdafitinib (a fibroblast growth factor receptor inhibitor) and enfortumab vedotin (an antibody-drug conjugate)2,3,4. However, no clinically approved biomarkers are available for predicting the responses to chemotherapy or immunotherapy. There is a critical need to generate informative preclinical models that can improve the understanding of the mechanisms driving bladder cancer progression and develop predictive biomarkers for different treatment modalities.
A major obstacle in bladder translational research is the lack of preclinical models that recapitulate human bladder cancer pathogenesis and treatment responses5,6. Multiple preclinical models have been developed, including in vitro 2D models (cell lines or conditionally reprogrammed cells), in vitro 3D models (organoids, 3D printing), and in vivo models (xenograft, carcinogen-induced, genetically engineered models, and patient-derived xenograft)2,6. Genetically engineered mouse models (GEMMs) are useful for many applications in bladder cancer biology, including analyses of tumor phenotypes, mechanistic investigations of candidate genes and/or signaling pathways, and the preclinical evaluation of therapeutic responses6,7. GEMMs can utilize site-specific recombinases (Cre-loxP) to control genetic deletions in one or more tumor suppressor genes. The process of generating desired GEMMs with multiple gene deletions is time-consuming, laborious, and expensive5. The overall goal of this method is to develop a rapid and efficient method of ex vivo Cre delivery for establishing bladder triple knockout (TKO) models from normal mouse urothelial cells carrying triple floxed alleles (Trp53, Pten, and Rb1)8. The major advantage of the ex vivo method is the fast workflow (1-2 weeks instead of years of mouse breeding). This article describes the protocol for harvesting normal urothelial cells with floxed alleles, ex vivo adenovirus transduction, organoid cultures, and in vitro and in vivo characterization in immunocompetent C57 BL/6J mice. This method can be further used to generate clinically relevant bladder cancer organoids in immunocompetent mice harboring any combination of floxed alleles.
All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Roswell Park Comprehensive Cancer Center, Buffalo, NY (1395M, Biosafety 180501 and 180502).
NOTE: Perform Steps 1-3 on the same day.
1. Dissection of mouse bladder
2. Dissociation of urothelial cells
3. Adenoviral transduction and plating organoids
CAUTION: Please be cautious when processing adenovirus. Laboratory personnel should follow biosafety level 2 practices and disinfect adenovirus with 10% bleach.
4. Organoid culturing and passaging
5. Organoid cells implantation into C57BL/6J mouse subcutaneously
6. Tumor collection and in vivo propagation
The workflow of adenovirus-Cre mediated gene deletions in mouse urothelial cells is shown in Figure 1A. The accompanying video demonstrates how the urothelial cells are disassociated from the fundus of the bladder and how the triple floxed cells are transduced ex vivo with Ad5CMVCre. Figure 1A showed that minimal submucosa and muscle cells were disassociated after enzyme digestion. To confirm the efficiency of adenovirus-Cre delivery, disassociated urothelial cells from triple floxed mice with membrane-targeted tdTomato/membrane-localized enhanced green fluorescent protein (mT/mG) were transduced with adenovirus Cre. Green fluorescent protein (GFP) was detected in nearly 100% of cells, indicating high efficiency of adenovirus transduction (Figure 1B).
The stable TKO organoids were established following standard organoid protocols (in vitro with more than 5 passages; Figure 2A). H&E and immunofluorescence (IF) of in vitro organoid sections (Figure 2A) demonstrated a morphology of high-grade urothelial carcinoma with positive expression of urothelial lineage markers CK5 and p6311. PCR analysis revealed recombined alleles in the TKO organoids, whereas unrecombined floxed alleles were only detected in urothelial cells untreated with adenovirus (Figure 2B). A total of 2 x 106 primary ex vivo organoids were injected into C57BL/6J for initial subcutaneous (SQ) tumor formation in 8 weeks. Then, SQ tumor (passage 1) was collected and passaged in mice. In general, 2-3 weeks were needed to form a 2 cm SQ tumor (passages 2-5) (Figure 2C). The immunohistochemistry (IHC) of TKO in vivo xenograft displayed positive expression of CK7 (patchy), CK5, and p63 and negative expression of CK8 and Uroplakin 3 (Figure 2D). To detect contamination of mesenchyme cells, the TKO tumors were stained for vimentin. H&E stain showed the tumor on the left and the capsule on the right demarcated by the yellow line. Positive vimentin was detected only in the tumor capsule or stroma (Figure 2E).
Figure 1: Adenovirus-Cre mediated gene deletions in mouse urothelial cells with loxP sites. (A) The workflow of TKO organoid generation via ex vivo adenovirus Ad5CMVCre. A short-time 30 min enzyme digestion was sufficient to disassociate urothelial cells from the underlying submucosa and muscularis propria. The disassociated urothelial cells were transduced with Ad5CMVCre for TKO organoids and subsequently injected into C57BL/6J mice. (B) Triple floxed urothelial cells with mT/mG (constitutive transgene expression of red fluorescent protein that converts to green fluorescent protein following Cre recombination) were transduced with Ad5CMVCre. Nearly 100% of cells were GFP positive, indicating highly efficient transduction and Cre recombination. Please click here to view a larger version of this figure.
Figure 2: Characterization of TKO organoids via ex vivo adenovirus-Cre recombinase. (A) The disassociated Trp53f/f: Ptenf/f: Rb1f/f urothelial cells were transduced with adenovirus (Ad5CMVCre) to generate TKO organoids (bright field). The TKO organoids were embedded with speciment processing gel and stained for H&E and immunofluorescence. (B) Genomic DNA was extracted from triple floxed urothelial cells (lane 2) and TKO organoids (lane 4) and amplified by PCR to detect floxed alleles or Cre recombined alleles of the Trp53, Rb1, and Pten. PCR bands were stained with ethidium bromide. Lane 1 and lane 3 were negative and positive recombined controls. (C) A gross TKO SQ tumor (passage 4) was shown on day 17 after SQ injection. (D) Tumor tissue sections from the TKO SQ xenografts were stained with H&E, IF (CK5, CK8), or IHC (CK7, p63, Upk3). Abbreviations: CK5 = Cytokeratin 5; CK20 = Cytokeratin 20; CK8 = Cytokeratin 8; CK7 = Cytokeratin 7; Upk3 = Uroplakin III. (E) Tumor tissue sections from the TKO SQ xenografts were stained with H&E and IF (vimentin). Please click here to view a larger version of this figure.
GEMMs have been the gold standards for cancer modeling initiated from normal cells, allowing the consequences of potential oncogenic perturbations (oncogene activation and/or loss of tumor suppressors) to be tested rigorously. Here, a rapid and efficient protocol is provided to generate bladder cancer organoids via ex vivo gene editing of normal mouse urothelial cells carrying floxed alleles in genes of interest. H&E and IHC staining demonstrate that TKO organoids exhibit histology consistent with high-grade urothelial urothelium, with squamous differentiation and a basal-like subtype both as in vitro organoids and as in vivo xenografts. The TKO organoids can be used in vitro for studying the effects of Trp53, Rb1, and Pten loss, drug screening, and the molecular assessment of signaling pathways. Rapid tumor formation in C57BL/6J mice will enable in vivo studies designed to evaluate treatment responses to systemic therapies like chemotherapy or immunotherapy.
The urothelium is composed of a layer of basal cells (positive for CK5 and p63), 1-2 layers of intermediate cells (Uroplakins and p63 positive), and umbrella cells (Uroplakins, CK8 and CK20 positive)6. A critical step in the method is the limited time (30 min) of tissue disassociation instead of a prolonged (4 hours) enzyme digestion, which allows minimal contamination of non-urothelial submucosa cells (Figure 1A). A longer disassociation time may increase the percentage of stromal cells, such as endothelial cells and fibroblast cells. The ex vivo transduction step of adenovirus is highly effective. After ex vivo adenovirus Cre transduction, GFP was visualized in nearly 100% of urothelial cells with mT/mG reporter alleles (Figure 1B).
There are limitations to the ex vivo method. First, the disassociated cells are not pre-selected before adenovirus transduction. For instance, cells are not differentiated for urothelial cells vs. non-urothelial cells, or luminal cells vs. basal cells. Cell sorting with EpCAM and/or CD49f can be used to sort pure epithelial cells and/or stem cells before ex vivo transduction12,13. Second, the adenovirus driving Cre expression with CMV promoter used in this protocol targets a wide range of cell types after tissue disassociation (urothelial vs. non-urothelial, basal vs. luminal cells). This nonspecific targeting may lead to a selection bias causing overgrowth of cells with the most oncogenic potential. Cell type-specific adenovirus vectors have been used topically in both lung cancer and bladder cancer GEMMs14,15,16,17. Future studies using ex vivo cell type-specific adenovirus (adenovirus with a CK8 promoter or a CK5 promoter) may address cell-of-origin questions of whether the TKO organoids are derived from basal or luminal cells18. CK20 or Upk2 promoter-specific adenovirus (not available to our knowledge) can also be designed to transduce umbrella cells in the urothelium. However, future studies are needed to examine the efficiency and specificity of these promoter-specific adenoviruses for ex vivo bladder infection. There may be a discrepancy in adenovirus-Cre efficiency between in vivo and ex vivo transduction due to the host environment18. Third, the ex vivo method is limited by the lack of tumor environment, which may impact the in vivo tumorigenesis and histomorphology. Despite these limitations, a major advantage of the ex vivo method is the fast workflow (1-2 weeks instead of years of mouse breeding). The second advantage of the ex vivo approach is that adenovirus-Cre (Ad5CMVCre) is commercially available, straightforward, and nearly 100% efficient for viral transduction and Cre recombination (Figure 1B). In contrast, alternative ex vivo methods using CRISPR editing or lentiviral transduction require further clone selection due to less efficient genomic editing5,13,19,20.
In summary, the ex vivo adenovirus approach described is convenient, fast, and efficient in generating bladder cancer models utilizing any combination of genes of interest (floxed alleles) related to human bladder cancer. These models can be combined with cell type-specific adenovirus and cell sorting to address the impact of the cell of origin on bladder cancer phenotype and to evaluate treatment responses to immunotherapies in a setting of defined genetic mutations.
The authors have nothing to disclose.
This research work was supported in part by NIH Grants, K08CA252161(Q.L.), R01CA234162, and R01 CA207757 (D.W.G.), P30CA016056 (NCI Cancer Center Core Support Grant), the Roswell Park Alliance Foundation, and the Friends of Urology Foundation. We thank Marisa Blask and Mila Pakhomova for proofreading the manuscript.
100 μm sterile cell strainer | Corning | 431752 | |
1 mL syringe | BD | 309659 | |
25G 1.5 inches needle | EXELINT International | 26406 | |
Adenovirus (Ad5CMVCre High Titer, 1E11 pfu/ml) | UI Viral Vector Core | VVC-U of Iowa-5-HT | |
C57 BL/6J | Jackson Lab | 000664 | |
Charcoal-stripped FBS | Gibco | A3382101 | |
Collagenase/hyaluronidase | Stemcell Technologies | 07912 | |
Dispase | Stemcell Technologies | 07913 | |
DPBS, 1x | Corning | 21-031-CV | |
L-glutamine substitute (GlutaMAX) | Gibco | 35-050-061 | |
Mammary Epithelial Cell Growth medium | Lonza | CC-3150 | |
Matrix extracts from Engelbreth–Holm–Swarm mouse sarcomas (Matrigel) | Corning | CB-40234 | |
Monoclonal mouse anti-CK20 | DAKO | M7019 | IF 1:100 |
Monoclonal mouse anti-CK7 | Santa Cruz Biotechnology | SC-23876 | IHC 1:50 |
Monoclonal mouse anti-p63 | Abcam | ab735 | IHC 1:100, IF 1:50 |
Monoclonal mouse anti-Upk3 | Fitzgerald | 10R-U103A | IHC 1:50 |
Monoclonal mouse anti-Vimentin | Santa Cruz Biotechnology | SC-6260 | IF 1:100 |
Monoclonal rat anti-CK8 | Developmental Studies Hybridoma Bank | TROMA-I-s | IF 1:100 |
HERAcell vios 160i CO2 incubator | Thermo Fisher | 51033557 | |
Polyclonal chicken anti-CK5 | Biolegend | 905901 | IF 1:500 |
Primocin | InvivoGen | ant-pm-1 | |
Recombinant enzyme of trypsin substitute (TrypLE Express Enzyme) | Thermo Fisher | 12605036 | |
Signature benchtop shaking incubator Model 1575 | VWR | 35962-091 | |
Specimen Processing Gel (HistoGel) | Thermo Fisher | HG-4000-012 | |
Surgical blade size 10 | Integra Miltex | 4-110 | |
Sorvall T1 centrifuge | Thermo Fisher | 75002383 | |
Y-27632 | Selleckchem | S1049 |