To allow highly sensitive detection of the disseminating human colorectal cancer (CRC) cells colonizing tissues, we herein show a protocol for efficient transduction of green fluorescent protein (GFP) lentiviral particles into PDX-derived CRC organoid cells prior to their injection into recipient mice, with stereo-fluorescence microscopic observation.
Despite current advances in human colorectal cancer (CRC) treatment, few radical therapies are effective for the late stages of CRC. To overcome this clinical challenge, tumor xenograft mouse models using long-established human carcinoma cell lines and many transgenic mouse models with tumors have been developed as preclinical models. They partially mimic the features of human carcinomas, but often fail to recapitulate the key aspects of human malignancies including invasion and metastasis. Thus, alternative models that better represent the malignant progression in human CRC have long been awaited.
We herein show generation of patient-derived tumor xenografts (PDXs) by subcutaneous implantation of small CRC fragments surgically dissected from a patient. The colon PDXs develop and histopathologically resemble the CRC in the patient. However, few spontaneous micrometastases are detectable in conventional cross-sections of affected distant organs in the PDX model. To facilitate the detection of metastatic dissemination into distant organs, we extracted the tumor organoid cells from the colon PDXs in culture and infected them with GFP lentivirus prior to injection into highly immunodeficient NOD/Shi-scid IL2Rγnull (NOG) mice. Orthotopically injected PDX-derived CRC organoid cells consistently form primary tumors positive for GFP in recipient mice. Moreover, spontaneously developing micrometastatic colonies expressing GFP are notably detected in the lungs of these mice by fluorescence microscopy. Moreover, intrasplenic injection of CRC organoids frequently produces hepatic colonization. Taken together, these findings indicate GFP-labelled PDX-derived CRC organoid cells to be visually detectable during a multistep process termed the invasion-metastasis cascade. The described protocols include the establishment of PDXs of human CRC and 3D culture of the corresponding CRC organoid cells transduced by GFP lentiviral particles.
Colorectal cancer (CRC) is the second leading cause of cancer deaths worldwide1. The insufficient response to conventional therapies of patients with advanced stage disease indicates the ineffectiveness of attempts to radically cure CRCs. To develop more effective therapeutic approaches, various preclinical mouse models of cancer that mimic the characteristics of CRCs have been established. Various CRC cell lines have been widely used to generate tumor xenografts due to their convenience and ease of manipulation. Long-term culture of cancer cell lines, however, often causes selection of unique cell populations that are quite proliferative under a particular culture condition, thereby resulting in unreliable outcomes and crucial limitations in preclinical drug development.
Without being cultured in vitro, patient-derived tumor xenografts (PDXs) have also been generated by implantation into animal models of human CRC tissues surgically dissected from patients2,3,4. PDXs are widely recognized as recapitulating the major histopathological features and genetic alterations originally present in tumors of the patients from which they were derived. Moreover, patient-derived tumor organoids composed of tumor cell clusters were established by culture under 3D conditions that closely mimicked the biological properties of the original tumors5,6. These tumor organoids were also applied to high-throughput drug screening, thereby allowing personalized therapies to be designed5. However, markedly heterogeneous CRC populations are assumed to be present within a tumor mass. Particular CRC populations might selectively proliferate and expand during the in vivo and in vitro series of passages of PDX and tumor organoids, respectively. This may also allow the overall gene expression profiles and epi/genetic status of the affected CRCs to change, thereby resulting in minimal resemblance to the parental CRC.
Patient-derived CRC organoids and those extracted from noncancerous human colon engineered to harbor combinations of oncogenic mutations have also been employed to investigate the hallmarks of human tumor cells exemplified by tumor invasion and metastasis6,7,8,9. However, the very low incidence of spontaneous metastasis arising from orthotopic implantation of patient-derived CRC into immunodeficient mice, has made it difficult to study the multistep process of the invasion-metastasis cascade that includes local invasion, intravasation, transport in the bloodstream, extravasation and colonization of distant organs4,10. Micrometastasis as represented by tumor cell deposition of ≤2 mm, formed by patient-derived CRC organoids, has often been overlooked on histopathological analysis of sections from affected distant organs in experimental mouse models. Visualization of spontaneous micrometastases has also been minimal in vivo due to the difficulty of efficiently introducing fluorescent markers into tumor organoid cells prior to their injection into recipient mice. In this study, we developed a protocol to efficiently transduce GFP lentivirus into PDX-derived CRC organoid cells in 3D culture prior to injection into recipient mice and to allow highly sensitive detection, employing stereo-fluorescence microscopy, of their colonization of different organs to form micrometastases.
The patient provided written informed consent and the project was approved by the Research Ethics Committee of the Juntendo University Faculty of Medicine. The mouse experiments were also approved by the Animal Research Ethics Committee of the Juntendo Faculty of Medicine.
1. Establishment of CRC PDXs in Immunodeficient Mice
Experimental procedures for establishing CRC PDXs (step 1) are outlined in Figure 1A.
2. Dissection of CRC PDXs from Mice
Experimental procedures for dissection of CRC PDXs (step 2) are outlined in Figure 1B.
3. Extraction of PDXs into the Cell Suspension for the CRC Organoid Culture
Experimental procedures for dissociation of CRC PDXs (step 3) are outlined in Figure 1B.
4. Generation of the CRC Organoids Cultured on Artificial Extracellular Matrix
Experimental procedures for the CRC organoid culture of the colon PDXs (step 4) are outlined in Figure 1B.
5. Generation and Enrichment of GFP Lentiviral Particles
Experimental procedures for generation and enrichment of GFP lentiviral particles (step 5) are outlined in Figure 1C.
6. Labelling of CRC Organoid Cells with GFP Lentiviral Particles Cultured on Artificial Extracellular Matrix
Experimental procedures for labelling the CRC organoid cells with GFP lentivirus (step 6) are outlined in Figure 1C.
7. Generation of Metastases by GFP-labeled CRC Organoids in Recipient Mice
Experimental procedures for generation of metastases using the GFP-labelled CRC organoids (step 7) are outlined in Figure 1C.
A primary colorectal adenocarcinoma diagnosed as moderately differentiated had been surgically resected from a 76 year-old female with TNM classification, stage IIIa, followed by post-operative chemotherapy. The primary CRC cells immunohistochemically stained positive for carcinoembryonic antigen, Ki-67, pan-cytokeratin and E-cadherin. Pieces of the resected tumor were also subcutaneously implanted into NOG mice to generate the colon PDX model. CRC organoid cells were then extracted for tissue culture from the colon PDX which had developed subcutaneously in these mice. The CRC organoids were capable of consistently forming multicellular clusters during a series of passages on artificial extracellular matrix. As tumor organoids histologically and genetically reflect the nature of primary tumors from patients5,6, we attempted to develop a mouse model that would allow in vivo detection of CRC organoid cells with high sensitivity during metastatic dissemination. To achieve this goal, CRC organoids were infected with GFP lentivirus, thereby resulting in nearly all organoids showing very bright GFP (Figure 2A). These organoids were then injected orthotopically and intrasplenically into secondary NOG mice prior to detection of GFP-positive cells in distant organs under a fluorescence microscope. We observed formation of GFP-positive primary tumors at the orthotopic site in all four mice examined (Figure 2B, left). Moreover, microscopic metastases were found in the lungs of three out of four mice examined at 2.5 months after orthotopic injections (Figure 2B, middle). Furthermore, liver metastases were observed in response to intrasplenic injection into all four mice examined at 1 month after the injections (Figure 2B, right).
Taken together, these findings indicate that high-resolution GFP fluorescence on PDX-derived CRC organoid cells allows their detection in both spontaneous micrometastases and those developed experimentally, in distant organs, when introduced into recipient mice. This high-sensitivity detection of the disseminating PDX-derived CRC organoids in vivo also offers a versatile model not only for studying the biology of tumor metastasis, but also for developing targeted therapies in the preclinical setting.
Figure 1: Schematic representation of generation of metastases by the PDX-derived CRC organoids labeled with GFP lentivirus in NOG mice. (A) Implantation of small pieces of the CRC tissue subcutaneously into NOG mice (step 1). The CRC tissue surgically dissected from the patient was cut into pieces and implanted subcutaneously into NOG mice. s.c.: subcutaneous implantation. (B) Generation of CRC organoids dissociated from PDXs (step 2–4). The developed CRC xenografts were minced (step 2) and transferred into a 15 mL tube containing the culture medium including collagenase (step 3). After incubation with slow agitation, the CRC cell suspension was filtered (step 3). Then, the organoid cell suspension in the CRC organoid medium was seeded onto an artificial extracellular matrix-coated plate and incubated overnight in a CO2 incubator (step 4). The CRC organoid cells attached to the artificial extracellular matrix were also coated with additional artificial extracellular matrix and incubated in a CO2 incubator (step 4). s.c.: subcutaneous implantation. (C) Generation of CRC organoids transduced by GFP lentiviral particles prior to employing injection into recipient mice (step 5–7). The GFP lentiviral particles were generated at a high titer (step 5). The PDX-derived CRC organoids grown on artificial extracellular matrix were directly harvested with a cell scraper and transferred into a microtube (step 6). After centrifugation, the cell pellet was re-suspended in PBS. The cell suspension was centrifuged and the cell pellet was dissociated. Then, the CRC organoid cell suspension was incubated with GFP virus stock in the CRC organoid culture medium on the artificial extracellular matrix-coated plate overnight in a CO2 incubator (step 6). The CRC organoid cells attached to the artificial extracellular matrix were coated with additional artificial extracellular matrix and incubated in a CO2 incubator to solidify the artificial extracellular matrix coating (step 6). The CRC organoids were then cultured for periods of 7–10 days to expand cell growth (step 6). To develop a spontaneous metastasis model, the dissociated 5 x 105 CRC organoid cells labelled with GFP suspended in 50 µL of PBS with 50% artificial extracellular matrix were injected orthotopically into NOG mice (step 7). To generate an experimental metastasis model, 4 x 104 CRC organoid cells labelled with GFP in 50 µL of PBS were injected intrasplenically into NOG mice (step 7). o.t.: orthotopic injection, i.s.: intrasplenic injection. Please click here to view a larger version of this figure.
Figure 2: High-resolution of GFP visualization detected in cultured GFP-labeled CRC organoids, primary tumors and micrometastases. (A) Nearly 100% of cultured CRC organoids are GFP positive. The images were captured using a stereo-fluorescence microscope. Scale bar = 250 µm. (B) GFP-positivity in the primary tumor and micrometastases in the lungs and liver. The dissociated GFP-expressing CRC organoid cell suspension was injected into the rectal submucosa (o.t. inj.) of NOG mice. The majority of tumor cells are shown to be positive for GFP in the primary tumor. GFP-positive micrometastases colonizing the lungs (indicated by an arrow) are also shown. Moreover, GFP-positive micrometastases are detected in the liver (indicated by an arrow), when the cell suspension was intrasplenically injected (i.s. inj.) into mice. Scale bar = 500 µm. Please click here to view a larger version of this figure.
Although the CRC PDX model has been widely employed to study primary tumor growth, whether this model is also applicable to investigating tumor metastasis has not yet been fully elucidated. Spontaneous metastases were also barely detectable in the liver and lungs of various reported colon PDX models4,10. To detect micrometastases with high sensitivity, we developed a protocol for transducing GFP lentiviral particles into PDX-derived CRC organoids prior to their orthotopic and intravenous injections into recipient mice. Of note, we were able to show that this method allows highly sensitive detection of CRC organoid-derived micrometastases, forming both spontaneously and experimentally, in distant organs.
Critical steps within the protocol include maintaining intact spheroid formation without induction of anoikis on the artificial extracellular matrix by gently pipetting the CRC organoids during the passaging series. The preparation of high titer lentivirus particles concentrated by ultracentrifugation is also important for achieving the nearly 100% GFP positivity of CRC organoids. Moreover, it is recommended that the GFP-labeled CRC organoids be injected within 10–14 days into recipient mice. Excessively long periods for expansion of the CRC organoids after infection may favor selection of weakly GFP-expressing organoids.
The injection of CRC organoids into the rectal submucosa of NOG mice showed their metastatic dissemination into the lungs, though not into the liver, presumably through the inferior vena cava. To generate liver metastases spontaneously, injection of CRC organoids into the colon with laparotomy would be required13.
Note that the incidence of B cell lymphomas is on occasion increased in NOG immunodeficient mice implanted with tissues including lymphocytes, infected with Epstein-Barr oncovirus, from human patients14. It is therefore reasonable to recommend that whether the emerging tumors originate from the implanted human CRC be determined by performing immunohistochemistry using human CRC cell markers, such as carcinoembryonic antigen and cytokeratin 20.
As we investigated CRC organoids derived from only one patient, more samples from larger numbers of patients would need to be examined to generalize our findings in the future. If the implanted PDX-derived CRC organoids show minimal growth at the orthotopic site and rarely form metastases at distant sites, investigators are encouraged to consider extracting CRC organoids from different patient-derived PDXs.
The highly sensitive detection of GFP-labeled micrometastasis achieved with this method encourages us not only to further investigate several unresolved biological issues regarding transition of micrometastasis to macrometastasis, formation of the stromal niche for metastatic colonization and acquisition of drug resistance, but also to evaluate the efficacies of novel anti-cancer drugs by employing PDX-bearing animals as preclinical models.
The GFP-based FACS-sorting of living CRC organoids extracted from primary and distant sites also has the potential to allow single cell-based examinations of the gene expressions, as well as the epigenetic and metabolomic profiles, of a tumor. Such findings may well lead to elucidation of the molecular mechanisms underlying metastatic dissemination of the CRC organoids.
The authors have nothing to disclose.
This work was supported by the Juntendo University Young Investigator Award (2013, 2014 and 2015) to Y.O., the Joint Project Award (2013 and 2014) to K. M., and Grants in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (16K15625 to Y. K. and 16K15598 to M.G.). We are especially thankful to all members of the Dept. of Coloproctological Surgery and Molecular Pathology for useful discussions and technical support. We also thank Dr. Hiroyuki Konno (Hamamatsu University Schoolof Medicine) and Dr. Hideki Kitajima (International University of Health and Welfare) for generous technical guidance in the surgical procedures for orthotopic implantation into mice and Dr. Yoshitaka Hippo (Chiba Cancer Center) for technical advice on performing the tumor organoid culture.
NOD/Shi-scid IL2Rγ null (NOG) mice | The Central Institute for Experimental Animals,Kanagawa, Japan | Breed 6-week-old male mice under germ-free and specific pathogen-free conditions | |
wound clips 2×10mm |
Natsume manufacturing, Japan | #C-21-S | Autoclave before use |
Hamilton syringe needle size:22 gauge |
Tokyo Science, Japan | Disinfect with 70% alcohol and sterile PBS. | |
6-well plate | BMBio | #92006 | |
12-well plate | BMBio | #92412 | |
15ml conical tube | Sumitono Bakelite | MS-57150 | |
50ml conical tube | Sumitomo Bakelite | MS-57500 | |
microtube | Eppendorf | #0030120086 | Autoclave before use |
Hemocytometer | Erma | #03-202-1 | |
40μm cell strainer | Corning | #352340 | |
Matrigel basement membrane matrix | Corning | #354234 | Store aliqupts at -20°C. Place on ice until use |
Collagenase type 1 | Sigma | #C1030 | 150 mg/ml collagenase type1 in 1×PBS. Store aliqupts at -20°C for up to 1 year |
Accutase | Innovate Cell Technologies | #5V2623A | Store at 4°C. |
DMEM/F-12 with GlutaMAX™ | Gibco | #10565018 | Store at 4°C. Warm at 37°C before use |
Cell banker 1plus | ZENOAQ | #628 | Store at 4°C. Use within 1 month |
Penicillin | Gibco | #15140122 | Store at 4°C. Use within 1 month |
Streptomycin | Gibco | #15140122 | Store at 4°C. Use within 1 month |
hEGF | PEPROTECH | #AF-100-15 | Store at -20°C. Add to medium on same day as use |
Y27632, a ROCK inhibitor | Wako | #253-00591 | Store at -20°C. Add to medium on same day as use |
Culture medium | Gibco | DMEM/F-12 with GlutaMAX™ supplement supplemented with 5% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin. Store at 4°C. Use within 1 month. |
|
CRC organoid culture medium with 1% or 5% FCS |
DMEM/F-12 with GlutaMAX™ supplement (Gibco #10565018) supplemented with 1% or 5% FCS, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 ng/ml hEGF and 10 µM Y27632, a ROCK inhibitor. Store at 4°C. Use within 1 month. |
||
the FuGENE 6 transfection regent | Roche | 11814 443001 | |
Minisart 0.45 µm filter | Sartorius stedim | 17598-K | |
5 ml polypropylene centrifuge tubes | Beckman Coulter | 326819 | |
PRRL-GFP vector | Gift from Dr. Robert A. Weinberg | ||
pCMV-VSV-G | Gift from Dr. Robert A. Weinberg | ||
pCMV-dR8.2 dvpr | Gift from Dr. Robert A. Weinberg | ||
the SW55Ti swinging bucket rotor | Beckman Coulter | ||
a Zeiss Axioplan 2 stereo-fluorescence microscope | Zeiss |