An in vitro model system was developed to capture tissue architectural changes during lung squamous carcinoma (LUSC) progression in a 3-dimensional (3D) co-culture with cancer-associated fibroblasts (CAFs). This organoid system provides a unique platform to investigate the roles of diverse tumor cell-intrinsic and extrinsic changes that modulate the tumor phenotype.
Tumor-stroma interactions play a critical role in the development of lung squamous carcinoma (LUSC). However, understanding how these dynamic interactions contribute to tissue architectural changes observed during tumorigenesis remains challenging due to the lack of appropriate models. In this protocol, we describe the generation of a 3D coculture model using a LUSC primary cell culture known as TUM622. TUM622 cells were established from a LUSC patient-derived xenograft (PDX) and have the unique property to form acinar-like structures when seeded in a basement membrane matrix. We demonstrate that TUM622 acini in 3D coculture recapitulate key features of tissue architecture during LUSC progression as well as the dynamic interactions between LUSC cells and components of the tumor microenvironment (TME), including the extracellular matrix (ECM) and cancer-associated fibroblasts (CAFs). We further adapt our principal 3D culturing protocol to demonstrate how this system could be utilized for various downstream analyses. Overall, this organoid model creates a biologically rich and adaptable platform that enables one to gain insight into the cell-intrinsic and extrinsic mechanisms that promote the disruption of epithelial architectures during carcinoma progression and will aid the search for new therapeutic targets and diagnostic markers.
Lung cancer is the leading cause of cancer-related mortality worldwide. Lung squamous cell carcinoma (LUSC), which is the second most common type of non-small-cell lung cancer (NSCLC) and accounts for approximately 30% of all lung cancer, is often diagnosed at advanced stages and has a poor prognosis1. Treatment options for LUSC patients are a major unmet need that can be improved by a better understanding of the underlying cellular and molecular mechanisms that drive LUSC tumorigenesis.
As with most human cancers, the pathogenesis of LUSC is characterized by the disruption of the intact, well-ordered epithelial tissue architecture2. During this process, proper apical-basal cell polarity, cell-cell and cell-matrix contacts are lost, permitting uncontrolled growth and invasive behavior of the tumor cells. It is now widely appreciated that the malignant features of cancer cells cannot be manifested without an important interplay between cancer cells and their local tumor microenvironment (TME)3. Key components in the TME including extracellular matrix (ECM), cancer-associated fibroblasts (CAFs) as well as endothelial cells and infiltrating immune cells actively shape the TME and drives tumorigenesis4. Nevertheless, our current understanding of how the tumor cells and these key components in the TME interact to drive tissue architectural changes during LUSC progression is very limited.
Three-dimensional (3D) culture is an important tool to study the biological activities of cell-intrinsic and extrinsic changes in regulating tissue architectural changes in both normal and diseased tissues5. 3D cultures provide the appropriate structural and functional context that is usually lacking in traditional two-dimensional (2D) cultures. The added dimensions of such systems more closely mimic tissue in vivo in many aspects of cell physiology and cellular behaviors, including proliferation, differentiation, migration, protein expression and response to drug treatment. In recent years, efforts from various labs have led to the development of in vitro 3D models for both the normal lung as well as NSCLC6,7,8. However, a model for lung squamous carcinoma that can recapitulate both the dynamic tissue architectural changes during tumorigenesis as well as incorporate key stromal components was unavailable.
Here, we describe the methods for establishing a novel 3-dimensional (3D) coculture system using primary PDX-derived LUSC cells (termed TUM622) and CAFs9,10. Both TUM622 and CAFs are derived from NSCLC patient with poorly differentiated tumors10. When embedded as single cells in ECM, a rare subpopulation of TUM622 cells have the capacity to form organoids with acinar-like structures that display proper apical-basal cell polarity. These acinar-like structures are hyperplastic, display heterogeneous expression of stem-like and differentiation markers similar to the original tumor while remaining non-invasive, and thus mimic the earliest stage of LUSC development. Importantly, we showed that the tissue architecture of the acinar-like structures could be altered by inhibition of cell-intrinsic signaling pathways with small molecule inhibitors or addition of key components in the ECM such as CAFs, the latter of which enhances acini formation and further provokes the acini to become invasive when in close proximity. Together, these data suggest that this 3D co-culture system of LUSC organoids provides a valuable platform for the investigation of the dynamic reciprocity between LUSC cells and the TME and could be adapted for monitoring the response of LUSC cells to drug treatment11.
1. Passaging and Culturing TUM622 Cells and CAFs in 2D Cultures
2. Plating TUM622 Cells in the Extracellular Matrix for 3D Culturing
3. 3D Coculturing of TUM622 Cells and CAFs in the Extracellular Matrix
4. Harvesting TUM622 Acini for RNA/Protein Extraction and Fluorescence-activated Cell Sorting (FACS)
5. Immunofluorescence of TUM622 Acini
6. Preparing 3D Culture Samples for Immunohistochemistry
7. 3D Cytotoxicity Assay for Compound Screening (Example for One 96-well Plate)
TUM622 and CAFs in 2D culture
Figure 1 presents the typical morphology of TUM622 cells and CAFs in 2D culture. TUM622 cells are rounded with large nuclei while CAFs are flat and elongated. TUM622 cells can reach 80%-90% confluency in culture. Further proliferation leads to more, but smaller cells aggregated in colonies that do not come into direct contact. In contrast, CAFs prefer to grow at higher cell density and will keep proliferating at full confluency if sufficient nutrients are provided.
Growth and morphology of TUM622 acini in 3D ECM
Figure 2 presents a time-course experiment of TUM622 cells seeded in 3D culture. Data show that single TUM622 cells are capable of forming organoids with acinar-like morphologies when embedded. Between days 5 and 7, a lumen becomes apparent in the acinar-like structures and remains hollow thereafter (Figure 2A). Each acinus, composed of a monolayer of cells surrounding the hollow lumen, displays proper apical-basal polarity similar to that of lung epithelium in vivo (Figure 2B). These acinar-like structures are hyperplastic and continue to grow as long as sufficient nutrients are provided. The culture can be maintained for up to 24 days before the ECM completely disintegrates (Figure 2C). Through limiting dilution assay (LDA) (data not shown), it is estimated that only a rare subpopulation of TUM622 cells (<0.02%) have the capacity to form acinar-like structures9.
Growth and morphology of TUM622-CAFs coculture
Figure 3 depicts the setup and representative results of TUM622-CAF cocultures. CAFs could be integrated into the coculture by either overlaying on top of the matrix or co-embedded with the TUM622 cells. Regardless of the setup, the presence of CAFs greatly enhanced the number and size of the spheroids formed (Figure 3B). Interestingly, when TUM622 acini come into close proximity with CAFs, they induce the acini to become invasive and migrate towards the CAFs, forming "tear-drop" like structures (Figure 3C). Note that TUM622 acini in monoculture do not display invasive behavior and only form "tear-drop" like structures when close to CAFs.
Representative immunofluorescent and immunohistochemistry staining results
Figure 4 shows representative results from immunofluorescence and immunohistochemistry staining of TUM622 acini after 10 days of culturing. Confocal images were taken at the equatorial plane of immunofluorescently stained TUM622 acini (Figure 4A). In contrast, each section from the immunohistochemistry sample may capture acini at different planes (Figure 4B). Both results showed heterogeneous expression of stem-like and differentiated cells within each acinus.
TUM622 3D cytotoxicity assay using a Wnt pathway inhibitor
Figure 5 shows the dose-response of TUM622 acini treated with XAV939, a tankyrase inhibitor (Figure 5A,B). XAV939 was added to the culture 1 day after plating and refreshed every 2 days for a total of 10 days. At the end of the experiment, the number of acini was quantified by an imager. Brightfield images at higher magnification were also acquired to capture the morphology of spheroids in control versus XAV939-treated wells (Figure 5C). Overall, XAV939 displays dose-dependent inhibition on acini formation and alters the tissue architecture of the spheroids formed. These results suggest that activation of the canonical Wnt pathway is required during TUM622 acinar morphogenesis.
Figure 1: TUM622 and CAFs in 2D culture. Representative bright-field images of TUM622 cells and CAFs cultured in 2D. Scale bar = 100 µm. This figure has been modified from Chen et al.9 and used with permission. Please click here to view a larger version of this figure.
Figure 2: Growth and morphology of TUM622 acini in 3D ECM. (A) Time course images of TUM622 cells cultured in basement membrane matrix over a 10-day period. Scale bar = 100 µm. (B) Immunofluorescence of TUM622 acini stained with apical-basal cell polarity markers, Golgi-enzyme (GM-130, green, apical) and Integrin alpha 6 (CD49f, basal, red). (C) Quantification of acini number (right y-axis, red) and the average size of acini (left y-axis, blue) plated in triplicate in a 24-well plate over 24 days in culture. Error bars represent SD. This figure has been modified from Chen et al.9 and used with permission. Please click here to view a larger version of this figure.
Figure 3: Growth and morphology of TUM622-CAFs coculture. (A) Schematic drawing of the setup of TUM622-CAFs coculture. CAFs are overlaid or co-embedded with TUM62 cells in ECM. After 6-12 days in coculture, TUM622 cells are able to form more and larger acini compared with mono-culture and invade the ECM when in close proximity and direct contact with CAFs. Note that the invasive phenotype could only be observed in the co-culture. (B) Brightfield image of TUM622 3D cultures in the presence or absence of overlaid CAFs after 8 days. Scale bars = 200 µm. (C) Brightfield images showing tear-drop shaped acini forming in cocultures regardless of CAFs are overlaid or co-embedded in the ECM. Scale bars = 200 µm. This figure has been modified from Chen et al.9 and used with permission. Please click here to view a larger version of this figure.
Figure 4: Representative immunofluorescent and immunohistochemistry staining results. (A) Antibody staining of acini with markers of stem/progenitor cells (CXCR4 and SOX2), mesenchyme (Vimentin), epithelial differentiation (Involucrin), apoptosis (Cleaved-Caspase-3) and proliferation (Ki67) in green, DAPI in blue, E-cadherin and Phalloidin in red. Scale bar = 50 µm. (B) Immunohistochemistry on FFPE sections of TUM622 acini. Scale bar = 100 µm (top) and 50 µm (bottom). This figure has been modified from Chen et al.9 and used with permission. Please click here to view a larger version of this figure.
Figure 5: TUM622 3D cytotoxicity assay using a Wnt pathway inhibitor. (A) Quantification of spheroid numbers in a 96-well plate where TUM622 cells were treated with dimethyl sulfoxide (Control) or XAV939. Each condition is assayed in triplicates. Error bars represent SD. (B) Whole well images from a 24-well plate taken with an imager showing the inhibitory effects of XAV939 on acini formation. (C) Representative brightfield images from the control vs. treated wells demonstrating the morphological changes caused by XAV939 treatment. Scale bars = 100 µm. This figure has been modified from Chen et al.9 and used with permission. Please click here to view a larger version of this figure.
Tumors are heterogeneous tissues composed of cancer cells coexisting side-by-side with stromal cells such as cancer-associated fibroblasts, endothelial cells and immune cells within the ECM. Together, these diverse components cross-talk and influence the tumor microenvironment, playing an active role in driving tumorigenesis, a process that involves progressive changes in tumor architecture. Ideally, an in vitro model of tumor development should be able to capture the dynamic tissue architectural changes observed in human tumors in vivo, the complex interplay of diverse cell types within the tumor microenvironment and at the same time permit experimental manipulation on both the tumor cells and components in the TME. Although much progress has been made in 3D cancer models in recent years, such models have not been readily forthcoming for LUSC. Most models reported to date only incorporate a few aspects of these important features. Here we report the methods for a 3D coculture system of LUSC that simultaneously captures key tissue architectural changes observed during LUSC development as well as dynamic interactions between tumor cells and major components of the TME, including the ECM and CAFs.
The ability of this system to more accurately model tissue architectural changes is based on the unique property of TUM622 cells in forming organoids with acinar-like morphologies when embedded in 3D EC. Formed from a self-renewing single cell, each acinus is composed of a monolayer of cells surrounding a hollow lumen. This monolayer of cells exhibits apical-basal cell polarity and remains non-invasive, resembling the tissue architecture of the lung epithelium. While TUM622 as a 3D mono-culture display hyperplastic growth, the addition of CAFs further enhances acinar morphogenesis and induces more and larger acini to form. Importantly, CAFs invoke dynamic tissue architectural changes in TUM622 cells when the two cell types come into close proximity, allowing the TUM622 cells to lose their apical-basal polarity and invade the matrix toward the CAFs. These phenotypic changes recapitulate both early hyperplasia as well as late invasive stages of LUSC.
Unlike many tumor spheroid models where each spheroid is formed by aggregation of many cells, each TUM622 acinus is derived from a single cell9. By in vitro LDA, it is estimated that only a minor subpopulation (≤0.02%) of TUM622 cells have such capacity9. Although rare, these cells could self-renew as evidenced by their capacity to undergo serial passaging in 3D as well as differentiate into a heterogeneous population of cells similar to that of the original tumor. Due to this unique feature of TUM622 cells, it is critical to ensure even distribution of single TUM622 cells within the ECM at the time of plating for successful culturing and downstream analysis. To achieve this goal, several key points need to be followed carefully in the protocol, including the determination of appropriate seeding density, keeping all tools and reagents cool during the mixing of cells and matrix to prevent premature solidification, avoiding the introduction of bubbles during the mixing process and allowing sufficient time for matrix to fully solidify before adding culture medium. Together, these precautions will help to achieve a more uniform matrix substrate and culture condition for all embedded cells.
Once successfully established, this culture can be used for a variety of downstream analyses to dissect the cell and biochemical process that regulate tumorigenesis. The number and size of acini formed in each well can be monitored over time with bright field imagers and used as a readout for the proliferative and self-renewal capacity of TUM622 cells. More detailed dynamics in the morphogenesis of each acinus could be observed with live-imaging on a confocal microscope, with or without various labeling dyes. The conditioned medium can be collected at multiple time points during the culture period for analyzing soluble factors that may mediate cell-cell or cell-matrix cross-talks. TUM622 cells extracted directly from the ECM using protocol 4 are suitable for RNA and protein extraction for gene expression analysis, flow cytometry quantification or FACS sorting based on cell surface markers. Alternatively, the cultures could be fixed for in situ immunofluorescence or immunohistochemistry studies to understand the spatial-temporal distributions of various markers. Although similar, immunohistochemistry complements immunofluorescence methods in that it allows the sampling of entire acini that may not be possible due to the limiting imaging-depth of the confocal objectives. For both of these methods, the time and temperature at which fixation and permeabilization are performed are critical, especially given that TUM622 cells are embedded in a dense matrix (>90% basement membrane matrix) in contrast to many other 3D cultures where matrix density is much lower. Therefore, attention to standardized and consistent fixation and processing is necessary to obtain replicative results.
Using this system as a platform, one can then investigate how cell-intrinsic changes in the tumor cells, as well as cell-extrinsic changes in the tumor microenvironment, influence epithelial architecture and model early events involved in carcinoma formation. For example, the roles of oncogenes or tumor suppressor genes in regulating tumor tissue architecture could be studied by gain- or loss-of-function experiment targeting the gene of interest in the tumor cells. Indeed, we demonstrated that over-expression of SOX2, which is commonly observed in LUSC, alters the phenotype of TUM622 cells as evidenced by a loss of hyperplasia in 3D and progression towards dysplastic growth9. On the other hand, one could compare normal versus cancer-associated fibroblasts in coculture settings, determine how matrix components or its stiffness impact acini growth/morphology/invasion, and if blocking certain cytokines could interfere with cell-cell communication and in turn affect tissue architecture and tumor progression. Importantly, all these assays could be performed in the presence or absence of certain therapeutic agents and be used as a tool to determine the drug response of LUSC cells with a multidimensional readout11. It is also important to note that this system is limited in regards to the pathways it could be used to interrogate, as only some but not all major signaling pathways regulate the growth and morphology of TUM622 organoids in culture (i.e., inhibition of Wnt but not Notch signaling affects acinar mophogenesis of TUM622 cells)9.
In summary, we demonstrate that this organoid system provides a unique platform for generating new insights into the dynamic interplay between LUSC cells and the tumor microenvironment during tumor progression. We anticipate that our model system will be a valuable platform for drug discovery and development. In this respect, screening novel anti-cancer therapeutics in a native tumor tissue context should aid in the selection and development of more effective therapeutics targeting LUSC.
The authors have nothing to disclose.
We thank Magali Guffroy, John Kreeger, and Stephani Bisulco of the Pfizer-Oncology Histopathology and Biomarker group for pathology/histology support and Michael Arensman for critical review of the manuscript. We also thank the Pfizer Postdoctoral Program and the Oncology R&D group, specifically Robert Abraham, Puja Sapra, Karen Widbin and Jennifer Tejeda for their support of the program.
Bronchial Epithelial Growth Medium | Lonza | CC-3170 | BEGM |
Cell Strainer 40um | ThermoFisher | 352340 | For passing TUM622 cells |
Cleaved Caspase 3 antibody | Cell Signaling Technology | 9661 (RRID:AB_2341188) | Rabbit |
CoolRack CFT30 | Biocision | BCS-138 | For 3D culture |
CoolSink XT96F | Biocision | BCS-536 | For 3D culture |
Cultrex 3D Cell Harvesting Kit | Bio-Techne | 3448-020-K | |
Cultrex (preferred for co-culture) | Bio-Techne | 3443-005-01 | For 3D culture |
CXCR4 antibody | Abcam | Ab124824 (RRID:AB_10975635) | Rabbit |
E-cadherin antibody | BD Biosciences | 610182 (RRID:AB_397581) | Mouse |
GelCount | Oxford Optronix | For Acini counts and measurements | |
GM130 antibody | BD Biosciences | 610822 (RRID:AB_398141) | Mouse |
Goat Serum | Vector Labs | S1000 (RRID:AB_2336615) | For Immunofluorescence |
Heat-inactivated FBS | Gibco | 10082-147 | For CAFs |
Histology sample gel | Richard Allan Scientific | HG-4000-012 | For Immunofluorescence |
Integrin alpha 6 antibody | Millipore Sigma | Mab1378 (RRID:AB_2128317) | Rat |
Involucrin antibody | Abcam | Ab68 (RRID:AB_305656) | Mouse |
Ki67 antibody | Abcam | Ab15580 (RRID:AB_443209) | Rabbit |
Lab-Tec II chambered #1.5 German Coverglass System | Nalge Nunc International | 155379 (2) | For 3D culture |
Lab-Tec II chambered #1.5 German Coverglass System | Nalge Nunc International | 155409 (8) | For 3D culture |
L-Glutamine | Gibco | 25030-081 | For CAFs |
Matrigel (preferred for mono-culture) | Corning | 356231 | For 3D culture |
p63 antibody | Cell Signaling Technology | 13109 (SRRID:AB_2637091) | Rabbit |
Pen/Strep | Gibco | 15140-122 | For CAFs |
ReagentPack Subculture Reagents | Lonza | CC-5034 | For TUM622 cell dissociation |
RPMI | ThermoFisher | 11875-093 | For CAFs |
Sox2 antibody | Cell Signaling Technology | 3579 (RRID:AB_2195767) | Rabbit |
TrypLE Express | Gibco | 12604-021 | For CAF dissociation |
Vi-Cell | Bechman Coulter | Automatic cell counter | |
Vimentin antibody | Abcam | Ab92547 (RRID:AB_10562134) | Rabbit |
β-catenin antibody | Cell Signaling Technology | 2677s (RRID:AB_1030943) | Mouse |