Organoid-fibroblast co-cultures provide a model to study the in vivo stem cell niche. Here, a protocol for esophageal organoid-fibroblast co-cultures is described. Additionally, whole mount imaging is used to visualize the fibroblast-organoid interaction.
Epithelial stem and progenitor cells contribute to the formation and maintenance of the epithelial barrier throughout life. Most stem and progenitor cell populations are tucked away in anatomically distinct locations, enabling exclusive interactions with niche signals that maintain stemness. While the development of epithelial organoid cultures provides a powerful tool for understanding the role of stem and progenitor cells in homeostasis and disease, the interaction within the niche environment is largely absent, thereby hindering the identification of factors influencing stem cell behavior. Fibroblasts play a key role in directing epithelial stem and progenitor fate. Here, a comprehensive organoid-fibroblast co-culture protocol enabling the delineation of fibroblast subpopulations in esophageal progenitor cell renewal and differentiation is presented. In this protocol, a method to isolate both epithelial cells and fibroblasts in parallel from the esophagus is described. Distinct fluorescence-activated cell sorting strategies to isolate both the esophageal progenitor cells as well as the fibroblast subpopulations from either transgenic reporter or wild-type mice are outlined. This protocol provides a versatile approach that can be adapted to accommodate the isolation of specific fibroblast subpopulations. Establishing and passaging esophageal epithelial organoid mono-cultures is included in this protocol, enabling a direct comparison with the co-culture system. In addition, a 3D clearing approach allowing for detailed image analysis of epithelial-fibroblast interactions is described. Collectively, this protocol describes a comparative and relatively high-throughput method for identifying and understanding esophageal stem cell niche components in vitro.
Organoids are used as 3D in vitro assays to characterize stem and progenitor cells, as well as to understand the signaling cues derived from the cellular components of the stem cell niche1,2,3,4. Mouse esophageal organoids were first described in 2014 and several papers have identified growth factors, like R-Spondin (RSPO), NOGGIN, and epidermal growth factor (EGF), needed to maintain and passage esophageal organoids5,6,7, arguing that similar signaling cues are required for in vivo progenitor cell renewal. However, growth factors are commonly added in non-physiological concentrations, resulting in organoid growth conditions which do not necessarily reflect the in vivo signaling environment.
Fibroblasts are heterogeneous stromal cell populations that support progenitor cell properties in many stem cell niches8. Combining epithelial progenitor cells and fibroblasts in the same organoid culture enables organoid formation in reduced concentrations of exogenously supplemented growth factors. Organoid co-culture systems from intestinal and hepatic epithelia are described, but a protocol to establish esophageal organoid-fibroblast co-cultures is still outstanding9,10,11.
In this protocol, two fluorescence-activated cell sorting (FACS) strategies for fibroblasts from the esophagus, using either transgenic PdgfrαH2BeGFP mice12 or wild-type mice with classical antibody staining are outlined. Different subpopulations of fibroblasts can be isolated using cell surface markers of choice, thereby providing flexibility to the protocol. In addition, a 3D fluorescence imaging technique preserving organoid morphology is used to characterize fibroblast-organoid interactions. Organoid clearing provides a quick method to increase the light penetration depth in the organoids, improving the visualization of organoid-fibroblast connections and enabling the recapitulation of the organoid structure in its entirety. This protocol combines esophageal organoid co-culture with a whole mount imaging strategy, enabling functional characterization of the fibroblast-organoid interaction.
Animal experiments for this study were approved by Stockholms Norra djurförsöksetiska nämnd (ethical permit no 14051-2019). Animals were housed in pathogen-free conditions according to the recommendations of the Federation of European Laboratory Animal Science Association.
1. Preparation
2. Dissection and separation of the esophageal epithelium and stroma
NOTE: Ensure that all instruments used for dissection and tissue processing are sterile. Prepare 2 mL of dissociation solution (see Table of Materials) in Hanks' balanced salt solution (HBSS) per three esophagi.
3. Isolation of esophageal progenitor cells
NOTE: The isolation of esophageal progenitor cells (step 2) and fibroblasts (step 3) can be performed simultaneously. Prepare a 50 mL tube of 1% FBS in HBSS (1% FBS).
4. Isolation of fibroblasts from the stromal layer
5. Establishment and culture of esophageal organoids
NOTE: Prewarm ERlow (organoid co-culture), ENR (organoid) medium (see Table 1 for description), and a 48-well plate at 37 °C. Place the thawed matrix (prepared in step 1) aliquot on ice. It is recommended to use the matrix provided here (see Table of Materials) for mouse esophageal organoid culture, since other brands of matrix negatively affect organoid forming efficiency.
6. Passaging of organoids
NOTE: Passaging of organoids grown in co-culture results in the loss of fibroblasts. Therefore, it is recommended to use ENR medium for all organoids when passaged. Prewarm ENR, PBS, and a 48-well plate at 37 °C.
7. Organoid processing for whole mount staining
NOTE: Coat the tips and tubes with 10% FBS in PBS before use to avoid organoids adhering to plastics. For pipette tips, it is sufficient to pipette once or twice up and down in 10% FBS/PBS solution before using the tip. For tubes, fill the tube with 10% FBS/PBS and then remove the solution.
The esophagus is divided into different layers: epithelium, lamina propria, submucosa, and muscularis externa (Figure 1A). Fibroblasts reside within the submucosa and lamina propria, referred to as the stroma. In this protocol, the muscularis externa is mechanically removed (Figure 1B), which does not lead to a loss of fibroblasts (PdgfrαH2BeGFP+) residing in the stroma (Figure 1C). Before dissociation, the epithelium is separated from the stroma resulting in two tissue segments (Figure 1B). Separating the two layers provides the opportunity to increase dissociation time for the more robust epithelial layer compared to the fragile stromal layer. In this way, an efficient isolation protocol yielding both viable epithelial progenitor cells as well as stromal fibroblasts is established (Figure 1B). Esophageal progenitor cells are sorted based on their high INTEGRIN-β4 and E-CADHERIN expression (Figure 1C,D).
Subpopulations of fibroblasts can be isolated by using distinct markers. In this protocol, a strategy for fibroblast isolation based on commonly used fibroblast markers PDGFRα and DPP4 (CD26) is provided. Isolation by either the PdgfrαH2BeGFP reporter expression or DPP4 antibody shows that around 50% of the isolated cells are fibroblasts (Figure 1E,F). Additionally, 70% of the PDGFRα+ fibroblasts are DPP4+, indicating that a largely overlapping, but not identical, fibroblast population is obtained. After isolating both epithelial and stromal cell populations, esophageal progenitor cells are either cultured alone or together with fibroblasts in a matrix dome. To study the contribution of fibroblasts to organoid formation, the co-culture is maintained in a growth factor reduced medium (Figure 1G).
Esophageal progenitor cells form organoids in the presence of EGF, NOGGIN and RSPO (ENR). Removing NOGGIN and reducing the amount of RSPO (25 ng/µL; ERlow) is sufficient to prevent organoid formation (Figure 2A). Interestingly, adding either DPP4+ or PDGFRα+ fibroblasts to the esophageal progenitor cells in the ERlow medium restores the organoid forming ability, demonstrating a supportive function for both fibroblast populations (Figure 2A–D). Visualization of the PdgfrαH2BeGFP transgene shows that fibroblasts are in close contact with the epithelial progenitor cells during organoid formation (Figure 2A). At day 6, PdgfrαH2BeGFP+ fibroblasts are still abundantly present in the co-culture. Fibroblasts are present throughout the dome, near and touching the organoids (full arrow), or attached to the organoids (arrowhead; Figure 2D).
Whole mount staining shows a 3D representation of the interaction of the fibroblasts with the organoids (Figure 3). While it is possible to perform the whole mount protocol without the use of a clearing solution, it decreases transparency and laser penetration of the organoid (Figure 3B, z-view). When mounting organoids, the spacer helps to maintain organoid morphology. In contrast, plating the coverslip directly on the organoids (without a spacer) flattens the organoids and results in loss of organoid structure (Figure 3A,B).
Both the DPP4+ and PDGFRα+ fibroblasts are found to be wrapped around the organoids (Figure 3C, Video1, and Video 2). Differentiation of esophageal organoids can be assessed using different markers. Figure 4 shows that the staining protocol provided is suitable for easier-to-stain keratins (KRT14/13) as well as more-difficult-to-stain transcription factors (TRP63/KLF4). The co-culture protocol generates organoids with a similar differentiation pattern, as demonstrated in vivo13,14 and as seen in organoids grown in ENR medium; KRT14+ or TRP63+ progenitor cells form the outer layer and KRT13+ or KLF4+ differentiated cells oriente inwards.
This protocol provides a tool to study the esophageal stem cell niche in vitro and visualizes the interaction between organoids and fibroblasts. By implementing a protocol for the isolation of fibroblasts using antibodies, the method is adaptable and can be used to study fibroblast subpopulations without the need of transgenic mice.
Figure 1: Isolation of progenitor cells and fibroblast subpopulations from the esophagus. (A) Schematic overview of the different layers in the esophagus. The stroma contains the lamina propria and submucosa. (B) Schematic overview of the isolation protocol. The muscle (muscularis externa) is mechanically removed using forceps; the remaining esophagus is cut open and incubated in thermolysin to separate the epithelial layer from the stroma. The epithelium and stroma are separated, mechanically minced, and enzymatically digested to single cell suspensions. Dissociated cells are then stained and prepared for FACS. (C) Cross section of the esophagus stripped from the muscularis externa showing PdgfrαH2BeGFP+ fibroblasts in the stroma. INTEGRIN-β4 (ITGβ4) and E-CADHERIN (ECAD) double positive cells are the epithelial progenitor cells of the esophagus. Scale bar = 100 µm. (D) Representative flow cytometry plot of epithelial cell isolation showing the percentage of live cells (upper panel) from all single cells. The lower panel shows the percentage of isolated ITGβ4+ ECAD+ progenitor (Prog.) cells from all live cells. (E) Representative flow cytometry plot of stromal cell isolation showing the percentage of live cells (upper left panel). Representative flow cytometry plots showing the percentage of isolated DPP4+ fibroblasts (Fibr.; upper right panel) and Pdgfrα+ fibroblasts (lower left panel) of all live cells. 70% of the Pdgfrα+ fibroblasts are also DPP4+ (lower right panel). (F) Representative flow cytometry plot of the stroma showing DPP4+ only cells (2.5%), DPP4+ PDGFRα+ cells (37.5%), and PDGFRα+ only cells (17.7%). The percentages are of all live cells. (G) Epithelial only cells are plated in a matrix dome in the presence of 50 ng/µL EGF, 100 ng/µL NOGGIN, and 250 ng/µL RSPO (ENR), or together with fibroblasts in the presence of EGF and a low concentration of RSPO (25 ng/µL). Please click here to view a larger version of this figure.
Figure 2: Representative results of organoid co-cultures. (A) Brightfield images showing growth of the organoids from day 1 to day 6. The brightfield images with the organoids co-cultured with PdgfrαH2BeGFP+ fibroblasts also show the nuclear eGFP signal. Scale bar = 25 µm. (B) Brightfield images of the whole matrix dome at day 6. The left column shows organoid co-cultures grown in the presence of Pdgfrα+ fibroblasts in ERlow or ElowRlow medium. The middle column shows organoid co-cultures grown in the presence of DPP4+ fibroblasts in ERlow or ElowRlow medium. The right column shows organoid mono-cultures grown in ENR medium. ENR medium = EGF (50 ng/µL), NOGGIN (100 ng/µL), and RSPO (250 ng/µL). ERlow = EGF and 25 ng/µL RSPO. ElowRlow = 5 ng/µL EGF and 25 ng/µL RSPO. Scale bar = 500 µm. (C) Graph showing the organoid forming efficiency (%) (i.e., the percentage of cells forming organoids in different culture conditions). Each dot represents a matrix dome and the bar represents the mean of all dots per condition. (D) Brightfield and fluorescent image of day 6 organoids co-cultured with PdgfrαH2BeGFP+ fibroblasts. PdgfrαH2BeGFP+ fibroblasts are present throughout the dome, attached to the organoids (arrowhead), and unattached but in contact with the organoids (full arrow). Scale bar = 250 µm. Please click here to view a larger version of this figure.
Figure 3: Whole mount staining protocol for the study of fibroblast-organoid interactions. (A) Schematic overview of the whole mount immunofluorescence protocol. AB = antibody. (B) Immunofluorescence picture of uncleared whole mount staining showing a decreased transparency and penetration of the laser light compared to the cleared organoids. The absence of a spacer results in flattening of the organoid and loss in organoid morphology. (C) Whole mount images of the co-cultured organoids show 3D surfaces of the organoids with VIMENTIN+ fibroblasts (Fibr.) wrapped around and in close contact with the organoid. 3D cross sections and 2D plane images show the lumen of the organoid. Scale bar = 50 µm. Please click here to view a larger version of this figure.
Figure 4: Whole mount images reveal distinct basal and suprabasal cell populations. (A) Whole mount staining of mono- and co-cultured organoids with PdgfrαH2BeGFP+ fibroblasts showing KRT14+ basal cells in the outer layer and KRT13+ differentiated suprabasal cells. Scale bar = 50 μm. (B) Whole mount staining of mono- and co-cultured organoids with PdgfrαH2BeGFP+ fibroblasts showing TRP63+ basal cells in the outer layer and KLF4+ differentiated suprabasal cells. Scale bar = 50 µm. Please click here to view a larger version of this figure.
Table 1: Table describing the organoid culture media components. Please click here to download this Table.
Video 1: PdgfrαH2BeGFP+ fibroblast wrapped around and in close contact with the organoid. The video accompanies the upper panel of Figure 3C. The scale bar in Figure 3C is 50 µm, and the organoid is ~120 µm in diameter. VIMENTIN is shown in white, E-CADHERIN in red, PdgfrαH2BeGFP in green, and DAPI in blue. Please click here to download this Video.
Video 2: DPP4+ fibroblast wrapped around and in close contact with the organoid. The video accompanies the lower panel of Figure 3C. The scale bar in Figure 3C is 50 µm, and the organoid is ~120 µm in diameter. VIMENTIN is shown in white, E-CADHERIN in red, and DAPI in blue. Please click here to download this Video.
The protocol presented here establishes an in vitro model for investigating functional esophageal epithelial-fibroblast interactions.
The epithelial layer is separated from the stroma, allowing for an optimized dissociation protocol for both the epithelial and stromal compartment. Despite optimization of the epithelial dissociation protocol, tissue clumps remain apparent. Pipetting up and down vigorously every 15 min decreases the number and size of clumps substantially. Other protocols use trypsin to further dissociate the epithelial layer5,6. Here, the use of trypsin, or increasing the dissociation time further, is not recommend, as this tends to result in decreased epithelial cell viability and organoid forming efficiency. In contrast to the epithelium, the stroma is easily dissociated, and 30 min in dissociation solution results in a single-cell suspension with ~90% fibroblast viability (Figure 1E). Excluding the epithelial-stomal separation step in the protocol increases the dissociation time substantially, resulting in a decreased fibroblast viability and a lower yield of epithelial cells. Additionally, separating the epithelium from the stroma provides an opportunity to determine cell numbers of each population and mix epithelial cells and fibroblasts from different mouse lines when setting up the co-cultures.
Studying fibroblast function on organoid growth is a commonly used method in stem cell biology 9,10,11,15,16. Established co-culture media are either DMEM supplemented with 10% fetal calf serum (FCS)9,15 or growth factor reduced medium10,16. In this protocol, the growth factor reduced medium is used to mimic the conditions in the in vivo stem cell niche, where fibroblasts are largely quiescent. FCS is a growth factor rich serum which results in activation and proliferation of fibroblasts in the co-cultures, likely corresponding to a fibroblast cell state distinct from the in vivo state. By excluding FCS and reducing growth factors, so that the medium alone (ERlow) does not support organoid growth and does not stimulate fibroblast proliferation, it is possible to isolate the effect of the fibroblasts on the organoid growth. In this medium, NOGGIN is removed and RSPO reduced to a minimum (10% RPSO). Both NOGGIN and RSPO have been demonstrated to be essential for esophageal organoid growth6. EGF was retained in the co-culture medium, as it does not support organoid growth by itself. However, fibroblasts are also capable of supporting organoid growth in an EGF-reduced medium (ElowRlow; Figure 2B,D).
Organoid co-cultures cannot be sustained through passaging as fibroblasts are lost during trypsinization. However, organoid passaging was included in the protocol since esophageal organoids can be maintained, expanded, and used for further experiment as mono-cultures. Passaged organoids from mono-cultures can be used to set up co-cultures with freshly isolated fibroblasts. A disadvantage of using primary cells is the number of mice needed to set up multiple organoid co-cultures. When focusing on small subpopulations of fibroblasts, the number of co-cultures obtained is limited. In other protocols, fibroblasts are first expanded in culture before using them to set up organoid co-cultures10. However, fibroblasts change morphology and identity during passaging, shown by using primary skin and cardiac fibroblasts17,18. Conventional 2D passaging of esophageal fibroblasts results in both morphology and phenotype changes, demonstrating that in vitro enrichment of fibroblasts is not suitable for co-cultures aiming to phenocopy the endogenous stem cell niche.
Whole mount staining provides a tool to maintain and visualize fibroblast-organoid interaction. It should be noted that, while not all organoids will have fibroblasts directly attached to them, most organoids are in contact with fibroblasts (see Figure 2C). To maintain epithelial-fibroblast interactions, it is important to handle the organoids with care and avoid vigorous pipetting, vortexing, and high-speed spinning. Optimal fixation is important to maintain 3D tissue architecture, as well as keep endogenous fluorescence. A 30 min fixation suffices to retain the H2BeGFP signal and is optimal for the antibodies used in this protocol, however this might vary between the fluorophores and antibodies used. Clearing of the organoids reduces light scattering and improves visualization of the whole 3D structure substantially. As the organoids are small, clearing is easy and fast; however, imaging whole organoids using laser-scanning confocal microscopy can be time-consuming, as multiple Z-stacks need to be made. Confocal microscopes, like the spinning disc, can be used to reduce imaging time.
Overall, esophageal organoids grown in the presence of fibroblasts provide a valuable tool to understand aspects of the esophageal stem cell niche. In addition, whole mount clearing provides an accessible method to visualize the interaction between fibroblasts and organoids.
The authors have nothing to disclose.
This study was supported by ERC StG TroyCAN (851241). E.E. is a Cancerfonden Postdoctoral Associate. M.G. is a Ragnar Söderberg Fellow and Cancerfonden Junior Investigator. We are grateful for the technical assistance from Karolinska Institutet core facilities, including the Biomedicum Flow Cytometry Core Facility, the Biomedicum Imaging Core (BIC), and the Comparative Medicine Biomedicum (KMB) animal facility. We thank members of the Genander lab for carefully reading and commenting on the protocol.
B-27 Supplement (50X), serum free | Thermo Fisher (Gibco) | 17504001 | |
Corning Matrigel Growth Factor Reduced (GFR) Basement Membrane Matrix | fisher scientific | 356231 | |
Dimethyl sulfoxide | Sigma-Aldrich | 276855-100ML | |
DMEM/F-12 | Thermo Fisher (Gibco) | 11320074 | |
DPBS | Thermo Fisher (Gibco) | 14190250 | |
Fetal Bovine Serum | Sigma-Aldrich | F7524 | |
GlutaMAX Supplement | Thermo Fisher (Gibco) | 35050061 | |
HBSS, no calcium, no magnesium, no phenol red | Thermo Fisher (Gibco) | 14175-129 | |
Normal Donkey Serum | Jackson Immuno | 017-000-121 | |
Penicillin-Streptomycin (10,000 U/mL) | Thermo Fisher (Gibco) | 15140122 | |
Triton X-100 solution | Merck | 93443-100ML | |
Trypsin-EDTA (0.25%), phenol red | Thermo Fisher (Gibco) | 25200-056 | |
Chemicals, Peptides, and recombinant proteins | |||
DAPI Solution | Thermo Fisher | 62248 | |
Dissociation solution: 0.25 mg/ml Liberase TM, 0.25 mg/ml Dnase in HBSS | |||
Dnase I | Sigma-Aldrich | 11284932001 | |
Formaldehyde, 37%, with 10-15% methanol | Sigma-Aldrich | 252549-1L | |
Liberase | Sigma-Aldrich | 5401127001 | |
N-Acetyl-cysteine | Sigma-Aldrich | A9165-25G | |
Noggin murine | Peprotech | 250-38 | |
RapiClear 1.47 | SunJin Lab | #RC147001 | |
Recombinant Mouse EGF Protein, CF | R&D systems | 2028-EG-200 | |
R-spondin-1 murine | Peprotech | 315-32 | |
SYTOX Blue Dead Cell Stain | Thermo Fisher | S34857 | |
Thermolysin | Sigma-Aldrich | T7902-25MG | |
Y-27632 dihydrochloride | Sigma-Aldrich | Y0503-5MG | |
Plastic & Glassware | |||
Corning Sterile Cell Strainers 40um | VWR | 15360801 | |
Corning Sterile Cell Strainers 70um | VWR | 431751 | |
Menzel Deckgläser/ cover slips | Thermo Fisher | Q10143263NR15 | |
SafeSeal reaction tube, 1.5 mL, PP | Sarstedt | 72.706 | |
Snap Cap Low Retention Microcentrifuge Tubes 0.6 mL | Thermo Fisher | 3446 | |
SuperFrost Slides | VWR | 631-9483 | |
Tools | |||
0.05 mm 4 circular well iSpacer | SunJin Lab | #IS204 | |
Dumont #5 forceps, biology tip | F.S.T | 11252-20 | |
ImmEdge Pen | VectorLaboratories | H-4000 | |
Spring Scissors Angled to Side Ball Tip 8mm Cutting Edge | F.S.T | 15033-09 | |
Instruments | |||
Confocal microscope Stellaris 5 | Leica | ||
Dissection microscope ZEISS Stemi 305 | Zeiss | ||
FACS ARIA III | BD Biosciences | ||
Conjugated Antibodies for FACS | |||
Alexa Fluor 647 anti-mouse CD104 Antibody Clone: 346-11A |
123608 | 123608 | |
APC anti-mouse CD26 (DPP-4) Antibody | H194-112 | H194-112 | |
PE/Cy7 anti-mouse/human CD324 (E-Cadherin) Antibody | 147310 | 147310 | |
Antibodies for Immunofluorescence | |||
CD104 (ß-integrin 4) Clone: 346-11A |
BioLegend | 553745 | |
Cytokeratin 14 | Acris Antibodies (AbD Serotec) | BP5009 | |
Cytokeratin13 Clone: EPR3671 |
abcam | ab92551 | |
E-cadherin (CD324) Clone: 2.40E+11 |
Cell Signaling Technology | 3195 | |
Keratin 5 Polyclonal Chicken Antibody, Purified Clone: Poly9059 |
BioLegend | 905901 | |
p63 Clone: 4a4 |
abcam | ab735 | |
Recombinant Anti-KLF4 antibod Clone: EPR20753-25 |
abcam | ab214666 | |
Vimentin | Sigma-Aldrich | AB5733 | |
Secondary antibodies | |||
Donkey anti-species* antibodies with fluorophore of choice | Jackson Immuno |