Here, we describe the isolation of enteric-glial cells from the intestinal-submucosa using sequential EDTA incubations to chelate divalent cations and then incubation in non-enzymatic cell recovery solution. Plating the resultant cell suspension on poly-D-lysine and laminin results in a highly enriched culture of submucosal glial cells for functional analysis.
The enteric nervous system (ENS) consists of neurons and enteric glial cells (EGCs) that reside within the smooth muscle wall, submucosa and lamina propria. EGCs play important roles in gut homeostasis through the release of various trophic factors and contribute to the integrity of the epithelial barrier. Most studies of primary enteric glial cultures use cells isolated from the myenteric plexus after enzymatic dissociation. Here, a non-enzymatic method to isolate and culture EGCs from the intestinal submucosa and lamina propria is described. After manual removal of the longitudinal muscle layer, EGCs were liberated from the lamina propria and submucosa using sequential HEPES-buffered EDTA incubations followed by incubation in commercially available non-enzymatic cell recovery solution. The EDTA incubations were sufficient to strip most of the epithelial mucosa from the lamina propria, allowing the cell recovery solution to liberate the submucosal EGCs. Any residual lamina propria and smooth muscle was discarded along with the myenteric glia. EGCs were easily identified by their ability to express glial fibrillary acidic protein (GFAP). Only about 50% of the cell suspension contained GFAP+ cells after completing tissue incubations and prior to plating on the poly-D-lysine/laminin substrate. However, after 3 days of culturing the cells in glial cell-derived neurotrophic factor (GDNF)-containing culture media, the cell population adhering to the substrate-coated plates comprised of >95% enteric glia. We created a hybrid mouse line by breeding a hGFAP-Cre mouse to the ROSA-tdTomato reporter line to track the percentage of GFAP+ cells using endogenous cell fluorescence. Thus, non-myenteric enteric glia can be isolated by non-enzymatic methods and cultured for at least 5 days.
Interest in the function of enteric glial cells (EGCs) has steadily increased due to their recognized roles in the gut integrity and homeostasis1,2. In addition, EGCs vary according to their location along the length of the GI tract3,4. EGCs release various trophic factors including glial cell-derived neurotrophic factor (GDNF), contribute to gut motility1,5 and respond to microbial byproducts6,7. Studies have indicated that the EGC population is heterogeneous and that their function varies depending on whether they are submucosal or reside within the myenteric plexus1,7. For example, EGCs within the submucosa contribute to tight junctions8. Differential GFAP expression and phosphorylation in EGCs have been linked to Parkinson's Disease, suggesting their possible link to the gut phenotype of this disorder9. Recently, it was observed that the loss of the nuclear protein menin in isolated cultures of EGCs from the proximal intestinal submucosa was sufficient to induce expression of the hormone gastrin10. As a result, it was proposed that EGCs might be the origin of duodenal gastrinomas, a type of neuroendocrine tumor10. Collectively, these examples underscore the relevance of studying the behavior and function of isolated EGCs in neuropathic disorders and cancer11.
The challenge in the field remains how to isolate and study either or both EGC populations in vitro. Lineage trace experiments demonstrated that EGCs in the submucosa and lamina propria originate from progenitor cells in the myenteric plexus7. Although there are several published isolation protocols available to generate cultures of myenteric EGCs12,13,14,15,16,17,18,19, none specifically targets isolation of the submucosal/lamina propria EGC population. Existing protocols for EGC isolation specifically use a combination of the mechanical separation or the microdissection of the smooth muscle combined with enzymatic dissociation, eventually discarding the mucosal cell layer.
The goal of this manuscript is to demonstrate the steps to non-enzymatically isolate primary EGCs from the lamina propria for in vitro studies. Since there are no markers that specifically distinguish myenteric EGCs from those in the submucosa, the spatial separation of the epithelial mucosa from the smooth muscle was exploited to isolate submucosal EGCs. In addition, by combining EDTA chelation with non-enzymatic dissociation, EGCs were isolated from the submucosa in contrast to the smooth muscle, which was discarded along with the associated inter-myenteric EGCs. Further separation of the submucosa and lamina propria EGCs occurred by culturing the cells on glial cell-friendly substrates, e.g., poly-D-lysine and laminin.
All animal experiments described were approved by the University of Michigan's Committee on the Use and Care of Animals.
1. Preparation of Sterile Poly-D-lysine (PDL) and Laminin Solutions
2. Coating Plates
NOTE: perform these steps under a sterile tissue culture laminar flow hood.
3. Preparation of Isolation Solutions
4. Removal of the Intestinal Segment
NOTE: Mice were allowed free access to the food and water prior to euthanization by the isoflurane drop method and removal of a vital organ. C57BL/6 mice greater than 8 weeks of age were used. Both genders were used without noticeable differences in the preparation.
5. Removal of the Epithelial Mucosa
6. Collection of Enteric Glial Cells from the Lamina Propria and Submucosa
7. Immunofluorescent Staining
8. Flow Cytometry
9. Preparing Mouse Tissues from the hGFAP-Cre:tdTomato Mice
NOTE: The Lox-STOP-Lox-tdTomato cDNA is expressed from the ROSA locus22,23 (Jackson Labs, #007914) and generates endogenous fluorescence when bred to a mouse expressing the Cre recombinase (hGFAP-Cre). In situ analysis is performed by generating frozen tissue sections.
10. Ca2+ Flux Imaging
Preps were considered unsuccessful if GFAP+ cells did not adhere and spread within 24 h (Figure 4A). The number of glial cells could not be determined until after 24 h when the cells adhered and showed evidence of spreading into flat aggregates (Figure 4B). Cells at the edge of the clusters tended to extend long processes and expressed classic glial markers, e.g., GFAP, S100b and p75NTR (Figure 4C, 4D)10,16. By the 2nd day, glial cells tightly adhered to the surface allowing the non-adherent population to be rinsed away with PBS. Most of the floating cells were epithelial and lamina propria cells along with cell debris. The presence of these floating cell clumps reduced the yield of adherent glial cells underscoring the importance of performing the EDTA incubations until the flow-through is nearly clear, signaling that the epithelial cells have been removed from the lamina propria core. In addition, centrifugation at speeds lower than 1,500 x g after incubating in the cell recovery solution did not effectively collect all of the viable cells into a pellet leading to reduced yields. Typically, 40,000 to 100,000 cells morphologically consistent with EGCs adhered firmly by day 3 in culture.
Immunohistochemical analysis with glial-associated antibodies, indicated that the cell clusters that remained adherent by day 3 were GFAP, S100b, and Sox 10 positive10,16 (Figure 5A). In the current study, this same degree of purity was observed by permeabilizing the cells and analyzing by flow cytometry (Figure 5B-F). Immediately after isolating the submucosal/Lamina propria glial cells, flow cytometry revealed that about 51% of the cells were GFAP+ (Figure 5B), suggesting the presence of GFAP-negative cell types, e.g., epithelial, hematopoietic, endothelial, neuronal cells, myofibroblasts known to exist in the epithelium and lamina propria. After 3 days in culture, the number of GFAP+ cells represented over 95% of the cell population analyzed after removing the original media, gently rinsing with PBS and then adding fresh media (Figure 5C). Interestingly, flow analysis revealed high and low GFAP protein-expressing populations (Figure 5C). Indeed, immunofluorescent analysis of the cells revealed that the periphery of the clusters tended to express higher levels of GFAP (Figure 4C, 4D). Taken together, the two types of analysis might indicate differences in EGC maturity or differentiation status. Collectively, the percentage of α-SMA+ (myofibroblast cell marker), E-cadherin+ (epithelial cell marker), and Pgp 9.5+ (neuronal cell marker) cells was less than 5% (Figure 5D-F). These results were consistent with the prior study performed using immunofluorescent labeling of EGCs showing about 93% GFAP+ cells 10. The flow analysis underscores the importance of allowing the cells to adhere to the plates since it allows removal of contaminating cell populations comprising approximately 5% of the cultures.
A goal in this study was to use a GFAP-activated reporter to facilitate flow cytometry of the cells for further analysis and to compare the degree of EGC enrichment using an endogenous fluorescent marker (Figure 6). The hGFAP-Cre mouse line was originally described by Messing and coworkers 20. In brief, the Cre recombinase was placed under the control of 2.2 kb of the human glial fibrillary acid protein (hGFAP) promoter. Thus, any cell transcribing this hGFAP promoter expressed the red fluorescent reporter tdTomato. The cell population labeled with the tdTomato reporter was visualized in situ using frozen sections of the intestine and colon (Figure 6A, 6B) before performing the EGC isolation procedures described above. After performing the EDTA/cell recovery isolation and culturing the cells for 3 days, EGCs from these mice were easily identified by their endogenous fluorescence (Figure 6C). Although GFAP+ cells arguably are more numerous in the proximal intestine4,21, tdTomato+ EGCs were also observed in the colon (Figure 6B). Using the tdTomato+ fluorescent reporter activated by the hGFAP-Cre also revealed a bimodal population of hGFAP-tdTomato+ cells (Figure 6D) as observed using the immunofluorescent analysis of GFAP protein (Figure 5C). Although it has been reported that not all EGCs express GFAP protein4, this point might reflect differences in the level of GFAP expression. On average, about 66% of the cells analyzed exhibited the highest level of tdTomato fluorescence. Therefore, the EDTA/cell recovery protocol could be used to isolate subsets of EGCs throughout the small intestine and colon labeled by the reporter to examine differences in gene expression.
This protocol generated a sufficient number of relatively pure GFAP+ glial cells for biochemical studies and western blot analysis10. To demonstrate glial cell responsiveness, a 3-day glial cell prep was treated with the hormones cholecystokinin (CCK) or gastrin to induce Ca2+ flux (Figure 7A-7B)10. The results showed that these GFAP+ adherent cells responded to extracellular agonists demonstrating their ability to exhibit known enteric glial function.
Figure 1: Set up and preparation of mouse intestines. (A) Picture of isolation set up on ice including extracted intestines. (B) Picture of 5 mL-syringe attached to a 20 G blunt end needle to flush out fecal contents. (C) Engineered end of a wooden cotton swab (~4 cm) soaking in the DPBS buffer. Please click here to view a larger version of this figure.
Figure 2: Steps used to clean intestines and remove the longitudinal muscle/myenteric plexus (LMMP). (A-C) Sliding an intestinal segment onto a wetted wooden stick. (D) Removal of adherent mesentery. (E) Nicking the intestine with a razor blade. (F) Removing the LMMP with a damp cotton swab. Please click here to view a larger version of this figure.
Figure 3: Sequential Flow-throughs after EDTA incubations. Shown are examples of the flow-throughs from 4 sequential incubations using three 7 cm intestinal segments incubated for 10 min with 5 mM EDTA/10mM HEPES in DPBS and then the triturated 20 times. Representative flow-throughs after pouring through a 100 µm nylon mesh strainer are shown. Please click here to view a larger version of this figure.
Figure 4: EGC images 24 h after plating. Light microscopic examination of resuspended cells 24 h after plating. (A) Shown is a representative example of a poor prep showing floating epithelial cell clumps (arrow) and debris. No adherent EGCs were observed after 24 h. (B) Shown is a representative example of an excellent prep 24 h after the cell isolation and resuspension in which patches of EGCs adhered to the PDL/Laminin coated plates. Images were captured on an inverted fluorescent microscope with a digital camera. Scale bars = 500 µm (A-B). (C) High power view of a patch of EGC cells stained with GFAP antibody (green) showing cells at the periphery with a higher intensity of labeling. Several of the cells showed double nuclei (arrowhead, DAPI, blue). (D) Colocalization of GFAP (green) with S100 (red) or p75NTR (red). Scale bars = 20 µm (C-D). The cells were mounted with an antifade reagent and DAPI (blue nuclei). Please click here to view a larger version of this figure.
Figure 5: Flow cytometry of enteric glial cells isolated from the duodenal lamina propria of the adult mouse. (A) Immunofluorescence of GFAP, S100B and Sox10 on EGC cultures after 3 days. Arrows indicate Sox 10 positive nuclei. Scale bars = 20 µm (used with permission from Sundaresan et al.10). (B) Percentage of GFAP positive cells before plating on coated plates (fluorescence intensity (or forward scatter, X-axis) versus side scatter (Y-axis). (C) Percentage of GFAP positive (D) α-SMA positive (E) E-cadherin positive (F) and Pgp 9.5 positive cells 3 daysafter plating. Gates were constructed to avoid dead cells and debris. Shown is the mean percentage of the fluorescent population relative to the number of cells (side scatter). Please click here to view a larger version of this figure.
Figure 6: Identification of enteric glial cells isolated from the duodenal lamina propria of the hGFAP-Cre+:tdTomato+ adult mouse. Endogenous tdTomato fluorescent (red) images were captured on a phase contrast fluorescent microscope with a digital camera in the (A) Intestine. (inset) Same as A except merged with DAPI image (blue nuclei). (B) Colon merged with DAPI. (C) Enteric glial cells (red) isolated from hGFAP-Cre:dtTomato+ mice and cultured for 3 days on PDL/Laminin substrate. Scale bars = 50 µm. (D) Flow cytometric analysis of tdTomato cells. Shown is the mean % of GFAP+ cells ± SEM for 3 different preparations (3 mice per prep). Please click here to view a larger version of this figure.
Figure 7: Video of Ca2+ fluxes after hormone treatment. EGCs plated on 24-well PDL and laminin coated plates were cultured in the glial growth media for 3 days. The cells were pre-loaded with 3 µM of Fura-2-AM for 20 min at 37 °C and then were treated with (A) 100 nM cholecystokinin (CCK) or (B) 100 nM gastrin. Please click here to download this file.
Components | Volume to be added | Final Concentration |
DMEM/F12 | 44.5 mL | _ |
Fetal Bovine Serum (FBS) | 5 mL | 10% |
Penicillin-Streptomycin | 500 µL | 100 IU/mL Pen |
Streptomycin | 100 µg/mL Strep | |
Gentamicin (50 mg/mL stock) | 20 µL | 20 µg/mL |
Table 1: Composition of Glial Cell Resuspension Media.
Components | Volume to be added | Final Concentration |
DMEM/F12 | 44 mL | _ |
Fetal Bovine Serum (FBS) | 5 mL | 10% |
Penicillin-Streptomycin (100x) | 500µL | 100 IU/mL (Pen) |
100 µg/mL (Strep) | ||
Gentamicin (50 mg/mL stock) | 20 µL | 20 µg/mL |
GDNF (10 µg/mL stock) | 50 µL | 10 ng/mL |
L-Glutamine (200 mM stock) | 500µL | 2 mM |
Table 2: Composition of Glial Cell Growth Media.
Components | Immunofluorescence Dilution | Flow Cytometry Dilution |
Chicken anti-GFAP | 1 to 500 | 1 to 2000 |
Rabbit anti-S100 | 1 to 500 | 1 to 500 |
Mouse anti-p75 NTR | 1 to 500 | 1 to 500 |
Goat anti-E-cadherin | 1 to 400 | |
Mouse anti-Pgp9.5 | 1 to 500 | |
Goat anti-a-Smooth Muscle Actin | 1 to 500 | |
Alexa Fluor 488 Goat Anti-Chicken IgY | 1 to 1000 | 1 to 1000 |
Alexa Fluor 568 Goat Anti-Mouse IgG | 1 to 1000 | 1 to 1000 |
Alexa Fluor 594 Donkey Anti-Rabbit IgG | 1 to 1000 | |
Alexa Fluor 488 Donkey Anti-Goat IgG | 1 to 1000 |
Table 3: List of Antibodies
EGCs play important roles in gut homeostasis, and it is essential to isolate and study them in vitro. In this protocol, a simple method for isolating EGCs from the lamina propria of the adult mouse intestine was introduced to study enteric glial function.
Removing the adherent mesentery and LMMP with a cotton swab removes some of the inter-myenteric glia residing between the longitudinal and circular muscle, increases the accessibility of buffers to the submucosal surface and removes much of the larger capillaries. The latter reduces the number of red blood cells contaminating the final cultures. The series of EDTA incubations strip away the epithelial mucosa exposing the lamina propria and submucosal glia, which can be fragile after exposure to the chelating solutions. Agitation either by shaking, vortexing or triturating (repetitive up and down pipetting) must be performed carefully and timed to avoid excessive damage to the cells. Trituration was used here because the technique resulted in a more consistent yield of adherent, viable glial cells. Shaking tended to introduce bubbles and reduce cell viability assessed morphologically by cell adherence to the coated plates. The tissue must be cut into pieces <0.5 cm to effectively triturate the tissue with the 5 mL pipette.
Generally, 7 cm segments from at least 3 mice is sufficient to generate enough EGCs to cover one 6-well plate at about 20–30% confluency, which can then be used for immunocytochemistry and qPCR. However, more cells might be required for biochemical methods such as western blots depending on the expression level of the gene studied. Although, type 1 collagen or Matrigel was briefly examined as an alternative to the poly-D-lysine/Laminin substrate, greater adherence of the epithelial cell population was observed, ultimately increasing contamination of the EGC cultures with other cell types. In addition, type 1 collagen did not adhere firmly to the plates and was easily dislodged from the surface when changing the media. Fibronectin is another substrate that has been used24, but was not tested here. Laminin apparently enhances the binding and differentiation of neural-crest derived cells25. However, lots of laminin can vary, impacting cell adherence and yields. Microbial contamination of the cultures occurred about 5% of the time and was kept to a minimum by the use of sterile reagents, technique and antibiotics. Use of the antifungal reagent was optional.
The major limitation of the protocol here is standardization of the agitation to remove the epithelium without damaging the underlying EGCs. Moreover, assessment of the preparation cannot be made immediately since it depends on cell adherence to the coated plates. Three areas to focus on when troubleshooting include: 1) Use of fresh PDL and laminin to coat the plates; 2) Minimize the time from dissection to beginning the EDTA incubations by increasing the number of assistants; 3) Quantify the time and method used to agitate the tissue to effectively remove the epithelium and then the EGCs. Extending the time in either of these steps increases EGC fragility and the likelihood that they will not recover when plated in growth media. Although the trituration method to dissociate the epithelium was used here, shaking and vortexing were also tested with inconsistent results perhaps because it is more operator dependent and difficult to quantify. Once plated the yield of adherent cells was greater if the cells were not disturbed for 16–24 h.
The method presented here was modified according to the original study by Smith et al.12, which focused on isolating the LMMP. The Smith approach was used here only to remove and discard the LMMP so that most of the glial cells isolated would originate from the submucosa and lamina propria. In addition, without enzymatic digestion, the myenteric EGCs are not readily liberated12. Nevertheless, since there are no specific markers for myenteric versus submucosal glia, the presence of glial cells from the inter-myenteric plexus cannot be excluded. Rosenbaum et al. reported flow cytometric analysis of EGCs expressing enhanced green fluorescent protein (EGFP) from the hGFAP promoter. They used very young mice (postnatal day 7) and isolated EGCs from the whole intestine by enzymatic digestion19. In addition, flow cytometric analysis was performed after two weeks of maintaining conditions that promoted free-floating neurosphere formation. Although the authors reported that the neurospheres highly expressed both GFAP and S100b, the floating cell aggregates also strongly expressed α-SMA, suggesting that expansion of the myofibroblast cell population encouraged gliosphere development in contrast to the flat sheet-like morphology described here. Therefore, while interesting, the study is not directly comparable to this protocol. By contrast, the morphology of the cells coupled with significantly less α-SMA+ cells suggest that the EGCs described in this protocol do not exhibit the 3-dimensional morphology during the 3 to 5-day culture period. Nevertheless, both the protocol here and the Rosenbaum method use a fluorescent reporter to identify and isolate GFAP+ cells that will improve investigator's ability to do live cell profiling.
In conclusion, the current protocol describes the isolation of enteric glial cells from the submucosa using a non-enzymatic approach. The entire protocol takes about 3 h from mouse dissection to initial plating on the pre-coated tissue culture plates. The most time intensive step in the protocol is the removal and preparation of the intestines for the first EDTA incubation. Preparing the plates and storing in DPBS is strongly recommended. The success of the preparation depends upon the efficient removal of the epithelium without damaging the underlying EGCs and adherence to the PDL/Laminin coated plates within 24 h. Primary EGCs are useful for in vitro studies such as biochemical analysis, Ca2+ fluxes and adenoviral transfections10. It is anticipated that the ability to isolate EGCs from either the submucosa or the LMMP will permit further studies to define the differences in EGC growth properties, differentiation and morphology using whole genome approaches.
The authors have nothing to disclose.
The authors wish to acknowledge support from R37 DK045729 (to JLM), R01 AR060837 (to HX) and the University of Michigan Gastrointestinal Research Center Molecular Core P30 DK034933.
Poly-D lysine (1 mg/ml stock) | Sigma | A-003-E | Dilute 1:10 |
Laminin (0.5 mg/ml stock) | Sigma | L4544 | Dilute to 10 µg/mL on ICE |
EDTA (0.5M) | Lonza | 51201 | Dilute 1:100 in DPBS |
HEPES (1 M) | Corning | 36216004 | Dilute 1:100 in DPBS |
Cell Recovery Solution | Corning | 354253 | |
Dulbecco's Phosphate Buffered Saline (DPBS) | HyClone | SH30028.02 | |
DMEM/F-12 | Thermo Fisher Scientific | 11320033 | |
Penicillin-Streptomycin (100X) | Life Technologies | 15140-122 | |
Gentamicin (50mg/mL stock) | Life Technologies | 15750060 | |
GDNF (10 µg stock) | Sigma | SRP3200 | |
L-Glutamine (200 mM stock) | Life Technologies | 25030-081 | |
Chicken anti-GFAP | Thermo Fisher Scientific | PA1-10004 | |
Goat anti-a-Smooth Muscle Actin | Abcam | ab112022 | |
Mouse anti-Pgp9.5 | Novus Biologicals | NB600-1160 | |
Goat anti-E-cadherin | R&D Systems | AF748 | |
Rabbit S100 | Abcam | ab34686 | |
Mouse p75 NTR | Millipore | MAB5592 | |
Alexa Fluor 488 Goat Anti-Chicken IgY | Invitrogen | A-11039 | |
Alexa Fluor 488 Donkey Anti-Goat IgG | Invitrogen | A-11055 | |
Alexa Fluor 568 Goat Anti-Mouse IgG | Invitrogen | A-11004 | |
Alexa Fluor 594 Donkey Anti-Rabbit IgG | Invitrogen | R-37119 | |
Prolong Gold antifade Reagent with DAPI | Thermo Fisher Scientific | P36931 | |
Fungizone (Amphotericin B) 250 µg/ml | Life Technologies | 15290-018 | |
L-Fura-2-AM | Invitrogen | F-14201 | |
CCK peptide | Anaspec, Fremont, CA | AS-20741 | |
Gastrin peptide (Gastrin-17) | Abbiotec, Bloomington, IN | 350188 | |
Nylon Mesh Celll Strainer (100 µm) | Fisher Scientific | 22363549 | |
Nylon Mesh Celll Strainer (40 µm) | Fisher Scientific | 22363547 | |
Disposable Serologic Pipet 5 ml | Fisher Scientific | 13-678-11D | |
0.25% Trypsin-EDTA (1X) | Life Technologies | 25200-056 |