Paneth Cells Are a Constitutive Source of IL-10 in Mouse Small Intestinal Organoids

Abstract

The healthy gut masks a dynamic balance between pro- and anti-inflammatory activities, largely due to microbial factors in the lumen. IL-10 is vital among the anti-inflammatory mediators, yet confirming constitutive versus stimulated secretion in any cell type is difficult due to the cellular complexity in the gut. Seeking to determine whether intestinal epithelial cells are programmed to constitutively make IL-10, we confirmed that IL-10 mRNA was present in enteroids from C57BL/6 mice and IL-10 protein was co-localized with a Paneth cell marker but not with markers for goblet or tuft cells. Paneth cells positive for IL-10 also possessed apical and basal IL-10RA, while cells negative for IL-10 had only basal IL-10RA, suggesting a possible autocrine role for IL-10. Indeed, Paneth cells in IL-10 gene knockout (IL-10KO) enteroids possessed lower levels of anti-microbial protein mRNAs, which could not be restored by adding IL-10. Enteroids passaged onto Transwell^® filters to form monolayers were treated with IL-10 and STAT3 phosphorylation was measured. Apically applied IL-10 resulted in a stronger STAT3 signal than basally applied cytokine. Our results indicate that a subpopulation of Paneth cells constitutively secrete IL-10 apically, which binds apical IL-10RA, impacting the expression of anti-microbial proteins unique to Paneth cells.

1. Introduction

A homeostatic state in the healthy gut is achieved by a number of physical barriers combined with active suppression of lymphocyte responses to microbial and ingested antigens. The gut epithelium, comprising multiple cell types aligned in a simple columnar epithelium, acts as the gate-keeper in this dynamic by providing a physical barrier, and also by secreting mucus (goblet cells), anti-microbial peptides (Paneth cells), complement (enterocytes) and by transporting immunoglobulins onto the lumen side of the monolayer [1]. Presumably, these secretory activities are stimulated through pattern recognition receptors, including toll-like receptors, formylated peptide receptors, and anaphylatoxin receptors, which enable the cells to respond directly or indirectly to challenges arising from the lumen [2]. Importantly, epithelial cells can be induced to secrete cytokines, some of which participate in homeostasis or launch inflammatory responses [3,4].

One cytokine discovered to be critical in maintaining gut homeostasis is IL-10. In fact, mice and humans lacking IL-10 or the IL-10 receptor will invariably develop colitis, and ileitis, in the case of gene knockout mice [5,6,7]. First discovered as a T lymphocyte cytokine and called cytokine synthesis inhibitory factor [8], IL-10 has since been reported to be secreted by multiple cell types. In the gut, macrophages stimulated by microbial products are a significant source of IL-10 [9,10], and IL-10 production by fibroblasts and cells within the epithelium has been reported [11,12]. Considering the cellular complexity of the gut and opposing anti-inflammatory versus inflammatory activities, it is difficult to prove whether IL-10 production is stimulated or constitutive in any cell type using in situ detection methods.

The gut epithelium is profoundly affected by IL-10, with IL-10 heightening epithelial integrity through multiple mechanisms [12], such as by promoting intestinal stem cell renewal [13] and the function of secretory cell types, including Paneth cells [14] and goblet cells [15]. Therefore, it is reasonable to hypothesize that epithelial cells may constitutively secrete IL-10, to bolster barrier properties and homeostasis. We sought to determine whether small intestinal epithelial cells constitutively produce IL-10 by examining cells in gut organoids, which are devoid of leukocytes and other parenchymal cells. Organoids have revolutionized intestinal epithelial cell research, since an organoid made from the small intestine (an enteroid) includes all the major polarized, differentiated cell types, in addition to dividing stem cells, something that has not been achieved by cell lines [16,17]. Using a combination of approaches, we show that a small population of Paneth cells in enteroids produce IL-10 and also possess the IL-10 receptor on the apical surface, implicating IL-10 in an autocrine loop in determining Paneth cell phenotypes.

2. Materials and Methods

2.1. Animals

Wildtype mice (C57BL/6J) and IL-10-gene knockout mice (B6.129P2-Il10tm1Cgn/J) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA), and then bred under specific pathogen-free conditions at the IWK Health Centre facility. Mouse genotyping was carried out with the KAPA Mouse Genotyping kit (Roche, Pleasanton, CA, USA) following the protocol provided by The Jackson Laboratory. IL-10GFP reporter mice of the C57BL genetic background, which have been described previously, were initially obtained from Dr. Richard Flavell (Yale University, New Haven, CT, USA), and then bred onsite [18,19]. For the establishment of enteroid cultures, primary tissues were isolated from male mice in the range of 6–14 weeks old. Experiments involving live animals meet the ARRIVE guidelines, and were approved by the institutional animal ethics committee, the University Committee on Laboratory Animals (protocol 20-012, 1 July 2020), which applies the guidelines of the Canadian Council on Animal Care.

2.2. Crypt Isolation and Organoid Cultures

Enteroid cultures were established using the protocol reported by Sato et al. (2009) with some modifications [16]. In brief, the entire small intestine dissected from a euthanized mouse was first flushed with chilled phosphate-buffer saline (PBS). The entire intestine was used to minimize region-specific sampling errors when taking material from the 2 strains of mice. The tissue was subsequently diced into ~2 mm pieces which were repeatedly washed with chilled PBS before incubation in 30 mL of ice-chilled 1 mM EDTA for 30 min, and then 30 mL of chilled 5 mM EDTA for 30 min. After the EDTA incubations, the tissue was resuspended in 25 mL of DMEM/F12 (Gibco, Waltham, MA, USA) and manually shaken to dissociate mucosal crypts, which were subsequently strained through a 100 μm cell filter for enriching crypts. The resulting crypt preparation was pelleted by centrifugation (130× g at 4 °C for 4 min) then resuspended in DMEM/F12, which was repeated. The crypts were finally seeded into wells of a 24-well plate at a density of 100–200 crypts per 50 µL basement membrane extract droplet (BME, comprising chilled reduced growth factor BME Cultrex^® (Trevigen, Gaithersburg, MD, USA) and DMEM/F12 in a 3:1 ratio). After 20 min at 37 °C, the droplets typically solidified, and 500 µL of IntestiCult™ (Stem Cell Technologies, Vancouver, BC, Canada) was added. The preparations were incubated at 37 °C with 5% CO₂. The media was changed on days 3 and 5 of culture. For the passaging of enteroids (on day 7 of culture, at a 1:5 ratio), droplets were dissolved and enteroids were dissociated into crypts by serial pipetting in ice-chilled DMEM/F12 + 5% FBS (Gibco) followed by centrifugation (300× g) and resuspension in 1 mL of chilled DMEM/F12, before being seeded back into a fresh 24-well plate in a BME droplet.

2.3. Cytokine Treatments of Enteroids

Enteroid media had 20 ng/mL murine recombinant IL-10 (Peprotech, Rocky Hill, NJ, USA) added, which was refreshed every second day. For experiments analyzed using Western blotting, a final concentration of 100 ng/mL of IL-10 in growth media was applied to enteroid cultures for 20 min.

2.4. RNA Isolation and RT-qPCR

To extract RNA from enteroids, 2 to 3 enteroid BME droplets were gently dissolved and then washed with ice-chilled PBS before being pooled. RNA was extracted using the ReliaPrep™ RNA Tissue Miniprep System (Promega, Madison, WI, USA) following the manufacturer’s instructions. RNA was assessed for purity and quantity using the NanoDrop™ (Thermo Scientific, Waltham, MA, USA) and stored at −80 °C. Complementary DNA (cDNA) was made from 1 µg RNA using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA), following the instructions provided with the kit. Subsequently, the PowerUp SYBR Green Master Mix (Applied Biosystems) was used for the amplification of the mRNAs of interest on a CFX96 Dx Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). The reaction temperature started with Uracil-DNA glycosylases’ activation at 50 °C for 2 min, followed by Dual-Lock Taq DNA polymerase activation at 95 °C for 2 min. Subsequently, for 40 cycles, the cDNA was denatured at 95 °C for 15 s, followed by an annealing step at 58 °C for 30 s. Additional reverse-transcription controls were run for the first replicate, and non-template controls were run for all replicates. The data were processed by the ΔΔCt analysis using the TATA-binding protein (TBP) signal as the internal standard, reported by Wang et al. (2010) to be the most stable qPCR reference gene for murine intestinal epithelial cells [20]. Primers were designed by Primer Blast to span an exon–exon junction, with Tm ranging from 59 °C to 62 °C, and containing 40–60% GC content. A list of primer sequences is provided in Supplemental Table S1.

2.5. Immunofluorescent Staining

An immunofluorescent staining protocol was adapted from O’Rourke et al. [21]. Specifically, enteroids in 20 µL BME droplets were cultured in chambers of a Millicell EZ SLIDE 8-well glass chamber (Millipore Sigma, Burlington, MA, USA) for 4 days. To prepare the enteroids for staining, the droplet was washed with ice-chilled PBS for 1 min, and then fixed in 300 μL chilled 4% paraformaldehyde in 1X PME buffer (10× PME buffer: 500 mM PIPES, 25 mM MgCl₂, 50 mM EDTA) for 20 min on a shaking platform. Subsequently, enteroids were permeabilized with ice-chilled 0.5% Triton X-100 in PBS for 20 min on the shaking platform. In experiments in which enteroids contained abundant dead cells in the lumen, the enteroids were incubated for 30 min in 100 mM glycine in PBS or Image-iT™ FX Signal Enhancer (Thermo Fisher, Waltham, MA, USA) at room temperature to quench autofluorescence by the dead cells. Enteroids were blocked with chilled 1% BSA in PBS for 30 min before being washed with washing buffer (0.2% Triton-X-100 and 0.05% Tween-20 in PBS) for 10 min. Primary antibodies were diluted in antibody signal enhancer (10 mM glycine, 0.05% Tween20, 0.1% Triton X-100 and 0.1% hydrogen peroxide in PBS) at a 1:50 ratio for the anti-IL-10 antibody (Thermo Fisher), 1:200 for anti-Lyz1 (Bioss, Woburn, MA, USA) and anti-Muc2 antibodies (Thermo Fisher). Enteroids were incubated overnight at 4 °C in the diluted primary antibody. Subsequently, enteroids were washed 3 times with washing buffer, for 10 min each time, before being incubated in a secondary antibody for 2 h at room temperature (1:400 for all secondary antibodies). A list of antibodies is provided in Supplemental Table S2. Enteroids were counterstained with 1 µg/mL DAPI in PBS for 5 min and washed 3 times with washing buffer, each time for 15 min, before mounting. To mount enteroids, chamber walls were detached from the slide and enteroids were mounted with Prolong™ Gold Antifade Mountant (Invitrogen, Carlsbad, CA, USA) under a coverslip which was sealed with nail polish. Stained enteroids were imaged using a LSM 710 confocal microscope (Carl Zeiss, Jena, Germany).

To stain monolayers, the Transwell^® insert was first detached using a scalpel then placed inside one of the chambers of a Millicell EZ SLIDE 8-well glass chamber (Millipore Sigma, Burlington, MA, USA). The monolayers were then processed for immunofluorescence following the protocol for 3D enteroids. Conjugated antibodies were diluted in antibody signal enhancer (10 mM glycine, 0.05% Tween20, 0.1% Triton X-100, and 0.1% hydrogen peroxide in PBS) at a 1:200 dilution for both occludin (Santa Cruz, Dallas, TX, USA) and KI67 (LifeSpan Biosciences, Seattle, WA, USA). Subsequent steps followed the immunofluorescence protocol for 3D enteroids.

2.6. Adapting Enteroids to Monolayer Cultures

We adapted a published protocol to achieve 2D epithelial cultures from enteroids [22]. Transwell^® inserts (6 mm, 0.4 μm pore size, Corning, Darmstadt, Germany) were treated with 40 µL of sterile type 1 collagen (Millipore) preparation consisting of 1 part 7.48 mg/mL collagen, 1 part 0.2% acetic acid, and 1 part 70% ethanol) and left to air-dry overnight under laminar flow (in a Biological Safety cabinet). High density enteroid cultures were established by dissociating 7-day enteroids into a suspension. The dissociated crypts were resuspended in fresh BME droplets up to the original volume of the source enteroids, and were then grown for 3 days before being dissociated again, this time including mechanical dissociation by passage through an 18-gauge needle. Prior to adding the cell suspension onto the Transwell^® insert, 500 µL of pre-warmed Intesticult™ was added to each Transwell^® bottom compartment. Three dissociated enteroid droplets were pooled together for each monolayer insert. Following an overnight incubation on the insert, the media was replaced with 200 µL of fresh Intesticult™ to remove unattached cells. Subsequently, the media was refreshed every 2 days.

2.7. Protein Extraction

Enteroids and monolayer cultures were pre-treated with cell-permeant phosphatase inhibitor (30µM phenylarsine in medium) for 15 min. Enteroids embedded in a BME droplet were gently washed with chilled PBS, then vigorously pipetted into chilled Cell Recovery Solution (VWR), before being left on a vortex for 20 min at 4 °C, to facilitate BME dissolution and enteroid dissociation. The dissociated enteroids were pelleted and washed in chilled PBS (200× g for 4 min each time); the pellet was then resuspended in 150 µL RIPA buffer (50 mM Tris HCl—pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 10 mM sodium fluoride, 1× freshly added Halt™ protease inhibitor cocktail (ThermoFisher), and freshly added phosphatase inhibitor cocktail (final concentration: 5 mM sodium orthovanadate, 10 mM sodium pyrophosphate decahydrate, and 10 mM β-glycerophosphate). To extract protein from monolayers after the phosphatase inhibitor pre-treatment, the Transwell^® insert was detached using a scalpel and immediately immersed in 150 µL of RIPA buffer. Lysis was conducted at 4 °C by vortexing the tube for 30 min, and the contents subsequently centrifuged at 13,000× g for 10 min, and the protein fraction was recovered in the supernatant. The protein/lysate was diluted in 4× SDS-PAGE loading buffer (25% glycerol, 12% SDS, 20% β-mercaptoethanol, 0.04% bromophenol blue in 150 mM Tris) (3:1 ratio, respectively) and boiled at 90 °C for 5 min before immediately being chilled on ice and stored at −80 °C.

2.8. Western Blotting

Proteins were separated by electrophoresis with Tris-Glycine running buffer (10× buffer contains 14% glycine, 3% Tris base, and 1% SDS in dH₂O) on either 8% SDS-PAGE gel for 2 h at 100 V or on gradient TGX Stain-Free™ Mini-PROTEAN precast gels (Bio-Rad Laboratories, Mississauga, ON, Canada) for 1 h at 120 V. Subsequently, protein was transferred from the gel to 0.45 µm low fluorescence PVDF membrane using a Trans-Blot Turbo RTA Transfer kit and Trans-Blot Turbo Transfer System. To enhance detection of low abundant proteins the membrane was incubated in SuperSignal™ Western Blot Enhancer (Thermo Fisher) for 30 min before incubation in blocking buffer TBS-T (20 mM Tris, 150 mM NaCl and 0.1% Tween^® 20 detergent) + 3% BSA) for 1 h. Subsequently, the membrane was incubated in a primary antibody (pSTAT3(TYR705), or STAT3, Cell Signaling Technology, Danvers, MA, USA) diluted 1:2000 in blocking buffer) overnight at 4 °C with gentle rocking. The membrane was washed 3 times in TBS-T, each time for 10 min with gentle rocking before being incubated with secondary HRP-linked goat anti-rabbit antibody for 1 h at room temperature (1:10,000 dilution, Cell Signaling Technology). The membrane was then washed 4 times with TBS-T, each time for 10 min, before being activated with either Clarity™ Western ECL Substrate (Bio-Rad Laboratories), or with SuperSignal™ West Atto ultimate Sensitivity Chemiluminescent Substrate (Thermo Fisher) for low-abundance protein. Proteins on membranes were imaged using a Bio-Rad ChemiDoc™ Touch Imaging System (Bio-Rad Laboratories, Mississauga, ON, Canada).

2.9. Statistical Analyses

Comparisons of 2 groups were conducted using unpaired two-tailed Student’s t-test. Comparisons of more than 2 groups were conducted first with one-way ANOVA, followed by Tukey’s post hoc test (1 independent variable), or two-way ANOVA (2 independent variables), followed by Sidak’s post hoc test. The significance of fold change was tested using a one-sample Wilcoxon test. All statistical tests and figures were created with GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). A p-value of less than 0.05 is considered statistically significant (* p < 0.05, ** p < 0.01, *** p < 0.001, ns—not significant). Specifically, for the IL-10 and IL-10RA transcript data, results were normalized to TBP and displayed as mean ± standard deviation (n = 3–4, one-way ANOVA followed by Tukey’s post hoc test). For cell type marker transcripts (LYZ1, MUC2, DLCK1, DEFA5, NOTCH1, ATOH1, HES1) all results were normalized to TBP and shown as mean ± standard deviation (n = 6–9, One way ANOVA followed by Tukey’s post hoc test). For pSTAT3 quantification in 3D enteroids, the results were normalized to STAT3 and shown as mean ± standard deviation (n = 3, ns - not significant, 2-tailed t-test). For pSTAT3 quantification in monolayers, the results were normalized to STAT3 signal and shown as mean ± standard deviation (n = 3, * < 0.05, ** < 0.01, ns—not significant, two-way ANOVA followed by Sidak’s post hoc test).

3. Results

3.1. Enteroid Culture Conditions and Intrinsic Expression of IL-10 and IL-10 Receptor mRNA

Enteroids provide a model recapitulating the intestinal epithelium, yet, because they are derived from whole intestine, risk being contaminated by leukocytes, particularly intraepithelial lymphocytes. It has been shown by others that screening for CD45 or CD3 is insufficient to confirm that leukocytes are absent, since these markers have been reported in data from single-cell mRNA sequences from intestinal epithelial cells [23]. Thus, we chose to work with enteroids after the first passage, at which point leukocytes are presumed to be lost. Budding commonly started as early as day 3, and most enteroids had budded by day 5 of culture. Notably, due to inevitable variations in growth factor gradients in the BME matrix, different enteroids within the same droplet were not necessarily homogenous in their developmental stages.

IL-10 and IL-10RA mRNAs were detected in WT enteroids at all time points tested, days 2, 4, and 6 of the culture, with the expression of IL-10 transcripts being significantly elevated in the day 4 cultures (Figure 1A). (We also used RT/PCR to confirm that our IL-10KO enteroids lacked mRNA for IL-10; our strategy is shown in Supplemental Figure S1). Based on the IL-10 mRNA data, we chose to analyze day-4 cultures more closely, on the presumption this is a crucial time point for differentiation and maturation of specialized epithelial cells. IL-10RA mRNA was detected, but levels were not significantly different over the culture period (Figure 1A). Since RT-qPCR results indicated the highest relative expression of IL-10 in day-4 enteroids, we used this period of development to investigate the source of IL-10 in WT and IL-10-GFP reporter enteroids using immunofluorescence. GFP fluorescence was undetectable in unstained IL-10-GFP reporter enteroids (bottom right panel in Figure 1B), possibly due to very low expression levels; therefore, the GFP protein was indirectly detected (and the signal was amplified) with an anti-GFP antibody followed by a suitably conjugated secondary antibody. GFP expression in IL-10-GFP reporter enteroids, and IL-10 protein in WT enteroids, were both detected on the apical side of cells in the crypt base (Figure 1B). To provide futher evidence that IL-10 is in the apical portion of cells in better resolution, we co-stained enteroids for IL-10 and phalloidin (stains apical F-actin, Supplemental Figure S2). IL-10RA protein was also detected in enteroids, in both WT and IL-10KO genotypes (we did not observe notable morphological differences between enteroids derived from WT and IL-10KO mice, Figure S3). While basal expression of IL-10RA can be seen at the crypt base of both WT and IL-10KO enteroids, apical expression was only detected in WT enteroids. An isotype control of the same host species (rabbit) was used as negative control, and no staining was detected (Figure 1B,C). Considering that IL-10 and IL-10RA are both expressed apically on cells found in the crypt base we hypothesized that there is a potential for autocrine activity in enteroids. We co-stained enteroids for IL-10 and IL-10RA and observed that both proteins indeed occurred in the same cell (Figure 1D). Interestingly, cells expressing apical IL-10RA also express basal IL-10RA, but are rarer in number compared to cells expressing only basal IL-10RA (Figure 1C,D), suggesting that this rare cell subtype with possible IL-10 autocrine activity may play a different role from the rest of the IL-10RA expressing cells present in the crypt base area.

3.2. Paneth Cells Produce IL-10 Apically for Autocrine Activity

Localized in the crypt, the cellular source of IL-10 is likely Paneth or stem cells, or possibly both. Using lysozyme (LYZ1) as the cellular marker for Paneth cells, we determined that IL-10 is expressed on LYZ1-positive cells in WT enteroids (Figure 2A). This finding was duplicated in enteroids from IL-10-GFP reporter mice; GFP was co-expressed with LYZ1, confirming the production of IL-10 in Paneth cells (Figure 2A). We detected 0–3 IL-10-positive cells per organoid. To rule out other differentiated cell types possibly expressing IL-10, IL-10 detection was combined with mucin-2 as marker for goblet cells, or doublecortin-like kinase 1 (DCLK1) as a marker for tuft cells. We determined that the epithelial IL-10 signal did not overlap with either of these cell types, indicating IL-10 expression is limited to Paneth cells (Figure 2A). Notably, only a very small number of LYZ1-positive Paneth cells were found to be positive for IL-10, suggesting that IL-10 production is perhaps limited to a particular developmental stage of Paneth cells, or a specialized subset of the Paneth cell (Figure 2B).

The cell type bearing the IL-10RA was also investigated using the cell-type specific markers. As predicted, apical IL-10RA was found on LYZ1-positive Paneth cells (Figure 2C). Contrasting the Paneth cell-specific staining of apical IL-10RA, basal IL-10RA was not limited to the crypt base, as scattered patches of basal IL-10RA staining were detected at regions outside the crypts. Some of the basal IL-1RA signal was localized to goblet cells, but not to tuft cells (Figure 2C).

3.3. Paneth Cell Markers Are Reduced in the Absence of Epithelial IL-10

Since IL-10 was found to be heightened in day-4 enteroids, we decided to investigate the effect of epithelial IL-10 upon cell types one day after the peak of IL-10 expression, to allow time for the cellular targets to respond to the endogenous IL-10. mRNA for anti-microbial proteins specific to Paneth cells, LYZ1 and Defensin alpha 5 (DEFA5), as well as mucin 2 (MUC2) for goblet cells, were measured in day-5 WT and IL-10KO enteroids. Both LYZ1 and DEFA5 mRNAs were significantly less in IL-10KO enteroids compared to WT enteroids (Figure 3A). Harvesting the entire enteroid, we cannot confidently determine whether the deficiency is in all Paneth cells or only the few that would be expressing apical IL-10RA, but, if it is limited to a subset of Paneth cells, then the deficiency is quite profound considering that much of the signal would be from cells unaffected by the loss of IL-10 signaling. To rule out a secretory cell-wide deficiency in IL-10KO enteroids, we measured MUC2 mRNA levels and found no significant difference between IL-10KO and WT enteroids. These results suggested that epithelial endogenous IL-10 likely plays a regulatory role specific to Paneth cell development.

We next tried to reverse the observed deficiencies of Paneth cell markers in IL-10KO enteroids by supplementing IL-10KO enteroids with recombinant IL-10. Recombinant IL-10 was added to IL-10KO enteroid cultures to a final concentration of 20 ng/mL beginning on day 1 of culture and replenished at each media change; however, IL-10 had no significant effect on the markers in IL-10KO enteroid (Figure 3A). This result indicated that stimulation with IL-10, at least through basal receptors, does not affect the cellular markers of Paneth cells (nor goblet cells) in IL-10KO organoids, recalling that cells in IL-10KO enteroids lacked apical IL-10RA.

To gain further insight into the developmental stage that epithelial IL-10 could be acting on, we examined cellular lineage markers that determine the balance between secretory and absorptive progenitors. NOTCH1 and HES1 induce absorptive progenitors, i.e., enterocytes, while ATOH1 induces the emergence of secretory progenitors such as goblet and Paneth cells [24]. We found no significant differences in the mRNA levels of these lineage markers between WT and IL-10KO enteroids, suggesting that IL-10 does not affect enteroids at this early lineage-determining stage (Figure 3B).

3.4. STAT3-Dependent IL-10 Signaling Occurs upon Stimulation of Apical IL-10RA Receptors

The cell-dependent polar expression of IL-10RA on the epithelium raised the possibility that different signaling pathways are activated through the receptors on either side of the cell. In many cells, IL-10RA predominantly signals through the phosphorylation of STAT3 [25,26]; as such, IL-10 signaling in WT and IL-10KO enteroids was examined by measuring phosphorylated STAT3 (pSTAT3) relative to unphosphorylated STAT3 in the same lysate. It was first noted that untreated WT and IL-10KO enteroids already have detectable pSTAT3 (Figure 4A). From there, over multiple repeats, levels of pSTAT3 in IL-10-treated enteroids were similar between WT versus IL-10KO genotype and untreated enteroids (Figure 4A). The effect of apical stimulation alone remained to be teased out.

To gain access to IL-10 receptors on both sides of the epithelium, we established 2D enteroid monolayers on Transwell^® filters (Figure 4B). Our 2D enteroid monolayer model reached 100% confluency with fully developed tight junctions (shown with greater occludin staining, Figure 4B) after 20 days of culture. However, because day-20 monolayers have few Ki67+ cells (Figure 4B), which could confound results compared to day-5 enteroids in the 3D culture, we conducted the IL-10 treatment experiment as soon as the monolayer reached confluency, despite the possibility that tight junctions may have been incompletely developed (Figure 4B). Nevertheless, upon apical exposure of monolayers to IL-10, the relative level of pSTAT3 in the WT monolayers was two-fold higher than that of the untreated monolayers prepared from the same batch of enteroids. This change was significantly higher than the pSTAT3 level observed in WT monolayers upon basal stimulation. No increase in STAT3 phosphorylation was observed in IL-10KO monolayers following either basal or apical stimulation with IL-10.

4. Discussion

Although IL-10 has long been considered the key regulator of intestinal epithelial, and therefore intestinal, homeostasis, the role of epithelial-derived IL-10 is not completely understood. Using an enteroid model, we were able to exclude leukocyte-derived sources and focus on epithelial-derived (endogenous) IL-10, while maintaining the spatial and temporal dynamics of the native epithelium.

Cells within enteroids constitutively express IL-10 and IL-10RA throughout enteroid developmental stages, detectable as mRNA and protein. We further show that some Paneth cells possess both apical IL-10 and apical IL-10RA. Given that Paneth cells are long-lived cells (>30 days in vivo) and undergo complex changes before reaching maturation [27], early enteroid cultures might only capture a snapshot of IL-10 and IL-10RA expression patterns in the Paneth cell population. The increase in IL-10 transcripts in day-4 cultures could be explained by a number of hypotheses, and the limited number of Paneth cells expressing IL-10 is compatible with expression being restricted to a specific developmental stage. The “budding” events in day-4 cultures indicate a transitional phase from proliferation to differentiation, and IL-10 production could be linked to this temporary phase and therefore be enhanced in the process. Alternatively, IL-10-positive Paneth cells might accumulate associated with new crypt formation. Finally, others have shown that Paneth cells are phenotypically diverse, depending on the region in the gut [23]. We grew enteroids from whole murine small intestine; IL-10-positive Paneth cells in our cultures could constitute a particular regional-specific subpopulation of Paneth cells, an idea that could be confirmed by growing enteroids from limited regions of the small intestine.

Regarding the impact of local IL-10 expression, we detected reduced expression of antimicrobial products LYZ1 and DEFA5 in IL-10KO compared to WT enteroids. Although reduced LYZ1 expression was not observed in vivo, the packing of lysozyme granules in Paneth cells of IL-10KO mice were reported to be abnormal compared to Paneth cells in WT mice [14,28,29]. Our finding of Paneth cell antimicrobial deficiencies in IL-10KO enteroids is compatible with those reports, and any specific differences may be due to the sensitivity of detecting products in enteroids versus whole intestine or other extrinsic factors in mice. Interestingly, the addition of IL-10 to either WT mice or WT ileum enteroids was shown to induce production of LYZ1 [30]. In our study, the addition of IL-10 to IL-10KO enteroids was shown to be ineffective at raising Paneth cell markers. This discrepancy is compatible with the lack of apical IL-10RA expression by cells in enteroids from IL-10KO mice, and, in vivo, other cell types may respond to the added IL-10 and indirectly provide signals to Paneth cells. Another consideration is that IL-10KO enteroids may lack or have altered expression of the second IL-10 receptor polypeptide, IL-10RB, which we did not characterize.

Although the lack of epithelial IL-10 was shown to affect the production of Paneth-cell-product mRNAs in our study, how IL-10 drives this physiological change remains undetermined. Our investigation of the lineage balance between secretory and absorptive progenitors showed insignificant differences between WT and IL-10KO enteroids. However, this observation does not rule out the impact of IL-10 on the upstream pathway of Paneth cell determination or development. It has been suggested that IL-10 may play a pivotal role in reaching complete cellular maturation, as IL-10KO mice were found with three-fold more abundant intermediate cells which co-express both Paneth cell and goblet cell featuressimilar to WT enteroids [14]. One candidate upstream-signaling molecule that could have such an impact is NF-κB. Indeed, epithelial-derived IL-10 has been shown to induce NF-κB [31], the activity of which was shown, in another study, to determine the cell fate of Paneth cells from goblet cells via the activation of the SOX9 transcription factor [32]. Perhaps the reason we did not see any changes in the levels of NOTCH1, HES1 or ATOH1 is because, in the process of Paneth cell fate determination, these markers regulate the early stage of secretory-versus-absorptive lineage commitment, which occurs upstream from Paneth cell-determining SOX9 activation [33]. Future examinations of the relationship between IL-10, NF-κB signaling, and Paneth cell development could shed light on this speculation.

Alternatively, the observed deficiency of Paneth cell markers in IL-10KO enteroids could be a result of proliferation-versus-differentiation imbalance, which aligns with previous reports of hyper-proliferation in IL-10KO crypts in vivo [30]. Addition of IL-10 to WT enteroids was shown to shift the cellular dynamics away from proliferation and towards differentiation [30]. Whether such effects were attributed to epithelial-specific sources of IL-10 or other exogenous sources remain undefined. It is noteworthy that our findings for Paneth cell IL-10 and apical IL-10RA are made in the absence of any microbes, implying that the events are intrinsic to the epithelium.

The polarity of IECs is critical for epithelial function, with the apical membrane facing the microbes and antigens in the lumen, in contrast to the basal membrane. As a result, the signaling direction across the epithelium is likely key to maintaining the balance of two seemingly paradoxical tasks: facilitating immune surveillance though antigen-sampling, while maintaining immune tolerance. It is therefore reasonable to predict that the same type of receptor on each polar extreme may have different signaling properties linked to different outcomes. We are not the first to report IL-10RA on the apical side of epithelial cells (although we show it among cells in enteroids), nor have we discovered a precedent showing that there may be differences in signaling by the same receptor from opposite sides of an epithelial cell [34,35]. The signaling of IL-10, through STAT3, was addressed on both sides of the epithelium using enteroid monolayers, after we observed that IL-10 did not induce STAT3 phosphorylation in 3D enteroids. We determined that STAT3-dependent signaling of IL-10 is more potent upon stimulation of apical, compared to basal, receptors in WT enteroid monolayers, despite our immunofluorescent data indicating that more cells possess basal IL-10RA. We entertain this as evidence of apical versus basal IL-10RA exerting a differential response, which presumably differentially impacts Paneth cell physiology. Our finding of greater STAT3 phosphorylation upon apical IL-10R stimulation is congruent with a past study, which reported increased STAT3-dependent IL-10 signaling associated with upregulation of apical IL-10R [36]. Interestingly, unlike WT enteroid monolayers, IL-10KO enteroid monolayers showed no significant STAT3 phosphorylation (over background) upon IL-10 addition, regardless of the direction of stimulation, which is compatible with the lack of apical IL-10 receptors in IL-10KO enteroids. It is noteworthy that WT and IL-10KO enteroids possess considerable constitutive pSTAT3, possibly due to other intrinsic STAT3-activators such as IL-6 or IL-15, cytokines reportedly produced by intestinal epithelial cells [3,37]. A limitation in our monolayer STAT signaling experiments was our presumption that differentiated cell type phenotypes were similar to cells in organoids [22].

In addition to Paneth cells, we detected IL-10 receptors on goblet cells. Unlike on Paneth cells, where the receptors were expressed on both polls, in goblet cells the receptors were limited to the basal side, suggesting that goblet cells are targets for IL-10 from non-epithelial cell sources. While several studies have shown a detrimental impact of IL-10/IL-10R deficiency on goblet cells, it remains elusive whether the impact is directly due to the ablation of IL-10/goblet cell crosstalk or indirectly due to changes in immune elements due to the absence of systemic IL-10. One study showed that IL-10 could directly act on goblet cells, mediating the protein-folding of MUC2 and endoplasmic reticulum homeostasis [38]. Moreover, recent studies have demonstrated that IL-10 is an important regulator of goblet cell homeostasis via STAT3-independent Notch1 signaling, as disruption of IL-10 signaling skewed the lineage toward excessive goblet cell numbers while diminishing the absorptive lineage [39,40]. Our results indicate that none of these processes are susceptible to apical IL-10 receptors on goblet cells.

5. Conclusions

IL-10 has been shown to play versatile roles in epithelia homeostasis, yet our understanding of any constitutive epithelial sources of IL-10 is incomplete. We identify that a subset or stage-specific population of Paneth cells make IL-10 in the absence of non-epithelial cell types and microbial factors. Our evidence points to apical secretion by IL-10-producing Paneth cells, which acts in an autocrine fashion to influence Paneth cells’ armament of anti-microbial factors.