IL-6 Signaling Regulates Small Intestinal Crypt Homeostasis

The intestinal epithelium is the most rapidly renewing tissue in the body, with the entire epithelium being replaced every 5–7 d. This renewal takes place by way of Lgr5-positive stem cells located at the base of intestinal crypts; stem cells proliferate, migrate along the crypt-villus axis, differentiate (into Tuft cells, enteroendocrine cells, Paneth cells, enterocytes, and goblet cells), and are shed into the gut lumen (1). Through the secretion of Wnts, epithelial Paneth cells play a major role in the maintenance of the crypt stem cell niche (2). Previous work has also shown that other growth factors and cytokines, as well as immune cells are key in modulating epithelial stem cell–driven tissue renewal during homeostasis (3–8). Understanding the mechanisms by which these pathways are regulated in the epithelium through autocrine (and paracrine) signaling is not fully understood.

Seminal work in the Drosophila gut has shown regenerative responses following infection are regulated by JAK/STAT signaling in gut epithelial stem cells through the release of enterocyte-derived Upd3, an IL-6–like cytokine (9, 10). In the mammalian gut both IL-6 and STAT3 have been shown to play a role in the proliferation of the colonic epithelium following injury and to promote the survival of epithelial cells (11–14) during inflammation and inflammatory bowel disease (15, 16).

IL-6 is a pleiotropic cytokine involved in a plethora of cellular and immune responses in health, disease, and cancer (17–19). IL-6 signaling involves the convergence of a number of signaling components (20). IL-6 first binds to a membrane-bound nonsignaling α-receptor IL-6 (mbIL-6R) located on the target cell. Next, this IL-6R/IL-6 complex binds to the ubiquitously expressed type I transmembrane transducer protein gp130, which results in activation of downstream signaling components JAK/STAT, ERK, and PI3K signaling pathways. Cells that express both the IL-6R and gp130 are responsive to IL-6; this is termed classic signaling and is traditionally associated with homeostasis. IL-6 can also signal via trans signaling, where the soluble IL-6 receptor, shed from the cell membrane via proteolytic cleavage of a membrane-bound precursor, binds to IL-6. This IL-6/soluble IL-6 receptor complex can then activate IL-6 signaling in any cell expressing gp130; this trans-signaling pathway is associated with inflammation and cancer (21).

The aim of this study was to determine whether IL-6 classic signaling could modulate small intestinal crypt homeostasis. This work demonstrates a previously unidentified role for autocrine IL-6 signaling in the maintenance of the crypt stem cell niche, through the differential expression of the IL-6 receptor and downstream STAT3 signaling in Paneth cells and the Wnt signaling pathway.

LGR5-EGFP-Ires-CreERT2 (Jackson Laboratory) or C57BL/6 mice aged 8–12 wk were used. Generation and genotyping of the LGR5-EGFP-Ires-CreERT2 allele has been described previously (1). All animal experiments were conducted in accordance with the Home Office Animals Scientific Procedures Act of 1986 with approval of the University of East Anglia Ethical Review Committee, Norwich, U.K. and under Home Office project license number 80/2545. Blocking Abs for IL-6 and IL-6 receptor or IgG controls (Bio X Cell) were administered to mice three times on alternate days by i.p. injection at a concentration of 58 μg/g. Animals were euthanized by Schedule One approved methods on day 6, and tissue processed immediately. In addition, small intestinal tissue from IL-6 knockout mice (B6.129S2-Il6tm1Kopf/J; Jackson Laboratory) was used to count the number of lysozyme-positive cells per small intestinal crypt.

Advanced DMEM/F12, GlutaMAX, B27, and N2 were purchased from Invitrogen. Murine recombinant epidermal growth factor, noggin, IL-6, and IL-22 were all obtained from PeproTech and mouse recombinant R-spondin 1 from R&D Systems. Growth factor–reduced Matrigel was purchased from VWR International.

Primary and secondary Abs were as follows: rat anti-BrdU (Abcam), mouse anti-lysozyme (Abcam), goat anti–E-cadherin (R&D), rabbit IgG, rabbit anti–IL-6 receptor, anti–IL-6 (Bio X Cell), rabbit anti-pSTAT3 Tyr705, pSTAT3 Y705 blocking peptide, and rabbit anti–cleaved caspase-3 (Cell Signaling). Immunolabeling was visualized by using an appropriate combination of species-specific Alexa Fluor–conjugated secondary Abs (488, 568, and 647 nm), raised in mouse, donkey, or goat (Invitrogen). FITC-conjugated Ulex europaeus lectin (UEA-1) was purchased from Sigma. Vectashield-mounting medium with DAPI was purchased from Vector Laboratories; STATTIC and IWP2 were purchased from Tocris Bioscience.

Small intestinal crypts were isolated from the proximal small intestine of LGR5EGFP mice as previously described (2). Briefly, the mouse small intestine was opened longitudinally, washed with PBS, cut into 2–4 mm pieces and incubated with 1 mM EDTA in PBS (pH 7.4) for 30 min at 4°C. Crypts were liberated by serial rounds of pipetting in ice cold PBS and removal of the crypt enriched supernatant; the solution was then filtered through a 70 μm cell strainer followed by centrifugation. Next, 50–100 crypts were embedded in a 200 μl droplet of growth factor–reduced Matrigel (VWR) and seeded on No. 0 coverslips (VWR) contained within a 12-well plate (Nunc). After polymerization at 37°C for 5–10 min, crypts were flooded with 0.5 ml of mouse crypt culture medium (MCM): advanced F12/DMEM containing B27, N2, n-acetylcysteine (1 mM), HEPES (10 mM), penicillin/streptomycin (100 U/ml), GlutaMAX (2 mM), epidermal growth factor (50 ng/ml), noggin (100 ng/ml), and R-spondin 1 (1 μg/ml).

In IL-6–neutralizing or IL-6 receptor–blocking Ab experiments, crypts were incubated for the entire 48 h period in MCM and appropriate Ab or respective IgG control before addition of BrdU (1 μM) for 17 h.

Following 48 h culture with MCM, crypts were stimulated with IL-6 at concentrations of 10 ng/ml to 1 μg/ml for 15 min to 17 h for pSTAT3 experiments, and for proliferation experiments placed into 10% MCM for 5 h, after which BrdU (1 μM) was added for 17 h. For STAT3 inhibition studies, crypts were preincubated with STATTIC (20 μM) or IWP2 (5 μM) for 1 h prior to addition of IL-6 (100 ng/ml) stimulation for 17 h (as described previously).

For pSTAT3 activation studies, the IL-6–neutralizing Ab was preincubated with IL-6 (100 ng/ml) for 1 h before addition to the crypts for 1 h. For receptor Ab studies, crypts were preincubated for 1 h with the IL-6 receptor–blocking Ab before addition of IL-6 (100 ng/ml) for 1 h. The same protocol was used for proliferation experiments, prior to the addition of IL-6 and BrdU (1 μM) 17 h.

For chronic IL-6 stimulation, crypts were incubated for 5 d with MCM containing 100 ng/ml IL-6 before addition of BrdU (1 μM) for 17 h; medium was changed every 2 d.

Mouse Lgr5EGFP small intestines were fixed in 4% paraformaldehyde, frozen in isopentane, and stored at −20°C. Following the culture period, small intestinal crypts were fixed with 4% paraformaldehyde. Next, 8–20 μm cryosections of small intestinal tissue (CM 1100C cryostat; Leica) or small intestinal crypts were permeabilized with 0.5% Triton–X (Sigma), blocked with 10% FBS, and stained with primary Abs diluted (1:100) in PBS overnight at 4°C, followed by the corresponding secondary fluorescence-conjugated Abs (1:200 in PBS) for 2 h as previously described (3, 4, 22).

Following immunofluorescent staining, samples were visualized by laser scanning confocal (510 META; Zeiss) or epifluorescence (Nikon Ti) microscopy. For confocal microscopy a ×40 1.3 NA oil objective or ×63 1.4 NA 0.75 mm WD oil immersion objective was used to obtain confocal images. Image stacks were taken at 1–3 μm intervals, which allowed selection of precise focal planes. A ×40 1.1 NA oil objective was used to image samples on the epifluorescent microscope.

For all experiments, between 20 and 100 crypts (in vivo) or organoids (in vitro) were counted per condition and each experiment was performed at least three separate times (i.e., at least three different mice were used).

ImageJ software was used to count the total number of DAPI-labeled cell nuclei in the crypt equatorial plane. The number of cells that were positive for BrdU or Lgr5EGFP was then expressed as a percentage of the total nuclei number. For semiquantitative analysis of pSTAT3 activation, regions of interest were drawn around the entire Paneth cell nuclei and the average fluorescence intensity of the region calculated.

Lgr5EGFP stem cells, lysozyme-, or UEA-1–positive Paneth cells were counted by taking a z-stack across the equatorial plane of the crypt and counting the number of nuclei in this stack that were positive for each marker. To determine the crypt length and villus height in vivo, at least 25 well-orientated crypts and villi were measured using the segmented line tool (Fiji 1.50e) from three animals in each treatment group.

RNA was isolated from freshly liberated or cultured small intestinal crypts using the Isolate II RNA mini kit (Bioline). Total RNA yield was determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific) and purity assessed by the ratio of absorbance at 260–280 nm. cDNA was generated from 0.5 μg RNA using the RT² First Strand cDNA synthesis kit (Qiagen).

End-point PCRs were performed in a final volume of 25 μl, comprising 200 nm forward and reverse primers, 200 μM dNTP, 0.5 U Taq polymerase, PCR buffer (Roche), 2.5 mM MgCl₂, 0.5 μl cDNA using a Tprofessional TRIO thermal cycler (Biometra) with the following thermal profile: one cycle at 94°C for 3 min, 30 cycles at 94°C for 25 s, 58°C for 30 s, and 72°C for 50 s, and 1 cycle at 72°C for 5 min. The primer sequences used are as follows; IL-6 forward primer: 5′-GCTACCAAACTGGATATAATCAGGA-3′; reverse primer: 5′-CCAGGTAGCTATGGTACTCCAGAA-3′. PCR products were run on a 2% agarose gel visualized by ethidium bromide staining.

Experiments were performed at least three times. Data are expressed as mean ± SE and significance determined by a Student t test or by one-way ANOVA with post hoc Tukey analysis, with p values < 0.05 considered significant.

First we determined the effect of exogenous IL-6 on the BrdU incorporation of small intestinal organoids through immunofluorescence microscopy (Fig. 1A). IL-6 caused a significant increase in the incorporation of BrdU at 10 and 100 ng/ml IL-6, but exhibited a dome-shaped response curve at higher concentrations (1000 ng/ml) (Fig. 1B). Using the specific STAT3 inhibitor STATTIC, we demonstrated that the increased BrdU incorporation induced by IL-6 was significantly reduced compared with IL-6 alone (Fig. 1C). There was no effect of STATTIC on basal proliferation, which was in contrast with the effects of chronic treatment with STATTIC, where a significant decrease in BrdU incorporation compared with control was observed (Supplemental Fig. 1A). Organoid survival following STATTIC treatment in the presence or absence of IL-6 was comparable to control (DMSO control 74.9 ± 3.9; STATTIC 73.7 ± 4.5; IL-6 63.1 ± 4.9; STATTIC + IL-6 69.5 ± 4.7). However, there was a small but significant (n = 3, **p < 0.01) effect of STATTIC on the percentage of caspase 3–positive cells per crypt compared with control (DMSO control 3.6 ± 0.4; **STATTIC 6.0 ± 0.7; IL-6 2.2 ± 0.4; STATTIC + IL-6 4.8 ± 0.7). A significant increase in BrdU incorporation was observed following chronic IL-6 stimulation (Supplemental Fig. 1B).

We next investigated the effect of IL-6 on pSTAT3 activation in the small intestinal epithelium using immunolabeling and visualization of pSTAT3 activation by confocal microscopy. Following 1 h stimulation with IL-6 (100 ng/ml), we observed pSTAT3 (green) in the nucleus (red) of UEA-1– (pink) and lysozyme-positive (blue) cells, which was abrogated by a blocking peptide to pSTAT3 (Fig. 2A). No fluorescent labeling was observed in IgG controls for the lysozyme Ab (Supplemental Fig. 1C). Following IL-6 stimulation, all UEA-1 lysozyme-positive cells in a crypt domain were positive for nuclear pSTAT3 (see projection Fig. 2A). All lysozyme-positive cells were UEA-1 positive and characteristic of Paneth cells (23). Although the majority of control organoids expressed no pSTAT3 (Fig. 2B), nuclear pSTAT3 activation in UEA-1–positive cells was observed in 9 out of 113 control organoids counted. Time course studies showed that maximal activation of nuclear (red) pSTAT3 (green) in UEA-1–positive (pink) Paneth cells occurred at 30 min, and persisted at 1 and 3 h poststimulation (Fig. 2B). Semiquantitative analysis showed the level of nuclear pSTAT3 activation remained significantly higher than control levels 1 and 3 h after IL-6 stimulation. Results also showed that pSTAT3 phosphorylation levels at 1 h were the same for 100 or 1000 ng/ml IL-6 (Fig. 2C).

The localization of the IL-6 receptor (red) in organoids was restricted to the basal membrane of lysozyme-positive (green) cells of the small intestine in vitro (Fig. 2D). Furthermore, in vivo labeling of the IL-6 receptor displayed the same distribution as in the in vitro crypt organoid cultures (Fig. 2E), with basal epithelial IL-6 receptor expression constrained to the lysozyme-positive cells (green) and on the membrane of their intracellular granules (Fig. 2F) with no fluorescence observed in the IgG control for lysozyme (Supplemental Fig. 2D). No IL-6R was expressed on Lgr5EGFP-positive stem cells (cyan). The small intestinal crypt in vitro model was therefore considered a relevant in vitro model to study IL-6 signaling in the small intestinal crypt epithelium.

Next, we demonstrated the ability to modulate the IL-6 STAT3 signaling axis in small intestinal organoids utilizing a receptor-blocking Ab to IL-6 or neutralizing Ab to IL-6. Either of these Abs in combination with IL-6 abrogated the IL-6–induced phosphorylation of STAT3 (Supplemental Fig. 1J–M).

The next step was to determine whether autocrine IL-6 signaling played a role in crypt homeostasis. We first used conventional PCR to demonstrate the expression of IL-6 mRNA in crypts in vivo using freshly isolated (0 h) and in vitro cultured (48 h) small intestinal crypts (Fig. 3A). Immunofluorescent studies showed that IL-6 protein (red) was localized inside epithelial cells toward the basal pole as well as extracellularly in discrete pools around the crypt basal membrane (white arrows, Fig. 3B). We next used a blocking Ab to the IL-6 receptor or a neutralizing Ab to the IL-6 cytokine in crypt organoid cultures and assessed basal BrdU incorporation (Fig. 3C, 3D respectively), and observed a significant decrease in the percentage of crypt nuclei incorporating BrdU compared with the respective IgG controls with both Abs.

We also confirmed that the Abs could functionally block IL-6 signaling in organoids. Incubation of either an IL-6 receptor–blocking or an IL-6 neutralizing–Ab with IL-6 significantly decreased BrdU incorporation of organoids compared with IL-6 stimulation alone (Supplemental Fig. 1D, 1E). Caspase 3 (used as a measure of cell death), and organoid survival was unaffected by either of these Ab treatments compared with IgG control organoids (Supplemental Fig. 1F–I).

To understand the effects of perturbation of the IL-6 signaling pathway in vivo, tissue was obtained from IL-6 knockout mice. No difference was observed in the number of lysozyme-positive cells per crypt, crypt length, and the number of nuclei per crypt in IL-6 knockout tissue compared with wild-type tissue (Supplemental Fig. 2A–C). Furthermore, using immunolabeling, low levels of pSTAT3 were detected in some UEA-1–positive cells at the base of the crypts in both wild-type and IL-6 KO small intestine (Supplemental Fig. 2J, 2K). We confirmed the specificity of the pSTAT3 Ab labeling by using a blocking peptide to pSTAT3, which abrogated the pSTAT3 signal in the UEA-1–positive cells (Supplemental Fig. 2I).

As intestinal stem cells drive the renewal of the crypt epithelium, we examined the effects of modifying the IL-6 signaling axis on crypt stem cells both in vitro and in vivo. In addition to the observed increase in the small intestinal organoid proliferation in vitro, we also demonstrated that the percentage of Lgr5EGFP-positive nuclei per organoid (Fig. 4A) was significantly increased compared with control. Furthermore, the average number of new buds per small intestinal organoid, also a measure of stem cell numbers, was significantly increased compared with control, an effect that was abrogated by STAT3 inhibition with STATTIC (Fig. 4B). The use of an IL-6R blocking Ab also caused a significant reduction in the budding of small intestinal crypts (Fig. 4C). Complementary in vivo studies using Lgr5EGFP mice treated with the IL-6 receptor–blocking Ab caused a significant reduction in the number of Lgr5EGFP cells per crypt, the number of crypt nuclei, crypt length, and villus height compared with IgG-treated mice (Fig. 4D–G respectively). Similar findings were obtained after mice were treated with an IL-6–neutralizing Ab with a significant reduction in the number of Lgr5EGFP cells per crypt, the number of crypt nuclei, and villus height compared with IgG-treated mice (Fig. 4H, 4I, 4K). No significant difference was observed in crypt length between mice treated with a IL-6–neutralizing Ab or control IgG (Fig. 4J). Furthermore, both the IL-6 receptor–blocking Ab and the neutralizing IL-6 Ab caused a significant reduction in the number of lysozyme- (Fig. 4L, 4N) and UEA-1–positive (Fig. 4M, 4O) cells per crypt compared with the respective IgG-treated mice. Representative confocal images from mouse in vivo Ab experiments including Lgr5EGFP, lysozyme, UEA-1, Ki67, and caspase 3 labeling are included in Supplemental Fig. 2E–H.

Paneth cells have been shown to be a major source of Wnts and the Wnt signaling pathway is a master regulator of epithelial renewal. We next determined whether IL-6 was affecting epithelial renewal or proliferation and stem cell numbers in vitro through this pathway by utilizing the Wnt inhibitor IWP2. We show that IWP2 abrogates the IL-6–induced increase in both proliferation and the average number of buds per organoid compared with control (Fig. 5). The percentage of caspase 3–positive cells and organoid survival in the presence of IWP2 and/or IL-6 was comparable to control (Supplemental Fig. 4).

This work highlights a previously unrecognized role for autocrine IL-6 signaling during homeostasis in the small intestine. IL-6 stimulation was shown to increase in vitro crypt cell proliferation and stem cell numbers through the STAT3 signaling pathway. IL-6 activated pSTAT3 specifically in the nucleus of the Paneth cell through differential expression of the IL-6 receptor located on the Paneth cell basal membrane. IL-6 was also expressed in the crypt epithelium. Neutralizing IL-6 or blocking the IL-6 receptor with Abs lowered basal organoid proliferation and crypt budding in vitro and reduced crypt stem and Paneth cell numbers in vivo. Use of the Wnt inhibitor IWP2 abrogated the IL-6–induced increase in organoid proliferation and crypt budding. These data suggest a role of autocrine IL-6 signaling and the Wnt signaling pathway in the regulation of crypt homeostasis.

Previous work using IL-6 receptors or IL-6 knockout mice demonstrated a role for IL-6 signaling in mouse models of inflammation and crypt epithelial cell survival and regeneration rather than homeostasis (11, 12, 14, 16, 24, 25). In agreement with other studies, we also showed that gut tissue from healthy IL-6 knockout mice was similar to tissue from wild type mice (Supplemental Fig. 2A–C). However, using a blocking Ab approach we demonstrated a previously unrecognized role for autocrine IL-6 signaling in the maintenance of crypt stem cell numbers (and Paneth cell numbers; see below) in vivo. These differences may be explained by redundancy effects known to occur in knockout mouse models, and highlights the importance of interrogating this pathway using a variety of methods as has been shown in other studies (26).

Other groups have also implicated a role for IL-6 in maintaining homeostasis (27–29) as well as ageing (30–32). Our findings suggest that classic signaling is involved in crypt homeostasis; we show that addition of IL-6 to the basal side of crypts in vitro causes pSTAT3 activation in Paneth cells only, which correlates with the expression of the IL-6 receptor on the Paneth cell basal membrane. It has been suggested that anti-inflammatory and regenerative properties of IL-6 most likely depend on IL-6 classic rather than trans signaling (20, 33). Our studies concur with this; if binding of soluble IL-6 receptor to the ubiquitously expressed gp130 was occurring in our system then this would permit all cells of the crypt epithelium to express nuclear pSTAT3 in response to IL-6, an effect that we did not observe. The relative contribution of classic IL-6 signaling versus trans signaling in crypt homeostasis versus inflammation remains the focus of future work.

Our findings demonstrated that IL-6 mediated its in vitro effects on crypt cell proliferation and stem cells through the pSTAT3 pathway. Crypt cell proliferation in response to IL-6 exhibited a bell-shaped concentration response curve, which has previously been described (34). The epithelial STAT3 signaling pathway has shown to be key for small intestinal stem cell survival (35) and regeneration during intestinal mucosal wound healing (9, 13). Using full growth factor media and chronic STAT3 inhibition, we also showed that epithelial STAT3 was important for in vitro crypt regeneration even in the absence of exogenous IL-6 (Supplemental Fig. 1A). Interestingly, we showed in vivo that low levels of nuclear pSTAT3 activation (Supplemental Fig. 2J) were present in most UEA-1–positive Paneth cells at the base of small intestinal crypts; there was also evidence of a small number of crypts expressing high pSTAT3 activation in Paneth cells (Supplemental Fig. 2K). These findings support the hypothesis that pSTAT3 signaling is dynamic and transient during homeostasis. In the future the development of fluorescent reporter systems would be required to study dynamic pSTAT3 signaling in real time.

The STAT3 signaling pathway can be activated by a number of cytokines including IL-6 and IL-22 (36, 37). Interestingly, activation of IL-22 rather than IL-6 was shown to be key in regulating epithelial STAT3 wound healing following dextran sodium sulfate–induced experimental colitis in mice (13) and in ameliorating gut inflammation (38). This highlights the complexity of STAT3 signaling and it is likely that finetuning of this pathway through different cytokines exerts different functional responses depending on the context. Indeed, we showed that IL-22 stimulation of organoids resulted in global epithelial pSTAT3 activation (Supplemental Fig. 3A), which was in contrast to the restricted activation in Paneth cells following IL-6 exposure. We also concur with previous findings (39) that IL-22 induces proliferation of both large and small intestinal organoids and affects organoid budding (Supplemental Fig. 3), and we show that IL-6 can cause proliferation of colonic organoids. Whether the same mechanisms exist in the human gut epithelium remains to be elucidated. The relative contribution of different cytokines in regulating the STAT3 pathway in gut homeostasis versus inflammation and in the small intestine versus the colon is the subject of future investigation.

Although Paneth cells are not an absolute requirement for survival, proliferation, and stem cell activity (40, 41), previous work has shown that they provide important regenerative signals for modulation of the crypt stem cell niche (2, 42). Future work will elucidate the identity of the Wnt pathway(s) involved in IL-6 and STAT3 signaling in small intestinal crypt homeostasis. This study has also identified a novel role for the IL-6 signaling axis in regulating Paneth cell numbers in vivo. Paneth cell metaplasia in the colon is a feature of inflammatory bowel disease (43, 44), in addition to elevated levels of IL-6 in the serum and tissues (45, 46). We speculate that alterations in the IL-6 signaling axis may be a contributing factor to this phenotype. Furthermore, in the clinic, some rheumatoid arthritis patients treated with the IL-6 Ab tocilizumab display adverse side effects (47, 48); the mechanism by which this occurs is unknown.

IL-6 can also activate other signaling pathways not studied in this paper (17). This highlights the importance of understanding the complex biology of IL-6 during health and disease. In this study, we show a previously unrecognized role for IL-6 signaling in health. Understanding the mechanisms of IL-6 regulation of crypt cell renewal has wider implications for the prevention or development of future treatments for inflammatory bowel disease. More targeted treatments for inflammatory disorders that retain classic IL-6 signaling and homeostatic levels of epithelial crypt renewal but inhibit trans IL-6 signaling (49) through specific cell types appears to be the ideal scenario for the future (50, 51).

We thank Prof. Alastair Watson, Simon Deakin, and staff from the Disease Modelling Unit at the University of East Anglia, as well as Dr. Paul Thomas from the Henry Wellcome Laboratory for Cell Imaging at the University of East Anglia.

The authors have no financial conflicts of interest.