Modulation of intestinal IL-37 expression and its impact on the epithelial innate immune response and barrier integrity

Background and Aims Intestinal epithelial cells separate the luminal flora from lamina propria immune cells and regulate innate immune responses in the gut. An imbalance of the mucosal immune response and disrupted intestinal barrier integrity contribute to the evolution of inflammatory bowel diseases. Interleukin (IL)-37 has broad anti- inflammatory activity and is expressed by the human intestinal epithelium. Mice ectopically expressing human IL-37 show reduced epithelial damage and inflammation after DSS-induced colitis. Here, we investigated the impact of IL-37 on the innate immune response and tight junction protein expression of mouse intestinal organoids and the modulation of IL37 expression in human intestinal organoids. Methods Murine intestinal organoids were generated from IL-37tg and wildtype mice. Human ileal organoids were generated from healthy young donors. Results Expression of transgene IL-37 or recombinant IL-37 protein did not significantly reduce overall proinflammatory cytokine mRNA expression in murine intestinal organoids. However, higher IL37 expression correlated with a reduced proinflammatory cytokine response in murine colonic organoids. IL37 mRNA expression in human ileal organoids was modulated by proinflammatory cytokines showing an increased expression upon TNF-α-stimulation and decreased expression upon IFN-gamma stimulation. Transgene IL-37 expression did not rescue TNF-α-induced changes in morphology as well as ZO-1, occludin, claudin-2, and E-cadherin expression patterns of murine jejunal organoids. Conclusions We speculate that the anti-inflammatory activity of IL-37 in the intestine is mainly mediated by lamina propria immune cells protecting intestinal epithelial integrity.


Introduction
The intestinal epithelium is the largest epithelial layer in the human body that is in contact with the environment (1).Intestinal epithelial cells (IEC) form the barrier between the intestinal luminal microbiome and the lamina propria immune cell compartment and mediate innate immune responses of the gut (2).Cytokines secreted by either resident or infiltrating immune cells modulate epithelial proliferation, apoptosis, and barrier function, and vice versa, epithelial cells themselves can initiate immune responses at the mucosal level (3).IEC express a variety of cytokine receptors and receptors to sense pathogens or pathogen-associated molecular patterns (2,4,5).Subsequent signaling pathways induce cytokine and chemokine secretion by IECs to recruit immune cells to the site of inflammation (2,(4)(5)(6) with destructive (7)(8)(9)(10)(11)(12) or beneficial effects (13-15) on the intestinal epithelial barrier.
The functional crosstalk between IEC and immune cells via cytokines and chemokines is tightly balanced in the healthy gut (3).In contrast, an impaired gut barrier integrity and imbalance of the mucosal immune response can contribute to inflammatory bowel diseases (16-21).The imbalance of mucosal immune responses may be caused by an exaggerated immune activation or the loss of antiinflammatory, protective immune mechanisms.As such, it has been shown that, i.e., the overproduction of interleukin (IL)-12 by macrophages (22), genetic variants of NOD2 overactivating NF-kB in monocytes (23), or alterations in TLR expression (20) are associated with Crohn's disease.On the other hand, reduced antiinflammatory mechanisms such as disturbed TGF-ß signaling in inflammatory bowel disease (IBD) patients (24), reduced number of regulatory T cells in peripheral blood in patients with ulcerative colitis (25) or IL-10 or IL-10 receptor deficiency in pediatric patients with early-onset enterocolitis (26) are examples of a harmful, overactivated immune response in the gut.
IL-37, an anti-inflammatory member of the IL-1 family, is attributed to balancing the intestinal immune response, and a homozygous loss-of-function variant has been shown to be associated with infantile-onset IBD (27).IL-37 acts via two different pathways.Extracellular IL-37 binds to the IL-18 receptor 1 (IL-18RI) and a single IL-1R-related molecule (SIGIRR), forming the tripartite complex and suppressing MyD88-dependent proinflammatory signals (28,29).Intracellular IL-37 interacts upon caspase-1 cleavage with Smad3, translocates to the nucleus, and reduces the production of proinflammatory cytokines (30,31).
The aim of this study is to analyze the impact of transgene IL-37 (tgIL-37) on innate immune response and TJ protein expression in murine intestinal organoids and to investigate the regulation of IL37 expression in human intestinal organoids.

Generation of human organoids
Human intestinal forceps biopsies were taken during routine endoscopies from the terminal ileum of two patients without intestinal inflammation (boy, 16 years; girl, 16 years) at Charite-Universitätsmedizin Berlin and collected in HBSS.Patients and legal guardians gave written consent (EA/134/19).
After sampling, ileum biopsies were incubated in EDTA (2.5 mM), followed by vigorous pipetting to isolate crypts, and then Matrigel was added for organoid culture as described (41).Plates were incubated at 37°C and 7% CO 2 .Organoid medium was prepared as described (Supplementary Table S1) and replaced every other day (41).Organoids were passaged every 10 to 14 days.

Generation of murine organoids, passage, and freezing
C57BL/6J mice expressing human IL-37 (IL-37tg) have been described previously (30).Age-matched mice from the congenic control line (WT) were used as controls.IL-37tg and WT mice were bred under specific pathogen-free conditions.Three 8-14-weekold mice of each genotype were used for organoid generation.
Jejunal organoids from middle jejunal tissue and colonic organoids from distal colonic tissue were generated as previously described with modifications (42,43).Gut segments were rinsed with ice-cold DPBS, opened longitudinally, cut into 10-mm pieces, and washed by vortexing several times until the supernatants became clear.Crypts were isolated as described (42) but with an incubation of 20 min for the jejunum and 90 min for the colon in small intestinal or colonic crypt isolation buffer (42).After incubation, biopsies were washed in DPBS.Crypts were then released by vigorously shaking the tissue segments in DPBS, and the supernatant was collected in 0.1% BSA/PBS and centrifuged at 300 ×g for 5 min at 4°C.Crypts were diluted to 200 crypts per µl in Matrigel (Corning Limited-Life Sciences, Amsterdam, The Netherlands, 354230) and seeded as 20 µl domes in 48-well plates.Culture plates were incubated for 30 min at 37°C.The respective organoid culture medium (Supplementary Table S2) was then added.The medium was supplemented with Y-27632 (Merck Millipore) for the first 2 days.Organoid cultures were maintained at 37°C and 5% CO 2 .The medium was changed every other day, and colonic organoids were passaged every 5-6 days and jejunal organoids every 4-5 days.A detailed description of organoid generation, passage, and freezing is described in the Supplementary material.

Stimulation of murine jejunal and colonic organoids
Jejunal and colonic organoids were thawed as previously described (42), cultured, and passaged at least once before experiments were performed.For experiments, only cultures from passages four to nine were chosen.Jejunal organoids were stimulated with TNF-a (Gibco, Thermo Fisher Scientific) and colonic organoids were stimulated with LPS (O55:B5, Sigma), as indicated in the figure legends.

Gene expression analysis
Organoids were harvested for mRNA isolation 4 h after stimulation.Culture medium was removed, and Matrigel domes were washed in DPBS.Organoids of four to five domes were resuspended in 1 ml of Trizol (Invitrogen, Carlsbad, CA, USA), incubated at room temperature for 10 min, and vortexed.mRNA was purified from the aqueous phase following the manufacturer's protocol, including DNAse digestion (GenElute ™ Mammalian Total RNA Miniprep Kit, Sigma).Reverse transcription into cDNA was done using 1 µg of total RNA (Applied Biosystems, Carlsbad, CA, USA).Quantitative PCR (qPCR) was performed using the SYBR Green Master Mix (Applied Biosystems).Primer sequences are summarized in Table 1.Measurements were quantified using Quant Studio Design and Analysis software (Applied Biosystems).Gene expression was normalized to Hprt expression and untreated control condition.

Immunofluorescence
For immunofluorescence microscopy, jejunal organoids were harvested with organoid harvesting solution 48 h after stimulation.As previously described (44), 5-mm slides of paraffinized agarose domes containing organoids were generated.After rehydration, antigen retrieval was performed in boiling Tris-EDTA Tween buffer.Slides were incubated in 10% donkey serum/DPBS containing 0.3 % Triton X-100 and with primary antibodies (Supplementary Table S4) overnight at 4°C.The next day, slides were stained with secondary antibodies (a-rabbit Alexa488 or amouse Alexa647) (Invitrogen; donkey; 1:1,000) for 1 h at room temperature and counterstained with DAPI.Image acquisition was performed with ×25 glycerol immersion on a Zeiss confocal LSM980 NLO microscope with Airyscan, and images were analyzed using Zen software.

Statistical analysis
Results are expressed as the mean values with 1 standard deviation (SD).A paired Student's t-test was applied to compare treated versus untreated conditions in one experimental group.An unpaired Student's t-test was used for comparisons between organoids generated from WT and IL-37tg mice.For data that were not considered as normally distributed using the Shapiro-Wilk test, the corresponding nonparametric tests (Wilcoxon signed-rank test or Mann-Whitney U test) were performed.For correlation analysis, statistical significance was tested with Pearson´s correlation and the Student's two-tailed t-test.Outliers were defined by ROUT (Q = 1%) and Grubbs (alpha = 0.05) test.All tests were performed using GraphPad Prism software (version 9.3.1).A p-value of 0.05 or less was considered significant.

Cytokine response and the expression of IL-37 and its receptor components in WT and tgIL-37 intestinal organoids
To corroborate whether IL37 impacts the immune response of intestinal epithelial cells, we stimulated intestinal organoids generated from IL-37tg and WT mice with TNF-a and IL-1b, as well as LPS and flagellin.Murine jejunal organoids were most potently stimulated by TNF-a and colonic organoids by LPS, inducing the highest mRNA expression levels of Cxcl1, Cxcl2, Ccl20, and Tnf within a 4-h stimulation period (Supplementary Figure S1; Figures 1A, B).
TgIL-37 expression was not associated with a significant change in proinflammatory cytokine mRNA expression in jejunal and colonic organoids before and after stimulation, despite a trend toward reduced Cxcl1, Cxcl2, and Tnf mRNA levels in colonic IL-37tg-derived organoids (Figures 1A, B).
Jejunal and colonic organoids generated from IL-37tg mice showed a trend toward higher IL37 expression after stimulation with TNF-a and LPS, respectively (Figure 1C).Similarly, IL37 expression was threefold induced by TNF-a in human ileal organoids but not by LPS or IL-1b, while IFN-gamma downregulated the expression of IL37 (Figure 1D).IL37 expression in human ileal organoids was gradually induced by TNF-a over a 24 -h period (Supplementary Figure S2).
Baseline IL37 expression was higher in murine colonic (CT mean: 32.4) versus jejunal organoids (CT mean: 34.4) (Supplementary Table S5).In colonic organoids, higher IL37 mRNA expression was associated with reduced Cxcl2 and Tnf mRNA expression, as indicated by the significant correlation between IL37-DCt-values of stimulated IL-37tg-derived organoids and their relative Cxcl2 and Tnf mRNA expression (Figure 1E).
The lack of correlation between IL-37 expression and the inflammatory response of jejunal organoids could be due to differences in anti-inflammatory IL-37 signaling.We therefore analyzed the mRNA expression of IL-37 receptor components in jejunal and colonic organoids.Il18r1 expression was eight times higher in WT colonic organoids compared to WT jejunal organoids before and 10 times higher after stimulation with TNF-a (Figure 1F).Baseline Il18r1 mRNA expression was higher in IL-37tg-derived colonic organoids compared to WT-derived colonic organoids (Supplementary Figure S3).Sigirr is highly expressed in both jejunal and colonic organoids (CT mean: 25.2 and 26.2) (Figure 1G), and the expression was two times higher in WT jejunal organoids compared to WT colonic organoids before and after stimulation (Figure 1F).The relation of Il18r1 and Sigirr mRNA expression between jejunal and colonic organoids was similar in IL-37tg organoids and WT organoids (Supplementary Figure S4).

Morphologic changes of WT and IL-37tgderived organoids after TNF-a treatment
Since IL37 expression did not markedly alter the intestinal epithelial immune response (Figures 1A, B), we evaluated the impact of IL-37 on the morphology of organoids derived from WT and IL-37tg mice during inflammation.Under untreated conditions, the morphology of WT and IL-37tg-derived murine organoids was similar (Figure 3A).Jejunal organoids grew as buds and colonic organoids in a spherical shape (Figure 3A).After TNFa treatment, jejunal WT and IL-37tg-derived organoids displayed spherical rounding (Figure 3B), while the gross morphology of colonic organoids remained similar after LPS treatment.The shape of organoids was similar in both genetic traits.

Gene
Forward primer Reverse primer

Downregulation and disassembling of intestinal barrier proteins in murine organoids after treatment with inflammatory stimuli are not modulated by tgIL-37 expression
Spherical rounding of small intestinal organoids after TNF-a treatment can be caused by increased TJ permeability (46).Therefore, we analyzed ZO-1, occludin, and claudin-2 mRNA and protein expression in organoids derived from WT and IL-37tg mice before and after respective treatments.There was no significant difference in TJ protein expression between WT-and IL-37-tgderived organoids on mRNA and protein levels before and after stimulation (Figures 4A, B, D).Changes in barrier protein expression and distribution, analyzed by immunofluorescence analysis, were s imilar in WT-and IL-37tg-derived organoids (Figure 4E).
Western blot analysis (Figures 4C, D) showed a trend toward reduction of ZO-1 in IL-37tg-derived organoids after TNF-a treatment but not in WT organoids.Occludin was reduced by 35% in WT-and by 36% in IL-37tg-derived organoids, and claudin-2 was not significantly reduced by 43% in WT-and by 51% in IL-37tg-derived organoids.Claudin-2 baseline expression is highly varied between single organoid lines, independent of the genotype (Supplementary Figure S5).
In the immunofluorescence analysis (Figure 4E), ZO-1, occludin, and claudin-2 displayed a strong signal at the apical side of epithelial cells, indicating expression within the TJs.Additionally, occludin showed basolateral staining and claudin-2 an intracellular staining.E-cadherin was highly enriched at the basolateral side and at the lateral membranes, indicating expression within the adhering junctions.ZO-1, occludin, and claudin-2 expression patterns were largely disrupted by TNF-a treatment.E-cadherin staining was restricted to the basolateral side and less detectable at the lateral membranes and at the apical sides after TNF-a treatment.Other than ZO-1, occludin, and E-cadherin, claudin-2 was not expressed homogenously within the organoids but was predominately located within crypt domains (Supplementary Figure S6).

Discussion
In this study, we show that IL-37 does not modulate the gross cytokine response of murine intestinal organoids after immune stimulation with TNF-a or LPS by using tgIL-37 mice or treatment by rhIL37.However, we show that IL37 expression is upregulated upon proinflammatory stimulation in murine and human organoids.More subtle analyses demonstrate a negative correlation between IL37 expression and proinflammatory cytokine response in murine colonic organoids.IL-37 did not alter the TNF-dependent downregulation and structural alteration of TJ proteins in murine jejunal organoids.
There was no significant change in proinflammatory cytokine mRNA expression by IL-37tg jejunal or colonic organoids after TNFa or LPS stimulation.In accordance with these results, we could also not observe differences in cytokine response after treatment with rhIL-37.This indicates that IL-37, expressed within the intestinal epithelium, does not grossly limit the immune response of IEC in an auto-or paracrine manner.Differences within the experimental approach might explain the contrasting results to the observation of Allaire and colleagues describing a downregulation of proinflammatory Ccl20 and Cxcl2 in flagellinand IL-1ß-stimulated murine colonic organoids by treatment with rhIL-37 protein (37).Accordingly, Günaltay et al. showed that CRISPR/Cas knockdown of IL37 increases the spontaneous chemokine expression in T84 IECs (36).We previously reported that hematopoietic-derived IL-37 from bone marrow transplantation is sufficient to ameliorate the severity of DSS-induced colitis in WT mice (32).This indicated the relevance of tgIL-37 expressed by immune cells to control mucosal inflammation.Since neither tgIL-37 nor rhIL-37 protein modulated the epithelial immune response in our organoid model, we speculate that predominantly extracellular IL-37, released from IECs at the site of tissue inflammation, contributes to the control of the immune response of intestinal mucosa-associated immune cells.
IL37 expression in IL-37tg mice is driven by a constitutively active CMV-promotor (30).However, like in healthy human tissue, baseline levels of IL-37 mRNA and protein are low in unstimulated IL-37tg mice, which is due to mRNA instability elements within the coding region of IL-37 (30, 47).IL37 expression is also low in resting jejunal and colonic organoids generated from IL-37tg mice, but we were able to show a trend toward increased levels after stimulation.In human ileal organoids, only TNF-a but not LPS or IL-1ß induced IL37 expression, while IFN-gamma reduced IL37 expression.This is consistent with observations on T84 epithelial colon carcinoma cells (38).Upregulation of IL37 by TNF-a is also in line with a high epithelial IL37 expression in IBD correlating with higher disease activity (38,48).In murine and human small intestinal organoids, there was only a minor inflammatory response to LPS, which might be caused by low expression of TLR4 in both and missing TLR4associated accessory protein expression in human organoids [Supplementary Figure S1 and as previously shown (49)].Insufficient immune stimulation could therefore explain the lack of IL37 induction after LPS treatment in human ileal organoids.The induction of IL37 by TNF-a in human ileal organoids was similar to that in murine jejunal organoids, corresponding to a high immune stimulation by TNF-a in murine and may be as well in human small intestinal organoids (Supplementary Figure S1).Increasing levels of tgIL-37 mRNA, induced by proinflammatory signals, correlated with a lower Cxcl2 and Tnf response in LPSstimulated colonic organoids.This negative correlation was not observed in jejunal organoids after stimulation.Since SIGIRR is evenly expressed throughout the murine gut (50), as similarly seen in our jejunal and colonic organoids, we speculate that a higher expression of the IL-37 receptor component Il18r1 (51-53) in colonic organoids sensitizes for autologous, anti-inflammatory IL-37 signaling.Indeed, polymorphisms in the IL18RI-IL18RAP locus are associated with adult and early-onset IBD (54-56), underlining the relevance of the IL-18RI for human disease activity to target both proinflammatory IL-18 and anti-inflammatory IL-37 immune pathways.In addition, tgIL-37 expression itself is associated with a markedly higher Il18r1 expression in our IL-37tg colonic organoids, suggesting a regulatory effect of IL-37 on bona fide Il18r1 expression (Supplementary Figure S3).Pretreatment with rhIL-37 was shown to induce SIGIRR expression in humans, LPS-stimulated PBMCs and M1 macrophages (45) indicating that IL-37 induces upregulation of its receptor components to increase its anti-inflammatory capacity.
TJ proteins claudin-2, ZO-1, or occludin are key determinants of epithelial integrity against mucosal inflammation in IBD (17,18,40,57).The impact of TNF-a on epithelial barrier function was extensively studied in 2D cell culture models (10)(11)(12)58).3D organoid models are a good representation of the epithelial architecture, with a villi-like structure and composition of a variety of epithelial subtypes.However, studies of TJ proteins and IL-37 in organoids are scarce.In our model, TNF-a and LPS caused significant changes in TJ protein expression and morphology in jejunal and colonic organoids.Nevertheless, overall TJ protein expression and organoid morphology were similar between WTand IL-37tg-derived organoids before and after stimulation.This indicates that tgIL-37 expression in IEC does not prevent the disturbance of TJ protein expression during inflammation.
In summary, our 3D intestinal organoid model enabled us to investigate the impact of IL-37 on IEC innate immune responses and intestinal barrier function, irrespective of epithelial immune cell activation.We show that IL37 expression is induced in intestinal epithelial cells but has no major impact on modulating IEC innate immune responses itself and does not prevent murine intestinal organoids from altered barrier protein expression after stimulation.We conclude that the protective effect of IEC-derived IL-37 on the intestinal epithelial barrier in murine models of IBD and potentially the human gut is more mediated by downregulating the release of proinflammatory mediators from lamina propria immune cells than intestinal epithelial cells.and their families for contributing to our research.We acknowledge financial support from the Open Access Publication Fund of Charité -Universitätsmedizin Berlin and the German Research Foundation (DFG).

Cxcl1 1
FIGURE 1Cytokine response and the expression of IL-37 and its receptor components in WT and tgIL-37 intestinal organoids.Jejunal and colonic organoids derived from IL-37tg and WT mice were stimulated for 4 h with murine TNF-a (10 ng/ml) or LPS O55:B5 (1,000 ng/ml)(37) in culture medium.mRNA expression of cytokines in (A) jejunal and (B) colonic organoids was measured by qPCR and expressed as fold change relative to the mean of unstimulated organoids of the respective mouse (2 −DDCT ).(C) Induction of IL37 mRNA expression in stimulated IL-37tg-derived organoids was measured by qPCR and expressed as fold change relative to unstimulated organoids (2 −DDCT ).(D) Human ileal organoids (il, n = 2, two donors) were stimulated for 24 h with LPS O55:B5 (1,000 ng/ml), human TNF-a (10 ng/ml), human IFN-gamma (20 ng/ml), and human IL-1ß (10 ng/ml) in the culture medium.mRNA expression of IL-37 was measured by qPCR and expressed as fold change relative to unstimulated organoids (2 −DDCT ).(E) Correlation of cytokine mRNA expression relative to mean of unstimulated colonic organoids of the respective mouse (2 −DDCT ) with IL37-DCTvalues of stimulated colonic organoids derived from IL-37tg mice.(F) mRNA expression of Il18r1 and Sigirr in jejunal and colonic WT organoids was measured by qPCR and expressed as fold change relative to mean of unstimulated jejunal organoids (2 −DDCT ).(G) Mean CT-values of Il18r1 and Sigirr qPCR duplicates in jejunal and colonic WT organoids.Each experiment was performed with organoid lines of two different mice and repeated three times (in total n = 6).Open bars: organoids derived from WT mice (WT); grey bars: organoids derived from IL-37tg mice (tg); no pattern: jejunal organoids (jj); striped pattern: colonic organoids (co).Each data point represents a single organoid line.Data are expressed as the mean with SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001.

3 4
FIGURE 3 Morphologic changes of WT and IL-37tg-derived organoids after TNF-a treatment.(A) Representative light microscope images of murine jejunal and colonic organoids in extracellular matrix on days 3 to 4 under baseline conditions.(B) Representative light microscope images of jejunal organoids in extracellular matrix with or without 48 h TNF-a (10 ng/ml) stimulation.One representative picture out of six independent stimulation experiments is shown.