Interleukin-10 regulates goblet cell numbers through Notch signaling in the developing zebrafish intestine

Cytokines are immunomodulatory proteins that orchestrate cellular networks in health and disease. Among these, interleukin (IL)-10 is critical for the establishment of intestinal homeostasis, as mutations in components of the IL-10 signaling pathway result in spontaneous colitis. Whether IL-10 plays other than immunomodulatory roles in the intestines is poorly understood. Here, we report that il10, il10ra, and il10rb are expressed in the zebrafish developing intestine as early as 3 days post fertilization. CRISPR/Cas9-generated il10-deficient zebrafish larvae showed an increased expression of pro-inflammatory genes and an increased number of intestinal goblet cells compared to WT larvae. Mechanistically, Il10 promotes Notch signaling in zebrafish intestinal epithelial cells, which in turn restricts goblet cell expansion. Using murine organoids, we showed that IL-10 modulates goblet cell frequencies in mammals, suggesting conservation across species. This study demonstrates a previously unappreciated IL-10-Notch axis regulating goblet cell homeostasis in the developing zebrafish intestine and may help explain the disease severity of IL-10 deficiency in the intestines of mammals.


INTRODUCTION
The intestine is a semi-permeable and highly regenerative tissue responsible for nutrient and water absorption while keeping the host protected from environmental and microbiota challenges. The control of these functions is orchestrated by complex cellular crosstalk between intestinal epithelial cells and tissue-resident immune cells, among others. In this regard, cytokines, and particularly IL-10, have been considered key mediators in the maintenance of intestinal homeostasis and function. 1 This has been demonstrated in mice defective in IL-10 signaling, which develop spontaneous colitis. 2 In humans, mutations in the IL-10 signaling genes IL10, IL10RA, and IL10RB are strongly associated with a very early onset of inflammatory bowel diseases (VEO-IBD), 3,4 which is characterized by clinical manifestations in children before 6 years of age, and a significant number of cases occurring in infants younger than 1 year. 5 VEO-IBD has a strong genetic component compared to the 5-10% familial cases observed in adult inflammatory bowel diseases (IBD). 6 In addition, VEO-IBD is phenotypically distinct from adult IBD and is characterized by extensive colitis, growth failure, and unresponsiveness to conventional therapies. 6 Common clinical manifestations in children with VEO-IBD include bloody diarrhea and mucous stool, 7 suggesting altered goblet cell activity.
Intestinal immune cells, including macrophages, B cells, and T cells, are the predominant IL-10 producers and responders. IL-10 signaling in intestinal macrophages is critical to maintaining intestinal homeostasis, as conditional deletion of IL-10Rα in CX3CR1 + macrophages is sufficient to trigger spontaneous colitis in mice. 8 Interestingly, intestinal epithelial cells also express IL-10 receptors and are therefore equipped to mount a cellular response upon IL-10R stimulation [9][10][11] In this regard, IL-10 deficiency in myeloid cells results in impaired regeneration of the epithelium following damage, whereas administration of recombinant IL-10 promotes intestinal epithelial cell proliferation in vivo. 12 Furthermore, IL-10 promotes intestinal stem cell (ISC) proliferation in organoid cultures, 13 suggesting an intrinsic role of IL-10 signaling in intestinal epithelial cells (IEC) differentiation. In line with this, specific IL-10R deficiency on IECs results in goblet cell hyperplasia in the colon of adult mice. 14 Altogether, the combined evidence not only shows that IL-10 signaling can modulate IEC differentiation but also raises the possibility that early-stage intestinal development is impaired in individuals with mutations in the IL-10 signaling pathway, which may lead to the phenotypes observed in VEO-IBD.
The intestinal tract develops at early embryonic stages from the endodermal germinal layer. It is initially formed as a flat tube of epithelial and mesenchymal cells that proliferate and stratify in a synchronic fashion, which results in the morphogenesis of the villi and crypts. At the bottom of the crypt, intestinal stem cells are constantly self-renewing and giving rise to specialized epithelial cells with either absorptive or secretory functions. The main drivers of intestinal stem cell differentiation include the orchestrated activity of signaling pathways such as Wingless/Int-1 (Wnt), bone morphogenetic proteins (BMP), and Notch, which antagonize each other to promote or attenuate cell replication. 15,16 Particularly, Notch signaling plays a critical function in stem cell maintenance and regulation of secretory cell differentiation across species. [17][18][19] In mice, Paneth and goblet cells are expanded after Notch inhibition, 20 and the lack of Notch target genes, such as hairy/enhancer of split 1 (Hes1) show a skewed epithelial cell differentiation towards the secretory lineage. 21,22 By contrast, continuous Notch activation by overexpression of the Notch intracellular domain (NICD) results in the expansion of the intestinal stem cell pool and impaired secretory-IEC differentiation. 23 Expression of the Notch1 and Notch 2 receptors and their main target genes Hes1, Hes5, and Hes6 have been detected in the dividing cells within the intestinal crypt. 24 Interestingly, the Notch pathway regulates T cell function, 25 including the production of IL-10 by Th1 cells. 26 However, whether IL-10 signaling regulates tissue differentiation by controlling Notch or other stem cell-related pathways remains unknown.
Zebrafish have been successfully used to model key biological processes associated with intestinal disorders, as physiological and cellular features are highly conserved between fish and humans. 27 Using zebrafish, we and others have addressed the impact of dietary compounds, 28 environmental pollutants 29 and IBD risk genes 30 in intestinal homeostasis, highlighting zebrafish as a powerful tool to investigate the function of genetics in intestinal development. The zebrafish ortholog for the human IL10 gene 31 shows a conserved immunomodulatory function. 32,33 Here, we took advantage of zebrafish to investigate the effect of IL-10 signaling in the developing intestine. Using CRISPR/Cas9 we generated Il10 mutants, which combined with histological and immunological stainings, transcriptional analysis and fluorescent transgenic reporters enabled us to investigate the intestinal cell composition of IL-10-deficient larvae. We found that IL-10 deficiency resulted in an increased number of goblet cells, which was independent of the microbiota. Mechanistically, the increase in goblet cells was associated with a decreased activity of intestinal Notch signaling, and ectopic Notch signaling activation restored alcian blue + goblet cell numbers in IL-10-deficient larvae. Moreover, IL-10 treatment in mouse small intestine (SI) organoids led to decreased frequencies of goblet cells and increased Notch activity, which suggests an evolutionarily conserved IL-10-Notch axis that regulates intestinal goblet cell differentiation across vertebrates.

RESULTS
Zebrafish il10 is expressed during early intestinal development The development of the larval zebrafish intestine, after the formation of the tube and lumen between 30 and 52 hours post fertilization (hpf), can be divided into an early stage of high proliferation and polarization of the intestinal epithelium (52-74hpf), followed by a later stage of tissue compartmentalization and the emergence of differentiated intestinal cell types (74-120 hpf). 34,35 To gain insights into the developmental stage in which Il10 might control zebrafish intestinal organogenesis, we analyzed il10 transcript levels in zebrafish larvae starting at 72 hpf or 3 days post fertilization (dpf) until the stage in which a functional intestine is appreciated (5dpf and 7dpf). Using qRT-PCR to analyze il10 transcript levels from dissected zebrafish larval intestines (Fig. 1a), we found higher transcript levels of il10 in 3dpf intestines, compared to 5dpf and 7dpf (Fig. 1b). Dissected intestines contain intestinal epithelial cells (IECs) among other intestine-associated cells that interact with IECs, therefore we sought to determine which group of cells was responsible for il10 expression. We then FACS-sorted epithelial cells versus the rest by using the zebrafish reporter line Tg(cldn15la:GFP) (Fig. 1c), in which the cldn15la promoter drives GFP expression exclusively in intestinal epithelial cells. 36 qPCR analysis from sorted cells showed that il10 expression was not detectable in GFP + cells but in GFPintestine-associated cells (Fig. 1d). To gain insights into a potential active IL-10 signaling in intestinal cells, we analyzed the expression of the orthologs for the receptors of il10, defined as il10ra and il10rb in the zebrafish database ZFIN. Whole-mount in situ hybridizations (WISH) showed that both il10ra and il10rb genes were expressed in the zebrafish developing intestine at 3dpf, while il10rb was predominantly expressed at 5dpf (Figs. 1e and S1). Altogether, these results suggest that il10 signaling may act in zebrafish intestinal cells during early developmental stages. Intestines from Tg(cldn15la:GFP) larvae were extracted from the body, disaggregated to cell suspensions, and FACS sorted based on the expression of GFP. GFP + and GFPcell collections were subsequently used for RNA transcriptomic analyses. d Transcriptomic expression analyses from sorted GFP + and GFPcells by qRT-PCR. Each dot represents an individual sorting experiment with around 30,000 cells collected (N = 6). e Whole-mount in situ hybridization for the genes il10ra and il10rb in 3dpf and 5dpf zebrafish larvae. Representative pictures of the stainings at each developmental stage are shown. Scale bar = 200 µm. One-way ANOVA with Fisher's LSD multiple comparisons test was performed in b, while two-tailed t-tests were performed in d (*p < 0.05; ***p < 0.001).
Altered immune and intestine-associated profiles in il10 mutant larvae Next, we aimed to determine whether il10 can modulate the development of the larval zebrafish intestine by using loss-offunction genetic approaches. Using CRISPR/Cas9, 37 we generated a 4 bp deletion in the exon 2 of the zebrafish il10 gene (il10 uu1751 ) (Fig. 2a), predicted to generate a premature stop codon (Fig. 2b). il10 uu1751/uu1751 mutant fish (hereinafter referred to as il10-Mut) did not show any apparent morphological or developmental defect while in homozygosis in larva (Fig. S2a), reached adulthood and were fertile (Fig. S3a, b). Analysis of selected type 1, 2, and 3 cytokines by qPCR from the whole larvae at 5dpf revealed that type 2 cytokines (il4 and il13) and the type 3 cytokine il22 were comparable whereas ifng1 (type 1) and il17a/f3 (type 3) were significantly elevated in il10-Mut larvae compared to wild-type (WT, il10 wt/wt ) larvae (Figs. 2c and S2b). Thus, Il10 deficiency results in altered pro-inflammatory cytokine levels in the whole larvae. We then sought to investigate whether Il10 deficiency affects the development of the intestinal tract. Absorptive and secretory cells, namely enteroendocrine and goblet cells, can be distinguished by 5dpf in zebrafish larvae. 18 To analyze enteroendocrine and goblet cells, we performed whole-mount immunofluorescence using the pan-secretory marker antibody 2F11. 18 By 5dpf, the il10-Mut larvae showed a higher number of secretory cells in the intestine compared to WT (Fig. 2d, e). Secretory cells in the zebrafish larvae are differentially distributed along the anterior-distal intestine, 34 WT il10-Mut WT il10-Mut with enteroendocrine cells being concentrated in the anterior intestine and mucin-producing goblet cells in the mid intestine ( Fig. 2f). Therefore, we further stratified 2F11 + cells based on location. We observed that the number of 2F11 + secretory cells in the mid-posterior, but not in the anterior segment, was significantly increased in il10-Mut compared to WT (Fig. 2g). The increase in secretory cells in the mid-posterior segment prompted us to hypothesize that goblet cells are increased in il10-Mut compared to WT larvae. In line with our hypothesis, whole-mount alcian blue (ab) staining labeling mucin production as an indicator of ab + goblet cell abundance in the mid intestine, displayed a higher staining area in 5dpf il10-Mut larvae compared to WT ( Fig. 2h-i). Area measurements were complemented with ab intensity and stained length (Fig. S2c), showing an overall increase of the ab + signal detected in il10-Mut larvae. Likewise, wholemount staining with fluorescently labeled wheat germ agglutinin (WGA), which binds to the N-acetylglucosamine present in the mucus produced by goblet cells 34 showed that the number of WGA + goblet cells in the mid intestine was increased in 5dpf il10-Mut larvae, compared to WT (Fig. S2d, e), further confirming that il10-deficiency results in goblet cell hyperplasia within the developing zebrafish embryo. To analyze if absorptive intestinal populations were affected in il10-Mut larvae, we used neutral red to stain a subset of mid intestine absorptive enterocytes with acidified lysosomes, namely lysosome-rich enterocytes (LREs). [38][39][40] Area, intensity and length analysis from neutral red-stained enterocytes in 5dpf larvae did not show differences between WT and il10-Mut (Fig. S2f, g), suggesting no impairments in the differentiation of absorptive intestinal lineages. Further, to determine whether the increased number of goblet cells in il10-Mut zebrafish was age-dependent, we analyzed different intestinal regions from >1-year-old WT and il10-Mut zebrafish (Fig. S3c). Ab staining on paraffin-embedded intestinal sections showed a higher number of goblet cells per villus area in il10-Mut larvae, compared to WT (Fig. S3d, e). Altogether, our results point towards a disbalance in both pro-inflammatory cytokine expression and the abundance of goblet cells in il10-deficient zebrafish.

The abundance of intestinal goblet cells depends on il10
To determine whether the phenotypes observed in the il10-Mut were dependent on Il10 signaling rather than off-target effects we performed rescue experiments using an in vitro-generated version of the zebrafish il10 mRNA. Injection of the full il10 mRNA at 1-cell stage resulted in a significant increase of il10 transcripts in the whole 5dpf larvae (Fig. 3a). il10 mRNA-injected il10-Mut 5dpf larvae showed a non-significant reduction of il17a/f3 transcript levels compared to il10-Mut, whereas ifng1 levels were comparable (Fig. 3b). In addition, the average ab-stained area was reduced in the intestines of il10 mRNA-injected il10-Mut larvae compared to il10-Mut, reaching comparable levels to those of WT larvae (Fig. 3c,   Days post-fertilization (dpf) Days post-fertilization (dpf) h Quantification of the ab-stained area of control and il17-injected larvae. N = 2 independent experiments. Each independent experiment is shown with a different symbol shape in a, b, d, f and h. One-way ANOVA was used for statistical analysis in a, b and d, whereas two-tailed student t-test were used in f and h (*p < 0.05; ***p < 0.001).
To assess the role of Il17 signaling on the increased goblet cell numbers observed in il10-Mut larvae, we analyzed the abundance of intestinal goblet cells in WT larvae injected with an in vitrogenerated il17 (il17a/f3) mRNA at 1-cell stage (Fig. 3e). While il17a/ f3 transcripts were high in 5dpf il17-injected larvae (Fig. 3f), ab analysis did not show differences in the ab-stained area between il17a/f3 mRNA-injected and control larvae (Fig. 3g, h), suggesting that the abundance of goblet cells does not depend on Il17 signaling in zebrafish larvae. To further confirm our findings, we measured the ab-stained area in a second CRISPR/Cas9generated mutant line for the il10 gene, this time containing a 17 bp insertion replacing 3 bp (+14 bp total insertion) in exon 3 (il10 uu1762 , referred to as il10-Mut2), which is predicted to generate a premature stop codon ( Fig. S4a, b). Similar to the il10-Mut, we observed increased staining for ab in 5dpf il10-Mut2 larvae compared to WT (Fig. S4c, d). Altogether, these results indicate that il10 modulates ab + goblet cell differentiation, but not cytokine expression, in zebrafish larvae.
Increased goblet cells in il10 mutant larvae are microbiotaindependent Previous findings suggest that goblet cell differentiation is dependent on the microbiota. 42 To investigate a potential role of the microbiota in the phenotypes observed in il10-Mut larvae, we established a protocol to reduce bacterial loads in zebrafish embryos (Fig. 4a). Briefly, 24-28 hpf embryos were treated with E3 water containing 0.004% sodium hypochlorite (bleach) for 5 min, and then they were incubated in the presence of an antibiotic cocktail (Abx, Ampicillin and Kanamycin) from 1dpf to 5dpf. We confirmed a reduction of total bacterial loads after bleach+Abx treatments in larvae by comparing bacterial 16S versus targeted zebrafish genomic DNA amplification levels by PCR (Fig. 4b).
Bleach + Abx-treated 5dpf il10-Mut larvae showed a reduced expression of both il17a/f3 and ifng1, compared to non-treated siblings (Fig. 4c), suggesting that cytokine modulation is microbiota-dependent. On the other hand, ab levels remained enhanced in il10-Mut compared to WT larvae, regardless of the reduced bacterial load (Fig. 4d, e). These results indicate that the increased ab + goblet cells observed in il10-Mut larvae are independent of the bacterial microbiota.
Impaired Notch signaling in the intestines of il10 mutant larvae To gain insights into the potential mechanisms by which il10 modulates goblet cell numbers in zebrafish larvae, we analyzed the activity of signaling pathways involved in goblet cell differentiation, such as Notch, 18 RAR, 28 and ARP/ASCL factors. 43,44 For this purpose, we dissected intestines from 5dpf WT and il10-Mut larvae (Fig. 5a) and we analyzed the expression of Goblet cell markers and key target genes for the above-mentioned signaling pathways by qRT-PCR. We did not observe differences in the expression of the RAR signaling target gene cyp26a1, the ARP/ ASCL-associated genes atoh1a and ascl1a, nor in the expression of the goblet cell markers agr2 and muc2.1 (Fig. S5a). In addition, we did not detect differences in the expression of olfm4.2, ortholog for the mammalian stem cell marker Olfm4, nor in the expression of the teleost intestinal progenitor marker sox9b (Fig. S5b). However, we found a reduction in the intestinal expression levels of her6 and her9, both Notch target genes described as the ortholog of human HES1 and HES5, respectively (Fig. 5b). To confirm our observations, we combined the il10-Mut fish with the Notch activity reporter Tg(tp1:EGFP), 45 which detects Notchresponding cells in the larval intestinal tract. 46 In vivo imaging of the intestines of 5dpf Tg(tp1:EGFP) larvae in a WT and il10-Mut background proved that the number of GFP + cells in the intestines of il10-Mut reporters was lower compared to WT (Fig. 5c,  d), thus demonstrating an impaired Notch activity in il10-Mut larval intestines. In mammals, intracellular IL-10 downstream signaling relies predominantly on the activity of the transcription factor Stat3, 47 and the zebrafish intestine was shown to contain Stat3-responsive cells with progenitor-like features. 48 To determine whether Stat3 activity is impaired in the intestines of il10-Mut larvae, we crossed il10-Mut with the Stat3 activity reporter Tg(7xStat3:EGFP). 48 Imaging and quantification of the number of intestinal GFP + cells showed no differences between WT and il10-Mut larvae (Fig. S5c, d), suggesting that Stat3 signaling is not affected in il10-Mut larvae. It has been previously shown that Notch signaling controls secretory cell differentiation in a time-specific manner, particularly during the phase of intestinal proliferation and polarization between 64-74 hpf. 49 We validated these findings by using the γ-secretase/Notch inhibitor DAPT at 2 different time points: from 48 to 72 hpf (hereinafter referred to as 2-3 dpf) and from 72 to 120 hpf (referred to as 3-5 dpf, Fig. S6a). We found that Notch inhibition during 2-3dpf led to an increased ab-stained area in WT larvae, while 3-5 dpf inhibition did not modulate ab levels (Fig. S6b, c). Additionally, we tested goblet cell abundance after inhibiting Stat3 signaling with the Jak2/Stat3 inhibitor AG-490 48 and found a reduced ab-stained area in AG-490-treated compared to control larvae, at both 2-3dpf and 3-5dpf time points (Fig. S6d-f). These results are opposite to our observations in il10-Mut larvae and suggest that Stat3 signaling is unlikely to play a role in the goblet cell hyperplasia observed in il10-Mut larvae. Therefore, we sought to determine whether boosting Notch activity between 2-3dpf restores ab-stained area to normal levels in il10-Mut larvae. For this purpose, we took advantage of the Notch activator Yhhu-3792 50 and tested its activity in zebrafish larvae. Incubation of WT larvae with 1 μM of Yhhu-3792 was sufficient to upregulate the expression of her6 transcripts (Fig. S7a,  b). We then analyzed goblet cell abundance in WT and il10-Mut larvae exposed or not to 1 μM of Yhhu-3792 at 2-3 dpf (Fig. 5e). Yhhu-3792 treatment of il10-Mut larvae resulted in reduced abstained area compared to vehicle-treated il10-Mut and comparable to WT larvae (Fig. 5f, g). As expected, Yhhu-3792 treatments in il10-Mut larvae at 3-5dpf did not result in changes in the abstained area (Fig. S7c-e). These results indicate that the increased abundance of goblet cells in il10-Mut larvae is restored by activating Notch signaling in a specific temporal window.
IL-10 enhances Notch signaling and restricts mucin production in mouse small intestine organoids To address whether our findings in zebrafish can be translated to mammals, we used a 3D murine small intestine (SI) organoid system which models key features of intestinal differentiation and development in vitro. 51 We isolated crypts from wild-type C57/Bl6 mice and we cultured them in EGF-, Noggin-, and R-Spondincontaining media (ENR medium; see methods) supplemented or not with recombinant murine IL-10 for 4 days (Fig. 6a). Crypt domains per organoid and organoid areas were comparable between treatments on day 4 (d4) of treatment (Fig. S8a-c). However, transcriptional analyses of d4 organoids showed that IL-10 treatment increased the expression of the stem cell-related gene Rgcc and the Notch target gene Hes1, compared to vehicletreated organoids (Fig. 6b). Similar to zebrafish, we did not observe differences in the expression of the secretory cell differentiation driver Atoh1 and the goblet cell marker Muc2 (Fig. 6b), nor in additional markers for intestinal stem cells (Olfm4) and other secretory cell types including enteroendocrine (Chga), tuft (Dclk1) and Paneth cells (Lyz1) (Fig. S8d). Further, we did not find differences in the expression levels of the intestinal differentiation marker Krt20 and the Wnt member gene Wnt3 (Fig. S8d). To analyze goblet cells abundance in control and IL-10treated organoids, we modified the zebrafish ab staining protocol to perform whole-mount staining in matrigel-embedded SI organoids (see methods). IL-10-treated SI organoids exhibited a reduced percentage of ab staining per organoid area when compared to vehicle-treated organoids (Fig. 6c, d). To confirm our findings and expand our analysis to additional intestinal cell populations, we performed flow cytometry analysis of d4 SI organoids generated from Lgr5-GFP mice 52 and stained with Epcam (Fig. S9a, b), together with WGA and CD24 to detect intestinal stem and secretory cells populations, as previously described. 53,54 IL-10 treatments did not affect the frequency of organoid ISCs (defined as live Epcam + Lg5-GFP + cells, Fig. 6e, f). However, IL-10 treatments reduced specifically the frequencies of goblet cells (defined as live Epcam + Lgr5-GFP -WGA hi CD24cells), while Paneth cells (live Epcam + Lgr5-GFP -WGA hi CD24 + cells) remained unaltered (Fig. 6g, h). Further, immunofluorescence stainings of d4 SI organoids with WGA and the Paneth cell marker Lysozyme (Lyz) 55 confirmed the presence of a group of WGA + Lyz -cells corresponding to goblet cells (white arrows on Fig. S9c), and its reduction after IL-10 treatments (Fig. S9d). Therefore, these results conclude that IL-10 signaling in intestinal epithelial cells modulates goblet cell frequencies across vertebrates through the activation of Notch signaling.

DISCUSSION
How IL-10 signaling controls intestinal immune homeostasis has been an active area of research over the past two decades. However, due to the viviparous nature of the sample sources such as humans and mice, many of these studies were performed in adult specimens, and studies understanding the role of cytokines in intestinal development remain elusive. The ex utero embryonic development and the transparency of the zebrafish larva made it an ideal model to investigate the role of cytokines in the development of intestinal mucosal immunology. Here, we combined zebrafish and mouse organoids to demonstrate that IL-10 signaling controls goblet cell homeostasis in a mechanism involving Notch signaling. Thus, we provided additional evidence showing that cytokines can act beyond leukocytes to orchestrate tissue development and that zebrafish is an excellent model organism to investigate the development and maturation of intestinal mucosal barrier functions.
The relationship between IL-10 signaling and goblet cell function has been a matter of study over the past years. An early study showed that Muc2 was predominantly overexpressed but aberrantly sulfated in the colon of IL-10 deficient mice. 56 This was further addressed in a later report, which showed alterations in the colonic mucus layer of IL-10 deficient mice, specifically a thicker adherent mucus layer which was however more permeable to bacteria 57 . In line with a role in orchestrating epithelial homeostasis, IL-10R signaling disruption specifically in intestinal epithelial cells (IECs) skewed IEC differentiation towards goblet cells over absorptive enterocytes. 14 Moreover, using mouse organoids treated with recombinant IL-10 it was proposed that goblet cell modulation could be a consequence of altered intestinal stem cell (ISC) maintenance and/or differentiation. 13 However, how IL-10 signaling might control goblet cell homeostasis, whether IL-10 signaling acts directly on ISC, and if such cytokine-IEC crosstalk occurs early in development remained elusive. Using zebrafish, our study proposes that IL-10 signaling regulates intestinal goblet cell homeostasis early in intestinal development in a mechanism involving Notch signaling and without affecting the ISC pool. The latter was investigated in murine organoids and zebrafish, using the Lgr5 reporter mice and zebrafish markers such as olfm4.2, the ortholog of the mammalian stem cell marker OLFM4 and a potential marker for intestinal stem cells in adult zebrafish. 58 In addition, we did not observe differences in the teleost intestinal progenitor marker sox9b, 48,59,60 nor in the proportion of absorptive neutral red + lysosome-rich enterocytes (LREs), suggesting that il10 deficiency in zebrafish leads to a specific increase in goblet cells in the larval mid intestine.
We found a temporal correlation between intestinal il10 expression and Notch activity in controlling secretory cell differentiation. Further, our data highlighted IL-10 signaling as a potential upstream regulator of Notch activity in the zebrafish intestine, which helped to explain the increase of goblet cells observed in il10-Mut larvae. Although the inhibition of Notch activity at the peak of intestinal il10 expression(2-3dpf) recapitulated the phenotypes observed in il10-Mut larvae, rescue experiments are needed to directly link Notch as a downstream effector of IL-10 signaling. To rescue Notch activity, a previous study used transgenic animals that overexpressed the Notch intracellular domain (NICD), leading to a permanent activation of Notch signaling. 42 In our approach, we used a novel synthetic Notch signaling activator Yhhu-3792 50 which allowed us to rescue Notch signaling specifically at the 2-3dpf time window. Since intestinal ab + goblet cells were restored to WT levels in il10-Mut after Yhhu-3792-mediated Notch activation at 2-3dpf, our data strongly suggest that early expression of il10 is key in controlling goblet cell homeostasis in zebrafish larvae through the regulation of Notch activity. Although Stat3 could serve as a potential link between IL-10 and Notch signaling, as previously reported on monocytes, 61 our data suggest a Stat3-independent mechanism linking both signals. An alternative mechanism could be PI3K/Akt signaling, previously described to be activated by IL-10, 62,63 or other members of the Stat family which can be activated by IL-10, such as Stat1 and Stat5. 47 However, further studies will be required to find the exact mechanisms which might explain the connection between il10 and Notch signaling in the intestine of zebrafish.
Our results from mouse small intestine (SI) organoids revealed that the IL-10-Notch axis is also present in mammals and further confirms its independence from the microbiota. Although we did not find differences in the transcript levels for secretory cell markers in organoids, it is possible that these changes may be visible at a protein level, as observed in zebrafish. In addition, SI organoids do not show elevated numbers of goblet cells under normal sterile growth conditions, thus the addition of inducers of goblet cell differentiation, such as type 2 cytokines or Notch inhibitors 64 would be needed to potentially visualize its regulation by IL-10. Nevertheless, our results from SI organoids match previous findings from single-cell transcriptomic analysis of mouse intestinal organoids, which showed an increased Rgcc expression after IL-10 treatments. 13 Given that both Rgcc and Hes1 are used as markers for intestinal progenitors in mice, 65 there is the possibility that IL-10 is modulating the intestinal stem cell (ISC) pool. However, we did not observe changes in Epcam + Lgr5-GFP + ISCs in our organoid settings after IL-10 treatments. Therefore, the decreased average in the ab-stained area from IL-10-treated organoids, as well as the decreased frequency of Epcam + Lgr5-GFP -WGA hi CD24cells observed by FACS, indicate that goblet cell homeostasis is specifically modulated by IL-10 in mice. Deeper analyses, focusing on specific intestinal cell populations, are required to better elucidate if IL-10 signaling impacts the specification of intestinal precursors to a secretory lineage or acts directly on goblet cells to control its function.
Our study might have implications for a better understanding of pediatric IBD, as mutations in IL-10 signaling pathway-associated genes are frequent in very-early onset (VEO) IBD cases, and dysregulation of mucus biosynthesis and goblet cell numbers are classic features of IBD. Thus, it is likely that defective IL-10 signaling may contribute to the severity of VEO-IBD by affecting early intestinal epithelial cell development, which may impact subsequent processes such as microbiota colonization. The fact that IL-10 can act beyond immune responses and directly control the differentiation of IECs raises several points to be considered in clinical studies. First, it suggests that aberrant immune responses are not the only driving cause for the development of VEO-IBD in patients carrying mutations in IL-10 signaling-related genes, which explains their resistance to immunosuppressive therapies 66 and improving defective intestinal development and differentiation might also be considered for therapeutics. 67 Second, although alterations in mucus and goblet cells are observed in active VEO-IBD cases, 7,68 there is no clear information about the intestinal cell composition before the onset of the disease and with no active inflammation, which could change the view on the causal factors and the physiological consequences of IBD pathogenesis. Third, despite the lack of clinical evidence directly linking IL-10 signaling, goblet cells, and VEO-IBD, data from experimental models support this idea, and further investigation needs to be conducted to better understand how these factors play a role in the initiation of the disease.
Thus, we propose the existence of an IL-10-Notch axis in the early development of intestinal epithelial cells which regulates mucin production, goblet cell differentiation, and is conserved across vertebrates.

MATERIAL AND METHODS Animals
Zebrafish (Danio rerio, AB strain) were reared and kept in the Karolinska Institutet (KI) Zebrafish Core Facility and the Institut Curie Animal Facility, according to standard protocols. Ethical permits used for zebrafish husbandry were Nr 5756/17 (to Eduardo Villablanca), 14049/19 (to KI Zebrafish Core Facility), and APAFiS #21197-2019062521156746-v2 (To Pedro Hernandez). Zebrafish embryos were collected by natural spawning of adults and were kept at 28°C in E3 water. The zebrafish line TgBAC(cldn15la-GFP) pd1034, 36 referred to in the text as Tg(cldn15la:GFP), was provided by Prof. Michael Bagnat (Duke University Medical Center, USA); the line Tg(Tp1bglob:eGFP) um14, 45 referred in the manuscript as Tg(tp1:EGFP) was kindly provided by Prof. Olov Andersson (Department of Cell and Molecular Biology CMB, Karolinska Institutet, Sweden); and the line Tg(7xStat3-Hsv.Ul23:EGFP), 48 referred in the manuscript as Tg(7xSta-t3:EGFP), was kindly provided by Prof. Francesco Argenton (Università degli Studi di Padova, Italy).
Wild-type mice (Mus musculus, C57/Bl6 strain) were purchased from Charles River Laboratories or Taconic and were kept in the Comparative Medicine facility from Karolinska University Hospital (AKM). The transgenic strain Lgr5-EGFP-IRES-creERT2, 52 referred to in the text as Lgr5-GFP, was maintained in a C57/Bl6 background at the Comparative Medicine facility from Karolinska Institutet (KM-B). All mice used for experiments were 6-10 weeks old. Animals were handled according to protocols approved by the Stockholm Regional Ethics Committee (Nr 3227-2017 and 6778-2020). All experiments were performed following the national and institutional guidelines and regulations.

Isolation of fluorescent intestinal epithelial cells from zebrafish larvae
To acquire the intestinal epithelial population, approximately 100 of 6dpf TgBAC(cldn15la:GFP) zebrafish larvae were collected for the FACS experiment, then intestines were dissected and placed into PBS on ice with a maximum dissection time of 2 h. Intestine dissociation was performed using TrypLE Express (Gibco, Cat. No. 12605028) for 1 h at 37°C, pipetting up and down every 10 min to support digestion. Digested samples were spun at 1500 g for 5 min at 4°C and washed twice with PBS 1× before being resuspended together with PBS 1× and 10% FBS (fetal bovine serum). Filtered cells were immediately subjected to FACS at Institut Curie Flow Cytometry Platform using the Sony SH800 Cell Sorter. Dead cells were excluded from analysis using the combination of Calcein Blue (Invitrogen, Cat. No. 65-0855-39) and Propidium Iodide viability stains (Sigma-Aldrich, Cat. No. P4864). Nontransgenic and single transgenic controls (pools of 10 dissected guts) were prepared as above and used for gating and compensation. RNA isolation was done using on average 30,000 GFP + or GFP − sorted cells with the Single-Cell RNA Purification Kit from Norgen Biotek Corp and reverse transcribed using Superscript IV Reverse Transcriptase (Life Technologies, Cat. No. 18090050) following the manufacturer's instructions.

Whole-mount in situ hybridization in zebrafish larvae
Portions of the coding sequences for the ZFIN-annotated orthologs of the human IL10RA (Gene name: il10ra, Gene ID: 777651, ENSEMBL ID: ENSDARG00000100383) and IL10RB genes (Gene name: il10rb, Gene ID: 619391, ENSDARG00000078042) were amplified by PCR using the primers from Table S1, and subsequently cloned into a pCRII-TOPO vector (Invitrogen, Cat. No. 450640). Probe and control RNA sequences were synthesized in vitro by using T7 and SP6 RNA polymerases and subsequently purified by Lithium chloride (LiCl) precipitation. Wholemount in situ hybridizations were performed in 3 dpf and 5 dpf zebrafish larvae as previously described 30 with the following modifications. Permeabilization was done by treating fixed larvae with 10 μg/mL of Proteinase K (Qiagen, Cat. No. 19131) prepared in PBS supplemented with Tween-20 0.1% and DMSO 1% for 30 minutes at room temperature. Prehybridization was done at 60°C for at least 3 h in Hybridization media (HM, Formamide 50%, Sodium citrate solution 5×, citric acid 9.2 mM, Heparin 50 μg/mL, RNAse-free yeast tRNA 500 μg/mL). Probe hybridization was done using 500 μL of HM containing 4 ng/mL of either probe or control RNA, and samples were incubated overnight at 60°C. Anti-dig-AP antibody (Roche, Cat. No. 1109327493) was diluted 1:3000 in PBS supplemented with Tween-20 0.1%, sheep serum 2%, and bovine serum albumin (BSA) 2 mg/mL. Development of probe hybridizations was done by diluting NBT-BCIP stock solution (Roche, Cat. No. 11681451001) 1:50 in alkaline phosphatase buffer (Tris-HCl 100 mM pH 9.5, MgCl 2 , NaCl, Tween-20 0.1%). After 3 h of developing, samples were washed extensively with PBS plus Tween-20 0.1%, fixed in PFA 4%, washed again in PBS Tween-20 0.1%, and subsequently dehydrated in Methanol and stored at −80°C for at least one night, to allow clearing of background signal. Larvae were rehydrated, transferred to Glycerol 85%, and imaged using a Nikon SMZ25 stereoscope equipped with a DS-Fi3 camera.

Generation of il10-Mut zebrafish lines
The zebrafish coding sequence for the ortholog of the human IL10 gene (Gene name: il10, Gene ID: 553957, ENSEMBL ID: ENSDARG00000078147) 31 was targeted by CRISPR/Cas9 37 with specific sgRNAs (Table S1). For the maintenance of wild-type and il10 mutant stocks, zebrafish embryos coming from the incross of heterozygous il10 individuals were raised, and genotyping was performed once they reached adulthood by fin-clipping a portion of the tail fin. The il10 uu1751 mutation created a restriction enzyme site for EcoNI, which is absent in wild-type fish. Carriers of the il10 uu1751 mutation were screened by PCR followed by enzymatic digestion with EcoNI. Zebrafish carrying the il10 uu1762 mutation were screened by PCR using combinations of primers in which the forward primer was specific for either the WT or mutant sequence. Primers used for genotyping are shown in Table S1.
Cloning of the zebrafish il10 and il17a/f3 genes and rescue experiments The coding sequences for the zebrafish il10 and il17a/f3 genes were amplified by PCR using specific primers for both genes (Table S1) and then cloned in a pCR2.1-TOPO vector (Invitrogen, Cat. No. 450641). In vitrotranscribed mRNA was synthesized using the T7 mMessage mMachine kit (Invitrogen, Cat. No. AM1344), following the manufacturer's guidelines. For injections, 1nL of 200 ng/μL of in vitro-transcribed mRNA was co-injected with 0.5% of Rhodamine-Dextran MW10000 (Invitrogen, Cat. No. D1824) in 1-cell stage wild-type or il10-Mut embryos. Embryos injected with 0.5% Rhodamine-dextran were used as control. At 2dpf, Rhodamine-positive embryos were selected and used for subsequent analysis.

Body-intestine isolation in zebrafish larva
Larvae from 3 dpf to 5 dpf were euthanized by overdosing them with MS-222 and intestines were isolated mechanically with the use of needles to pierce the tissue and extract the intestines. Around 10 intestines and their respective bodies were collected for analysis.

RNA extraction and qRT-PCR from zebrafish samples
For zebrafish samples, up to 10 tissue samples (i.e. whole larvae, intestines, or body carcasses) were pooled and collected in Trizol reagent (Invitrogen, Cat. no. 10296010), and tissues were homogenized by pipetting the samples repeatedly through 23G and 27G needles. Total RNA was extracted following Trizol manufacturer's instructions. Synthesis of cDNA was performed using the iScript cDNA synthesis kit (BioRad, Cat. No. 1708841). Quantitative PCR was performed using iTaq Universal SYBR Green supermix (BioRad, Cat. No. 1725124), as previously reported. 69 Primers used for zebrafish qRT-PCR analysis are found in Table S2.

Bleach and antibiotic treatments in zebrafish larvae
Zebrafish embryos between 24-28 hpf were treated with chlorine hypochlorite (0.04%) for 5 minutes and then washed twice with sterile E3 medium for 5 min. After washing, embryos were left in sterile E3 medium containing Ampicillin 100 μg/mL and Kanamycin 5 μg/mL and placed at 28°C in isolated containers. Media was renewed every day in sterile conditions until the day of sample collection. For the quantification of bacterial loads, genomic DNA from 2 larvae per condition was extracted by (heat-basic), and PCR against 16S All Bacteria 30 and zebrafish gDNA (for il23r gene) was performed. PCR products were run in a 2% agarose gel, and band intensities were measured in Fiji/ImageJ (NIH).

Alcian blue, neutral red stainings and quantifications in zebrafish larvae
Stainings for Neutral red (Sigma, Cat. No. N4638) and Alcian blue (Sigma, Cat. No. B8438) in 5dpf zebrafish larvae were performed as previously described. 70 Alcian blue-stained larvae were additionally treated with H 2 O 2 1.5%/KOH 0.5% to remove pigments before imaging. Larvae were mounted in a lateral position using 1% low-gelling point agarose (Sigma, Cat. No. A9414), and RGB images were acquired using a Nikon SMZ25 stereoscope equipped with a DS-Fi3 camera. Images were cropped to keep the midintestine section. Automatic unbiased analyses of the Alcian blue-stained area in the intestine of individual larvae were performed in Fiji/ImageJ software (NIH) using "Colour Deconvolution 1.7" and selecting the "Alcian blue & H" vector to identify the Alcian blue-stained area. For the automatic detection of the Neutral red-stained area, the function "Color Threshold" was used. The length of the alcian blue-and neutral red-stained regions were measured manually in randomized images.

Immunofluorescence staining in zebrafish larvae
Immunostaining was performed on whole larvae at 5 days post fertilization. Paraformaldehyde at 4% was used to fix zebrafish larvae overnight at 4°C. The samples were then washed with distilled water. For whole-mount 2F11 immunostaining, fixed larvae were permeabilized with cold 100% acetone for 20 min at 4°C before being washed three times with PBST (PBS supplemented with 0,5% Triton-X-100

Fluorescent reporter imaging and analysis
The fluorescent reporters for Notch activity Tg(tp1:GFP) and Stat3 activity Tg(7xStat3:EGFP), in either wild-type or il10 uu1751 genetic background, were imaged at 5dpf in an SMZ25 stereoscope equipped with a CoolLed laser set and a DS-Fi3 camera (Nikon), focusing on the intestinal region. Z-stacks were merged in Fiji/ImageJ software, and the number of GFP + cells in the intestinal region was counted manually in Fiji/ImageJ.

Dissection, sectioning, and staining of adult zebrafish intestines
Zebrafish older than 1 year of age were euthanized with an overdose of MS-222. Intestines were collected and fixed in neutral buffered formalin (Sigma, Cat. No. HT-501128) overnight at room temperature, and subsequently transferred to ethanol 70% for at least 24 h. Tissues were dehydrated and embedded in paraffin, and 10 µm sections were collected and stained with alcian blue. Briefly, tissue sections were deparaffinized and rehydrated to PBS, and incubated in 3% alcian blue solution (Sigma) for 5 min. After staining, tissues were counterstained with nuclear fast red (Sigma, Cat. No. N3020). Images were acquired in a Nikon SMZ25 stereoscope and quantifications of goblet cells per villus area were performed using QuPath 0.2.3 (University of Edinburgh, UK).

Mouse intestinal organoid cultures
Mouse organoids were generated from crypts derived from the entire small intestine (SI). Briefly, SIs from WT mice were collected and flushed with PBS, cut opened longitudinally, and subsequently cut into five pieces of similar size. Tissue pieces were placed in PBS and were vigorously shaken to remove mucus. Tissue pieces were then transferred to cold PBS-EDTA 10 mM and incubated for 1 h in ice. After incubation, SI villi were removed by gentle scraping of the luminal side using two glass slides. Then SI crypts were scraped from the tissue pieces by applying stronger pressure with the glass slides and collected in recipient tubes filled with cold PBS. Collected crypts were centrifuged for 5 min at 300 × g and 4°C, and then quantified using an upright microscope by placing 10 µL of crypt solution on a glass slide. The basic culture medium (ENR) contained advanced DMEM/F12, 1× penicillin/streptomycin, 1x Glutamax (Thermo Fisher Scientific), 10 mM HEPES (Thermo Fisher Scientific), 1x B27 supplement (Life Technologies, Cat. No.17504044), 1× N2 supplement (Life Technologies, Cat. No. 17504048), 1mM N-acetylcysteine (NAC, from Sigma, Cat. No. A9165) and was supplemented with 50 ng/mL of recombinant murine epidermal growth factor (EGF, from R&D, Cat. No. 2028-EG), 250 ng/mL recombinant murine R-Spondin (R&D, Cat. No. 3474-RS) and 100 ng/mL recombinant murine Noggin (Peprotech, Cat. No. 250-38). SI crypts were resuspended in 30-40% basic culture medium with 60-70% Matrigel (Corning, Cat. No. 356231) and 20 µl containing approximately 500 crypts were plated in a pre-warmed flat-bottom 48-well plate (Sarstedt, Cat. No. 83.3923). The plate was placed at 37°C and allowed to solidify for 15 min before 200 µl of ENR medium (containing the different stimuli) was overlaid. The medium was replaced every 2 days with fresh medium and cultures were maintained at 37°C in fully humidified chambers containing 5% CO 2 . During the first 2 days of culture, the ENR medium was supplemented with 10 µM of the ROCK inhibitor Y-27632 (Sigma, Cat. No. Y0503). For organoid in vitro stimulation, 25 ng/mL of recombinant murine IL-10 (Peprotech, Cat. No. 210-10) diluted in 0.1% bovine serum albumin (BSA, Sigma) was added in the ENR medium for the entire duration of the organoid cultures. Control organoids were supplemented with a similar volume of 0.1% BSA vehicle. Each condition was plated in technical triplicates. On day 4 of culture, crypt domains per IL-10-treated or vehicle-treated organoids were quantified in 2-3 wells/condition Each dot in the quantification plot represents one mouse and an average of 2-3 technical replicates.

Alcian blue stainings on mouse SI organoids
The Alcian blue protocol used for the staining of mouse SI organoids was adapted from the protocol used for whole-mount Alcian blue stainings in zebrafish. Organoids were collected after 4 days of culture, washed 5 times with cold PBS-BSA 0.1%, and replated in a new Matrigel stock in 24-well plates, to remove debris from the initial culture. After Matrigel gelification, organoids were fixed in PFA 4% for 45 minutes, washed twice with PBS-BSA 0.1%, and washed twice with acidic ethanol (70% Ethanol supplemented with 1% HCl). Organoids were stained with Alcian blue solution (Sigma, Cat. No. B8438) for 1 h at room temperature. After the incubation, stained organoids were washed extensively with acidic ethanol and were left overnight in acidic ethanol at 4°C, to remove excessive staining from the Matrigel. Once the Matrigel became clear, organoids were washed with PBS-BSA 0.1% and imaged using the Nikon SMZ25 system.

RNA extraction from mouse organoids and qRT-PCR
After 4 days of culture, treated organoids were harvested and Matrigel was removed by consecutive washes with cold PBS-BSA 0.1%. Cleaned organoids were resuspended in RLT-plus buffer supplemented with 1% β-Mercaptoethanol, and RNA extraction was performed using the RNeasy extraction kit (Qiagen, Cat. No. 74136), following the manufacturer's instructions. Similar to zebrafish samples, synthesis of cDNA was performed using the iScript cDNA synthesis kit (BioRad), and quantitative PCR was performed using iQ SYBR Green mix (BioRad). Primers used for mouse qRT-PCR analysis are found in Table S2.

Statistical analysis
Quantitative data were analyzed using GraphPad Prism v8.0 software (GraphPad, San Diego, CA, USA). Group comparisons were considered statistically significant when they reached a p-value below 0.05. R.A. Morales et al.