Therapeutic zinc targets dysregulated GC-C signaling and restores ileal defects in a preclinical model of familial diarrheal disease

Abstract

Hyperactivating mutations in guanylyl cyclase C (GC-C) are monogenic causes of early-onset inflammatory bowel disease, familial diarrheal syndrome and congenital secretory diarrhea. The mechanisms linking elevated cGMP levels to immune imbalance remain poorly defined. Here, using a preclinical model of a disease-associated GC-C mutation, we observe pleiotropic alterations in the small intestinal epithelium. Transcriptomic and functional analyses revealed impaired Paneth and goblet cell differentiation, compromised barrier integrity, heightened epithelial permeability, and increased proinflammatory cytokine levels. Intestinal organoids from mutant mice exhibited amplified cGMP responses to GC-C ligands and defects in secretory lineage specification, confirming cell-autonomous mechanisms. Strikingly, oral zinc administration suppressed aberrant GC-C activity, normalized cGMP levels and restored barrier function. These findings highlight the central role of epithelial cGMP signaling in coordinating barrier integrity and immune–epithelial interactions, and identify zinc as a tractable therapeutic strategy for GC-C–mediated intestinal disorders.

Graphical Abstract

NEW & NOTEWORTHY

Activating mutations in GUCY2C, which encodes the receptor guanylyl cyclase C (GC-C), cause early-onset diarrheal disease and gastrointestinal inflammation. Knock-in mice carrying a familial diarrheal syndrome mutation exhibited impaired gut barrier function. Mutant organoids showed defective secretory lineage specification associated with reduced Wnt3 expression. Zinc administration, which lowers epithelial cGMP levels by inhibiting GC-C, reversed most pathological changes in both mice and organoids.

INTRODUCTION

Maintenance of intestinal homeostasis requires coordinated signaling between epithelial cells that form the barrier, the microbiome, and immune cells that provide surveillance at the mucosal interface. Disruption of this crosstalk is a hallmark of inflammatory bowel disease (IBD), where barrier defects, dysbiosis, altered secretory cell function, and aberrant immune responses converge to drive chronic inflammation (1–3). A large number of monogenic causes of IBD in the pediatric population have been identified (4, 5). Most of them are associated with immune cell dysfunction (6) and fewer with genes expressed in intestinal epithelial cells (5). There is a paucity of preclinical models that mimic different monogenic causes of Crohn’s disease (CD) (7, 8). Therefore, a mechanistic understanding of the underlying causes of inflammation mediated by individual genes that can lead to directed therapeutic strategies, is lacking.

The intestinal receptor guanylyl cyclase C (GC-C), activated by the endogenous peptides guanylin and uroguanylin, regulates fluid and ion secretion through cGMP signaling (9–11). Bacterial heat-stable enterotoxins are a major cause of diarrheal disease and are super-agonists of GC-C (12). In addition to its canonical role in electrolyte transport, GC-C signaling influences barrier integrity (13), and host–microbiota interactions (14, 15). Activating mutations in the human GUCY2C gene have been linked to familial diarrhea syndrome and congenital secretory diarrhea (16–19), with gut abnormalities continuing to adulthood (20). These mutations either lead to augmented cGMP production following ligand-mediated activation of GC-C, and/or show enhanced basal cGMP production even in the absence of ligand (17). Affected individuals have chronic diarrhea that is of early onset, and is associated with increased susceptibility to inflammatory bowel disease, small-bowel obstruction, and esophagitis. In pediatric cases, more severe GUCY2C mutations present with intrauterine onset of diarrhea, and newborns show prominent abdominal distension due to dilated fluid-filled loops, which require surgical intervention (17). Mechanisms by which aberrant cGMP signaling mediated by GC-C hyperactivation perturbs epithelial–immune homeostasis remain unclear.

In earlier studies from the laboratory, we have described the generation of a mutant mouse harboring the first reported hyperactivating mutation in GUCY2C that was associated with familial diarrhea disease and CD (21). The disease-associated mutation was at position S840 that was converted to an isoleucine present in the guanylyl cyclase domain of human GC-C. The equivalent mutation, S839I, in mice mimicked phenotypes seen in humans (16), and molecular dissection of changes in the colon of these mice revealed an inflammatory signature and dysbiosis similar to that seen in patients with ulcerative colitis (21). This was the first report of epithelial cGMP produced by GC-C as a mediator of colonic dysfunction associated with diarrhea and susceptibility to colitis.

A severe consequence of activating mutations in GC-C is ileal inflammation, which is diagnosed as CD in pediatric and adult populations (16, 17). Here, we investigated the impact of the S839I mutation in GC-C on small intestinal physiology and immunity. Using complementary in vivo and organoid models, we demonstrate that hyperactive GC-C disrupts secretory lineage specification, compromises barrier function, culminating in heightened permeability and proinflammatory cytokine production. Strikingly, oral zinc supplementation to mutant mice inhibited aberrant GC-C activity and restored barrier and immune function. Our findings identify aberrant GC-C signaling as a driver of epithelial–immune dysregulation and reveal zinc as a readily deployable therapeutic strategy for GC-C–mediated intestinal disorders.

MATERIALS AND METHODS

Maintenance of mice.

Experiments on the S839I mice (C57BL6/N) described earlier (21) agreed with the Institutional Animal Ethics Committee of the Indian Institute of Science (Approval CAF/Ethics/547/2017 and CAF/ETHICS/086/2024). All animals were bred and housed as described earlier (21) in the same vivarium. Chow was procured from Altromin International (Germany) and contained ~ 24% protein, 6% oil, and 3% dietary fiber. Mice of both sexes were used for experiments unless specified. Mice were between 6–8 weeks of age, with a body weight between 20–25 g, and were considered healthy at the time of sacrifice. Wild type mice were bred in the colony and mice of similar age and sex to the S839I mice were used for experiments. Mice of similar age, sex and genotype were randomized when selected for experimentation. The number of mice used for experiments were determined using G* (22; RRID:SCR_013726. Investigators were not blinded during the study, except during some data analysis.

Transcriptomic analysis of ileal tissue.

Approximately 1 cm of the terminal ileum was collected from 4 eight-week-old female wild type and S839I mice, and RNA prepared (21). 3 μg RNA was used for RNA Sequencing. The raw sequence data were generated using Illumina HiSeq by Clevergene Biocorp Pvt Ltd, Bengaluru, India. The raw data generated was analyzed in the Ingenuity Pathway RNAseq Analysis Portal (Qiagen) after uploading the Fastq files. Filters placed were to restrict the biotype to protein-coding genes only, and with an expression of ≥ 2 in S839I mice to get a list of differentially expressed genes.

Data was analyzed by Ingenuity Pathway Analysis (IPA; Qiagen; RRID:SCR_008653) to identify direct and indirect relationships. We constrained IPA to analyze the data according to genes expressed in the intestine known to be associated with ileitis, ileal inflammation and CD disease. 345 genes were downregulated, and 298 genes upregulated in the terminal ileum of S839I mice with fold changes < 2 and >2 and false discovery rate of 0.1 Networks included endogenous chemicals, and causal networks were defined to include inflammation of the ileum and CD disease, with genes expressed in colon cancer cell lines and small and large intestine. Upstream Regulators were filtered for genes and proteins with a z-score cut-off at ~ 2.

To identify cell-specific gene expression, analysis was performed at the Broad Institute Single Cell Analysis portal (https://singlecell.broadinstitute.org/single_cell) using data from Haber et al. (23). Expression in individual cell types was compared in heat maps to the fold change in expression seen in S839I mice in comparison to wild type mice.

Reverse transcription and quantitative PCR (RT-qPCR)

RNA was prepared and reverse transcribed, using 2 μg of RNA. Real-time PCR was performed using SYBR Premix Ex Taq (Tli RNase H Plus) on a CFX96 Touch real-time PCR detection system (Bio-Rad, USA). The housekeeping genes, glyceraldehyde 3-phosphate dehydrogenase (Gapdh), was used for normalization of the real-time PCR data. The sequences of the primers used for real-time quantitative PCR were obtained from those validated at the PrimerBank database (24) and are shown in Supplemental Table 1.

Periodic acid Schiff’s (PAS) and Alcian staining

1 cm of the terminal ileum was collected from 6 to 8 weeks old mice and processed as described earlier (21). Sections were stained with 0.5% Periodic acid solution followed by rinsing in water and staining with Schiff’s reagent for 15 min. The sections were rinsed in water and stained with haematoxylin for 2 min. Stained sections were imaged using an Axio-Observer Z1 microscope (Carl Zeiss Microimaging, Germany).

To visualize the mucus, 1 cm of the terminal ileum was collected from 6–8 weeks-old mice and fixed in Carnoy’s solution (60% ethanol, 30% chloroform, 10% glacial acetic acid) for 2 h at room temperature. Tissues were transferred to 100% ethanol after 2 h till further processing. Tissues were serially immersed in 100% ethanol, xylene, and finally, paraffin using an automated tissue processing system (Leica Biosystems, Germany). Tissues were embedded in paraffin, and 5 μm sections were prepared, dewaxed and rehydrated by serial immersion in reducing concentrations of ethanol and finally rinsed in distilled water. PAS staining was performed as described above. For Alcian Blue staining, tissue sections were incubated in Alcian blue solution (pH 2.5) (Merck) for 5 min and washed in distilled water. Tissue sections were stained for 20 s in haematoxylin and images were acquired using an Axio-Observer Z1 microscope (Carl Zeiss Microimaging, Germany)

Transmission Electron Microscopy (TEM)

Ileum tissues were fixed in 3% glutaraldehyde (0.1 M sodium cacodylate buffer, pH 7.4) overnight at 4°C, rinsed in buffer, and post-fixed with 1% osmium tetroxide for 2 hr. Following dehydration through an ethanol series (50–100%) and propylene oxide clearing, samples were infiltrated with epoxy resin (propylene oxide:resin 1:1 for 2 hr, pure resin overnight) and polymerized in silicone molds at 60°C for 48 hr. Semi-thin sections (1 μm) cut via Leica Ultracut UCT ultramicrotome (glass knife) were toluidine blue-stained for light microscopy screening. Ultrathin sections (70 nm, diamond knife) mounted on 200-mesh copper grids underwent double staining: 2% uranyl acetate (10 min) followed by Reynold’s lead citrate (5 min in CO₂-free chamber with NaOH pellets), with thorough rinsing (0.02 N NaOH, distilled water) to prevent precipitates. Grids were analyzed using an FEI Tecnai T12 TEM (ThermoFisher Scientific) at 80–120 kV, with artifact mitigation via chloroform vapor relaxation and strict stain filtration (0.22 μm). TEM was performed in the facility at the Christian Medical College, Vellore, Tamil Nadu.

Immunofluorescence for staining tissue sections

Antigen retrieval was performed on tissue sections (5 μM) by incubation in 10 mM sodium citrate buffer (pH 6.0) containing 0.05% Tween-20 in a microwave at 85 °C for 10 min. The slides containing the tissue sections were washed in Tris buffer containing 0.025 % Triton X-100 for permeabilization, blocked in buffer containing 1% BSA, and incubated with primary antibody overnight at 4° C. The tissue sections were washed twice for 5 min each in TBS and incubated in a mix of the secondary antibody and Hoechst diluted in blocking buffer for 1 h at room temperature. Anti-lysozyme antibody (Abcam; RRID:AB_10861277), anti-mucin antibody (RRID:AB_1523987) and anti-rabbit IgG Alexa Fluor 488 (Invitrogen; RRID:AB_2535850) were used at a dilution of 1:1000. Hoechst (Invitrogen) was used at a dilution of 1:5000. The tissue sections were rinsed and imaging was performed in an Axio-Observer Z1 microscope (Carl Zeiss Microimaging, Germany). Image analysis was performed using Fiji (RRID:SCR_002285).

Measurement of small intestinal permeability.

6 to 8 weeks old mice were fasted for 6 h with access to water ad libitum. Fasted mice were orally gavaged with 200 μl of 4 kDa FITC-dextran dissolved in PBS (80 mg/ml). Mice were sacrificed after 30 min, blood was collected by cardiac puncture, serum collected and the amount of 4 kDa FITC-dextran in the serum was measured by fluorescence spectroscopy (excitation: 428 nm, emission: 585 nm) and interpolated against a set of known concentrations of 4 kDa FITC-dextran (0 – 16 ng/μl) diluted in PBS

Measurement of cytokines in ileal extracts

8 weeks-old mice were sacrificed by CO₂ asphyxiation, and 3 cm of the terminal ileum was harvested and snap-frozen in liquid nitrogen. The frozen tissue was crushed to a powder in liquid nitrogen using a mortar and pestle. The powdered tissue was transferred to homogenization buffer containing 50 mM HEPES pH 7.5, 100 mM NaCl, 5 mM EDTA, 1 mM DTT, 5 μg/ml SBTI, 5 μg/ml leupeptin, 5 μg/ml aprotinin, 2 mM PMSF, 10 mM sodium orthovanadate, 1 mM sodium pyrophosphate, 10 mM β-glycerophosphate, and 20 mM sodium fluoride. The extract was subjected to sonication at 60 cycles, 60% amplitude (each pulse 5 sec) (IKA Labortechnik, Germany), followed by centrifugation at 12,000 g for 60 min. The supernatant containing the cytosol was collected, and protein estimated by the modified Bradford assay (25). Total levels of the cytokines – Ifnα, Ifnγ, Ifnλ3, Tnfα, Il6, Il1β, Il18, and Il10, were measured in the soluble fractions of the tissue extracts (50 μl of neat extract) by ELISA kits (Invitrogen) following the manufacturer’s protocol. The amount of cytokine was normalized to the total protein present in the soluble fraction of the tissue extract.

Generation of mouse small intestinal organoids

Small intestinal organoids were generated from wild type and S839I C57BL/6 mice (both male and female mice) aged ~ 6 weeks. Around 10 cm of the small intestine from the ileo-cecal junction was taken and processed according to the protocol provided by Stem Cell Technologies (https://www.stemcell.com/how-to-establish-mouse-intestinal-organoid-culture.html). Organoids were cultured in IntestiCult Organoid Growth media (Stem Cell Technologies; Catalog No 06005) in phenol red free, growth Factor reduced Matrigel (1:1 medium; Corning Cat. No. 356231) domes at 37° C in a humidified incubator in the presence of 5 % CO₂ and ambient O_2. Drops of 50 μl were cultured in 24-well tissue culture dishes (Corning; Catalog No 3524) in 500–700 μl of Organoid Growth Media. Crypts formed spheroids after 24 h and started budding to form organoids after 72 h,. Media was changed every 72 h, and organoids were passaged every 5–7 days, by harvesting the organoids in chilled Cell Dissociation Reagent (Stem Cell) as per the manufacturer’s protocol. A split ratio of 1:4 was used, and RNA prepared from organoids 4–6 days after seeding, after at least 2–3 passages. Passaging was performed using Gentle Cell Dissociation Reagent (Stem Cell Technologies; Catalog No. 100–0485) as per the manufacturer’s protocol. Organoids at passage numbers less than 10 were used for all experiments.

Fluid secretion into the lumen of the organoids causes swelling, which was studied by time-lapse microscopy. Organoids were treated with agonists (Sp-8-pCpt-cGMPS (Biolog) or uroguanylin (Bachem)) and imaged.

Organoids were preincubated with zinc sulphate (10μM) for 4h at 37° C and 5% CO₂. Media was removed and fresh media containing uroguanylin (10⁻⁶ M) was added. Time-lapse microscopy was performed immediately using a Zeiss Cell Discover 7 microscope, maintaining CO₂ at 5% and humidity. Image analysis was performed using Zeiss Zen Lite (RRID:SCR_023747). To quantify the volume of the organoid lumen, the length and width of the lumen were measured, and the volume was calculated using the following formula: volume = 4/3π × (length/2) × (width/2) (26). The relative volume at 4h was normalized to the volume at 0 min (set as 1.0).

Isolation of RNA and RT-qPCR from mouse SI organoids

Matrigel drops containing ~ 200 organoids were dissociated by pipetting 1 ml of chilled Advanced DMEM:F12 directly on top of the dome, pipetted up and down 15 times to release the organoids from the dome and centrifuged at 290 g for 5 min at 4 °C. The supernatant was decanted, and the pellet containing the intact organoids was process for RNA isolation (Qiagen RNEasy Mini Kit). RNA (2 μg) was subjected to RNAseq analysis performed by Azenta Life Sciences. RNA was prepared from multiple cultures and wells, across different passages, but never more than 10 passages.

Sequence reads were analyzed to remove possible adapter sequences and nucleotides with poor quality using Trimmomatic v.0.36 (RRID:SCR_011848). The trimmed reads were mapped to the Mus musculus GRCm38 reference genome available on ENSEMBL (RRID:SCR_002344) using the STAR aligner v.2.5.2b (RRID:SCR_004463). Unique gene hit counts were calculated by using featureCounts (RRID:SCR_012919) from the Subread package v.1.5.2 (RRID:SCR_009803). Only unique reads that fell within exon regions were counted. The gene hit counts table was used for downstream differential expression analysis using DESeq2 (RRID:SCR_000154). The Wald test (https://www.sciencedirect.com/topics/mathematics/wald-test) was used to generate p-values and log2 fold changes. Genes with an adjusted p-value < 0.05 and absolute log2 fold change > 1 were called as differentially expressed genes for each comparison. Cell specific genes were analyzed for differential expression as described above and heat maps prepared.

Complementary DNA preparation (with 1 μg RNA), and RT-qPCR were performed as described above from RNA prepared from multiple organoids cultures prepared over a period of 4 months.

Estimation of cGMP and cAMP levels in mouse SI organoids

Matrigel domes containing ~ 300 organoids were dissociated by the addition of 500 μl of cold Dulbecco’s Phosphate-Buffered Saline (DPBS) directly onto the dome and pipetting up and down to promote the detachment of organoids from the dome matrix. The resultant suspension was centrifuged at 290 g for 5 minutes at a temperature of 4° C, and the pellet resuspended in 100 μl of 0.1 N hydrochloric acid (HCl). The sample was heated at 95° C for 5 minutes, centrifuged at 17,000 g for 5 minutes and the supernatant collected. Cyclic GMP and cyclic cAMP were quantified using an enzyme-linked immunosorbent assay (ELISA) kit from Cayman Chemicals (USA). The levels of cyclic nucleotides were normalized to protein estimated in the 0.1N HCl extracts by the Bradford method.

Immunofluorescence of mouse organoids

Organoids were recovered from Matrigel drops in cold DMEM:F12 with gentle pipetting, centrifuged at 200g for 5 minutes and the pellet fixed in 4% PFA. Permeabilization was achieved in PBS containing 2.5% Triton-X100, 2.5% bovine serum albumin, with gentle agitation for 2–3h. IgG Alexa Fluor 488 (Invitrogen; RRID:AB_2535850) were used at a dilution of 1:1000. Hoechst (Invitrogen) was used at a dilution of 1:5000. Organoids were then incubated overnight with anti-lysozyme (Abcam; RRID:AB_10861277) at 1:100 dilution at 4°C, with gentle rocking in permeabilization buffer. Organoids were washed with PBS containing 1% Triton-X100 followed by incubation with Alexa Fluor 555 Goat anti-Rabbit IgG (Thermo Fisher Scientific Cat# A32732, RRID:AB_2633281)) for 3h at room temperature. Organoids were washed 3 times, treated with Fluoromount with DAPI counterstain (Thermo Fisher Scientific Cat # 00-4959-52), placed on polylysine-coated slides, and imaged on a Leica SP8 confocal microscope.

Maximum intensity projections were generated from all z-stack slices for both lysozyme and DAPI channels using ImageJ. Thresholding was applied to identify the regions of interest, and the corresponding positive areas were quantified. The lysozyme-positive area was normalized to the DAPI-positive area and expressed as a percentage using the formula: (Lysozyme area / DAPI area) × 100.

Western blot analysis of organoids

Organoids (~ 500) were harvested from the Matrigel drop in chilled DMEM:F12, centrifuged at 200g for 5 mins and the pellet lysed directly in 50 μl Cell Lysis Buffer (Cell Signaling Technologies; 5803). Protein was estimated by the Bradford method and 10 μg total protein subjected to SDS gel electrophoresis on a 10% polyacrylamide gel. Western blotting was performed after transfer of proteins to a PVDF membrane and probed with anti-mouse β-catenin antibody (Cell Signaling Technology Cat# 9562, RRID:AB_331149) or β-actin antibody (Cell Signaling Technology Cat# 4967, RRID:AB_330288) to normalize for protein loading. Blots were developed with Clarity Max ECL reagent from BioRad.

Zinc supplementation

6–8 weeks old mice were supplemented with 25 mM zinc sulphate in drinking water for 15 days. Bowel frequency, fecal water, fecal sodium, epithelial levels of cGMP, and small intestinal permeability were measured after 15 days of zinc sulphate supplementation. Ileal extracts were prepared after 15 days of zinc sulphate supplementation and cytokines were measured, as described earlier. Serum was collected from mice and used for estimation of zinc in circulation (27) using the QuantiChrom Zinc Assay Kit (Bioassay Systems).

Measurement of bowel frequency, fecal water and sodium content

Bowel frequency was measured in mice in the morning hours by placing an individual mouse in a clean cage without bedding material and noting the number of fecal pellets passed in 10 min. For measurement of fecal water content, fresh fecal pellets were collected in pre-weighed 1.5 ml microcentrifuge tubes. Tubes were weighed (wet weight) and then subjected to lyophilization for 12 to (14) h. The weight of the tube containing dry faeces (dry weight) was recorded, and the percentage water content in the faeces was calculated using the following formula:

$$\%\text{water}\text{content}=\left[\left(\text{wt}\text{of}\text{wet}\text{feces}-\text{wt}\text{of}\text{dry}\text{feces}\right)/\text{wt}\text{of}\text{wet}\text{feces}\right]\times 100$$

Measurement of fecal sodium

Fresh feces were collected from mice in pre-weighed 1.5 ml microcentrifuge tubes, subjected to lyophilization, and the dry weight was noted. Double distilled water was added to each sample (100 mg/ml), and the fecal pellets were homogenized. The samples were vortexed for 30 sec, followed by centrifugation at 3,000 g for 5 min. Sodium ions in the supernatant was measured using a sodium ion meter (Horiba Japan) and normalized to the weight of the dry faeces.

In vitro guanylyl cyclase assays

10 μg of membrane fraction was first incubated for 30 min on ice in assay buffer (60 mM Tris-Cl pH 7.5 and 500 μM IBMX) in the presence of varying concentrations of zinc sulphate. Following the incubation, 1 mM Mg-GTP and 10 mM MgCl₂ (source of free Mg²⁺) were added to the reaction mix and incubated at 37 °C for 10 min. The concentration of free Mg²⁺ in the reaction was calculated using (http://www.stanford.edu~cpatton/maxc.html; RRID:SCR_00045). The reaction was stopped by the addition of ice-cold 50 mM sodium acetate buffer (pH 4.75) followed by boiling at 95 °C for 5 min. The reaction mixture was centrifuged at 12,000 g for 10 min, and the amount of cGMP present in the supernatant measured by radioimmunoassay (21).

Isolation of intestinal epithelial cells (IECs) for cyclic GMP estimation

Mice were sacrificed by CO₂ asphyxiation, and 10 cm of the terminal ileum was harvested, flushed in ice-cold Hank’s balanced salt solution (HBSS), cut longitudinally open, and submerged in 10 ml of IEC dissociation buffer (HBSS containing 10 mM HEPES, 1 mM EDTA, 71.5 mM β-mercaptoethanol, 500 μM IBMX, 10 mM sodium orthovanadate, 1 mM sodium pyrophosphate, 10 mM β-glycerophosphate, and 20 mM sodium fluoride) in a 50 ml conical tube and stored on ice. The tissues were incubated at 37 °C, 100 rpm for 45 min, and vortexed for 30 sec. The tissue pieces were removed gently. The tubes were centrifuged at 3000 rpm for 10 min at 4 °C. The pellet containing the IECs was washed twice with ice-cold PBS and finally resuspended in buffer containing 10 mM sodium orthovanadate, 1 mM sodium pyrophosphate, 10 mM β-glycerophosphate, and 20 mM sodium fluoride. An aliquot of the cell suspension was taken for protein estimation by the Bradford assay. The remaining cells were harvested by centrifugation, resuspended in 0.1 N HCl, and heated at 95 °C for 5 min. The mixture was then centrifuged at 17,000 g for 10 min at 4 °C. The supernatant was collected, and cGMP levels were estimated using a cGMP ELISA kit and normalized to the amount of protein taken for the cGMP ELISA

Statistical Analysis

All data were analyzed with GraphPad Prism 10 (RRID:SCR_002798), and specific analyses used to test for significance are detailed in the Legends to the Figures.

RESULTS

RNA sequencing of the terminal ileum of S839I mice reveals disrupted immune pathways and epithelial secretory lineages

To obtain a global picture of differential gene expression in the ileum of knock-in mice, we performed RNAseq of the entire tissue and analyzed the data with Ingenuity Pathway Analysis (IPA; Figure 1A and Supplemental Figure 1A). We found that canonical pathways associated with cellular immune response, specifically the Th1 pathway, were significantly misregulated in S839I mice (Supplemental Figure 1B). Disease severity in CD patients has also been associated with an increased Th1 response (28, 29).

Figure 1. Bulk RNAseq of the terminal ileum reveals defects in secretory cells with reduced secretion of anti-microbial peptides and mucin.

(A) Heatmap showing differentially expressed protein-coding genes in the terminal ileum of S839I mice compared to wild type (WT) mice. (B) Heatmap showing predicted upstream regulators in the terminal ileum of S839I mice. Values in individual cells are the z-scores for the predicted regulators.

(C) Expression patterns of the genes shown were analyzed on the single cell portal of the Broad Institute (https://singlecell.broadinstitute.org/single_cell) using the data from Haber at al (23). Note that Cftr, Dra1 and Nhe3 are expressed in distinct epithelial cell types in the small intestine. (D) IPA analysis to determine changes in expression of genes in specific cell types that are seen in S839I mice in comparison to wild type mice. (E) Heatmap showing the expression of Paneth cell-specific markers obtained from the ileal transcriptome analysis. (F) RTqPCR of Paneth cell-specific markers normalized to Gapdh. Each dot represents RNA prepared from an individual mouse. Data were analyzed by an unpaired, two-tailed t-test with Welch’s correction. The mean ± SD and p values are shown from 4 male and 4 female mice of each genotype. (G) Upper panels: representative images from those obtained from three mice per genotype, of Periodic acid Schiff’s (PAS) staining of PFA-fixed terminal ileum (1 cm from the ileocecal junction). Red and black arrowheads indicate goblet cells and Paneth cells, respectively. Lower panel: Representative images of immunofluorescence staining against lysozyme in PFA-fixed terminal ileum. Dotted lines outline the crypt bottom. Individually stained sections from three mice were examined for the number of Paneth cells per crypt, with at least 10 crypts examined per mouse. Data are shown in the graph and analyzed using an unpaired t-test with Welch’s correction. (H) Transmission electron microscopy of the ileum. Intact longitudinal sections of intestinal crypts, clearly displaying both the lumen and the base, were selected and cropped, with auto-contrast adjustments applied using ImageJ. Crypt boundaries are marked with black dotted lines, while Paneth cells are outlined with yellow dotted lines. A representative image is shown, taken from three mice per genotype. Scale bar = 20 μm. Cells containing dark granules were counted in three mice of each genotype and the data is shown in the graph. Data was analyzed using an unpaired t-test with Welch’s correction. (I) Representative images of PAS and Alcian blue- stained sections of the terminal ileum, after fixation in in Carnoy’s solution. The number of goblet cells were counted in at least 10 crypts across 3 mice and plotted in the graph. Data was analyzed using an unpaired t-test with Welch’s correction

The upstream regulator analysis feature of IPA identifies transcriptional regulators and predicts the activation state of a transcriptional (or upstream) regulator based on the direction of change of the differentially expressed genes. Positive z-scores of some predicted upstream regulators associated with inflammation were observed (i.e., Ripk2, Parp1, Ifng), indicating activation in S839I mice. Regulators with anti-inflammatory effects, such as Il10ra and Spry2, were inhibited (Figure 1B). These suggest a proinflammatory milieu within the ileum of S839I mice during homeostatic conditions. Notably, regulators essential for mucosal repair (Tgfβ1 and Yap) were suppressed in S839I mice, suggesting impaired regenerative capacity in these animals. This inhibition seen in S839I mice could contribute to the greater susceptibility of mutant mice to injury models such as DSS, which we have described earlier (21).

GC-C, its ligands, and downstream effectors are localized in distinct cells in the gut, as seen from single-cell RNA seq of the mouse small intestinal epithelium (Figure 1C) (23). Mature enterocytes and goblet cells express the highest levels of GC-C, guanylin (Guca2a) and uroguanylin (Guca2b), though stem cells and transit amplifying (TA) cells show higher expression of Cftr, a downstream target of Prkg2. In our earlier study (21), RTqPCR analysis showed that guanylin and uroguanylin were downregulated in the S839I mice, while the expression of other genes remained similar. This pattern was consistent in the RNAseq analysis performed here. Transcript levels of Dra1 (Slc26a3), an anion exchanger protein also shown to be associated with diarrhea (30), were unchanged (Log2FC 0.2; FDR 0.6) in the RNAseq analysis. It is interesting to note that the highest expression of Dra1 is in cells distinct from those expressing the highest levels of Cftr (23). The highest level of Nhe3 (Slc9a3), another target downstream of cGMP and Prkg2, is seen in the mature proximal enterocytes (enterocytes from the duodenum and jejunum) and enteroendocrine cells, suggesting that these cells are responsible for sodium absorption into the gut. Our earlier results demonstrated that Nhe3 levels were reduced in the small intestine of S839I mice (21), raising the possibility that the sodium-enriched feces seen in S839I mutant mice and patients harboring activating mutations in GUCY2C could also be caused by reduced Nhe3 expression.

Since GC-C is expressed, albeit at a lower level, in stem cells and in immature or progenitor enterocytes (Figure 1D), we hypothesized that elevated cGMP levels could alter epithelial cell fates and functions. Indeed, our analyses showed that key genes expressed specifically in different cell types were misregulated (Figure 1D and Supplemental Figure 2). Paneth cells play a key role in enhancing gut barrier function (31, 32). Transcripts that are abundant in Paneth cells (e.g. defensins, Retnlb and Mmp7) that help establish an antimicrobial barrier at the epithelial cell surface were markedly downregulated in S839I mice (Figure 1E). This was further confirmed by RTqPCR analyses performed with independent animals (Figure 1F). However, Lyz1 (lysozyme) expression was unaltered. Importantly, Mucosal pentraxin 2 (Mptx2), which controls the expression of several genes expressed in Paneth cells (33), was downregulated, suggesting an overall compromised functioning of Paneth cells in S839I mice.

Mature Paneth cells are located at the crypt base of the small intestine. We performed histological analysis of the terminal ileum of wild type and S839I mice, which revealed reduced PAS-stained cells in the villi and the crypts, indicating a reduction in neutral, fucosylated glycoconjugates. (Figure 1G, black arrows; Supplemental Figure 3). Secretory granules stained with antibodies directed against lysozyme, a major antimicrobial peptide, showed a reduction (Figure 1G) and TEM images demonstrated fewer electron-dense granular Paneth cells, and fewer granules in cells from S839I mice (Figure 1H). The decrease in lysozyme protein in the absence of altered mRNA levels (Fig. 1F) suggests that Paneth cell maturation is compromised. CD patients show Paneth dysfunction, especially in the terminal ileum (34), highlighting the fact that elevated cGMP in the gut of the S839I mouse model contributes to CD-like symptoms reported in patients harboring activating GUCY2C mutations.

Goblet cells are responsible for the secretion of glycosylated mucins, which constitute the mucus layer that protects the intestinal lining (35). We observed a significant reduction in PAS-stained goblet cells (neutral mucins) along the villi in S839I mice (Figure 1I), black arrows). Fixing tissue in Carnoy’s solution followed by Alcian blue (to stain acidic mucins; (Figure 1I) revealed a significant reduction in the amount of mucus production in S839I mice (Supplemental Figure 3). These changes are similar to the thinning of the mucus layer seen in adult and pediatric CD patients who also have reduced goblet cells in the terminal ileum (36).

Thus, hyperactive GC-C and elevated cGMP levels impair secretory functions of Paneth and goblet cells in the ileum, resembling the changes seen in CD patients.

Compromised barrier function in S839I mice triggers inflammatory cytokine production

A reduced mucus layer is proposed to allow closer interaction between the gut microbiome and the epithelial cell, leading to increased epithelial barrier permeability and inflammation. To investigate whether epithelial barrier integrity is affected in the S839I mice, we orally gavaged them with 4 kDa FITC-dextran and collected blood after 30 min, when the dye would have migrated along the small intestine but would not have reached the cecum or colon. The concentration of FITC-dextran in the serum of S839I mice was significantly higher than that in wild type mice (Figure 2A), indicating increased small intestinal permeability in S839I mice. While there was no difference in the permeability seen between male and female wild type mice, male S839I mice showed a markedly higher permeability than females. This could be due to the protective role played by estrogen in regulating epithelial barrier permeability (37).

Figure 2. The ileum of S839I mice shows increased epithelial barrier permeability and higher levels of proinflammatory cytokines.

(A) The serum levels of 4 kDa FITC-dextran after oral gavage as a determinant of small intestinal permeability. Each dot represents an individual mouse, and the mean ± SD is shown. Data was analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted p values are shown. n = 6 male and 6 female mice of each genotype. (B) Levels of the indicated cytokines were measured in the soluble fraction of terminal ileal tissue extracts by ELISA. Each dot represents an individual mouse. Data were analyzed by an unpaired, two-tailed t-test with Welch’s correction. The mean ± SD and P values are shown. n = 5 male mice of each genotype.

We hypothesized that the disrupted epithelial barrier integrity in S839I mice could allow paracellular translocation of luminal antigens, thereby inducing chronic inflammation. We measured the levels of several cytokines in the soluble fraction of ileal extracts prepared from wild type and S839I mice by ELISA. The pro-inflammatory cytokines Ifnα, Ifnγ, Ifnλ3, Tnf, Il-1β, Il-18, and Il-6 were elevated in the ileal extracts of S839I mice, suggesting an elevated Th1-type response in the gut, validating our findings from the RNA seq data (Figure 1 A–B and Supplemental Figure 1B). No changes were seen in levels of the anti-inflammatory cytokine IL-10 (Figure 2B). Thus, these observations suggest that the S839I mice have increased epithelial barrier permeability, resulting in a pro-inflammatory environment within the intestine.

Ileal organoids derived from S839I mice reveal epithelial cell-intrinsic secretory lineage defects

We prepared ileal organoids from wild type and S839I mice and first explored whether the expression of genes in the GC-C signaling pathway was altered in S839I organoids in comparison to that seen in the intact gut. While expression levels of GC-C (Gucy2c) were not significantly different between the two sets of organoids (Figure 3A), transcripts of guanylin (Guca2a) and uroguanylin (Guca2b) were significantly lower in S839I organoids, as observed earlier from RTqPCR analysis of whole tissue (Figure 3A; (21)). This could reflect the reduced number of goblet cells in S839I organoids where guanylin and uroguanylin are expressed (Fig. 1C). Prkg2 and Cftr were expressed to similar levels in wild type and S839I organoids while levels of Dra1 were increased, in contrast to what was seen in ileal tissue. Moreover, Nhe3 expression was also increased in organoids in contrast to ileal tissue (21), suggesting that signals from non-epithelial cells are regulators of Nhe3 and Dra1 expression in the gut. Transcript levels of the sodium-dependent transporter Sglt1 (Slc5a1) were unchanged, while Slc10a2 transcripts were decreased. These results indicate that the regulation of expression of these transporters is nuanced and does not reflect an increase or decrease in a particular cell type.

Figure 3. Ileal organoids from S839I mice show increased responses to cGMP-mediated swelling.

(A) RTqPCR of GC-C signaling pathway genes in organoids derived from the terminal ileum of WT and S839I mice. All transcript levels were normalized to Gapdh. Each dot represents organoids harvested from an individual well. Mean is shown, and p values were obtained by an unpaired two-tailed t-test with Welch’s correction. (B) Organoids were cultured in domes and treated with uroguanylin (10⁻⁶ M) for 6 h. Organoids (~ 300) were then harvested, lysed in 0.1N HCl, and samples taken for measurement of cGMP, cAMP by ELISA and protein. A similar number of organoids were left untreated and used to estimate basal levels of cyclic nucleotides. Values shown are the mean,± SD, and p values were obtained by an unpaired two-tailed t-test with Welch’s correction. Each dot represents data from an individual experiment. (C) Organoids were treated with Sp-8-pCpt-cGMPS or uroguanylin and images were taken of the same organoid at the times indicated. Shown are representative images taken from at least 4 independent organoids. Experiments were repeated at least thrice. (D) WT or S839I organoids were treated with the indicated concentrations of uroguanylin, and the volume of each organoid was measured 4h after initiation of treatment. The change in volume was calculated (11) and plotted. Each dot represents measurements from multiple organoids across experiments repeated twice, Data were analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted p values are shown.

Treatment of organoids with the stable toxin peptide resulted in pronounced swelling (Supplemental Figure 4). We wanted to focus, however, on the effect of the endogenous ligand uroguanylin on organoids, since this would represent the scenario in the gut of patients harboring the S840I mutation. We measured cGMP levels in organoids with and without uorguanylin stimulation. Basal levels of cGMP were higher in S839I organoids, which reflects the expression of a hyperactive GC-C in the epithelial cells or the inhibition of cGMP-hydrolyzing phosphodiesterases (Figure 3B). Notably, cGMP levels upon stimulation with uroguanylin were higher in S839I organoids, a finding consistent with earlier observations that the S839I mutation (and the S840I mutation in human patients) resulted in higher ligand-stimulated cGMP production (11, 16). We also measured cAMP in organoids, and while levels in unstimulated organoids were similar, a significant increase was seen on the addition of uroguanylin, in both wild type and S839I organoids (Figure 3B). This could reflect an inhibition of the cAMP-hydrolyzing phosphodiesterase 3 activity by elevated levels of cGMP, which can undergo inhibitory phosphorylation by PKG1 (but not PKG2 to date) on a specific serine residue (S654) in PDE3 (38).

To test cyclic nucleotide-mediated signaling in organoids, we treated them with forskolin (which activates adenylyl cyclases and elevates cAMP levels (39), Sp-8-pCpt-cGMPS, a cell-permeable analog of cGMP that is a potent activator of PKGII (40) or uroguanylin. Treatment of wild type and S839I organoids with either forskolin or Sp-8-pCpt-cGMPS led to their swelling, which is reflective of fluid secretion from the epithelial cells, mediated by cAMP and cGMP, into the organoid lumen (Figure 3C and Supplemental Movies 1 and 2). Swelling was induced to equivalent extents in wild type and S839I organoids. Treatment with uroguanylin, however, showed more pronounced swelling in organoids prepared from S839I mice (Figure 3C; Supplemental Movie 3 and 4). This could be a consequence of enhanced cGMP production in S839I organoids, greater activation of Cftr, inhibition of Nhe3 and higher Dra1 expression in S839I organoids. No difference in basal swelling was seen in the absence of agonists (Figure 3C, 0h), indicating that the increase in basal levels of cGMP in S839I organoids (Figure 3B) was insufficient to induce significant swelling in the absence of agonists.

We previously reported that the S839I mutation in mouse GC-C decreases the EC50 for uroguanylin (11), which could also account for the diarrhea seen in patients following GC-C activation by endogenous concentrations of uroguanylin (16, 17). We treated organoids with varying doses of uroguanylin and measured their diameters before and 4h after treatment. We calculated the volume increase of organoids treated with different concentrations of uroguanylin as described earlier (26) and saw a greater increase in organoid volume in S839I organoids than in organoids prepared from wild type mice (Figure 3D). Therefore, at similar levels of guanylin and uroguanylin in the gut, fluid and ion secretion from epithelial cells would be enhanced in patients harboring the S840I mutation, resulting in the diarrhea they experience.

We prepared RNA from organoids and investigated the misregulation of genes expressed in differentiated cell types by RNAseq analysis. Several genes across different cell types were misregulated (Figure 4A). We confirmed some of these changes by RTqPCR in RNA prepared from multiple organoid cultures across different passages. No changes in the transcript levels of the stem cell marker Lgr5 were seen (Figure 4B). As observed in the intact ileal tissue, Mptx2 (33), Lyz1, Lyz2 and several defensins were downregulated (Figure 4B). Similarly, Atoh1, which determines secretory cell fate (41) is significantly downregulated in S839I organoids (Figure 4B). In accordance with this predicted defect in the secretory cell lineage, Muc2 and Agr2 (goblet cell), Chga and Chgb (enteroendocrine) and Alox5p (tuft cell) expression were also decreased, while Rbp2 (proximal enterocyte) levels were higher (Figure 4B).

Figure 4. Secretory cell differentiation is altered in S839I organoids.

(A) Heat-map of cell specific gene expression. Data was obtained from an RNAseq analysis of wild type and S839I organoids. (B) RNA and cDNA were prepared from organoid cultures and subjected to RTqPCR using primers directed towards the indicated genes in specific cells of the intestinal epithelium. All transcript levels were normalized to Gapdh. Each dot represents organoids harvested from an individual well across at least 8 independent experiments. Mean is shown, and p-values were obtained by an unpaired two-tailed t-test with Welch’s correction. (C) Organoids were probed with antibodies to lysozyme and visualized by confocal microscopy. Data shown represent maximum intensity projections generated from all z-stack slices for both lysozyme and DAPI channels from two captured images of wild type and S839I organoids. The graph shown below represents the analysis of staining of lysozyme normalized to DAPI, with each dot representing data from images of distinct organoids. (D) RTqPCR of genes linked to the Wnt and Notch signaling pathways. Each dot represents organoids harvested from individual wells across at least 6 independent experiments. Mean is shown, and p-values were obtained by an unpaired two-tailed t-test with Welch’s correction. (E) Western blot analysis of β-catenin levels in independent preparations of protein extracts from organoids. The blot was normalized to β-actin levels and expression levels are shown in the graph. Values shown are the mean,± SD, and p values were obtained by an unpaired two-tailed t-test with Welch’s correction

Staining of organoids with anti-lysozyme antibodies showed a reduction in spots that represent Paneth cells in S839I organoids in comparison with wild type organoids (Figure 4C; Supplemental Figure 5). To study the pathways that could account for the compromised maturation of secretory cells, we monitored the expression of genes in the Wnt and Notch signaling pathways in organoids. Wnt3, Tcf7, and Sox9 show downregulation (Figure 4D), while Tcf7l2, the main transcription factor that regulates stem cell function (42) showed no change, in agreement with no alterations in Lgr5 transcripts. Sox9 is a target of Wnt/β-catenin signaling and is necessary for Paneth cell specification. Moreover, Wnt3 produced by Paneth cells regulates the function of adjacent stem cells. Thus, reduction of Wnt3, Sox9 and Tcf7, another direct downstream target of Wnt/β-catenin signaling, reflects overall impaired Wnt signaling and reduced Paneth cell number/function in the S839I organoids. No change in total levels of β-catenin by western blot analysis was seen in organoids (Figure 4E.

Interestingly, Hes1, a downstream target of Notch signaling was also downregulated, but so was Atoh1 (Figure 4B), which is usually expressed when Hes1 expression is low (Figure 4D) (43). This indicates that both Notch and Wnt pathways are misregulated in mutant organoids and it is perhaps the crosstalk between these two pathways that is modulated by elevated cGMP in enterocytes.

Zinc supplementation suppresses GC-C activity and restores gut homeostasis

Currently, there are no pharmacological inhibitors of GC-C to reverse the disease symptoms observed in patients with hyperactivating mutations in GUCY2C. Oral zinc supplements in the form of zinc sulfate, zinc acetate, or zinc gluconate are widely used to treat diarrhoea and IBD (44). Several years ago, we showed that zinc can inhibit the guanylyl cyclase activity of the recombinant intracellular domain of GC-C in vitro (45). We therefore studied the effects of zinc sulphate on the activity of full-length mouse GC-C in vitro. Membrane fractions prepared from HEK293E cells expressing wild type or S839I GC-C were incubated with varying concentrations of zinc, followed by in vitro guanylyl cyclase assays. The IC₅₀ for S839I GC-C was ~30 times lower than that of the wild type receptor (Figure 5A), suggesting that the S839I mutation alters the structure of mouse GC-C in a way that makes it more conducive for zinc-mediated inhibition. We therefore hypothesized that zinc supplementation in S839I mice might reverse some of the deleterious gut-associated effects seen in these mice.

Figure 5. Zinc administration to S839I mice relieves diarrheal symptoms by inhibiting guanylyl cyclase activity.

(A) In vitro guanylyl cyclase assay with membrane fractions prepared from HEK293E cells stably expressing either WT or S839I mouse GC-C in the presence of varying concentrations of zinc sulphate (10⁻⁸ M to 10⁻² M). The data shown are pooled across two independent experiments. Inset values indicate the mean IC₅₀ ± SD. (B) Organoids from S839I mice were incubated with ZnSO₄ (10μM) for 4h prior to the addition of uroguanylin (10⁻⁶M) or Sp-8-pCpt-cGMPs (20 μM). Incubation was continued for an additional 18h before imaging. Data shown are organoids representative of experiments repeated twice. (C) Zinc levels in the serum of WT and S839I mice were either left untreated or treated with 25 mM zinc sulphate in drinking water for 15 days. Each dot represents an individual mouse, and the mean ± SD is shown. Data was analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted p-values are shown. 5 male and 5 female mice of each genotype per condition were used across two independent experiments. (D) Cyclic GMP levels in ileal epithelial cells of WT and S839I mice, either left untreated or treated with 25 mM zinc sulphate in drinking water for 15 days. Each dot represents an individual mouse, and the mean ± SD is shown. Data was analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted p values are shown. Data is from 5 female mice of each genotype per condition and across two independent experiments. (E-G) Wild type and S839I mice were either left untreated or treated with 25 mM zinc sulphate in drinking water for 15 days. Estimation of bowel frequency (E) in wild type and S839I mice was performed by noting the number of fecal pellets passed in 10 min between 8 am and 9 am. Fecal water (F) and fecal sodium (G) content in WT and S839I mice were obtained by measuring fecal weight and sodium before and after lyophilization. Each dot represents data from a single mouse. Five male and 5 female mice of both genotypes were used from experiments performed twice. Mean ± SD are shown, and adjusted p values have been obtained from a two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli. (H-J) Representative images of PAS-stained Paneth cells (H), PAS-stained goblet cells (I), and immunofluorescence against lysozyme in the terminal ileum of S839I mice, either left untreated or treated with 25 mM zinc sulphate in drinking water for 15 days. Stars indicate Paneth cells at the crypt bottom (H) and goblet cells along villi (I). Dotted lines indicate the crypt outline in (J). 5 male and 5 female mice per genotype were tested across 2 experiments.

We first tested this possibility using organoids. Swelling was monitored in S839I organoids treated with uroguanylin with or without treatment with zinc sulfate. Swelling was significantly reduced following pre-treatment of organoids with zinc (Figure 5B). To confirm that this swelling was a result of GC-C inhibition of uroguanylin-mediated GC-C activation, we treated organoids with Sp-8-pCpt-cGMPS and observed swelling even in the presence of zinc (Figure 5B). Therefore, zinc-mediated inhibition of the guanylyl cyclase activity of GC-C was a viable means of reducing fluid secretion from intestinal epithelial cells.

Encouraged by this observation, we administered wild type and S839I mice with 25 mM zinc sulphate in drinking water for 15 days. Zn²⁺ levels in the plasma of mice were measured. In untreated animals, there was a trend towards lower circulatory zinc levels in S839I mice (Figure 5C). Administration of zinc did not increase circulatory levels in wild type mice, but a significant increase was seen in zinc-administered S839I mice (Figure 5C).

Levels of cGMP levels were significantly reduced in S839I mice after zinc supplementation (Figure 4D) but remained unaltered in wild type mice. Notably, post zinc supplementation, cGMP levels in S839I mice were still higher than that in wild type mice (Figure 5D). Diarrhea-like symptoms such as bowel frequency, fecal water, and fecal sodium content were all reduced in S839I mice following zinc supplementation (Figure 5E–G). Importantly, mature Paneth cell abundance and mucin-enriched goblet cells were restored following zinc treatment in the ileum of mice (Figure 5H–4J), suggesting a restoration of barrier integrity. Therefore, zinc administration is an efficient way to reduce cGMP levels in the epithelial cell and consequently restore their optimum functioning and reduce GC-C-mediated diarrhea in mice.

Zinc supplementation is reported to improve intestinal epithelial barrier function and reduce inflammation in a mouse model of TNF-induced intestinal inflammation that was dependent on glucocorticoids and the aryl hydrocarbon receptor, AHR (46, 47). Indeed, zinc supplementation to S839I mice reduced small intestinal epithelial barrier permeability to the levels seen in wild type mice (Figure 6A). Moreover, the levels of Ifnα, Tnfα, and Il6 were reduced in ileal extracts of S839I mice post zinc supplementation (Figure 6B) (48). However, Ifnγ levels increased in the ileum of both wild type and S839I mice, an effect reported earlier following zinc administration to mice (Figure 6B) (46).

Figure 6. Zinc administration reduces intestinal permeability, proinflammatory cytokines, and alters the fecal microbiome in S839I mice.

(A) The serum levels of 4 kDa FITC-dextran were measured 30 min after oral gavage of WT and S839I mice, either left untreated or treated with 25 mM zinc, as a determinant of small intestinal permeability. Each dot represents an individual mouse, and the mean ± SD is shown. Data was analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted P values are shown. Three male and 3 female mice of each genotype per condition were used across two independent experiments. (B) Levels of the indicated cytokines were measured in the soluble fraction of terminal ileal tissue extracts of WT and S839I mice. Each dot represents an individual mouse, and the mean ± SD is shown. Data was analyzed using two-way ANOVA with a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, and adjusted p values are shown. Three male and 3 female mice of each genotype per condition were used across two independent experiments. levels. The Kruskal-Wallis H test was applied to calculate statistical significance (p value).

DISCUSSION

Our study demonstrates that hyperactivation of GC-C fundamentally alters intestinal homeostasis by disrupting epithelial differentiation and immune surveillance. Using a disease-associated GC-C mutation in a mouse model, we establish a direct mechanistic link between dysregulated cGMP signaling, barrier dysfunction, and CD-like pathology. Our earlier findings revealed that the S839I mouse results in loss of overall homeostasis, fluid-ion imbalance, gut microbiota dysbiosis, and susceptibility to colitis (21). In this study, we focused on cell intrinsic and extrinsic changes that occur in the small intestine, since patients harboring activating mutations in GC-C present with ileal inflammation associated with CD (16).

RNAseq analysis of the terminal ileum revealed misregulation of several genes that regulate inflammatory pathways, signatures observed in CD patients (Figure 1B) (49). A recent study, which integrated transcriptomic analysis from both preclinical and clinical models of intestinal inflammation, identified conserved molecular mechanisms between preclinical mouse models and human IBD, that included inflammatory circuits seen in the S839I mice (8). We therefore posit that much insight into changes in patients with hyperactivating mutations in GC-C can be gained from a careful analysis of the preclinical model we have developed.

A key finding of this study is that GC-C hyperactivation diverts epithelial lineage allocation away from secretory fates and toward absorptive enterocytes. This underscores a conserved role for cGMP in epithelial fate decisions, reflective of the evolutionary conservation of GC-C from fishes to mammals. Both tissue and organoid analyses revealed reduced Paneth and goblet cell function, with diminished antimicrobial peptide and mucus production, respectively (Figure 1 and Figure 3 and Supplemental Figure 2). Secretory cell differentiation is normally governed by Wnt and Notch-Delta signaling, with Atoh1 driving progenitor cells toward this lineage (50). Further specification depends on transcription factors such as Sox9 for Paneth cells (51) and Spdef for goblet cells (32, 52). Interestingly, PKGII can phosphorylate SOX9 and reduce its expression in human glioma cells (53) thereby preventing target gene activation. This suggests a mechanism by which high levels of GMP could inhibit the transcription of genes required for Paneth cell specification. Paneth cells are stable over time and are not impacted by the ongoing inflammation. Patients who harbor a higher percentage of abnormal Paneth cells show dysbiosis (54). Given that Paneth and goblet cells are the major sources of antimicrobial peptides and mucus, their dysfunction is likely to contribute to the dysbiosis observed in S839I mice (15).

Gut microbial dysbiosis, coupled with reduced mucus layer and loss of epithelial barrier integrity appears to trigger the immune dysregulation in the S839I mice (Figure 2 B–E). Proinflammatory cytokines were elevated in the mutant ileum (Figure 2). Oral zinc supplementation attenuated GC-C activity, normalized cGMP levels, and reversed multiple pathological features, including diarrhea, barrier leakage, proinflammatory cytokine production and Paneth cell function (Figures 4 and 5). Zinc’s dual ability to restore epithelial and Paneth cell function positions it as an attractive candidate for patients with GC-C–mediated enteropathies, who exhibit an altered microbiome that may contribute to the disease phenotype (55). The heightened sensitivity of the mutant receptor to zinc inhibition further suggests that activating mutations may create a therapeutic window not present in wild-type GC-C, due to structural alterations that remain to be characterised.

Our findings also reveal a tractable therapeutic strategy for patients harboring hyperactivating mutations in GUCY2C. Results presented here highlight cGMP signaling as a critical integrator of epithelial and immune homeostasis in the gut. While GC-C has long been appreciated for its role in electrolyte balance, our study uncovers its broader influence on epithelial lineage specification and mucosal immunity. We identify aberrant GC-C activation as a driver of epithelial–immune dysregulation and establish zinc supplementation as a simple, clinically relevant intervention. These findings not only clarify the pathophysiology of GUCY2C-related disorders but also open new avenues for therapeutic modulation of cGMP signaling in intestinal inflammation.

SUPPLEMENTAL MATERIAL

Supplemental Videos S1-S4

Supplemental Figs 1–5

Supplemental Table 1

All data supplements can be accessed here: https://doi.org/10.6084/m9.figshare.32306769

Figure 7.

DATA AVAILABILITY

Accession numbers for the RNA seq analyses are E-MTAB-14556 (Organoid RNAseq) and E-MTAB-15690 (Ileal RNA seq) available at Array Express (https://www.ebi.ac.uk/biostudies/arrayexpress).