Glucose-dependent insulinotropic polypeptide receptor signaling alleviates gut inflammation in mice

We previously determined that GIPR agonism regulates BM hematopoietic responses to 5FU and Pam3CysSerLys4 (Pam3CSK4), whereas loss of thedysregulated the hematopoietic response to 5FU but not to Pam3CSK4 or LPS (). Analysis of the effect of these treatments on a subset of immunoregulatory gene expression profiles in the gut revealed that [D-Ala]-GIP (hereafter referred to as GIP) did not modulate the immune response to LPS or Pam3CSK4 within the ileum and jejunum (data not shown). However, treatment with GIP downregulated cytokine gene expression in the SB of mice treated with a moderate dose of 5FU (150 mg/kg, injected twice, 1 week apart) (), a dosing regimen originally selected to interrogate hematopoiesis (). Levels of IL-1β () and IL-10 () mRNA transcripts were reduced in the duodenum () and jejunum () of mice treated with GIP and 5FU; however, levels of TNF-α (), IFN-γ (), and chemokine receptor-2 () were not different (, A and B; supplemental material available online with this article;). The immunoregulatory effects of GIP were most evident in the distal SB as,,, andmRNAs were downregulated by GIP in the ileum of 5FU-treated mice (). The ileal transcript level of() and the protein concentrations of IL-1β, IL-10, keratinocyte chemoattractant/human growth-regulated oncogene (KC/GRO), TNF-α, IL-6, and IFN-γ within the ileum and circulation were not different (, D and E). Mice cotreated with GIP and the moderate dose of 5FU had reduced body weight, but no differences in SB weight or the SB weight-to-length ratio (). Furthermore, spleen weight of 5FU-treated mice was lower, irrespective of GIP treatment. Histology analysis showed a reduction in crypt depth in the ileum of vehicle-and GIP-treated mice exposed to 5FU, indicative of mild gut injury, though there were no differences between groups in villus height or crypt density (, B and C). Gipr Il1b Il10 Tnf Ifng Ccr2 Il1b Il10 Ifng Ccr2 Tnf ¹² Figure 1, A–D ¹² Figure 1B Figure 1C Supplemental Figure 1 Figure 1D Supplemental Figure 1C Supplemental Figure 1 Supplemental Figure 2A Supplemental Figure 2 2 https://doi.org/10.1172/jci.insight.174825DS1↗

Given that the selected dose of 5FU, initially chosen to study hematopoiesis (), resulted in limited intestinal damage and inflammation, the experiment was repeated using more frequent injections of 5FU to enhance the severity of gut injury and inflammation (). Administration of 60 mg/kg/day of 5FU over 4 consecutive days led to body weight loss in both vehicle- and GIP-treated mice (), but it did not perturb SB weight, length, or gut permeability (). However, more frequent 5FU administration induced intestinal injury characterized by blunting of villus height and a reduction in crypt density in the ileum (). GIPR agonism increased villus height and crypt depth in the vehicle-treated and 5FU-treated mice, respectively (). GIPR agonism also attenuated the extent of decreased cellular proliferation in the 5FU-treated mice as assessed by the number of Ki67cells in the ileum (). GIP treatment also reduced neutrophil activation and the number of macrophages within the ileum in response to 5FU, as evidenced by a reduction in the number of neutrophil elastase(NE) and CD68cells, respectively (). Furthermore, GIPR agonism attenuated the 5FU-induced upregulation of several proinflammatory genes within the ileum, including the immune cell markers adhesion G protein-coupled receptor E1 () and, and the cytokines,,, and(). There was no GIP treatment effect on gene expression of lymphocyte antigen 6 family member G (),, S100 calcium binding protein A8 (), or S100 calcium binding protein A9 () (). ¹² Figure 2A Figure 2B Figure 2, C and D Figure 2, E and F Figure 2F Figure 2G Figure 2, H and I Figure 2J Figure 2J + + + + Adgre1 Cd68 Il1b Il6 Tnf Ifng Ly6g Ccr2 S100a8 S100a9

We next used the same protocol to examine the effect of GLP-1R agonism using semaglutide (Sema) and GIPR–GLP-1R coagonism using TZP on 5FU-induced intestinal inflammation (). While both treatments led to similar reductions in body weight, SB weight, and the SB weight/length ratio (, B and C), only TZP significantly attenuated 5FU-induced neutrophil activation within the ileum (). Neither TZP nor Sema treatment modified the effect of 5FU on villus height, crypt depth, or crypt density (, D and E). Similarly, there was no Sema or TZP treatment effect on the number of mucosal Ki67and CD68cells (, F and G). Expression levels of,,,,,, andshowed no change following either treatment (). However, the expression level ofwas downregulated, whereaswas upregulated in the TZP-treated mouse group compared with 5FU treatment alone (). Supplemental Figure 3A Supplemental Figure 3 Figure 2, K and L Supplemental Figure 3 Supplemental Figure 3 Supplemental Figure 3H Supplemental Figure 3H + + Adgre1 Cd68 Il1b Il6 S100a8 S100a9 Tnf Ifng Ly6g

To assess the role of physiological GIPR signaling in the intestinal response to 5FU, we analyzedmice. In mice exposed to the intermittent doses of 5FU (), the gene expression levels for,,, chemokine ligand 1 (),,, andwere upregulated in the ileum of 5FU-treatedmice (). Moreover, ileal protein content of IL-1β, IL-10, IL-6, and TNF-α was increased in 5FU-treatedmice (). The plasma concentration of the proinflammatory cytokine IL-1β was also elevated in 5FU-treatedmice (). There was no consistent genotype effect observed on gene expression levels of inflammatory markers within the proximal SB (i.e., the duodenum and jejunum) (, A and B). Circulating levels of IL-10, KC/GRO, TNF-α, IFN-γ, and IL-6 were not different between groups (). Furthermore, mouse body weights and SB biometry were not different betweenandwith or without 5FU administration; however, spleen weight was reduced in all 5FU-treated groups (). Histological analysis of the ileum revealed reductions in crypt depth in response to 5FU, but there was no genotype effect on crypt depth, villus height, or crypt density (, B and C). Gipr –/– Il1b Il10 Tnf Cxcl1 Ifng S100a8 S100a9 Gipr –/– Gipr –/– Gipr –/– Gipr +/+ Gipr –/– Figure 3A Figure 3B Figure 3C Figure 3D Supplemental Figure 4 Supplemental Figure 4C Supplemental Figure 5A Supplemental Figure 5

Repeated daily exposure to 5FU () induced significant injury in the mouse ileum characterized by villus blunting as well as a reduction in crypt depth and crypt density ().mice exhibited higher sensitivity to 5FU-induced gut injury indicated by a further decrease in villus height and crypt depth (). Bothandmice had lower body weight and SB weight after 5FU exposure (, A and B); however,mice had a higher SB weight/length ratio compared with the 5FU-treatedmice ().mice also had upregulated,,,, andmRNA transcripts in the ileum (). 5FU treatment dysregulated the expression levels of,,,, and, but there was no discernible genotype effect in response to 5FU (). Similarly, there was no difference in gut permeability, cellular proliferation (Ki67cell count/ring), neutrophil activation (NEcell count/ring), and the number of macrophages (CD68area/ring) inversusmice exposed to high-dose 5FU (, C–G). Figure 4A Figure 4, B and C Figure 4, B and C Supplemental Figure 6 Supplemental Figure 6B Figure 4D Figure 4D Supplemental Figure 6 Gipr –/– Gipr +/+ Gipr –/– Gipr –/– Gipr +/+ Gipr –/– Ly6g Adgre1 Il1b Il6 Tnf Cd68 Ifng Ccr2 S100a8 S100a9 Gipr Gipr + + + +/+ –/–

Previous studies demonstrated that increased adipose tissue inflammation inmice could be attributed to loss of immunosuppressive GIPRmyeloid cells in the BM that contributed to adipose tissue macrophage populations (,). Accordingly, we assessed whether BM-derived-expressing cells modulate gut inflammation induced by 5FU. BM was transplanted fromordonor mice expressing the CD45.2 allele into irradiated WT recipient mice expressing the CD45.1 allele. The resulting WTand WTmice were then treated with 5FU (). Gipr Gipr Gipr –/– Gipr +/+ –/– + BM-Gipr+/+ BM-Gipr–/– ^{11 12} Supplemental Figure 7A

Efficiency of BM reconstitution in recipient mice was determined by analysis of the percent of CD45.1and CD45.2(from total CD45cells) in peripheral blood, revealing that 90% of the cells were CD45.2(). Gene expression within the BM showed WTmice exhibited ablation ofexpression versus WTmice (). However,expression within the ileum was not downregulated in response to BM-specificdeletion (). Interestingly, 5FU treatment upregulated expression of both BM and ilealin WTmice (). Similarly, ilealexpression was upregulated in response to 5FU treatment in both WTand WTmice (). Plasma GIP levels were not different between groups (). Tissue biometry and histological analysis of the ileum revealed no genotype effects on spleen weight, SB weight, crypt depth, and density after 5FU treatment (, B–D). However, villus height was blunted in 5FU-treated WTbut not in WTmice (). + + + + BM-Gipr–/– BM-Gipr+/+ BM-Gipr+/+ BM-Gipr+/+ BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– Figure 5A Figure 5B Figure 5C Figure 5, B and C Figure 5C Figure 5D Supplemental Figure 7 Supplemental Figure 7D Gipr Gipr Gipr Gipr Gip

Intriguingly, in the absence of 5FU, WTmice exhibited lower ileal,, andmRNA transcripts compared with WTmice (). However, mRNA biomarkers of inflammation, including,,,,,,,, and, were not dysregulated in the ileum of 5FU-treated WTcompared with 5FU-treated WTmice (and). Similarly, ileal protein expression levels of IL-1β and TNF-α were reduced in vehicle-treated WTmice compared with vehicle-treated WTmice, whereas the levels of IL-1β, IL-10, TNF-α, KC/GRO, and IL-6 protein were not different between 5FU-treated groups (). Consistent with the protein cytokine expression within the ileum, circulating concentrations of IL-1β were reduced in the vehicle-treated WTmice compared with vehicle-treated WTmice (). Plasma concentrations of TNF-α were increased in 5FU-WTcompared with 5FU-treated WTmice (). Circulating KC/GRO was elevated in all 5FU-treated mice independent of genotype, while circulating IL-6 was only elevated in the 5FU-WTcompared with vehicle- WTmice (). IL-10 and IFN-γ plasma concentrations were not different between groups (). Therefore, while there are some modest genotype effects on gut and plasma inflammatory markers, knocking out the BMdid not completely phenocopy the extent of 5FU-induced gut inflammation observed inmice. BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– BM-Gipr+/+ BM-Gipr–/– BM-Gipr–/– –/– Il10 Ifng Ccr2 Il1b Il10 Tnf Cxcl1 Ifng Ccr2 Il6 S100a8 S100a9 Gipr Gipr Figure 5E Figure 5E Supplemental Figure 7E Figure 5F Supplemental Figure 7F Supplemental Figure 7F Supplemental Figure 7F Supplemental Figure 7F

We next interrogated whether BM-derived-expressing cells modulate the extent of 5FU-induced gut inflammation by transplanting BM from WT CD45.1 donor mice intoorCD45.2 recipient mice (). After transplantation, mice were designatedor, representing mice with or withoutdeletion in all tissues excluding the BM. Ninety percent of the CD 45cells in the peripheral blood of the recipient mice expressed the CD45.1 allele (). BMexpression was restored inrecipient mice and was not different frommice, indicating successful BM reconstitution (). However, ilealexpression remained ablated in theversusmice, suggesting minimal contribution of BM-derived-expressing cells to local gutexpression (). Ilealexpression was upregulated in response to 5FU exposure inmice (). However, plasma GIP levels were not different in response to treatment or genotype (). Gipr Gipr Gipr Gipr Gipr Gipr Gipr Gipr Gipr +/+ Gipr Gipr Gipr Gipr Gipr Gip Gipr –/– +/+ +/+BM-WT –/–BM-WT + –/– –/–BM-WT +/+BM-WT –/–BM-WT Supplemental Figure 8A Figure 6A Figure 6B Figure 6C Figure 6C Figure 6D

Tissue biometry showed elevated SB weight in thecompared with themice treated with 5FU (). Histological analysis showed that 5FU-treatedmice had modestly higher ileal villus height compared with 5FU-treatedmice, but no differences were observed in crypt depth or density (, C and D). Within the ileum,mice treated with 5FU exhibited upregulated gene and protein expression of the proinflammatory cytokine/TNF-α and the chemokine/KC/GRO compared with the vehicle-treated groups; an effect that was ameliorated in themice treated with 5FU (). Similarly, protein expression, but not gene expression, of IL-1β was decreased in themice compared withtreated with 5FU (). Furthermore, plasma levels of KC/GRO, IFN-γ, and IL-6 were lower in 5FU-treatedversusmice (). Gene and protein expression of/IL-10,/IL-6,,,, andin the ileum were not different between genotypes (, E and F). Circulating IL-1β, TNF-α, and IL-10 concentrations were not different between 5FU-treated groups (and). Gipr Gipr Gipr Gipr Gipr Tnf Cxcl1 Gipr Gipr Gipr Gipr Gipr Il10 Il6 Ifng Ccr2 S100a8 S100a9 –/–BM-WT +/+BM-WT –/–BM-WT +/+BM-WT +/+BM-WT –/–BM-WT –/–BM-WT +/+BM-WT –/–BM-WT +/+BM-WT Supplemental Figure 8B Supplemental Figure 8 Figure 6, E and F Figure 6, E and F Figure 6G Supplemental Figure 8 Figure 6G Supplemental Figure 8G

Collectively, these findings implicate BM-derived-expressing cells as important modifiers of the extent of gut inflammation. Since WT BM does not influence localexpression within the gut ofmice, these findings suggest an indirect role for-expressing BM-derived cells in modulating local gut-tissue inflammation. Gipr Gipr Gipr Gipr –/–

To ascertain the relative abundance and potential localization ofmRNA transcripts along the gastrointestinal tract, we compared relativemRNA expression in multiple tissues and gut segments.expression was identified in the hypothalamus, brainstem, duodenum, jejunum, ileum, colon, lung, heart, and adipose tissue (). Levels ofmRNA transcripts were highest in the hypothalamus and brainstem, followed by adipose tissue (). Within the gut, levels ofmRNA transcripts were comparatively low and were highest in the jejunum (). A similar trend was observed using GIPR reporter mice (). GIPR-tdTomato expression was detected among all gut segments but was highest in the jejunum in comparison with the ileum and colon (). To localize endogenousexpression within the gut using complementary approaches, we analyzed different jejunal subcompartments (i.e., mucosa, submucosa, and muscle layers). The epithelial cell marker Villin () and the glial cell marker glial fibrillary acidic protein () were not enriched in the submucosal layer, confirming minimal mucosal or muscle layer contamination (). The submucosa was enriched for the stromal cell marker Sialomucin () ().mRNA expression was enriched in the submucosal layer, which contains the lamina propria and crypts (). We next isolated the epithelial layer from the lamina propria and muscle across gut segments using EDTA dissociation. Adequate epithelial cell separation from the lamina propria was confirmed via analysis ofwhich was selectively enriched in the epithelial layer, whereaswas enriched in the remaining lamina propria and muscle layer within all gut segments ().mRNA transcripts were enriched within the lamina propria and muscle of the proximal (i.e., duodenum and jejunum) but not distal SB (i.e., ileum) ().expression in the lamina propria was further delineated by costaining-tdTomato–expressing cells with the epithelial cell marker E-cadherin (CDH1) and the immune cell marker CD45 showing that the receptor is not localized to either of these cell types (). Gipr Gipr Gipr Gipr Gipr Gipr Cre.tdTomato/+ Gipr Vil1 Gfap Cd34 Gipr Vil1, Gfap Gipr Gipr Gipr Figure 7A Figure 7A Figure 7A Figure 7B Figure 7C Figure 7C Figure 7C Figure 7D Figure 7D Figure 7, E and F

Next, submucosa cells were extracted by tissue digestion from all segments of the small intestine ofand littermate controlmice. Among CD45immune cells, CD11bCD3T cells, CD11bMHCIIB cells, and CD11bmyeloid cells were all low for-tdTomato signal ().-tdTomato fluorescence signals were detected in some CD31CD45ECs, but not among CD45CD31nonimmune/EC cells (). GIPR has been previously localized to CD146mesenchymal cells and pericytes in adipose tissue and the CNS (,). Accordingly, we next examined whethermRNA transcripts were higher within intestinal CD146fractions, enriched for mesenchymal cells, isolated using magnetic cell separation. Notably, SB CD146populations were enriched for(). These cells also had higher expression of the pericyte marker platelet-derived growth factor receptor β () and the EC marker platelet EC adhesion molecule (), and they were relatively depleted for Protein tyrosine phosphatase receptor type C (), which encodes for CD45 (). These findings reveal thatexpression within the gut is not enriched within immune cells of the lamina propria; rather, it is predominantly localized to CD146cells, which include pericytes and ECs (). Gipr Cre.tdTomato/+ tdTomato fl/fl Gipr Gipr Gipr Gipr Pdgfrb Pecam1 Ptprc Gipr + – + – + + + – – – + + + + Figure 7G Figure 7G ^{21 22} Figure 7H Figure 7H ²³

To further refinelocalization within the gut, we analyzed publicly available single-cell RNA-Seq (scRNA-Seq) data from the mouse ileum (); however,expression was not detected in this dataset, although pericytes coexpressingandwhich encodes for CD146 (,), displayed a very lowsignal (). In the human gut cell atlas (),was detected in epithelial cells, plasma cells, T cells, myeloid cells, and 2 subsets of mesenchymal cells (). In the mesenchymal cells, a subset ofandpericytes express(). Coupled with the enrichment ofin mouse gut CD146cells, our data reveal consistent/expression in mouse and human gut pericytes. Gipr Gipr Pdgfrb Mcam, Gipr GIPR MCAM PDGFRB GIPR Gipr Gipr GIPR ^{24 23 25} Supplemental Figure 9 ²⁶ Supplemental Figure 10 Supplemental Figure 10 + + +

Male mice were used in these experiments due to the much greater sensitivity of female mice to 5FU-induced gut injury, resulting in much greater weight loss and illness in the animals. To date, all the actions described for GIP in animals and humans have been ultimately conserved in both males and females.

Mice were housed at The Centre for Phenogenomics animal facility at 21°C on a 12-hour light/dark cycle with ad libitum access to water and a standard rodent chow diet (18% kcal from fat, 2018 Harlan Teklad). All GIPR gain-of-function experiments were carried out in male mice on a C57BL/6J background received from The Jackson Laboratory (no. 000664). Animals were given i.p. injections with 24 nmol/kg [DAla]-GIP (Chi Scientific) or vehicle (phosphate-buffered saline [PBS]) twice daily (9 am and 5 pm) for a total of 8 days with 2 i.p. doses of 150 mg/kg 5FU (Mount Sinai Hospital Pharmacy) given at day 1 and day 7. Then, mice were sacrificed on day 8 as previously described for the interrogation of hematopoiesis (). The same GIPR gain-of-function experiment was repeated with a more severe 5FU protocol utilizing 60 mg/kg/day of 5FU over 4 consecutive days to induce greater gut injury. Mice were treated with either vehicle or 24 nmol/kg [DAla]-GIP twice daily for 5 consecutive days starting 1 day prior to the onset of the 5FU protocol. Similarly, to study the effects of GLP-1R agonism and GLP-1R/GIPR coagonism on the modulation of 5FU-induced gut injury, mice were treated with a once-daily s.c. injection of 10 nmol/kg of Sema (Ozempic, Novo Nordisk), 3 nmol/kg of TZP (Mounjaro, Eli Lilly), or vehicle for 5 consecutive days. Mice were cotreated with 4 daily doses of 60 mg/kg of 5FU (). On day 5, 24 hours after the last 5FU injection, all mice were sacrificed for blood and tissue collection. The GIPR loss-of-function studies were similarly performed using both the old and new 5FU protocols in mice with whole-bodyand WT () mice that were generated, bred, and validated as previously described (,). 2 2 −/− ¹² Supplemental Figure 3A ^{37 43} Gipr Gipr +/+

For the localization ofin the gut, GIPR reporter mice () were generated by crossingmice obtained from Frank Reimann () with B6.Cg-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze/J mice obtained from The Jackson Laboratory (no. 007909), enabling the detection of cells currently expressingor originating fromexpressing cells. Gipr Gipr Cre.tdTomato/+ Gipr Gipr Gipr cre/+ ⁴⁴

To study the contribution of hematopoietic or BM-derived GIPR to the immune-regulatory response to 5FU administration, 8-week-old WT B6.SJL-PtprcPepc/BoyJ CD45.1recipient males obtained from The Jackson Laboratory (no. 002014) were irradiated with 1,100 cGy, split into 2 equal doses separated 4 hours apart. Following this, the tail vein was injected with of 5 × 10congenic (CD45.2) BM cells from C57BL/6Jordonor males, as previously described (,). C57BL/6J CD45.2orrecipient males were irradiated and then transplanted with BM cells harvested from WT B6.SJL-PtprcPepc/BoyJ CD45.1donor males, following a similar protocol. The degree of reconstitution was analyzed by flow cytometry analysis (Gallios, Beckman Coulter) of tail vein blood ~4 weeks after transplantation using CD45.1-PE-Cy7, CD45.2-APC, and CD45.2-FITC antibodies added to the lymphocyte-myeloid and monocyte-neutrophil panels as previously described (). At 8–16 weeks after BM transplantation, mice were treated with 2 doses of 5FU (150 mg/kg) a week apart before being sacrificed 24 hours after the second 5FU dose. Mice were fasted for 4–5 hours before they were sacrificed. a b + 6 + –/– +/+ + –/– +/+ a b + Gipr Gipr Gipr Gipr ^{12 45 12}

To test for gut permeability, mice were daytime fasted (5–6 hours) on the last day (day 4) of 5FU injections then administered an oral gavage of 1 mg of OVA suspended in sterile water. Three hours after oral gavage, 5 μL of tail blood was collected from each mouse using heparin-coated capillary tubes. The blood was treated with 10 μL of PBS containing 0.5% Tween 20 and 50 mmol/L EDTA and was then centrifuged for 5 minutes. Plasma was collected and frozen at –80C for future analysis. To assess the OVA plasma concentration that leaked out of the gut after injury, antibody-conjugated carboxylate modified (CML) beads (Thermo Fisher Scientific) were added to the plasma samples and incubated in a 96-well U-bottom plate overnight at 4°C on a plate shaker to capture the plasma OVA antigen. The OVA-CML complex was later pelleted and detected using a primary rabbit anti-OVA polyclonal antibody (GTX21221; GeneTex, 10 μg/mL), and a secondary phycoerythrin-conjugated-F(ab')2 fragment donkey anti–rabbit IgG polyclonal antibody (711-116-152; Jackson Immunoresearch Laboratories, 0.5 μg/mL). The beads were then resuspended with FACST buffer (1× PBS [311-425-CL, Wisent], 2% heat-inactivated FBS [090150, Wisent], 2 mmol/L EDTA [B10093-34, Em Science], and 0.05% Tween 20 [P1379, MilliporeSigma]) and quantified by flow cytometry as previously described (). ° ⁴⁶

Mice were sacrificed by COinhalation, blood was collected by cardiac puncture, and tissues were dissected, weighed, and immediately frozen in liquid nitrogen. All blood samples for measuring plasma cytokines and total GIP were collected from tail vein into lithium-coated Microvette tubes (Sarstedt) and mixed with a 10% volume of TED (5000 kIU/mL Trasylol [Bayer], 32 mM EDTA, and 0.01 mM Diprotin A [MilliporeSigma]). Samples were kept on ice, and plasma was collected shortly afterward by centrifugation (12,000) and stored at −80°C. 2 g

Plasma and ileal protein concentrations of TNF-α, IL-10, IL-1β, IL-6, KC/GRO, and IFN-γ were measured using the V-PLEX Proinflammatory Panel 1 Mouse Kit (Meso Scale Discovery, K15048D) as per the manufacturer's instructions. Ileal protein lysates were extracted by homogenizing tissues in a lysis buffer (50 mM Tris [pH 8], 1 mM EDTA, 10% glycerol, 0.067% Brij 35) supplemented with protease inhibitors (MilliporeSigma) using a TissueLyzer II system (Qiagen). Plasma total GIP was analyzed using an ELISA kit as per the manufacturer instructions (Crystal Chem, 81517).

The gut was dissected and flushed with PBS. Then, the entire SB weight and length were measured. For histology measures, two 2 cm segments of the ileum were collected and fixed in 10% formalin for 24 hours before being transferred to 70% ethanol and stored at 4°C for future processing. Samples were then embedded in paraffin. Paraffin-embedded tissue blocks were sectioned into 4 μm–thick slices and mounted onto charged slides (Assure, Epic Scientific). For gut histology, sections were stained with H&E using standard protocols. Sections were scanned using the Hamamatsu Nanozoomer. Using the QuPath-0.3.2 imaging software, crypt depth was measured as crypt base to tip and villus height was measured as villus base to tip of an average of 10–20 longitudinally, well-orientated crypt/villus units per mouse. Crypt density was measured as the total number of crypts/ring and the average from 2–4 ring sections per mouse was calculated.

Sections were deparaffinized and subjected to heat-induced epitope retrieval using citrate buffer (pH 6.0) in a pressure cooker. After retrieval, the sections were incubated with Peroxidase Block (Bloxall; Vector, SP-6000, lot no. ZJ1129) for 10 minutes, followed by washing in TBS-T. The sections were then treated with 2.5% normal horse serum (ImmPRESS HRP horse anti–rabbit IgG Polymer Kit, Vector, MP-7401, ZL0314) for 20 minutes to block nonspecific binding. Subsequently, the sections were incubated with anti–rabbit monoclonal antibody to Ki67 (Abcam, ab16667, GR3341233-19) at a 1:250 dilution in Antibody Diluent (Agilent, S3022, 1172069) for 1 hour at room temperature. Following TBS-T washes, the sections were incubated with ImmPRESS-HRP horse anti–rabbit IgG Polymer Reagent (ImmPRESS HRP Kit, Vector, MP-7401, ZL0314) for 30 minutes. After another rinse in TBS-T, the sections were treated with ImmPACT DAB Peroxidase (HRP) Substrate (ImmPACT DAB Substrate Kit, Peroxidase, Vector, SK-4105, ZK1018) until chromogen development was complete; they were then washed with distilled water. The sections were counterstained with Mayer's Hematoxylin (Chaptec, HIY0085-500, C150) for 20 seconds and rinsed under warm running water. Finally, the tissue sections were air dried for 20 minutes and cover slipped using Permount.

To measure neutrophil activation and macrophage number, antigen retrieval was performed by boiling slides in 1× TE buffer (pH 9.0). The ileum sections were stained with either anti-NE antibody (Cell Signaling Technology, E8U3X, rabbit mAb, 90120; 1:400 dilution) or anti-CD68 antibody (Cell Signaling Technology, CD68 [E3O7V], rabbit mAb, 97778; 1:150 dilution), and the signal for all sections was detected using SignalStain Boost IHC Detection Reagent (HRP, rabbit) (Cell Signaling Technology; 8114P) and developed using the ImmPACT DAB Substrate Kit, Peroxidase (HRP) (Vector Laboratories, SK-4105). All sections were counterstained with hematoxylin.

All sections were scanned using the Hamamatsu Nanozoomer. Using the QuPath-0.3.2 imaging software, the number of Ki67and NEcells was counted, and the total positive area for CD68 was averaged over 4–6 ring sections per mouse. + +

For the extraction of total RNA, tissue samples were homogenized in TRI Reagent (Molecular Research Center) using a TissueLyser II system (Qiagen). mRNA was then chloroform extracted, precipitated using isopropanol, washed with 75% ethanol, and reconstituted with DEPC-treated water. First-strand cDNA was synthesized from DNase I–treated total RNA using the SuperScript III and random hexamers (Thermo Fisher Scientific). Reverse transcription reactions were performed for 10 minutes at 25°C, 50 minutes at 50°C, and an additional 15 minutes at 70°C. Gene expression levels were quantified by quantitative PCR (qPCR) using a QuantStudio System and TaqMan Gene Expression Master Mix and Assays (Thermo Fisher Scientific) (). Gene expression levels were calculated as 2relative to the housekeeping genes,, oras indicated. Supplemental Table 1 –ΔCT Tbp Ppia Rpl32

Lamina propria cells were isolated as previously described () with minor modifications. Briefly, the entire small intestine was cleaned, flushed with HBSS without calcium or magnesium (HBSS, 311-512-CL, Wisent), and cut into 0.5 cm pieces. Gut pieces were transferred to a predigestion solution containing 5 mM EDTA, 5 mM DTT (R0861, Thermo Fisher Scientific), and 2%FBS in HBSS+ 10 mM HEPES (15630-080, Thermo Fisher Scientific) before being shaken at 37.2at 37°C for 20 minutes. The gut tissue pieces were vortexed briefly, and the supernatant was discarded. The EDTA washes were repeated 2 times. A third wash was performed with HBSS+ 10 mM HEPES. Tissues were then collected using a 100 μm strainer, minced, and incubated at 37°C for 30 minutes in a digestion solution containing DNase I (200 KU/mL; MilliporeSigma) and Collagenase D (400 Mandl units/mL; Roche) () in HBSS with magnesium and calcium + 10 mM HEPES. The tissues were gently sheared with a syringe needle and strained sequentially through 70 and 40 μm strainers, and single cells were resuspended with a MACS buffer for magnetic cell separation (Miltenyi Biotech). ^{7 47} –/– –/– –/– v/v g

Cell suspensions of digested lamina propria and muscle from all small intestinal segments were incubated on ice with fluorochrome-conjugated antibodies in a FACS buffer. The following antibodies were used to stain the different cell populations: CD45 APC-Cy7 (clone 30-F11, BD Biosciences), CD11b PE-Cy7 (clone M1/70, BioLegend), CD31 Percp-Cy5.5 (clone 390, BioLegend), CD3 FITC (clone 145-2C11, BioLegend), and MHCII BV 421 (clone M5/114.15.2, BioLegend). Multiparameter flow cytometry analyses were performed using a FACSCanto II machine (BD Biosciences). Flow cytometry analysis was performed using FlowJo software (BD Biosciences).

Magnetic cell separation was performed using CD146 (LSEC) MicroBeads (Miltenyi Biotech, 130-092-007) as per the manufacturer's instructions. Both the supernatant, containing the CD146fraction, and the precipitant, containing the CD146fraction, were collected and stored in TRI Reagent at –80°C for later RNA extraction and gene expression analyses. – +

For in vivo imaging system (IVIS) studies, duodenum, jejunum, ileum, and colon were collected and imaged immediately after euthanasia. Regions of interest from the images obtained were identified and quantified as average radiance using Living Image software 4.0. (Spectral Instruments Imaging).

Each segment (i.e., duodenum, jejunum, and ileum) of the small intestine was removed, opened longitudinally, and rolled with the mucosa outward to image the entire tissue in one segment as previously described (). Tissues were then fixed using 4% PFA for 24 hours, dehydrated in 30% sucrose, and subsequently embedded in OCT freezing media. Sections of approximately 18 μm were obtained using a cryostat (Thermo Fisher Scientific) and blocked with a buffer containing 2% BSA for 1 hour. Sections were stained with CD45 monoclonal Ab (Invitrogen, YW62.3, MA1-80090) at dilution 1:100 and secondary antibody goat anti–rat AF647 (Abcam, AB150167) dilution 1:200 or Ecad monoclonal antibody (BD Biosciences, 610182) dilution 1:100 and secondary antibody donkey anti–mouse AF488 (Jackson ImmunoResearch, 715545150) dilution 1:200 and then mounted with fluorescence mounting medium containing DAPI. Images were taken with a ZEISS Confocal Microscope LSM700 (Micro Imaging GmbH, ZEISS). Image processing was performed with ZEN 2011 SP7 software (ZEISS) calculated by subtraction of the background from each slide, and an average was calculated. ⁴⁸

Published scRNA-Seq data of the mouse ileum () and the human gut cell atlas () were reanalyzed for the expression of GIPR. For the mouse data, Uniform Manifold Approximation and Projection (UMAP) plots were generated with a standard pipeline and default parameters using Seurat 4.1.0 (). Scanpy was used to generate the gene expression plots for the human gut cell atlas (). ^{24 26 49 50}

Data are represented as the mean ± SD. Statistical comparisons were made by 1- or 2-way ordinary ANOVA followed by Tukey or Dunnett post hoc tests as indicated in the figure legends using GraphPad Prism version 8 software. Values considered outliers using Grubbs' test were excluded from analysis.≤ 0.05 was considered statistically significant. P

All animal experiments were approved by the Animal Care Committee of the Mount Sinai Hospital and the Animal Care Use Committee of the Sourasky Medical Center.

Values for all data points in graphs are reported in thefile and. Supporting Data Values Supplemental Table 2