Six- to eight-week-old male C57BL/6J (wild type; WT), Cxcr6^gfp/+ (15), Cd1d^−/− (16), Jα18^−/− (17), and LysM^eGFP (18) mice (all on a C57BL/6J background) were bred and kept under specific pathogen-free conditions at the Monash Medical Centre Animal Facility. Following transportation, mice were acclimatized for a minimum period of 7 days before use. All mice were housed in groups of no more than 5 animals in each cage after weaning, in a 12-h light–dark cycle and a temperature-controlled environment. Water and food pellets (Irradiated Rat and Mouse, Specialty Feeds, Australia) were provided ad libitum, and cages were changed weekly, unless otherwise specified. All experimental procedures were performed in accordance with protocols approved by the Monash Medical Centre B Animal Ethics Committee (MMCB/2015/15).

Adult male mice were provided with 2% (w/v) dextran sodium sulfate (DSS; molecular weight 36–50 kDa, MP Biomedicals) in autoclaved drinking water for a maximum of 7 days, and weighed and monitored daily. Clinical scoring included general symptoms of activity (0: normal, 1: isolated, 2: huddled/inactive, 3: moribund), coat (0: normal, 1: rough, 2: unkempt, 3: severe hair loss/bleeding), and dehydration (0: none, 1: skin less elastic, 2: skin tenting, 3: skin tenting and eyes sunken), and colitis-specific symptoms of percentage weight loss from the previous day (0: none, 1:0–3%, 2:3–6%, 3:6–10%, 4:>10%), stool consistency (0: normal, 1: soft, 2: loose, 3: liquid, 4: diarrhea), and stool blood content (0: none, 1: visible blood in stool). Control mice were provided with autoclaved water without DSS (H₂O controls). To aid monitoring, cage bedding and water were not changed during the period of DSS induction of colitis.

In experiments where required, blockade of antigen presentation by CD1d was achieved by intraperitoneal (i.p.) injections of 200 µg rat anti-mouse CD1d-blocking antibody (α-CD1d; 1B1, BD Biosciences) or its isotype control antibody (A95-1, BD Biosciences) at day 0; depletion of iNKT cells was achieved through i.p. injection of NKT14 [200 µg, NKT Therapeutics Inc. (19)] at day 0; depletion of neutrophils was achieved by injections of 500 µg anti-mouse Ly6G (αLy6G; 1A8, Bio X Cell) from 1 day prior to the start of DSS administration, and every second day thereafter.

At the experimental endpoint, mice were anesthetized with isoflurane, and blood collected using cardiac puncture. A midline incision was made to isolate and excise the colon, and colon length was measured. The distal colon was collected and fixed in 10% formaldehyde for 24 h, paraffin-embedded, cut into 4 µm sections, and stained using hematoxylin and eosin (H&E). Images of the colon sections were taken at 4× magnification using the Leica DM LB widefield microscope and MC120 HD camera (Leica), and scored following blinding and randomization. Parameters for histology scoring includes the level of tissue involvement (0: none, 1: mucosa, 2: mucosa and sub-mucosa, 3: sub-mucosa-transmural, 4: transmural), level of inflammation (0: none, 1: mild, 2: moderate, 3: severe, 4: severe with gastrointestinal-associated lymphoid tissue involvement), involvement of crypt and epithelium (0: none, 1: surface, 2:2/3 basal, 3: crypt and goblet cell loss, 4: crypt and goblet cell destruction with hyperplasia, surface epithelial destruction), and level of lamina propria/sub-mucosa edema (0: none, 1: mild, 2: moderate, 3: moderate-severe, 4: severe). Following collection of the distal colon for histology, the rest of the colon was used for flow cytometric analysis, or sectioned into two parts for myeloperoxidase assay (middle section) and RNA isolation (proximal section).

Fecal samples were collected directly from mice into individually labeled tubes in sterile conditions. The samples were immediately snap frozen in liquid nitrogen, and stored at −80°C. Extraction of fecal DNA was performed using the Isolate II Genomic DNA Kit (Bioline, USA) in accordance with manufacture’s protocol. PCR (30 cycles), using 50 ng of fecal-derived DNA as template, was performed using Q5 DNA polymerase (New England BioLabs) with a primer set selected to amplify the V3–V4 region of the gene encoding the 16S rRNA (forward: ACTCCTACGGGAGGCAGCAG; and reverse: GGACTACHVGGGTWTCTAAT). Equal quantities of each amplicon were pooled, and sequencing was performed on an Illumina MiSeq (2 × 300 bp), following the method detailed by Fadrosh et al. (20). Data analysis was performed using QIIME 1.9.1 software (21), and sequences were joined using the fastq-join method. The maximum allowed percentage difference within the overlapping region was zero. Sequences were de-multiplexed using the QIIME split library protocol, keeping only sequences with a Phred quality score >20. The data set was inspected for chimeric sequences using Pintail (22). Operational taxonomic units (OTUs) were picked using Uclust algorithm (23), and the taxonomy was assigned against the GreeneGene database (24). OTUs with less than 0.01% abundance or those assigned to Cyanobacteria were filtered out. Further data analysis was done using Calypso (25). Other than UniFrac and alpha diversity measures that used rarefied data, all statistical analysis was carried out using OTU table that was log 2 transformed and Cumulative Sum Scaling (CSS) normalized (26).

Serum collected at the end of experiments was stored at −80°C until use for CBA. Procedures for CBA were performed as per manufacture’s protocol (BD Biosciences). Samples were incubated with beads specific for IFN-γ, IL-2, IL-4, IL-6, and IL-17, and measured using a Navios flow cytometer (Beckman Coulter). The median fluorescent intensity for each sample was calculated, and quantified relative to a standard curve to obtain the concentration of each cytokine per sample.

Total RNA from colon tissue was extracted using TRI-reagent (Sigma) as per the manufacturer’s protocol. Since DSS has been shown to disrupt the activity of reverse transcriptase, the precipitated RNA samples were then purified of DSS through a lithium chloride purification step (27), and treated with DNase (Invitrogen). For flow cytometry-sorted CD45⁺Ly6G⁺ neutrophils (BD FACSARIA III), RNA isolation with DNase treatment was performed using the RNeasy Mini kit (Qiagen), as per the manufacturer’s protocol. The purified RNA was reverse transcribed into cDNA using SuperScript III synthesis system (Invitrogen). Quantitative PCR was performed using Power SYBR Green PCR Master Mix (Applied Biosystems), targeting expression of 18S, Chil3 (Ym-1), Cxcl1, Cxcl2, Il1b, Il18, Mrc1 (CD206) Nos2 (iNOS), Tgfb, and Tnfa (primer sequences can be found at Table S1 in Supplementary Material). All samples were run in triplicate and normalized to 18S. Expression of each gene was expressed as fold change relative to WT H₂O control, or to WT DSS for experiments involving the depletion of neutrophils.

Isolated colons were washed in phosphate buffered saline (PBS) to remove fecal matter, then cut into small pieces and washed in Hank’s balanced salt solution (HBSS, Life Technologies). These samples were incubated in HBSS with 10% fetal calf serum (FCS, Life Technologies) and 5 mM EDTA, then digested in 0.5 mg/mL collagenase D for 90 min, and passed through a 70 µm mesh. The cells were resuspended in 40% isotonic Percoll (Sigma), and layered over 80% isotonic Percoll. Gradients were resolved by centrifugation for 20 min at 1000 g with no brake, after which leukocytes were collected from the interface. Total colonic leukocyte numbers were counted by hemocytometer. For the spleen, splenocytes were gently pushed out of a needle-punctured spleen, and resuspended in FACS buffer (10% FCS and 5 mM EDTA in PBS). Red blood cells were lysed, and splenocytes were filtered through a 70 µm mesh.

Cell viability was determined using 7-Aminoactinomycin D, and leukocyte populations enumerated using fluorochrome-conjugated monoclonal antibodies against CD45 (30-F11, eBioscience), CD3 (145-2C11, BD Biosciences), Ly6C (AL-21, BD Biosciences), and Ly6G (1A8, Biolegend). Cells were analyzed on a BD FACSCanto II (BD Biosciences) and data analyzed using FlowJo (v10.0.7, Tree Star Inc.), or sorted using Influx sorter (BD Biosciences) or FACSARIA III (BD Biosciences). Populations were defined as CD45⁺ leukocytes, with further separation into CD3⁺ T cells, Ly6C^hiLy6G⁻ monocytes, and Ly6G⁺ neutrophils.

To examine the effect of DSS-induced intestinal inflammation on iNKT cell trafficking in the liver, intravital scanning disk confocal microscopy of the intact liver was performed on Cxcr6^gfp/+ mice, as previously described (28). Mice were anesthetized by i.p. injection of an anesthetic cocktail consisting of 150 mg/kg ketamine hydrochloride and 10 mg/kg xylazine, and the tail vein was cannulated to administer fluorescently labeled antibodies and/or additional anesthetic as required. Mice were placed in a right lateral position on an adjustable microscope stage. A lateral abdominal incision along the costal margin to the mid-axillary line was made to exteriorize the liver, and all exposed tissues were moistened with saline-soaked gauze to prevent dehydration. As positive control for iNKT cell activation, mice were injected with agonist alpha-galactosylceramide (α-GalCer; 2 μg/mouse i.p.) 4 h prior to imaging. Tissue was imaged using spinning disk confocal microscopy (Perkin Elmer). FITC and PE were excited at 488 and 567 nm respectively; in rapid succession and visualized with a 530/60 and 624/40 nm band pass filter, respectively. Typical exposure time for each excitation wavelength was 300 ms. A 512 × 512 pixel image was acquired every 6 s for 10 min. Andor IQ software (Andor) was used to drive the microscope.

The video files were imported into ImageJ, and adjustments to brightness and contrast were made to individual files to allow the best visualization and measurements of cells. Cells were tracked using the MTrackJ plugin [Biomedical Imaging Group Rotterdam (29)] to find a bright centroid within a 51 × 51 pixel box. The selection box is placed on the cell for every frame that the cell is clearly visible in. The displacement of each cell relative to its initial position was automatically calculated, then averaged. The measurements for all cells in at least two fields of views were averaged to form the data point for one individual mouse.

To examine the neutrophil–endothelial cell interactions in the colon during DSS-induced colitis, intravital multiphoton microscopy of the intact colon was performed on anesthetized LysM^eGFP mice, with tail vein cannulation as described. Body temperature was maintained using a heat pad. Mice were placed in a supine position, and the colon exteriorized through a midline incision. All exposed tissues were kept moist with saline-soaked gauze to prevent dehydration. The colon was carefully placed between a custom 3D-printed platform and coverslip setup, which was modified from a previously described setup (30). Vacuum grease was applied in between the 3D platform and the coverslip, allowing the colon to sit within a sealed chamber, enabling continued superfusion with a solution of 10⁻² M atropine in saline.

At least three post-capillary venules of approximately 35–60 µm in diameter were chosen per mouse. Images and videos were acquired using a multiphoton microscope (Leica SP5), equipped with a 20× water-dipping objective (NA 1.0) and a MaiTai pulsed infrared laser (SpectraPhysics) set to an excitation wavelength of 810 nm. A 512 × 512 pixel image was acquired every 1.5 s for 10 min. Labeling of the vasculature was performed by intravenous administration of 2.5 µg PE-conjugated anti-CD31 (clone 390, eBioscience).

The video files were imported into FIJI (v1.51, NIH) and converted into composite-color images. Adjustments to brightness and contrast were made to individual files to allow optimal visualization and measurements of blood vessels and cells. Multiple measurements of the width of blood vessels were taken and averaged, while the length of the vessel was measured along its center, both using the segmented line tool. The number of adherent, intravascular neutrophils was normalized to the vessel surface area and time of recording. Neutrophils were considered adherent if they interacted with the vessel wall for 30 s or more (31). The number of extravascular neutrophils was normalized to the area of colon that was visible in the field of view. The measurements from all fields of view from one mouse were averaged with the final data representing neutrophil numbers from one mouse.

For experiments assessing effects of atropine on the rate of blood perfusion in the colonic vasculature, FITC-conjugated microbeads were injected intravenously (i.v.), and images were acquired at 330 ms intervals for at least 30 s over three different fields of views. The effect of atropine superfusion was assessed with increasing concentrations of atropine (10⁻⁶–10⁻² M) in the same mouse.

Adult mouse colon was separated into 0.5–1 cm segments and embedded in O.C.T. embedding compound. 8 µm sections were taken using a cryostat, fixed using 4% paraformaldehyde and washed using distilled water and TBST. Blocking solution containing 0.5% Triton X-100 was added for 15 min followed by an incubation with FITC-conjugated anti-CD45 (1:100; eBioscience, San Diego, CA, USA) and PE-conjugated anti-FoxP3 (1:50; eBioscience, San Diego, CA, USA) antibodies for 48 h at 4°C. Samples were mounted with ProLong™ Diamond Antifade Mountant with DAPI (Life technologies) and left to cure for 2 h. Samples were imaged the same day using a Nikon C1 (Upright) confocal microscope equipped with a 40× objective lens and running NIS Elements Software (Nikon. Tokyo, Japan). Images were analyzed using ImageJ 2.0 software. CD45⁺FOXP3⁺ cells were considered as regulatory T cells (Tregs).

At the end of the model of colitis, 500 mg/kg of 4.4 kDa fluorescein-isothiocyanate-labeled dextran (FITC-dextran, Sigma) was orally gavaged into mice. After 4 h, mice were anesthetized by isoflurane, and cardiac puncture was performed to collect blood for serum. Samples were stored at −80°C until use. Serum fluorescence was read at gain 75, excitation 485 nm, and emission 520 nm using an Infinite M1000 Pro (Tecan) plate reader and Magellan 7.2 (Tecan) software. Sample concentrations of FITC-dextran translocation were determined relative to that of a standard curve using Prism (GraphPad) software.

Student’s t-test was used to compare the mean between two groups, except for gene expression data, where Mann–Whitney’s U-test was performed. Two-way ANOVA with matched time points was used to compare the differences between two groups over time. All data are represented as mean ± SEM, or with a line at the median for gene expression data. Comparisons were considered significant if the p-value is <0.05.

To investigate the role of iNKT cells in colitis, DSS colitis was induced in wild type (WT) and iNKT cell-deficient mice (Jα18^−/− mice). Both WT and Jα18^−/− mice on normal water had natural weight gain and developed no clinical symptoms, as scored by the DSS-associated disease activity index (DAI) throughout the 7 days of the experimental period. However, administration of DSS caused progressive weight loss in WT mice starting from day 3, when normalized to the weight of H₂O controls (Figure 1A, non-normalized data in Figure S1A in Supplementary Material). Additionally, DSS-treated WT mice also showed observable clinical symptoms as measured by the DAI such as loose blood-containing stools (Figure 1B) and shortening of the colon length (a surrogate marker of inflammation; Figure 1C). In contrast, Jα18^−/− mice treated with the same concentration of DSS demonstrated reduced and delayed onset of weight loss (Figure 1A), as well as lower DAI (Figure 1B) and significantly less colon shortening compared to their WT counterparts at the end of the model (Figure 1C).

Blinded histological assessment of colons revealed no differences between WT and Jα18^−/− mice in the control H₂O group. Following DSS treatment, both WT and Jα18^−/− mice showed increased inflammation (Figures 1D,E). Consistent with the DAI data, colons from Jα18^−/− mice exhibited significantly lower histology scores compared to WT mice (Figure 1E). Intriguingly, these DSS-induced differences in pathology between WT and Jα18^−/− mice were independent of increased intestinal permeability, which was similar in WT and Jα18^−/− mice following DSS treatment (Figure S1B in Supplementary Material). Taken together, these results demonstrate that Jα18^−/− mice have less severe colon inflammation following DSS-induced colitis, and suggest that iNKT cells play a key role in contributing to the colonic inflammation and pathology in this model.

A recent study showed that mice deficit of all types of NKT cells (Cd1d^−/− mice) possess a pro-colitogenic gut microbiota, and the disruption of gut epithelial integrity by DSS exposes the host to the luminal contents, inducing a pro-inflammatory response (32). This raises the possibility that mice with a selective deficiency in iNKT cells may harbor a different microbiota to WT mice, influencing the development of disease. To investigate this, we compared the fecal microbiota composition of naïve WT and Jα18^−/− mice. We found there were no differences in the richness of microbial communities in the fecal content between WT and Jα18^−/− mice as measured by α-diversity (Figure S2 in Supplementary Material), but there was a significant difference in the microbial communities between these two strains of mice based on unweighted (different membership; p = 6.6 × 10⁻⁴, Figure 2A) and weighted (different abundance, p = 6 × 10⁻⁴, Figure 2B) UniFrac and Adonis statistics.

Comparison of the fecal microbiota at the phylum level in WT and Jα18^−/− mice showed significantly elevated relative abundance of Tenericutes in Jα18^−/− mice, along with a marked reduction of Deferribacteres and Actinobacteria in these mice (Figure 2C). Moreover, at the family level, there was an increased abundance of Rikenellaceae, Turicibacteraceae, Bifidobacteriaceae, and Prevotellaceae (Figure 2D); and a markedly reduced frequency of OTUs belonging to the families of Lachnospiraceae, Lactobacillaceae, Bacteroidaceae, Ruminococcaceae, Desulfovibrionaceae, Coriobacteriaceae, Dehalobacteriaceae, and Defferibacteriaceae in the fecal microbiota of Jα18^−/− mice (Figure 2D). The genera that were found to be significantly more abundant in Jα18^−/− microbiota include Bifidobacterium, PSBM3, Olsenella, and Prevotella (Figure 2E). Intriguingly, we clearly observed a distinction of microbial composition between the strains at the level of genera (Figure 2E) and OTU (species; Figure 2F). Furthermore, differential phylotypes were determined using t-test on normalized data with FDR corrected p values <0.05 considered significant. At the OTU level (often considered as a species level; clustered at 97% similarity), there were 411 (out of 686) OTUs significantly different (FDR p < 0.05) between the fecal microbiota of WT and Jα18^−/− mice at p < 0.01 and a total of 513 OTUs with FDR corrected p < 0.05 (see MG-RAST library mgl621715). In WT mice, the fecal microbiota consisted of bacterial species belonging to the genera Mucispirillum, Ruminococcus, Bacteroides, Clostridium, and Lactobacillus, and these have previously been positively associated with inflammation and colitis (33–35) (Figure 2G). In contrast, the LEfSe analysis demonstrated the major drivers of difference in the microbiota shaped by the absence of iNKT cells were microorganisms previously associated with anti-inflammatory properties, including Bifidobacterium, PSBM3, Alistipes, Olsenella, and Prevotella (Figure 2G). Taken together, these findings indicate the composition of the colonic microbial communities is significantly affected by the absence of iNKT cells.

Next, we investigated whether iNKT cells and the gut microbiota shaped by it have a role in regulating colonic inflammation following DSS-induced colitis. Leukocyte recruitment, in particular that of neutrophils, is a histological hallmark of disease (1). Therefore, we examined the gene expression of pro-inflammatory cytokines and chemokines that support the recruitment of leukocytes in the colons of WT and Jα18^−/− mice. There were no significant differences in the expression of Il1b, Cxcl1, and Cxcl2 between WT and Jα18^−/− mice in the H₂O control groups (Figures 3A–C). Following administration of DSS, there were significant increase in the expressions of Cxcl1 and Cxcl2 in WT mice following DSS treatment (Figures 3B,C). In Jα18^−/− mice, the expressions of Il1b, Cxcl1, and Cxcl2 were also significantly higher after DSS treatment (Figures 3A–C). We also examined the gene expression of the anti-inflammatory cytokine, Tgfb, and found a DSS-induced reduction in WT mice was not observed in Jα18^−/− mice (Figure 3D). There were no changes in the expression of Tnfa and Il18 between WT and Jα18^−/− mice (Figure S3 in Supplementary Material).

In extension to the local changes, we examined whether the absence of iNKT cells and the accompanying alterations in microbiota composition were associated with systemic inflammatory response following DSS-induced colitis. Although iNKT cells have previously been shown to be present at low levels in the colon (36), the liver contains a high abundance of iNKT cells at baseline (37), and is the site where luminal content derived from the gut must undergo filtration before entry into the systemic circulation (38). Therefore, it is feasible that the iNKT cell-dependent effect on inflammation we observed in the colon was dependent on hepatic iNKT cells. To assess whether DSS induction activated hepatic iNKT cells, iNKT cell-reporter (Cxcr6^gfp/+) mice were administered with DSS and liver intravital microscopy was performed. At baseline, hepatic iNKT cells had a mean displacement of 15 µm within the liver sinusoids over 10 min (Figure 3E). Consistent with published reports, the administration of a potent iNKT cell agonist, α-GalCer, resulted in a significant reduction in displacement 4 h after administration (Figure 3E). In contrast, we found no such reduction in the migratory displacement in mice following 7 days of DSS treatment (Figure 3E; Figure S4 in Supplementary Material), suggesting that hepatic iNKT cells were not activated or affected following DSS-induced colitis.

Similarly, when we examined the serum level of cytokines that iNKT cells are known to produce (IFN-γ, IL-4, and IL-17), we observed no changes between baseline and post-DSS nor between WT and Jα18^−/− mice, except for a significant increase of IFN-γ in WT mice after DSS treatment (Figures 3F–H). In addition, there were no differences in the systemic levels of IL-2 and IL-6 in DSS-treated WT and Jα18^−/− mice (data not shown). Taken together, these findings indicate that the local colonic milieu in WT and Jα18^−/− mice, rather than systemic inflammation, supports the recruitment of leukocyte following DSS-induced colitis.

To compare the colonic leukocyte infiltrate in WT and Jα18^−/− mice following DSS-induced colitis, we isolated leukocytes from the colon lamina propria and quantitated total leukocytes, neutrophils, monocytes, and T cells using flow cytometry (Figure S5 in Supplementary Material). At baseline, there were similar numbers of total immune cells in the colons of WT and Jα18^−/− mice (Figure 4A). However, in the presence of DSS-induced colitis, Jα18^−/− mice had increased gut immune cell numbers compared to WT DSS animals (Figure 4A). More specific immune cell analysis demonstrated that this increase was due to the greater number of neutrophils (Ly6G⁺; Figure S6A in Supplementary Material; Figure 4B), monocytes (Ly6C^hi/Ly6G⁻; Figure 4C), and T cells (CD3⁺; Figure 4D). Also evident when assessing the overall immune make-up of CD45⁺ cells was that although DSS-induced colitis significantly increased the proportion of neutrophils in the gut of Jα18^−/− mice to a greater extent than in WT mice (Figure S6A in Supplementary Material), there were no significant differences between the proportions of monocytes and T cells between post-DSS WT mice and Jα18^−/− mice (Figures S6B,C in Supplementary Material). These findings suggest iNKT cells play a key role in dampening neutrophil recruitment during colonic inflammation.

To provide further support for the role of iNKT cells in this response, we also administered 2% DSS to Cd1d^−/− mice, which lack all types of NKT cells. Similar to the Jα18^−/− mice, total leukocyte infiltration was increased in Cd1d^−/− mice compared to WT mice following DSS induction (Figure S6D in Supplementary Material). The proportion of neutrophils in both WT and Cd1d^−/− mice following DSS induction was also increased when compared to H₂O controls (Figure S6E in Supplementary Material). Furthermore, post-DSS Cd1d^−/− mice also exhibited elevation of neutrophil recruitment to the colon at day 7 of the disease compared to WT mice (p = 0.05 Figure S6F in Supplementary Material). These results clearly demonstrate that iNKT cells, rather than other types of NKT cells, play an important role in the regulation of neutrophil recruitment at day 7 of DSS-induced colitis.

Our findings so far show that DSS-induced colitis promotes neutrophil recruitment in both WT and Jα18^−/− mice. However, whether the enhanced neutrophil recruitment in Jα18^−/− mice is contributed exclusively by the absence of iNKT cells, or in combination with the altered composition of the gut microbiota remains unclear. To address this issue, we examined the effect of iNKT cell-depletion on neutrophil recruitment in a single mouse strain consisting of the same microbiota. To do this, we utilized NKT14, an antibody specific for the invariant chain of the T cell receptor, which has been shown previously to deplete iNKT cells in mice (19). To validate the efficacy of NKT14 for iNKT cell depletion, we injected WT mice with NKT14 at day 0, then used flow cytometry to analyze leukocyte populations in the spleen and blood at days 3, 5, and 7. A single dose of NKT14 was sufficient to induce a sustained reduction in iNKT cells over 7 days in both spleen and the circulation (representative plots in Figure S7 in Supplementary Material, data in Figures S8A,B in Supplementary Material), without altering T cell proportions or numbers (Figures S8C–F in Supplementary Material).

To elucidate the effect of iNKT cell-depletion on neutrophil infiltration, we injected neutrophil reporter (LysM^eGFP) mice with NKT14 at day 0 of the DSS model and examined the effect at baseline (day 0), day 3, 5, or 7 using colon intravital microscopy (Figure 5A; Videos S1 and S2 in Supplementary Material). For these experiments we used and validated a new approach for colon intravital microscopy of the colonic microvasculature modified from published methodology (30), avoiding the use of tissue glue, which enables the colon to be in a condition that is much more physiological (Figure S9A in Supplementary Material). In LysM^eGFP mice, circulating neutrophils are detectable as round cells brightly fluorescent via GFP (Figures 5B–F). This allowed the differentiation between, and evaluation of, adherent intravascular and transmigrated extravascular neutrophils. At baseline, there were minimal intravascular or extravascular neutrophils in the colon (Figures 5B,G; Figure S9B in Supplementary Material). The number of intravascular neutrophils increased with the progression of DSS-induced colitis (Figures 5C–E,G). Interestingly, depletion of iNKT cells resulted in significantly higher numbers of intravascular neutrophils at 3 and 5 days after DSS induction (Figures 5F,G). The number of extravascular neutrophils also increased with the progression of DSS-induced colitis, however, there were no significant differences following depletion of iNKT cells (Figure S9B in Supplementary Material). These experiments reveal that iNKT cells have the capacity to regulate neutrophil recruitment, regardless of the composition of the microbiota, and particularly in the earlier stages of DSS-induced colitis development.

To examine the mechanisms by which iNKT cells regulate neutrophil recruitment, we examined the role of antigen presentation via CD1d (39). Antibody-mediated inhibition of CD1d-dependent antigen presentation resulted in a significant increase in total leukocyte numbers in the colon at day 7 of DSS-induced colitis (Figure 5H). Similar to our findings in Jα18^−/− mice, there was a significant increase in neutrophil numbers in the colon of post-DSS mice treated with α-CD1d antibody (Figure 5I), though no differences were seen in the proportion of recruited neutrophils (Figure S9C in Supplementary Material). Therefore, by removing the potential effect of different microbiota composition between WT and Jα18^−/− mice, our findings demonstrate that iNKT cells function to suppress the recruitment of neutrophils in a CD1d-dependent manner following DSS-induced colitis.

Neutrophils have the capacity to adopt different phenotypes including non-traditional anti-inflammatory functions that may be important in reducing intestinal inflammation (40, 41). However, whether the phenotypes and functions of neutrophils in WT and Jα18^−/− mice are different is currently unknown. To investigate this, neutrophils were isolated from the colons of WT and Jα18^−/− mice at day 7 of DSS-induced colitis, and the expression of anti-inflammatory markers were assessed. Neutrophils that infiltrated the colons of Jα18^−/− mice demonstrated increased expression of Chil3 (Ym-1) and Nos2 (iNOS) (Figure 6A), indicating a shift to an anti-inflammatory phenotype (42, 43). To test whether this difference in neutrophil phenotype is important in decreasing intestinal inflammation and confer disease protection, mice were removed of neutrophils using neutrophil depleting antibody and treated with DSS (Figure S10 in Supplementary Material). We have shown that post-DSS Jα18^−/− mice had reduced weight loss, significantly lower DAI, longer colons, and lower histology score when compared to post-DSS WT mice (Figures 1 and 6B–E). However, after depletion of neutrophils (comparing WT + αLy6G vs. Jα18KO + αLy6G), the difference between WT and Jα18KO is abrogated (Figures 6B–E).

Next, we examined local changes in gene expression in the colon. Depletion of neutrophils resulted in a significant reduction in the expressions of colonic Il1b, Cxcl1, and Cxcl2 irrespective of strain (Figures 6F–H). On the other hand, expression of Tgfb was significantly higher in αLy6G-treated Jα18^−/− mice than in WT mice (Figure 6I). TGF-β has been shown to induce the differentiation of Tregs (44), thus we used immunofluorescence to assess changes in CD45⁺Foxp3⁺ Tregs in the colon. Intriguingly, only in mice without neutrophils and iNKT cells did we observe a significant increase in colonic Tregs (Figure 6J). Taken together, these results indicate that neutrophils recruited to the inflamed colon in the absence of iNKT cells exhibit an anti-inflammatory phenotype. Therefore, in the absence of iNKT cells, neutrophils have important roles in immune regulation that ultimately confers protection to DSS-induced intestinal inflammation.