Rab32‐related antimicrobial pathway is involved in the progression of dextran sodium sulfate‐induced colitis

Inflammatory bowel disease (IBD) is a multifactorial disease involving defective immune responses against invasive microbiota. Genes associated with innate immune responses to microbes have been highlighted in the pathogenesis of IBD. To determine the role of Rab32 in the pathogenesis of IBD, we administered dextran sodium sulfate (DSS) to CD11c+ cell‐specific Rab32 knockout (CD11c‐Cre+Rab32f/f) mice to induce colitis. Rab32 deficiency in CD11c+ cells resulted in more severe disease progression and increased mortality. Histopathological analysis showed extensive damage to the colon mucosa in DSS‐treated CD11c‐Cre+Rab32f/f mice, including more severe damage to the epithelial layer and crypts, as well as more inflammatory cell infiltration. The pro‐inflammatory cytokines IL1A, IL1B, IL6, and CSF3 and chemokines CXCL1 and CXCL2 were significantly increased, and the frequency of CD11b+Ly6G+ neutrophils was higher in CD11c‐Cre+Rab32f/f colitis mice. Furthermore, CD11c+ cells deficient for Rab32 exhibited a significant increase in bacterial translocation in inflamed colon tissue. The present data demonstrate that Rab32 knockout in CD11c+ cells aggravates the development of DSS‐induced colitis and suggest that the Rab32‐related antimicrobial pathway is involved in the pathogenesis of IBD.

Inflammatory bowel disease (IBD) is a multifactorial disease involving defective immune responses against invasive microbiota. Genes associated with innate immune responses to microbes have been highlighted in the pathogenesis of IBD. To determine the role of Rab32 in the pathogenesis of IBD, we administered dextran sodium sulfate (DSS) to CD11c + cell-specific Rab32 knockout (CD11c-Cre + Rab32 f/f ) mice to induce colitis. Rab32 deficiency in CD11c + cells resulted in more severe disease progression and increased mortality. Histopathological analysis showed extensive damage to the colon mucosa in DSS-treated CD11c-Cre + Rab32 f/f mice, including more severe damage to the epithelial layer and crypts, as well as more inflammatory cell infiltration. The pro-inflammatory cytokines IL1A, IL1B, IL6, and CSF3 and chemokines CXCL1 and CXCL2 were significantly increased, and the frequency of CD11b + Ly6G + neutrophils was higher in CD11c-Cre + Rab32 f/f colitis mice. Furthermore, CD11c + cells deficient for Rab32 exhibited a significant increase in bacterial translocation in inflamed colon tissue. The present data demonstrate that Rab32 knockout in CD11c + cells aggravates the development of DSS-induced colitis and suggest that the Rab32-related antimicrobial pathway is involved in the pathogenesis of IBD.
Inflammatory bowel disease (IBD) is a prevalent chronic inflammatory disorder in the digestive tract that comprises two main types, Crohn's disease (CD) and ulcerative colitis. Abdominal pain, diarrhoea and rectal bleeding are common symptoms of IBD patients [1,2]. Since the 1950s, the incidence of IBD has increased, and currently, the prevalence of this disease is up to 0.5% of the general population in Western countries [3,4]. The pathogenesis of IBD is multifactorial, including environmental factors, genetic susceptibility, dysregulated immunity and infectious microbes [5]. Genomic studies have identified more than 200 loci containing IBD susceptibility genes, focusing on critical pathways involved in innate immunity, autophagy, lymphocyte differentiation and chemotaxis [6,7].
Among these loci, a number of genes associated with microbial sensing (NOD2, REL, IRF5, RELA, PTPN22, CARD9 and RIPK2) and clearance (ATG16L1, NCF4I and RGM) have been highlighted [8][9][10][11]. In the gastrointestinal tract lamina propria, the intestinal dendritic cells (DCs) and macrophages (MPs) provide first-line defence against invading microorganisms [12]. Disruptions in the DCs and MPs lead to bacterial translocation across the intestinal mucosa and subsequent inflammation [13]. Accumulated evidence in IBD patients has demonstrated that dysfunctional bacteria clearance ability increases bacterial invasion in mucosa biopsies [14,15]. However, more mechanisms underlying the innate immune response against microbial invasion during IBD progression remain unexplored.
Recently, Rab32 was identified as an essential player in the clearance of intracellular pathogens both in DCs and MPs [16][17][18]. Moreover, Rab32 has been found to be frequently methylated in IBD and IBD-related neoplasms [19,20]. In this study, we examined the role of Rab32 in a dextran sodium sulfate (DSS)-induced colitis model. We found that Rab32 deficiency in CD11c + cells exacerbates the progression of colitis in mice with increased neutrophil infiltration and bacterial invasion.
CD11c-specific Rab32-deficient mice (CD11c-Cre + Rab32 f/f ) were generated according to our published paper [16]. Rab32 tm1a(KOMP)Wtsi mice were initially crossed with Flpe mice and then crossed with CD11c-Cre mice to delete the second exon of Rab32. All of the mice were bred under specific-pathogen-free conditions. All animal experiments were conducted on protocols that were approved by the Medicine Animal Care Committee of the Army Medical University (Chongqing, China).

Histological analysis
The caecum and colon were fixed with paraformaldehyde and stained with haematoxylin and eosin (HE). The colonic inflammation degree was evaluated based on a previously validated scoring system [25][26][27] as follows: infiltration (0 points = rare infiltration, 1 point = increased number of inflammatory cells in the mucosa and 2 points = confluence of inflammatory cells extending into the submucosa), crypt damage (0 points = none, 1 point = basal 1/3 damage, 2 points = basal 2/3 damage, 3 points = only surface epithelium intact and 4 points = entire crypt and epithelium damage) and percentage of area involved (0 points = none, 1 point = 1-25%, 2 points = 26-50%, 3 points = 51-75% and 4 points = 76-100%). The colon was equally divided into four segments. The proximal colon score was determined as the average of the upper two segments, and the distal colon score was calculated as the average of the lower two segments.

Cell isolation and flow cytometry analysis
Intraepithelial (IEL) and lamina propria lymphocytes (LPL) were isolated from the colon as previously described [28]. Briefly, the colon was flushed with 5 mL of PBS to remove the contents; then, it was opened longitudinally and cut into~0.5-cm pieces. To harvest IEL, the tissues were incubated twice with 5 mM EDTA containing 2% FBS at 37°C for 20 min and the supernatant was collected. To isolate LPL, the tissues were digested twice in 0.5 mg collagenase D (Roche)/100 mL RPMI 1640 containing 5% FBS at 37°C for 30 min and the supernatant was subsequently collected. 40%/80% Percoll (GE Healthcare, Uppsala, Sweden) step-gradient was used to enrich IEL and LPL. Mesenteric lymphoid nodes (MLNs) were mechanically dispersed through a stainless steel 200-lm wire mesh in PBS, and the cells were washed after centrifugation for 5 min at 500 g. The cells were incubated with FITC-conjugated CD3, PE-conjugated CD4, PerCP-cy5.5-conjugated CD8a, APC-cy7-conjugated CD11b and APC-conjugated Ly6G (all antibodies from eBioscience, San Diego, CA, USA) antibodies at 4°C for 30 min, and DAPI was used to stain for dead cells at 4°C for 5 min. The cells were subsequently detected on the BD FACSVerse platform and analysed with FLOWJOV10 software (San Carlos, CA, USA).

Real-time qPCR
RNA was extracted from the colon tissues using an RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Total RNA (1 lg) was reverse-transcribed into cDNA with a reverse transcription kit (Sigma, St. Louis, MO, USA) and quantified with SYBR Green-based real-time qPCR with specific primers for IL1A, IL1B, IL6, IL 17, IFNc and TNFa (synthesised at BGI, China). The inflammatory-related genes were summarised from the published studies. All the designed primers for these genes (135 genes) were optimised and organised into the inflammatory response qPCR array primer plate (Table S1). The qPCR array reaction was performed in a 384-well optical plate with a 2.0 lL total reaction volume by liquid handling with Echo 550 (Labycte, Sunnyvale, CA, USA).
A16S ribosome DNA (rDNA) qPCR was used to quantify the bacteria of the colon, MLNs and spleen as previously described [29]. Briefly, the spleen, MLN and colon were separated, and the colon was flushed with 5 mL of sterile PBS to remove the contents. DNA was extracted using a DNA isolation kit (Qiagen) and amplified using primer sets for 16S rDNA of universal bacteria. The sequences of the primers used for real-time qPCR are shown in Table S2.
The CT value of all the qPCR products was normalised to housekeeping genes, and the quantitated relative expression levels were determined using DDC T analysis. In the qPCR array analysis, the limma package in R was used to identify genes that were differentially expressed between experimental conditions [30].

Bacterial culture
The colon, MLNs and spleen were separated from DSSinduced colitis mice on day 8. The colon was flushed with 5 mL of sterile PBS to remove the colon contents. Then, the colon tissues, MLN and spleen were homogenised in sterile PBS and plated onto BBL Brucella agar plates (containing 5% horse blood) under sterile conditions as previously described [29]. The plates were incubated for 48 h at 37°C under anaerobic conditions, followed by colonyforming unit (CFU) counting.

Statistical analysis
Statistical analysis was performed by using GRAPHPAD PRISM 6 software (Graphpad software Inc., San Diego, CA, USA) to calculate unpaired Student's t-test or Log-rank test. All P-values < 0.05 were considered to indicate significant results and were marked with asterisk (*P < 0.05; **P < 0.01; ***P < 0.001).

Rab32 knockout in CD11c + cells aggravates the development of DSS-induced colitis
To investigate the roles of Rab32 in CD11c + cells in DSS-induced colitis, we generated CD11c + cell-specific Rab32-deficient (CD11c-Cre + Rab32 f/f ) mice (Fig. 1A). Deletion of Rab32 in CD11c + cells was confirmed by PCR and western blot analysis of Rab32 in BMDCs at the genomic and protein levels (Fig. S1). DSSinduced colitis is a classical disease model with many similarities to the clinical manifestations of colitis in human IBD patients. After orally administering 2.5% DSS in drinking water for 8 days, DSS-induced weight loss was significantly higher in CD11c-Cre + Rab32 f/f mice than in WT mice (Fig. 1B). CD11c-Cre + Rab32 f/f mice weighed 73.9 AE 3.5% of their original body weight, while WT mice weighed 85.2 AE 3.5% of their original body weight. To more comprehensively evaluate the disease progression, we monitored mice daily with a disease activity index (DAI) that includes weight loss, general appearance, diarrhoea and rectal bleeding. The weight loss in CD11c-Cre + Rab32 f/f mice began on day 4, and diarrhoea with bleeding was clearly detected on day 7, while weight loss began on day 5 in the WT mice, and little bleeding was observed. The DAI score of CD11c-Cre + Rab32 f/f mice was significantly higher than that of control mice (P < 0.001, Fig. 1C). Moreover, the mortality of the CD11c-Cre + Rab32 f/f mice reached over 90% when post continuously treated with 2.5% DSS but reached only approximately 50% in the WT mice (Fig. 1D). These data demonstrate that Rab32 deficiency in CD11c + cells results in more severe disease progression and increased mortality.
Consistent with the clinical score data, the colon of CD11c-Cre + Rab32 f/f mice was shorter than that of control mice (44.4 AE 3.6 mm versus 63.1 AE 4.7 mm, respectively, P < 0.001, Fig. 1E,F). Histological scores revealed a significantly greater degree of tissue injury in CD11c-Cre + Rab32 f/f mice than in WT control mice after DSS treatment (Fig. 2), although no significant differences were observed between CD11c-Cre + R-ab32 f/f and WT mice administered water (Fig. S2).
Increased pro-inflammatory cytokine expression and neutrophil infiltration in the inflamed colon tissue of CD11c-Cre + Rab32 f/f mice Cytokines, including interleukins and interferons, play a critical role in the pathogenesis of IBD and are dramatically upregulated in the DSS model [31]. The inflammatory cytokines IL1A, IL1B, IL6, IL17, TNFa and IFNc were investigated first. IL1A, IL1B and IL-6 were significantly increased in CD11c-Cre+Rab32 f/f mice (Fig. 3), while the cytokines IL17, TNFa and IFNc showed no changes (Fig. S3). Using a homemade inflammatory response qPCR array, the expression of 135 inflammatory-related genes, including cytokines, chemokines and chemokine receptors, and transcriptional factors was assessed (Table S1). The expression of chemokines CXCL1, CXCL2 and CSF3 was significantly increased in CD11c-Cre+Rab32 f/f mice (Fig. 4A,B). However, the cytokines and transcription factors associated with the T-cell response and differentiation, such as IL4, IL12, IL10, Cre + Rab32 f/f mice and WT mice treated with DSS or water (n = 10 mice/group). The survival of mice was recorded daily, and the survival rate was calculated using the Kaplan-Meier test. (E, F) Length and gross imaging of colons dissected from the CD11c-Cre + Rab32 f/f and WT mice administered with DSS or water. All data are shown as the mean AE SD. Significances were calculated using unpaired Student's t-test or Log-rank test. *P < 0.05, **P < 0.01 and ***P < 0.001. CD11c-Cre + Rab32 f/f colitis mice were microscopically examined at 40X and 200X, and histology scores were analysed for 6-8 mice from each group. Scale bar: 500, 100 lm. All data are shown as the means AE SD. Significances were calculated using unpaired Student's t-test. **P < 0.01 and ***P < 0.001. Fig. 3. Aberrant expression of pro-inflammatory cytokines in CD11c-Cre + Rab32 f/f colitis mice. Total RNA was extracted from the colons tissues of untreated and DSS-treated mice on day 8. Expression of the pro-inflammatory cytokines (A) IL1a, (B) IL1b and (C) IL6 was measured by qPCR (n = 6-8 mice/group). All data are shown as the means AE SD. Significances were calculated using unpaired Student's ttest. *P < 0.05, **P < 0.01 and ***P < 0.001. Granzyme A/B, STAT4, GATA3, RORC and FOXP3, showed no differences. These data suggest that the severe colonic inflammation in CD11c-Cre + Rab32 f/f mice may not be caused by dysfunctional T cells.
Intraepithelial and LPL were isolated and analysed by FACS. The frequency of CD3 + , CD4 + and CD8 + T cells showed no significant differences between the CD11c-Cre + Rab32 f/f and WT mice (Fig. S3). The expression of CD80, CD86 and MHCII also showed no differences in CD11c + BMDCs after LPS stimulation (Fig. S4). However, the frequency of CD11b + Ly6G + neutrophils was higher in CD11c-Cre + Rab32 f/f colitis mice than in the WT colitis mice (Fig. 4C,D). No significant differences were observed between CD11c-Cre + R-ab32 f/f and WT mice administered water (Fig. S5). The MP chemokines CXCL1/CXCL2 were reported to control the early stage of neutrophil recruitment during tissue inflammation [32][33][34]. These data indicate that the severe tissue damage of CD11c-Cre + Rab32 f/f colitis mice was related to the abundant neutrophil infiltration, which were recruited by the increased expression of chemokines in the tissues.
CD11c-Cre + Rab32 f/f mice are more susceptible to bacterial invasion Neutrophils are the critical components in the innate immune response against the invasion microbes and can be recruited by bacteria-derived products [35,36]. During DSS administration, the epithelial layer is destroyed, and bacteria infiltrate the intestinal mucosa [29]. Therefore, we assessed the bacterial infiltration in the DSS colitis mice. By quantifying the bacterial 16S rDNA with real-time qPCR, we found that CD11c-Cre + Rab32 f/f and WT mice administered water showed no differences in the level of bacterial invasion. However, after DSS administration, bacterial 16S rDNA copies were significantly increased in the MLN and colons. Under anaerobic culture, the colon tissues and MLN of CD11c-Cre + Rab32 f/f mice showed more bacterial growth. No significant differences were observed in the spleen between the two groups through qPCR of 16S rDNA and the anaerobic culture (Fig. 5A,B). These data indicated that commensal bacteria could be resisted in a healthy state of CD11c- (C, D) The frequencies of neutrophils (CD11b + Ly6G + ) in isolated colonic IEL, LPL and MLNs from mice in the indicated groups were determined by FACS, 5-8 mice in each group. All data are shown as the means AE SD. Significances were calculated using unpaired Student's t-test. *P < 0.05, **P < 0.01 and ***P < 0.001.
Cre + Rab32 f/f mice, but when the mucosal was damaged by DSS, the deficiency of Rab32 in CD11c + cells causes more bacterial translocation in inflamed colon tissue and MLN.

Discussion
Accumulating studies have highlighted the importance of innate immune responses to microbes in the pathogenesis of IBDs. Blood monocytes isolated from IBD patients showed decreased phagocytosis and killing of bacteria [37]. Recent studies determined that MPs isolated from IBD patients showed decreased reactive oxygen species production [38,39]. Defects in autophagy impaired the killing of adherent-invasive E. coli either in DCs or in MPs [11,40]. Deletion of autophagy-related gene ATG16L1 in CD11c + cells results in more severe DSS-induced colitis in mice [41]. In this study, we found that Rab32 deficiency in CD11c + cells aggravates the progression of DSS-induced colitis, with a more severe inflammation appearance, colonic tissue damages and higher mortalities. The pro-inflammatory cytokines (IL1A, IL1B and IL-6), neutrophils and bacterial infiltration were significantly increased in CD11c-Cre + Rab32 f/f mice with colitis induced by DSS. These data suggest that the CD11c positive cells play an important role in the process of DSS-induced colitis by defending the translocation of microbiota. Although the CD11c is a traditional surface marker to identify DCs, a subset of MPs in gastrointestinal tract express low to intermediate amounts of CD11c [42]. Both types of CD11c positive cells are known to have bacterial processing capability. We do not rule out which subset of CD11c + cells with a deletion in Rab32 contribute to the effects we observed. Future studies using deletions placed behind alternative promoters may clarify the picture.
Massive amounts of neutrophils were recruited to the colon tissues of IBD patients [43]. Although the specific role of neutrophils during colitis has not been clearly delineated, some reports suggest that neutrophils are rapid responders to eliminate the translocated microbes across the intestinal epithelial layer [44]. Our data support the hypothesis because Rab32 deficiency in CD11c + cells resulted in a more severe colitis progression and increased the frequency of CD11b + Ly6G + neutrophils in inflamed tissue. It has been reported that MP chemokines CXCL1/CXCL2 The colon tissues, MLN and spleen were homogenised and incubated onto BBL Brucella agar plates for 48 h at 37°C under anaerobic conditions, followed by CFU counting (n = 8-10 mice/group). All data are shown as the means AE SD. Significances were calculated using unpaired Student's t-test. *P < 0.05, **P < 0.01 and ***P < 0.001.
contribute to the recruitment of neutrophils at an early stage of tissue inflammation [45]. In this study, proinflammatory cytokines IL1A, IL1B, IL6, and CSF3 and chemokines CXCL1 and CXCL2 were significantly increased in CD11c-Cre + Rab32 f/f mice. It is possible that the deficiency of Rab32 in CD11c + cells recruits more neutrophils by chemokines and growth factors to cause more serious tissue damage [46,47].
Rab32 has been well characterised in intracellular vesicle trafficking processes involving lysosome-related organelle biogenesis, apoptosis and autophagy [48][49][50]. Recent studies have demonstrated the critical role of Rab32 in bacteria restriction and infectious diseases [51,52]. Genome-wide association studies identified IL23R and Rab32 as new susceptibility genes of leprosy, a disease caused by Mycobacterium leprae [53]. Rab32 and its guanine nucleotide exchange factor BLOC-3 are essential to prevent the growth of the human-restricted salmonella enterica serovar Typhi (S. Typhi) in mice [17]. In our previous study, Rab32-Phb-Phb2 complex-associated membrane recaptures escaped Listeria to control Listeria replication in cytosol [16]. Here, the deficiency of Rab32 in CD11c + cells significantly increased the bacterial translocation in inflamed colon tissue after the mucosa was damaged by DSS. The data indicate the Rab32-related antimicrobial pathway is involved in the progress of colitis.
In summary, the present data demonstrated that Rab32 in CD11c + cells protects against the development of DSS-induced colitis by preventing bacterial invasion and suggested that Rab32 might contribute to the pathogenesis of human IBD and IBD-related colorectal carcinoma.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig. S1. Analysis of the CD11c-Cre + Rab32 f/f mice. BMDCs and BMDMs generated from WT and CD11c-Cre + Rab32 f/f mice. (A) PCR analysis with primers amplified from exon1 and exon 3 of Rab32 with genomic from BMDCs. (B) The expression of Rab32 protein in BMDCs generated from WT and CD11c-Cre + Rab32 f/f mice were analysed by Western blot. (C) The expression of Rab32 protein in BMDMs of WT and CD11c-Cre + Rab32 f/f mice were analysed by Western blot. Fig. S2. Histology analysis of colon tissue in the WT and CD11c-Cre + Rab32 f/f mice administered water. HE-stained sections of the (A) caecum, (B) proximal colon and (C) distal colon from WT and CD11c-Cre + Rab32 f/f mice were microscopically examined at 40X and 200X, and histology scores were analysed for 7-8 mice each group. Scale bar: 500, 100 lm. All data are shown as the means AE SD. Fig. S3. The expression of pro-inflammatory cytokines and frequencies of T cells in the colon tissue. (A) Total RNA was extracted from the colons tissues of untreated and DSS treat mice to analyse the expression of pro-inflammatory cytokines IL17, IFNc and TNFa with qPCR. (B) The frequencies of CD3 + , CD4 + and CD8 + T cells in the isolated colon from mice in the indicated groups were determined by FACS (n = 3 mice/group). All data are shown as the means AE SD. Fig. S4. The maturation of BMDCs generated from WT and CD11c-Cre + Rab32 f/f mice after stimulated with LPS. On Day 7 in culture, BMDCs were stimulated with 1 lgÁmL À1 LPS for 24 h. (A) The surface marker MHC II molecule, CD80 and CD86 were analysed by FACS. (B) Proportions of MHC II hi , CD80 + , and CD86 + BMDCs were calculated. The experiments were repeated 3 times. All data are shown as the means AE SD. Fig. S5. The proportion of neutrophils infiltrated in the colon of the WT and CD11c-Cre + Rab32 f/f mice administered water. The frequencies of neutrophils (CD11b + Ly6G + ) in isolated colonic IEL, LPL and MLNs from mice in the indicated groups were determined by FACS (n = 5-8 mice/group). All data are shown as the means AE SD. Table S1. The primers of the inflammatory response qPCR array. Table S2. Sequences of primers used for qPCR amplification.