Dietary Chitin Particles Called Mimetic Fungi Ameliorate Colitis in Toll-Like Receptor 2/CD14- and Sex-Dependent Manners

Disease progression in DSS-elicited colitis is evident in weight loss and higher clinical scores for diarrhea, bloody stool, and hunching positions; animals given oral CMPs demonstrated improvement in these clinical parameters as well as in colon histological scores (12). We therefore determined whether dietary CMPs produce the most effective anti-IBD activity among several types of chitin preparations. To minimize effects of potential sex-biased gene expression, we used male mice for the initial screening approach. As shown in Fig. 1, we confirmed that oral CMPs protected against the disease-associated weight loss (Fig. 1A) and improved the clinical scores (Fig. 1B). However, disease parameters were not improved by soluble forms of chitin or glucosamine (Fig. 1C and D) or de-CMPs (Fig. 1E and IF). Interestingly, mice given LCBs were protected from acute colitis, as indicated by reduced weight loss and lower clinical scores; in fact, the LCB treatment offered greater protection on these parameters than the CMP treatment (Fig. 1A and B).

We next determined that oral LCBs as well as CMPs reduced the histology scores on day 9, improving the colon inflammation (Fig. 2A). To further assess the colon inflammation improved by LCBs and CMPs, we measured inflammatory mediators, which play key roles in colitis. Tumor necrosis factor alpha (TNF-α) and IL-6, both proinflammatory cytokines in DSS-induced colitis (34), were slightly increased in colitis compared to their levels in the noncolitis group (Fig. 2B). However, neither LCBs nor CMPs reduced these proinflammatory cytokine levels. DSS-elicited colitis reduced gamma interferon (IFN-γ), which, together with TNF-α, induces epithelial barrier dysfunction (35) but also activates anti-inflammatory indoleamine-pyrrole 2,3-dioxygenase 1 (IDO1) and IL-18bp genes in intestinal epithelial cells (12, 36); neither LCBs nor CMPs increased the IFN-γ levels. In contrast, there was no effect of DSS-colitis on IL-1β, which propagates inflammation in DSS-colitis (37) but induces group 3 innate lymphoid cells to produce anti-inflammatory IL-22 (38). Treatment with LCBs, but not CMPs, slightly increased IL-1β levels. Figure 2B further shows that relative to a group without colitis or chitin treatment, colitis mice with vehicle treatment had significantly lower colon IL-10 and IL-22 levels, both of which are anti-inflammatory in DSS-colitis (39). IL-22 is reported not only to heal mucosal inflammation but also to maintain barrier integrity (38). These IL-10 and IL-22 levels were restored to normal levels in LCB-treated mice. However, CMP treatment did not restore the levels of these anti-inflammatory cytokines (Fig. 2B). Chitinase 3-like 1 (CHI3L1), which is a mucosal chitin-binding lectin regulating a chitin-mediated interaction between epithelial cells and microbes (4) and anti-inflammatory in DSS-colitis (40), was increased by LCB and CMP treatments. Overall, our results indicated that anti-inflammatory cytokines, such as IL-10, IL-22, and IL-1β, and chitin-binding CHI3L1, were increased by LCB treatment, whereas proinflammatory cytokines TNF-α and IL-6 were not significantly reduced by LCBs.

We then measured these mediator levels in serum samples harvested on day 9. We detected IL-6 and CHI3L1 in the saline-treated colitis group (see Fig. S1 in the supplemental material). Interestingly, serum IL-6 levels were significantly reduced more than 70% by LCB and CMP treatments (Fig. S1). CHI3L1 levels were not modified by LCBs or CMPs. However, other mediators (TNF-α, IL-1β, IFN-γ, IL-10, and IL-22) were difficult to detect (data not shown); analysis at an earlier or later time point might be necessary for consistent results.

A previous study (4) suggested that dietary CMPs suppress an interaction between colonic mucosa and microbes important in the systemic invasion of local bacteria in DSS-elicited colitis. We confirmed that the total number of bacteria associated with the colon (Fig. 3A) as well as the liver (Fig. 3B) and spleen (Fig. 3C) was significantly reduced in the CMP-treated group. LCB treatment also reduced the number of bacteria translocated into these tissues (Fig. 3A to C).

Previously, we identified that TLR2 and CD14 are essential for CMP-induced classical macrophage (M1) activation of peritoneal macrophages (29). TLR2 KO and CD14 KO mice are known to develop more severe colitis than wild-type (WT) mice in response to DSS treatments, indicating that both TLR2 and CD14 play protective roles (41, 42). Therefore, we used 3%, instead of 4%, DSS and explored whether these chitin-binding proteins were responsible for anti-inflammatory effects of LCBs using male mice deficient in TLR2 and male mice deficient in CD14.

In TLR2 KO mouse colitis, body weight loss was not prevented (Fig. 4A) and the clinical score was not improved by dietary LCBs (Fig. 4B). The saline group in TLR2 KO colitis appeared to express relatively higher levels of colon TNF-α, IFNγ, IL-10, and IL-22 than those in WT colitis (Fig. 4C versus Fig. 2), although it is unclear whether the increased cytokine levels were due to lower DSS concentration and/or TLR2 deficiency. However, it is particularly important that these inflammatory mediator levels were not modified by dietary LCBs in TLR2 KO colitis (Fig. 4C), implying that dietary LCB-induced production of anti-inflammatory cytokines is TLR2 dependent.

Similarly, LCBs did not improve any disease parameters in the CD14 KO colitis. In fact, LCB treatment worsened the disease parameters (Fig. 5A and B), although LCB treatment increased levels of anti-inflammatory IL-10, IL-22, IFN-γ, and IL-1β (Fig. 5C). It is unclear why dietary LCBs worsened CD14 KO colitis. In this regard, Buchheister et al. (42) demonstrated that the intestinal barrier function is not affected by CD14 deficiency but is reduced in CD14 KO colitis. Alternatively, Wang et al. (43) demonstrated that increased inflammatory cytokines, including IFN-γ, induce intestinal epithelial barrier dysfunction by upregulating myosin light-chain kinase expression. Further studies are needed to elucidate how colon cytokines induced by dietary LCBs regulate CD14-mediated tight-junction permeability.

In male colitis, anti-inflammatory activities of chitin preparations addressed above showed that dietary LCBs resulted in better anti-inflammatory effects than CMPs and that the effects were dependent on host CD14 and TLR2. We next examined whether LCBs are more effective than CMPs in treating female colitis. Figure 6 shows that CMPs and LCBs had comparable anti-inflammatory effects, as detected by significantly reducing both body weight loss (Fig. 6A) and clinical scores (Fig. 6B) in colitis in WT female mice. Furthermore, neither CD14 KO (Fig. 6C and D) nor TLR2 KO (Fig. 6E and F) female mice showed any anti-inflammatory effects of LCBs. Taken together, our results indicated that LCBs were fully effective for TLR2/CD14-dependent anti-inflammatory activity in both sexes, whereas CMPs were fully effective in females but less effective in males.

The anti-inflammatory effects of LCBs and CMPs addressed above suggested that the activity of orally administered LCBs and CMPs can be influenced by additional gastrointestine (GI)-specific factors that are possibly sex dependent. We hypothesized that orally administered LCBs and CMPs are digested in the mouse GI tract by chitinases, specifically AMC, and that the resulting reduction in particle sizes facilitates phagocytosis of LCB remnants; stomach AMC activity is sex dependent. To test this hypothesis, we extracted AMC from mouse stomach and colon using a method previously described (44). As shown in Fig. 7A, levels of stomach AMC preparations isolated from saline- and LCB-treated colitis groups were greater in males than in females. Moreover, AMC levels in the colon samples in these two groups were also higher in males than in females (Fig. 7B). Serum AMC levels in the male groups were higher than those in female groups (Fig. 7C). These sex-dependent AMC levels in the stomach, colon, and serum samples were found in mice without colitis or LCB treatment (Fig. 7A to C).

CMPs are known to induce phagocytosis-mediated M1 activation in a TLR2/CD14-dependent manner; LCBs, larger than macrophages and difficult to phagocytose, induce no M1 activation (29). We therefore determined whether AMC-digested LCBs induce M1 and whether CMP-induced M1 activation is reduced by AMC treatment. Figure 7D shows that, as demonstrated previously (29), LCBs induced no TNF-α production by RAW 264.7 cells. AMC digestion increased LCB-induced TNF-α production (Fig. 7D). In contrast, untreated CMPs (buffer control) are known to induce higher production of TNF-α (29); the production of TNF-α by CMPs was slightly but significantly reduced when CMPs were digested by AMC (Fig. 7D). All these effects of AMC were reduced by bisdionin F (BF), an AMC inhibitor. Furthermore, the mechanism inducing TNF-α production by digested LCBs was also dependent on TLR2 and CD14, since peritoneal macrophages isolated from TLR2 KO and CD14 KO mice did not respond to the digested LCBs (Fig. 7E). Taken together, the results suggested that stomach AMC digested dietary chitin particles, which became smaller but were still recognized by cells expressing TLR2 and CD14.

CMPs (1 to 10 μm) were prepared as previously described (29). The quality of CMP preparations was evaluated on the basis of the following criteria: (i) the capacity to induce splenic macrophages in vitro to produce TNF-α and IL-12 but not IL-10 (24–26); (ii) more than 95% of particles had 1- to 10-μm diameters, determined by flow cytometry calibrated by 1-, 10-, and 48-μm latex beads (Polysciences) (24, 27); (iii) no detectable endotoxin (<0.03 endotoxin units/10 mg chitin/ml) determined by the Limulus amebocyte lysate assay; and (iv) minimal surface deacetylation of chitin particles determined by attenuated total reflection Fourier transform infrared (ATR-FTIR) difference spectroscopy (29). LCBs were prepared as a by-product of CMP preparation (24) and met criteria for size, endotoxin content, and surface deacetylation. To prepare de-CMP, CMPs were deacetylated with 1 N sodium hydroxide for 4 h at 37°C, followed by neutralization and washing (29). Oligosaccharide chitin and glucosamine (both from YSK, Yaizu, Japan) were also used for this study.

C57BL/6 WT, TLR2-deficient (KO), and CD14 KO mice were purchased from the Jackson Laboratory (Bar Harbor, ME). These mice were maintained under specific Helicobacter-free conditions at Florida Atlantic University. Animal care and procedures of the experiments in this study were approved by the IACUC at FAU, protocol number A15-34.

For the DSS-induced colitis model, acute colitis was elicited after mouse body weight reached 19 to 21 g (around 8 weeks old) by administration in drinking water of 4% DSS (molecular mass, 35 to 45 kDa; MP Biomedicals, Irvine, CA) for WT mice and 3% DSS for CD14 KO and TLR2 KO mice; after 5 days, they were switched to untreated drinking water. For chitin treatments, groups of 6 mice were given oral gavage with 0.5 ml saline containing 2 mg each chitin preparation listed above daily for 15 days, starting 7 days before DSS treatment. Vehicle control groups were given 0.5 ml saline. Mice were euthanized 9 days after DSS treatment; blood, spleens, livers, and colons were harvested. The colon was cut open longitudinally along the main axis, luminal contents were collected, and the remaining colon was washed with ice-cold saline.

(i) Body weight and clinical score. Colitis mice were weighed and monitored daily for appearance and signs of bloody stools (score of 0 or 1), diarrhea with soft stools (0 or 1), and hunching posture (0 or 1) as clinical parameters previously described (4).

(ii) Histologic score. Histological evaluation of colitis was performed using hematoxylin-and-eosin-stained sections of colon tissue as previously described (4). A total histologic score ranging from 0 to 6 represents the sum of the epithelial nonulcerative region alteration score and the infiltration score. Epithelium nonulcerative region alteration scores were the following: 0, normal morphology; 1, loss of goblet cells in small areas; 2, loss of goblet cells in large areas. Infiltration scores were the following: 0, no infiltration; 1, infiltration around crypt basis; 2, infiltration reaching the lamina muscularis mucosae; 3, extensive infiltration reaching the lamina muscularis mucosae and thickening of the mucosa with abundant edema; 4, infiltration of the lamina submucosa.

(iii) Colon cytokine measurement. Proteins were extracted from colons, and levels of inflammatory mediators IL-1β, IL-6, IL-10, IL-22, IFNγ, TNF-α, and CHI3L1 were measured by specific enzyme-linked immunosorbent assays (ELISAs) (BioLegend, San Diego, CA, and BD, Franklin Lakes, NJ). Protein levels were measured by Pierce protein bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, Rockford, IL).

(iv) Bacterial CFU. Bacterial translocation assay was performed as previously described (4). Tissues were homogenized in Hanks’ balanced salt solution. Serial dilutions of the tissue homogenates were plated on Luria-Bertani (LB) agar plates and incubated at 37°C overnight. Numbers of CFU were counted.

Stomach tissue isolated from normal WT mice was homogenized in 10 volumes (one stomach in 3 ml) of ice-cold Tris buffer (20 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1:500 protease inhibitor cocktail; P8340; Sigma-Aldrich) using a Teflon/glass homogenizer (51). The homogenates were centrifuged at 12,000 rpm for 20 min at 4°C, and the supernatants were filtered through a 0.22-μm filter (Mustang E-membrane; Pall, Nottingham, MD) to remove endotoxin, collected, and kept at −80°C. Mouse AMC and protein levels were measured by ELISA (Nordic Biosite, Plymouth Meeting, PA) and Pierce protein BCA assay.

LCBs at 1 mg/ml in 0.1 M glycine-HCl buffer, pH 2.0, were incubated with stomach AMC at 10 ng/ml. Some groups treated with AMC were further supplemented with BF, an AMC inhibitor (Sigma), at 10 mM. After incubation at 37°C for 60 min, digested LCBs were washed with saline. CMPs were also digested in an identical manner. To measure M1 induction by digested LCBs and CMPs, RAW 264.7 macrophages (5 × 10⁵ cells/ml) were stimulated with digested LCBs and CMPs at selected doses for 18 h at 37°C. Peritoneal macrophages (2.5 × 10⁵ cells/ml), isolated from WT, TLR2 KO, and CD14 KO mice (29), were also stimulated with the digested LCBs and CMPs. TNF-α in the culture supernatants was measured by ELISA (BioLegend).

Statistical significance was evaluated by Student’s t test for parametric data, and results are shown as the means ± standard deviations (SD). Nonparametric data (e.g., clinical data) were analyzed by using the Mann-Whitney U test. Differences were considered statistically significant at a P value of ≤0.05.