ABSTRACT
The symbiotic relationship between the gut microbiome and the host provides a nutrient-rich environment for gut microbes and has beneficial effects on host health. Although the composition of the gut microbiome is known to be influenced by both host genetics and environmental factors, host effects on the activities and functions of the gut microbial communities remain poorly understood. Intestinal epithelial cells exert front-line responses to gut microbes and contribute to maintaining a healthy intestinal homeostasis. Here, seeking to elucidate whether intestinal epithelial cells modulate Lactobacillus rhamnosus GG (LGG) functions, we examined the production of p40, an LGG-derived secretory protein that protects intestinal epithelial cells against inflammation. We found that growth medium conditioned with colonic epithelial cell-derived components promotes p40 protein synthesis and secretion by LGG and enhances LGG-stimulated protective responses in intestinal epithelial cells. Furthermore, when LGG was cultured with the colonic luminal contents from healthy mice, p40 production was upregulated but was attenuated with luminal contents from mice with intestinal inflammation. Importantly, the colonic epithelial cell-derived components potentiated LGG-produced p40 levels in a mouse model of colitis and enhanced LGG-mediated amelioration of intestinal inflammation in this model. Notably, we found that colonic epithelial cell-secreted extracellular vesicles participate in communicating with LGG and that heat shock protein 90 (HSP90) in these vesicles might mediate the promotion of p40 production. These results reveal a previously unrecognized mechanism by which the anti-inflammatory effect of LGG is reinforced by intestinal epithelial cells and thereby maintains intestinal health.
INTRODUCTION
The human gastrointestinal tract harbors a broad range of microbiota with highly diverse composition and redundant metabolic activities. The microbe-host interaction establishes a mutually beneficial system that provides a nutrient-rich environment allowing for microbiota survival and growth. This interaction contributes to maintaining host homeostasis, such as through supporting functions of the gastrointestinal tract and the nervous system, as well as protective immune responses, partially via metabolites and functional factors derived from the microbiota (1, 2). One well-known characteristic of the human microbial community is high interpersonal differences in taxonomic composition (3, 4). Studies in germfree zebrafish and mouse models have shown that the recipient host influences the composition of the transplanted gut microbiota to its native consortium, indicating that factors from the host selectively affect the establishment of the microbial community composition in the host (5). Findings from recent studies have revealed that host genetics and environmental factors, such as diet, nutrient availability, immunological responses, and disease states, shape the composition of the gut microbiota (5, 6). Identifying the effects of host factors on the functions of the gut microbiota under normal and disease conditions is currently an interesting research area.
Intestinal epithelial cells along the mucosal surface exert front-line responses to the gut microbiota and contribute to the maintenance of the symbiotic relationship between the gut microbiota and the host (7). Increasing evidence indicates that extracellular vesicles (EVs) secreted by both the apical and basolateral surfaces of intestinal epithelial cells are important intercellular messengers for maintaining intestinal homeostasis (8, 9). Major histocompatibility complex class II molecules in intestinal epithelial cell-secreted EVs are an important mediator of communication between intestinal epithelial cells and dendritic cells for antigen presentation (10). Annexin A1-containing EVs secreted by intestinal epithelial cells play roles in colonic wound repair (11). Furthermore, intestinal epithelial cell-secreted EVs have been shown to exert antibacterial effects (12, 13). However, the involvement of host-derived EVs in directly regulating the microbe-host relationship of mutualism remains poorly defined.
EVs are composed of complex cargoes, including transmembrane and cytosolic proteins, lipids, and nucleic acids (14). Moreover, EVs are important messengers for intercellular and interorganismal communication, modulating cell motility and polarization as well as immune responses (15). Although the mechanisms of EV biogenesis are not well understood, EVs have been reported to be secreted through multivesicular bodies in the endosomal pathway or through budding off the plasma membrane (16). Studies have identified the Rab family of small GTPases as critical docking factors for multivesicular bodies. For example, Rab27a and Rab27b play roles in exosome secretion without influencing the secretion of soluble proteins (17). Moreover, Rab27-dependent exosome production mediates the maintenance of immunological homeostasis against inflammatory stimuli (18).
We aimed to pursue mechanistic studies to explore the effects of intestinal epithelial cells on regulating microbiota functions under physiological conditions as well as the impact of intestinal inflammation on the mutual relationship between the gut microbiota and the host. We previously cloned and characterized a Lactobacillus rhamnosus GG (LGG)-derived secretory protein, p40. p40 preserves barrier function, inhibits cytokine-induced apoptosis, and upregulates mucin production in intestinal epithelial cells; in addition, it stimulates expression of a proliferation-inducing ligand in intestinal epithelial cells, resulting in IgA production and consequently preventing and ameliorating experimental colitis in mice. p40 exerts these effects through transactivation of the epidermal growth factor receptor (EGFR) in intestinal epithelial cells (19–23). In this work, we utilized production of p40 by LGG as a model to investigate the mutual relationship between intestinal epithelial cells and LGG. Results from this work demonstrate that intestinal epithelial cell-derived EVs reinforce p40 production by LGG, thereby enhancing the protective effect of LGG against colitis. This finding provides novel mechanistic insights into the microbe-host interaction of mutualism.
RESULTS
Intestinal epithelial cell-derived components promote p40 production and the beneficial effects of LGG.In this study, we examined the level of p40 produced by LGG and the degree of LGG-modulated cellular responses to evaluate the effects of intestinal epithelial cells on LGG function. In addition, we evaluated another LGG-derived protein, p75, which is shed from LGG. p75 has less of an effect than p40 on regulating intestinal epithelial cellular responses (22). The gastrointestinal epithelial cells utilized in this study included young adult mouse colonic epithelial cells (YAMC), mouse small intestinal epithelial (MSIE) cells, and immortalized stomach epithelial (ImSt) cells, which were generated from the immortomouse strain (24). We cocultured gastrointestinal epithelial cells in fetal bovine serum (FBS)-free RPMI medium (106 cells in 5 ml medium/plate) with LGG (multiplicity of infection [MOI] of 100:1) for 3 h at 37°C. Simultaneously, conditioned medium (CM) was prepared from a 3-h culture of these three cell lines in FBS-free RPMI medium (106 cells in 5 ml medium/plate). CM was used to culture the same amount of LGG as that for culture with cells. FBS-free medium was used to maintain cells in a basal state. Levels of p40 and p75 in culture supernatants were analyzed by Western blotting using anti-p40 and anti-p75 antibodies (22). Compared with LGG cultured in FBS-free RPMI medium, coculture of LGG with these three cell lines increased the levels of p40 but not p75 in supernatants, and we observed greater stimulation of p40 production in YAMC and MSIE cells than in ImSt cells (Fig. 1A). Because not all bacteria in the gastrointestinal tract directly interact with intestinal epithelial cells, we asked whether epithelial cell-derived components in CM affect p40 production by LGG. CM from these three cell lines, compared with RPMI, upregulated the level of p40, but not p75, in supernatants. YAMC-CM had a greater effect on p40 production than MSIE-CM or ImSt-CM (Fig. 1A). Thus, colonic epithelial cell-derived components have a greater effect on p40 production than those derived from stomach and small intestinal epithelial cells.
Intestinal epithelial cell-derived components promote p40 production by LGG. (A) LGG (2 × 107 CFU/ml) was cultured at 37°C for 3 h with ImSt, MSIE, and YAMC cells (106 cells/plate) in 5 ml of serum-free RPMI (left) or in CM prepared from a 3-h culture of the indicated cells (106 cells/plate in 5 ml of serum-free RPMI) (right). LGG cultured in serum-free RPMI at 37°C for 3 h was used as a control. (B and C) LGG was cultured in CM, which was prepared by culture of YAMC in serum-free RPMI for the indicated times. (D and E) Intestinal flush samples from the small intestine (SI) and the colon (Col) of 8- to 10-week-old WT and Rag2–/– mice were incubated with LGG (2 × 107 CFU/ml) for 3 h at 37°C. p40 and p75 levels in culture supernatants (A, B, D, and E) and in LGG (C) were examined by Western blot analysis. Images in panels A to C are representative of at least 5 independent experiments. Images in panels D and E are representative of 5 WT and 5 Rag2–/– mice.
The effect of YAMC-CM on promoting p40 production in supernatants was positively related to the YAMC culture time for generating CM (Fig. 1B), indicating that more components that stimulate p40 production are present in CM after longer culture of intestinal epithelial cells. These results were further supported by the observation of higher p40 levels in LGG cultured in YAMC-CM prepared from YAMC cultured for 24 h rather than <24 h (Fig. 1C). Interestingly, YAMC-CM prepared from YAMC cultured for 3 to 24 h stimulated p75 levels in LGG (Fig. 1C). These results suggest that components of intestinal epithelial cells promote protein synthesis of p40 and p75 and secretion of p40 but not shedding of p75.
Notably, the intestinal epithelial cell culture conditions were maintained with a cell survival rate higher than 90% to exclude the influence of apoptotic and necrotic fragments on the function of CM. In addition, we assessed p40 and p75 production by LGG from 3-h cultures because there was no difference in growth rate between LGG cultured in CM versus RPMI for 3 h; thus, we concluded that the p40 and p75 protein levels under different culture conditions were from comparable numbers of LGG.
To provide in vivo evidence that intestinal epithelial cells regulate microbiota function, we isolated the colonic contents by flushing the small intestine and the colon of wild-type (WT) and Rag2 knockout (Rag2–/–) mice, which have no mature B cells or T cells. In agreement with the result that CM from colonic epithelial cells had a greater effect on p40 production by LGG than did CM from small intestinal epithelial cells (Fig. 1A), colonic mucosa contents stimulated greater p40 production than did small intestinal contents (Fig. 1D). Colonic flush samples from WT and Rag2–/– mice stimulated p40 production by LGG to a similar degree (Fig. 1E), suggesting that B or T lymphocytes in the intestinal tract are not sources of factors promoting p40 production by LGG.
Because p40 has been shown to stimulate protective cellular responses through activation of EGFR in intestinal epithelial cells (21, 23), we asked whether YAMC-CM treatment increases the effect of LGG on maintaining intestinal epithelial homeostasis. LGG was cultured in YAMC-CM prepared from 24-h YAMC cultures (YAMC-CM-LGG) and in FBS-free RPMI (RPMI-LGG) for 3 h at 37°C. Compared with LGG and RPMI-LGG, YAMC-CM-LGG stimulated higher levels of EGFR activation in YAMC (Fig. 2A). In agreement with this result, YAMC-CM-LGG showed higher levels of activity in preventing cytokine-induced apoptosis in HT-29 cells, as detected by Western blot analysis of a caspase-3 cleavage product (Fig. 2B). H2O2 treatment induced the disruption of tight junctions in T84 cells on the basis of the intracellular localization of the tight-junction protein ZO-1. Immunostaining showed that YAMC-CM-LGG treatment resulted in greater maintenance of ZO-1 membrane localization in H2O2-treated T84 cells (Fig. 2C). These results suggest that colonic epithelial cell-derived components promote the function of LGG in protecting intestinal epithelial cells.
Intestinal epithelial cell-derived components promote the protective effects of LGG on intestinal epithelial cells. YAMC-CM was prepared from 24-h cell culture. LGG was cultured in RPMI (RPMI-LGG) or YAMC-CM (YAMC-CM-LGG) for 3 h at 37°C. (A) YAMC were treated with LGG, RPMI-LGG, or YAMC-CM-LGG at 107 CFU/ml for 1 h. (B) HT29 cells were treated with TNF (100 ng/ml), IL-1α (10 ng/ml), and IFN-γ (100 ng/ml) for 6 h in the presence or absence of LGG, RPMI-LGG, or YAMC-CM-LGG. Cellular lysates were collected for Western blot analysis of levels of phosphorylated ERGF (P-EGFR) and total EGFR (T-EGFR) (A) and cleavage caspase-3 (B). β-Actin was used as a protein loading control. The relative density was determined by normalization of the band density of P-EGFR to that of T-EGFR and cleavage caspase-3 to that of β-actin in the same sample. The fold change (shown under P-EGFR and cleavage caspase-3 blots) was calculated by comparison of the relative density in each sample to that of P-EGFR in nontreated YAMC (A) and cleavage caspase-3 in HT29 cells treated with cytokines only (B). (C) T84 cells were treated with H2O2 (20 μM) for 3 h in the presence or absence of RPMI-LGG or YAMC-CM-LGG. The cotreatment was present during H2O2 treatment. Cells were fixed for immunostaining of ZO-1. Data are representative of at least 3 independent experiments.
Perturbation of the symbiotic relationship between the host and the intestinal microbiota is associated with several diseases in humans, such as inflammatory bowel disease (IBD) (25, 26). We have demonstrated that production of p40 mediates LGG-regulated protective responses in intestinal cells and aids in prevention of colitis (21, 22). We further investigated whether treatment of LGG with intestinal epithelial cell-derived components might enhance the effect of LGG on preventing colitis in mice.
First, we examined whether YAMC-CM treatment of LGG affects p40 production in vivo. Mice were administered YAMC-CM-LGG or RPMI-LGG by gavage. The p40 level in the colonic mucosa 24 h after gavage was determined by Western blot analysis of colonic mucosal lysates. Compared with RPMI-LGG-treated mice, mice with YAMC-CM-LGG treatment had elevated p40 in the colonic mucosa (Fig. 3A), suggesting that colonic epithelial cell-derived components promote p40 production by LGG in vivo.
Production of p40 by LGG in mice under normal and colitis conditions. (A) Mice were administered 108 CFU YAMC-CM-LGG or RPMI-LGG once by gavage. Colonic mucosal lysates were collected at 24 h after LGG treatment for Western blot analysis of p40. p40 relative density was determined by normalization of the density of the p40 band to that of the β-actin band for the same mouse. The average relative density of untreated mice was set at 1. The fold change was calculated by comparison of the p40 relative density in each mouse to this average. (B) WT mice were treated with 3% DSS in drinking water for 4 or 7 days. Injury/inflammation scores are shown. (C) Colonic contents were collected from mice treated with DSS for 7 days and were incubated with LGG for 3 h at 37°C. The p40 level in the supernatants was examined. The average density of p40 in the water group was set as 1. The fold change was calculated by comparison of the p40 band density in each mouse to this average. (D) Mice were treated with DSS for 4 days and then administered 108 CFU LGG through gavage. Feces were collected on the indicated days. (E) p40 levels in feces were examined by Western blot analysis. The fold change of p40 in each mouse was calculated by comparing the p40 band density after treatment (day 2 in control and days 4 and 5 in DSS groups) to that before treatment of the same mouse (day 0). (F) For quantitative analysis of the amount of LGG in mouse feces, DNA was isolated from feces for qPCR analysis using LGG-specific primers. The expression of LGG-specific genes from each mouse was compared to the feces-based LGG concentration curve to obtain the amount of LGG in feces. Data were analyzed with two-tailed t tests and are shown as means ± SD. n.s., not significant. All mice were 8 to 10 weeks old. Each symbol represents one mouse.
Second, we investigated whether the p40 level is altered in intestinal inflammation. The dextran sulfate sodium (DSS) mouse model of acute colitis is characterized by increased epithelial injury and the production of inflammatory cytokines (27). DSS induces intestinal injury and inflammation in mice (Fig. 3B). The colonic contents were isolated from mice treated with DSS and then incubated with LGG. The colonic contents from mice with colitis exerted less of an effect on stimulating p40 production by LGG (Fig. 3C).
Third, we evaluated the effects of intestinal inflammation on p40 production by LGG in mice. After colitis was induced, the mice were subjected to LGG gavage (Fig. 3D). LGG gavage resulted in less p40 in mice with colitis than in mice under normal conditions (Fig. 3E). The decreased p40 level in colitis may have been due to dysbiosis, which might have decreased the numbers of bacteria producing p40. Host factors that promote p40 production might also be altered in colitis. Therefore, we further examined the amount of LGG in the feces by quantitative PCR (qPCR) analysis using LGG-specific primers, as previously reported (28). The results of the qPCR analysis showed no differences in the amounts of LGG in feces from mice receiving LGG gavage under the normal and the colitis conditions (Fig. 3F). Our previous studies found that no LGG was detected in feces of mice maintained in our animal facility using this qPCR method (29). In the current experiment, mice with and without colitis received the same amounts of LGG and exhibited the same amounts of LGG in feces 24 h after gavage, indicating that there is no significant difference in the amount of LGG in these two groups of mice. These results suggest that intestinal inflammation affects p40 production by LGG in mice.
Given our findings that p40 production by LGG was promoted by YAMC-CM treatment in vitro and in mice, and that p40 levels decreased under inflammatory conditions, we asked whether treatment of LGG with YAMC-CM has physiological consequences on intestinal inflammation. YAMC-CM-LGG and RPMI-LGG were administered to mice with DSS-induced colitis (Fig. 4A). The body weight loss in the DSS group on the first and second days after discontinuation of DSS treatment was decreased by YAMC-CM-LGG treatment but not by RPMI-LGG treatment (Fig. 4B). On the second day of recovery after DSS treatment, mice receiving DSS and cotreatment of DSS and RPMI-LGG showed injury and colitis with massive colon ulceration, crypt damage, and severe inflammation. These abnormalities were less pronounced in mice with cotreatment with YAMC-CM-LGG (Fig. 4C). The injury/inflammation scores in the DSS group (26.2 ± 7.72) and RPMI-LGG cotreatment group (24.8 ± 5.02) were significantly decreased by YAMC-CM-LGG cotreatment (16.2 ± 5.61, P < 0.05) (Fig. 4D). Furthermore, increased proinflammatory cytokine production is a hallmark of DSS-induced colitis (27). In our experiments, mRNA levels of tumor necrosis factor (TNF) and keratinocyte chemoattractant (KC; also known as chemokine C-X-C motif ligand, or CXCL1) were increased in the DSS and the DSS and RPMI-LGG cotreatment groups but were significantly decreased in mice with YAMC-CM-LGG cotreatment (Fig. 4E).
YAMC-CM-treated LGG ameliorates DSS-induced intestinal injury and inflammation in mice. (A) Three percent DSS in drinking water was administered to mice for 4 days to induce colonic injury and colitis. Mice were gavaged with YAMC-CM-LGG or RPMI-LGG LGG at 108 CFU/day as indicated. (B) Body weight was recorded. The percentage of body weight compared to that at day 1 of the same mouse is shown. (C and D) Colon sections were stained with H&E for light microscopic assessment of epithelial damage and inflammation. The inflammation/injury scores are shown. (E) mRNA was isolated from the colonic tissues for real-time PCR analysis of expression levels of the indicated cytokines. The average cytokine mRNA expression level in the water group was set as 100%, and the mRNA expression level of each mouse was compared to this average. Data were analyzed using one-way ANOVA. All data are presented as means ± SD. *, P < 0.05 compared to the water group. #, P < 0.05 compared to the DSS-treated mice. Each symbol represents one mouse.
The effects of treatment of LGG with YAMC-CM on prevention of colitis was further examined by using a 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced, interleukin-12 (IL-12)-driven Th1 immune response-mediated colitis model (30). We administered YAMC-CM-LGG and RPMI-LGG to mice before induction of colitis (Fig. 5A). TNBS induced histological changes, including disruption of the epithelial monolayer and inflammatory cell infiltration: the inflammation score was 4.29 ± 1.50 in the untreated group, 4.40 ± 1.52 in the RPMI-LGG-treated group, and 2.00 ± 1.41 in the YAMC-CM-LGG-treated group (P = 0.024 and P = 0.032 compared with no-LGG and RPMI-LGG groups, respectively) (Fig. 5B and C). The TNBS-mediated increase in TNF and gamma interferon (IFN-γ) expression in the colonic mucosa was prevented by YAMC-CM-LGG but not RPMI-LGG treatment (P < 0.05) (Fig. 5D).
YAMC-CM treatment enhances the effect of LGG on ameliorating TNBS-induced colitis in mice. (A) Colitis was induced by TNBS. Ethanol treatment was used as a control. Mice were gavaged with YAMC-CM-LGG or RPMI-LGG LGG at 108 CFU/day as indicated. (B and C) Colon sections were stained with H&E for light microscopic assessment of inflammation. The inflammation/injury scores are shown. (D) RNA was isolated from the colonic tissues for real-time PCR analysis of the indicated cytokine mRNA expression levels. The average cytokine mRNA expression level in the no-LGG group with ethanol treatment was set as 1, and the mRNA expression level of each mouse was compared to this average. Data were analyzed using one-way ANOVA. All data are presented as means ± SD. *, P < 0.05 compared to the no-LGG group with ethanol treatment. #, P < 0.05 compared to the RPMI-LGG group with TNBS treatment. Each symbol represents one mouse.
Collectively, these results indicate that treatment of LGG with intestinal epithelial cell-derived components increases p40 production and the protective effects of LGG on intestinal epithelial cells and aids in prevention of colitis. Thus, host factors from intestinal epithelial cells are involved in maintaining the mutual relationship between the gut microbiota and the host.
EVs secreted from intestinal epithelial cells mediate the effects of intestinal epithelial cells on promoting p40 production by LGG.We next investigated how YAMC communicate with LGG in delivering components that stimulate p40 production. Gastrointestinal epithelial cells have been found to secrete EVs (8, 9). We asked whether YAMC-secreted EVs mediate the communication between YAMC and LGG. Neutral sphingomyelinase is an enzyme involved in biosynthesis of ceramide, which mediates secretion of EVs. GW4869 is a drug that inhibits ceramide synthesis by blocking neutral sphingomyelinase activity and consequently decreases exosome secretion (31, 32). Thus, we utilized GW4869 to examine the involvement of intestinal epithelial cell-secreted EVs in the regulation of p40 production by LGG. GW4869-treated YAMC and YAMC-CM prepared from GW4869-treated YAMC showed decreased effects on the stimulation of p40 but not p75 in supernatants in a concentration-dependent manner (Fig. 6A). Furthermore, p40 and p75 production was not affected when GW4869 was added to CM (Fig. 6B), suggesting that GW4869 had no direct effect on LGG regarding p40 production.
Intestinal epithelial cell-secreted EVs mediate epithelial cell and LGG communication and promote p40 production by LGG. (A) To prepare CM, YAMC (106 cells/plate) were cultured in 5 ml of FBS-free RPMI for 24 h in the absence (Cont) and presence of GW4869 at the indicated concentrations (GW-CM) and dimethyl sulfoxide (DMSO-CM, 1:2,000 dilution, which is the same dilution rate as that for 10 μM GW4869). Cell culture medium was collected and centrifuged. YAMC-CM and YAMC in 5 ml of FBS-free RPMI with the indicated treatment were cultured with LGG (2 × 107 CFU/ml) at 37°C for 3 h. (B) LGG was cultured in Cont-CM supplemented with GW4869 (10 μM). (C) The expression level of Rab27a in YAMC infected with shRNA Rab27a or nontargeting shRNA was detected by Western blot analysis. Data from two clones of each infection are shown. LGG was cultured with CM from the indicated cells. (D) The method used to isolate EVs. (E) TEM analysis of EVs isolated from YAMC-CM. (F) The size of EVs was analyzed by a nanoparticle tracking system. (G) LGG was cultured in RPMI with EVs containing proteins labeled by green fluorescent dye at 37°C for 2 h. LGG was fixed for immunostaining of p40 (red). DNA was labeled with DAPI (blue). (H) LGG was cultured at 37°C for 3 h in 5 ml of YAMC-CM and in RPMI with EVs secreted from 5 ml of YAMC-CM. p40 and p75 production in culture supernatants was examined using Western blot analysis. Data are representative of at least 3 separate experiments.
Rab27a has been shown to regulate exosome secretion (17). We infected YAMC with lentiviral short hairpin RNA (shRNA) against Rab27a to generate a stable cell line with decreased Rab27a expression (Fig. 6C). CM from YAMC with decreased expression of Rab27a had decreased effects on p40 production (Fig. 6C). Together, these data suggest that EVs secreted from intestinal epithelial cells play roles in the regulation of p40 production by LGG.
We isolated EVs from YAMC-CM by successive centrifugation (Fig. 6D) and analyzed EVs through transmission electron microscopy (TEM) and nanoparticle tracking analysis. The sizes of EVs were 30 to 150 nm in diameter (Fig. 6E and F). To examine the interaction between EVs and LGG, EVs with proteins labeled by green fluorescent dye were cultured with LGG for 2 h. The proteins in EVs were found to be transferred to LGG (Fig. 6G), suggesting that EVs are involved in communication with LGG. Notably, EVs secreted by YAMC significantly stimulated p40 production (Fig. 6H).
We further analyzed EVs from colonic and small intestinal contents by using a nanoparticle tracking system. EVs from colonic flush samples were larger, but lower in density, than those from small intestinal contents (see Fig. S1A to C in the supplemental material). Fluorescently labeled proteins in EVs from colonic contents were transferred to LGG (Fig. S1D). EVs isolated from colonic contents had a greater effect on p40 production by LGG than small intestinal contents (Fig. S1E). These results suggested that biogenesis of EVs from colonic and small intestinal epithelial cells differs and potentially leads to different functions in the regulation of p40 production by LGG.
Together, these results suggest that EVs secreted by intestinal epithelial cells participate in communication with LGG and promote the production of p40 by LGG.
HSP90 in intestinal epithelial cell-derived EVs promotes p40 production by LGG.A series of studies were applied to identify components in intestinal epithelial cell-secreted EVs that play roles in promoting p40 production. First, we found that heat treatment and trypsin digestion abolished the effects of YAMC-CM on p40 production (Fig. S2). Because YAMC-released EVs participate in regulation of p40 production (Fig. 6), these results suggest that proteins in EVs contribute to the regulation of p40 production. Second, the protein contents in YAMC-released EVs were analyzed by liquid chromatography-tandem mass spectrometry (Table S1). iTRAQ-based quantitative proteomic analysis was performed to identify YAMC-derived proteins that were transferred into LGG during culture in YAMC-CM. Biological process analysis indicated that these proteins were involved in maintaining cell structure, motility, integrity, and adhesion, regulating cell signaling and transcription, and serving as molecular chaperones, extracellular matrix, and enzymes (Table S2). All YAMC-derived proteins transferred to LGG that were identified by iTRAQ analysis were included in EVs (Table S1). Based on this functional analysis, two categories of proteins in EVs have potential functions in regulation of protein production in LGG, which are chaperones and enzymes. Third, we searched which proteins identified in YAMC-derived EVs have bacterial homologs and function in protein production in bacteria. We identified significant candidate proteins in EVs, which are molecular chaperones, especially heat shock protein (HSP90α) and HSP90β (33, 34). Thus, this work was focused on elucidating whether HSP90 in EVs contributes to promoting p40 production by LGG.
HSP90 is secreted by cells in EVs (35) and functions as an ATPase-driven molecular chaperone modulating the folding, stability, transportation, and degradation of client proteins (36). Most bacteria have one homolog of HSP90, known as high-temperature protein G. HSP90 from bacteria hydrolyze ATP and undergo conformational changes similar to those from mammalian cells (37). The localization of HSP90 in EVs was determined by dot blotting of EVs in the presence and absence of permeabilization by Tween 20. HSP90 was found in EVs without permeabilization, indicating that HSP90 is present on the surfaces of EVs (Fig. 7A). The contribution of HSP90 in EVs to the regulation of p40 production was further studied. Geldanamycin inhibits the function of HSP90 by binding to the ADP/ATP-binding pocket. To avoid direct interaction of geldanamycin with LGG, we pretreated YAMC with geldanamycin overnight and then removed the geldanamycin-containing medium and added new medium. Cells were cultured for another 3 h and used to prepare CM. CM from geldanamycin-treated YAMC did not stimulate p40 production (Fig. 7B). Furthermore, geldanamycin added to CM from control cells also decreased p40 production (Fig. 7B). Remarkably, geldanamycin inhibited EV-stimulated p40 production (Fig. 7C). Geldanamycin does not affect the effects of endogenous chaperones in LGG, because p40 production was not affected when LGG was cultured in MRS in the presence of geldanamycin (Fig. 7D). These data suggest that inhibition of the activity of HSP90 in EVs blocks the effect of EVs on p40 production. The direct effect of HSP90 on regulating the function of LGG was further studied. Recombinant human HSP90α and HSP90β were cultured with LGG in RPMI. Both HSP90α and HSP90β promoted p40 production (Fig. 7E). Together, these results indicate that HSP90 mediates EV-upregulated p40 production by LGG.
HSP90 in intestinal epithelial cell-secreted EVs mediates upregulation of p40 production. (A) EVs were isolated from YAMC-CM. Indicated numbers of EVs were dotted on nitrocellular membrane and blotted using an anti-HSP90α/β antibody in the presence (+) and absence (−) of 0.1% (vol/vol) Tween 20 (detergent). (B) CM was prepared from YAMC treated with geldanamycin (GDM-CM) or from untreated cells (Cont-CM). LGG was cultured in GDM-CM or in Cont-CM with the indicated concentrations of GDM. (C) LGG was cultured with EVs in RPMI with the indicated concentrations of GDM. (D) LGG was cultured in MRS in the presence or absence of GDM (10 μM). (E) LGG was cultured in RPMI with the indicated concentrations of human recombinant HSP90α and HSP90β. LGG (2 × 107 CFU/ml) was cultured under the conditions used for panels B to E at 37°C for 3 h. p40 production in supernatants was examined using Western blot analysis. Data represent at least 3 independent experiments.
DISCUSSION
The intestinal epithelium is an integral component in mucosal defense for maintaining intestinal homeostasis. Although previous studies suggest that bacterial functions are altered in the gastrointestinal tract—for example, genes of L. plantarum that are related to metabolism and stress functions are induced in the gastrointestinal tract (38)—knowledge of the specific effects of the host on the function of the microbiota is limited. Findings from this study reveal a previously unrecognized function of intestinal epithelial cells in promoting production of p40, a functional factor of LGG that protects the integrity of the intestinal epithelium and prevents intestinal inflammation. In addition to LGG, we reported that p40 is produced by other Lactobacillus strains, such as L. casei 393 (22). p40 production by L. casei 393 was also promoted by YAMC-CM (see Fig. S3 in the supplemental material). These results support a mutual relationship between the host and the gut microbiota. We found that the protein level of p75 was increased in LGG cultured with YAMC-CM. However, the level of p75 in LGG culture supernatants was not upregulated by CM. p40 is a secretory protein, and p75 is shed from LGG. Therefore, CM might exert several roles in protein production by LGG, protein synthesis, and secretion. Furthermore, higher levels of p40 were found in the LGG culture supernatants, but not inside LGG, under stimulation with YAMC-CM prepared from 3-h and 6-h YAMC cultures, suggesting that most p40 was secreted into the supernatant. Although YAMC-CM prepared from 24-h YAMC culture had greater effects on p40 synthesis and secretion, the degree of YAMC-CM-stimulated p40 synthesis might be higher than that of secretion; therefore, p40 accumulates inside LGG. We also performed qPCR analysis for measuring transcriptional regulation of p40 in LGG by YAMC-CM. YAMC-CM did not affect p40 gene expression in LGG (Fig. S4).
We found that HSP90 in EVs mediates intestinal epithelial cell-stimulated production of p40 by LGG. Our studies reveal a novel role of intestinal epithelial cell-derived HSP90 in EVs in regulating LGG function and provide a preliminary framework for understanding how the symbiotic relationship of mutualism between gut microbiota and the host is established through HSP90. Our study further suggests that HSP90 is on the surface of YAMC-released EVs. It is known that different isoforms of HSP90 are localized in different cellular compartments. Both HSP90α and HSP90β have been characterized as abundant proteins in the cytosol and nucleus, whereas the other isoform, Grp94, is in the endoplasmic reticulum (39). However, several groups reported that a pool of HSP90 is localized on the surface of cells to exert unique properties. For example, cell membrane-associated HSP90 in macrophages and dendritic cells (40) and in monocytes (41) plays roles in immune responses. HSP90 on the cell surface of fibroblasts (42) and neuronal cells (43) promotes migration. Therefore, we anticipate that HSP90 can be associated with the membrane of EVs released from YAMC. Notably, other components in EVs, including nuclear acids and lipids, may play roles in regulating p40 production. Therefore, our future studies will aim to strengthen the characterization of the composition of epithelium-derived EVs and to define other components that are responsible for the regulatory effects on the microbiota.
It is known that both the endosomal pathway (32) and budding from the plasma membrane (44) are involved in release of EVs with different sizes. Different cell types may produce vesicles with distinct contents of proteins, lipids, mRNA, DNA, noncoding RNA, and surface characteristics. These repertoires of vesicle may also be affected by the physiological and pathological status of cells and the isolation approaches (15, 45). The protein contents of EVs released by YAMC identified in our study show high similarity to those in EVs released from cultured intestinal epithelial cells in the published paper (9). Therefore, we predict that protein contents in EVs identified in our study should represent the characteristic protein profile in intestinal epithelial cell-derived EVs under normal conditions.
Bacterial composition varies both qualitatively and quantitatively between the stomach and the colon, and most commensal bacteria are harbored in the colon (46). This microbial homeostasis may be related to motility and the secretion of acid and digestive enzymes in different portions of the gastrointestinal tract. Our studies demonstrate that components derived from the colonic epithelial cells have more potent effects on regulating LGG function than do those from small intestinal and gastric epithelial cells. This evidence supports the physiological relevance of the colon as a suitable ecological niche to support the function of the gut microbiota.
It is important to understand the mechanisms underlying regulation of protein synthesis, translocation, degradation, and secretion in LGG by factors derived from intestinal epithelial cells. Changes in protein abundance in LGG regulated by intestinal epithelial cells were identified by iTRAQ protein quantification analysis (Table S3). We found that YAMC-CM upregulated p40 and other LGG proteins that mediate several functions of LGG, including bacterial transporter function, purine metabolism, amino acid synthesis, protein synthesis and secretion, enzyme activity, and glycolysis (Table S4). Our future studies will focus on identifying how these pathways in LGG regulated by CM contribute to p40 synthesis and secretion.
Dysbiosis of the gut microbiota is associated with several diseases in humans, including IBD (25, 26). Therefore, strategies for manipulating the microbiota for disease prevention and treatment have clinical applications. To develop microbiota-targeted therapies, an understanding of the mechanisms underlying the mutual reinforcement of host-microbiota interactions is crucial. A key point from this study is that intestinal epithelial cell-secreted EVs mediate the regulatory effects of intestinal epithelial cells on LGG. EVs have previously been reported to mediate communication between the microbiota and intestinal epithelial cells. For example, Cryptosporidium parvum stimulates intestinal epithelial cells to produce exosomes and also stimulates luminal secretion through the activation of TLR4 signaling. These exosomes directly bind to the C. parvum surface and kill bacteria through antimicrobial peptides, such as beta-defensin 2 and cathelicidin-37 (12). In addition, enterocyte microvillus-derived vesicles contain alkaline phosphatase and cluster on the surface of native luminal bacteria, where they prevent the adherence of enteropathogenic E. coli to epithelial monolayers and limit bacterial population growth (13). Therefore, EVs secreted by intestinal epithelial cells might interact with the gut microbiota.
This study reveals that p40 is downregulated in the colonic mucosa under inflammatory conditions. The colonic contents from mice with colitis induced by DSS and TNBS (Fig. S5) showed less effect on stimulating p40 production by LGG. The fecal abundance of LGG detected by qPCR analysis was comparable in control and DSS-treated mice after LGG gavage. These results suggest that the growth of LGG is not affected by the inflammatory conditions in this mouse model. The low p40 levels in mice with colitis may be due to a decreased ability of LGG to produce p40. However, it should be noted that the amount of DNA detected by qPCR analysis could be from live and dead bacteria. Further analysis, such as the propidium monoazide (PMA) treatment method, might increase the efficacy of PCR amplification of DNA from viable bacteria (47). Stress has been reported to modulate protein synthesis in LGG. Proteomic and transcriptomic analyses demonstrated that bile stress induces changes in protein abundance in LGG, including stress responses and cell envelope-related functions, such as fatty acid composition, cell surface charge, and the thickness of the exopolysaccharide layer. Notably, p40 gene expression and protein synthesis are inhibited by bile acids (48). Stress responses in the microbiota may contribute to downregulation of p40 under inflammation. Another possibility is that intestinal inflammation alters the biogenesis of EVs in intestinal epithelial cells or lymphocytes. EVs are shed from sites of intestinal inflammation in patients with IBD, who have mRNA and protein profiles distinct from those of healthy individuals. These EVs have proinflammatory effects on the colonic epithelium (49). Another report has shown that serum exosomes from DSS-treated mice contain 56 differentially expressed proteins, including acute-phase proteins and immunoglobulins, and trigger a proinflammatory response in macrophages in vitro (50). This evidence indicates that the composition of EVs is altered under inflammation and supports the hypothesis that alteration of the composition of EVs generated under the inflammatory condition abolishes the regulatory effect of intestinal epithelial cells on p40 production.
The evidence of decreased p40 levels in colitis supports a rationale for the therapeutic potential of treating IBD by increasing the function of LGG via intestinal epithelial cell-secreted EVs. To increase the regulatory effects of EVs on LGG functions, an efficient strategy is encapsulation of LGG with EVs. Encapsulation is a technology for preserving probiotic viability and improving mucoadhesion of probiotics to the intestinal tract under adverse environmental conditions (51). Our future work will utilize citrus pectin LC950 with an esterification system developed by our group (52) and galacturonic acid by another group (53) to encapsulate LGG with EVs to improve the efficacy of LGG treatment by EVs for the prevention and treatment of intestinal inflammation.
In summary, our findings reveal a previously unrecognized role of epithelial cell-secreted EVs in establishing the mutual relationship between intestinal epithelial cells and the microbiota, thereby reinforcing the microbiota functions for benefiting the host. This knowledge provides mechanistic insights supporting the development of a microbiota-based therapy for maintenance of intestinal health and prevention of inflammatory disorders.
MATERIALS AND METHODS
Cell culture.ImSt, MSIE, and YAMC cell lines were generated from the stomach, small intestine, and colon from 6- to 8-week-old C57BL/6J mice harboring a thermolabile mutation (tsA58) under the control of an IFN-γ-inducible H-2Kb promoter and a temperature-sensitive simian virus 40 (SV40) large T antigen (immortomouse) (24). The functional expression of the SV40 large T antigen was induced by culturing cells in vitro in RPMI 1640 medium containing IFN-γ at 33°C, a permissive temperature for the function of the tsA58 mutation. Expression of this gene is required for cell proliferation. Cells do not proliferate when the temperature is raised to a nonpermissive temperature (37°C) or in the absence of IFN-γ. These cells have been characterized on the basis of the expression of keratin, an epithelial cell marker, and the production of brush border peptidases and a disaccharidase (24). Cells were sorted using an antibody to E-cadherin, an epithelial cell marker.
ImSt cells, MSIE cells, and YAMC were maintained in RPMI 1640 medium supplemented with 10% FBS, 5 U/ml of mouse IFN-γ, 100 U/ml penicillin and streptomycin, 5 μg/ml insulin, 5 μg/ml transferrin, and 5 ng/ml selenous acid at 33°C (permissive temperature) under 5% CO2.
Lentiviral infection.The mouse shRNA Rab27a (5′-GTGCGATCAAATGGTCATGCC-3′) was cloned into the pLKO.1 vector (54) obtained from the RNAi Consortium (TRC). YAMC were infected with lentiviral vectors containing shRNA Rab27a and nontargeting shRNA in the presence of 8 μg/ml Polybrene. Cells with stable expression of shRNA were selected by addition of 5 μg/ml puromycin to the medium on the second day after infection. Western blot analysis of cellular lysates was performed to determine the levels of Rab27a protein expression.
Preparation of CM.ImSt cells, MSIE cells, and YAMC at passages 18 to 25 were cultured in FBS-free RPMI 1640 medium (2 × 105 cells/ml medium, equal to 106 cells/plate in 5 ml of RPMI) with confluence at 80% to 90% for 3 to 24 h at 37°C (nonpermissive temperature) in the presence or absence of a cell-permeable inhibitor of neutral sphingomyelinase, GW4869 [N,N-bis(4-(4,5-dihydro-1H-imidazol-2-yl)phenyl)-3,3′-p-phenylene-bis-acrylamide dihydrochloride (Sigma-Aldrich, Inc.)]. The culture supernatants then were collected.
Cells were treated with an HSP90 inhibitor, geldanamycin (BioViotica), at concentrations from 0.5 to 5 μg/ml in FBS-free RPMI 1640 medium for 12 h at 37°C. The cells then were cultured in fresh FBS-free medium for 12 h at 37°C, and culture supernatants were collected.
CM was prepared by successive differential centrifugation of culture supernatants at 300 × g for 10 min, 2,000 × g for 20 min, and 10,000 × g for 30 min. To exclude the influence of apoptotic and necrotic fragments on CM, cells were maintained in culture with a survival rate higher than 90%.
Collection of intestinal luminal contents from mice.All animal experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee at Vanderbilt University Medical Center. WT C57BL/6J and Rag2–/– mice on a C57BL/6J background were purchased from the Jackson Laboratory.
The small intestine and the colon were isolated from 8- to 10-week-old WT and Rag2–/– mice. After feces were removed, the small intestine and the colon were washed with 10 ml of cold phosphate-buffered saline (PBS). Luminal contents were collected by flushing with 10 ml of cold RPMI three times. The flushed material was centrifuged at 500 × g for 15 min and then 2,000 × g for 20 min and filtered through a 0.2-μm filter.
Preparation and characterization of EVs.EVs were isolated from CM and flushed intestinal luminal contents by centrifugation at 100,000 × g for 8 h. EVs were washed with PBS. The size and concentration of EVs were measured through TEM and nanoparticle tracking analysis.
For TEM, EVs were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer for 30 min at room temperature, incubated on Formvar carbon-coated grids for 1 min, and then negative stained with 2% uranyl acetate for 30 s. Imaging was performed on a Philips/FEI T-12 microscope.
For nanoparticle tracking analysis, EVs were diluted in PBS and analyzed through nanoparticle tracking with a NanoSight LM10 system (NanoSight Ltd., Amesbury, UK) configured with a 405-nm laser and a high-sensitivity sCMOS camera (OrcaFlash 2.8, Hamamatsu C11440; NanoSight Ltd.). Analysis was performed with NTA software (version 2.3, build 0006 beta 2). Three videos of 60-s duration were recorded for each sample. Calibrations were performed with polystyrene latex microspheres with sizes of 100, 200, and 400 nm.
LGG treatment.LGG (American Type Culture Collection 53103) was incubated in Lactobacillus MRS broth at 37°C until an optical density at 600 nm (OD600) of 0.5 to 1.0 was reached. LGG was precipitated from MRS broth and washed with PBS. LGG then was cultured with YAMC, MSIE cells, and ImSt cells (106 cells/plate) in 5 ml of serum-free RPMI. The total amount of LGG was 108 CFU in each plate (MOI of 100:1). LGG was also cultured with CM, EVs in RPMI, recombinant human HSP90α and HSP90β (Sigma-Aldrich, Inc.) in RPMI, and flushed intestinal luminal contents at 37°C for 3 h. An HSP90 inhibitor, geldanamycin (BioViotica), was used in LGG culture with CM and EVs. Culture of LGG in RPMI was used as a control. Culture supernatants and LGG proteins isolated using a MicroRotofor lysis kit (Bio-Rad Laboratories, Inc.) were prepared for examining the levels of p40 and p75. LGG cultured with YAMC-CM prepared from 24-h cultures of YAMC alone (YAMC-CM-LGG) and culture with RPMI (RPMI-LGG) for 3 h were used to treat cells and mice.
Cell treatment.YAMC at passages 18 to 25 were serum starved (0.5% serum in RPMI medium) at 37°C overnight and then treated with RPMI-LGG and YAMC-CM-LGG at an MOI of 100:1 for 1 h. Total cellular lysates were collected and used to detect EGFR activation by Western blot analysis.
HT29 cells (ATCC HTB-38), a human colorectal adenocarcinoma cell line, were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS. HT29 cells were treated with TNF (100 ng/ml), IL-1α (10 ng/ml), and IFN-γ (100 ng/ml) in DMEM containing 1% FBS for 12 h to induce apoptosis, in the presence or absence of RPMI-LGG or YAMC-CM-LGG at an MOI of 100:1. Total cellular lysates were collected for detection of cleaved caspase 3 by Western blot analysis.
T84 cells (ATCC CCL-248), a human colorectal carcinoma, were cultured in a 1:1 mixture of DMEM–Ham's F-12 medium with 2.5 mM glutamine, 5% FBS, and 100 U/ml penicillin and streptomycin at 37°C under 5% CO2. T84 cells in normal culture medium were treated with H2O2 (20 μM) for 3 h with or without treatment with RPMI-LGG or YAMC-CM-LGG at an MOI of 100:1. Cells were fixed for immunostaining.
HT29 and T84 cells were used in these experiments within 10 to 20 passages after purchase from the ATCC.
Mouse treatment and collection of intestinal tissues and feces.WT C57BL/6J mice aged 8 to 10 weeks were used to detect p40 production in vivo and to determine the effects of YAMC-CM treatment of LGG on prevention of colitis.
Colitis was induced by DSS or TNBS treatment in mice. In the DSS-induced injury and acute colitis model, mice were administered 3% DSS (molecular weight, 36 to 50 kDa; MP Biomedicals) in drinking water for 4 to 7 days. Mice were euthanized at the end of DSS treatment or 2 days after DSS treatment, and intestinal injury and acute colitis were evaluated. Mice were given drinking water as controls. For the TNBS-induced acute colitis model, mice were intrarectally treated with 100 μl of 70 mM TNBS in 50% ethanol. Control mice received 100 μl of 50% ethanol intrarectally. Mice were euthanized 4 days after TNBS treatment.
Mice with or without colitis were administered YAMC-CM-LGG or RPMI-LGG by gavage at a dosage of 108/day, once per day, as outlined in Fig. 4A and 5A.
The entire colon was rolled with the Swiss roll method for preparation of formalin-fixed paraffin-embedded tissues. Hematoxylin and eosin (H&E)-stained sections were used for light microscopy examination to assess colon injury and inflammation. Samples from the entire colon were examined by a pathologist blinded to treatment conditions.
The scoring system was generated on the basis of the pathological characteristics of models of colitis. DSS-induced colitis is characterized by colon epithelial cell injury-driven inflammation with elevated proinflammatory cytokines, including TNF, IL-6, KC, and IFN-γ (27). TNBS-induced colitis is mediated by IL-12-driven Th1 immune responses, including increased TNF and IFN-γ production (30). The scoring system for DSS-induced injury and inflammation includes (a) degree of inflammation (scale of 0 to 3), (b) crypt damage (0 to 4), (c) percentage of area involved by inflammation (0 to 4), (d) crypt damage (0 to 4), and (e) depth of inflammation (0 to 3) (27). The total score is determined by (a + b) × c + d × e. The scoring system for TNBS-induced inflammation includes (a) lamina propria mononuclear cell and polymorphonuclear cell infiltration, (b) enterocyte loss, (c) crypt inflammation, and (d) epithelial hyperplasia. Each category was scored from 0 to 3 (55, 56). The total score is determined by a + b +c + d + e.
Colonic tissues were collected for preparation of colonic contents by flushing and isolation of RNA. Feces were collected. Because the feces from mice with diarrhea were wet, the feces were dried at –20°C overnight, weighed, solubilized in PBS at a ratio of 1:5 (wt/vol), and homogenized with a TissueLyser. Supernatants were used for Western blot analysis.
qPCR analysis.To examine the amount of LGG in feces, DNA was extracted from feces using a ZR fecal DNA Miniprep kit (ZYMO Research Corporation). qPCR was performed using the LGG strain-specific primers 5′-CGCCCTTAACAGCAGTCTTC-3′ and 5′-GCCCTCCGTATGCTTAAACC-3′, which have been previously described (28), and IQ SYBR green supermix (Bio-Rad Laboratories) according to the manufacturer’s instructions. Because no detectable LGG in feces from WT mice in our animal facility was found by DNA fingerprint analysis and conventional PCR analysis with LGG-specific primers (29), LGG (104 to 108/g feces) was mixed with feces to generate a feces-based LGG concentration curve. All DNA samples were analyzed in triplicate.
Real-time PCR assay.Total RNA was isolated from colonic mucosal tissues with an RNA isolation kit (Qiagen) and was treated with RNase-free DNase. Reverse transcription was performed with a high-capacity cDNA reverse transcription kit (Applied Biosystems). Real-time PCR was performed with TaqMan gene expression master mix and the following primers: Mm00443259 (TNF), Mm99999071 (IFN-γ), and Mm00433859 (KC). We did not find any changes in glyceraldehyde-3-phosphate dehydrogenase (GAPDH) protein levels in colonic mucosa by Western blot analysis in mice with DSS- and TNBS-induced colitis (data not shown). Thus, the relative abundance of GAPDH mRNA was used to normalize the levels of the mRNAs of targeted genes. All cDNA samples were analyzed in triplicate.
Immunostaining.Cultured cells were fixed with 4% paraformaldehyde overnight at 4°C, permeabilized with 0.2% Triton X-100 for 5 min, and blocked with 3% goat serum albumin for 1 h. Cells then were incubated with rabbit anti-ZO-1 antibody (Invitrogen Corporation, Carlsbad, CA) overnight at 4°C and with Cy3-conjugated goat anti-rabbit IgG secondary antibody (Cell Signaling Technology) for 1 h at room temperature. Cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and observed through fluorescence microscopy.
Proteins in isolated EVs were labeled with green fluorescent dye with an ExoGlow-Protein EV labeling kit (System Biosciences), according to the manufacturer’s instructions, and cultured with LGG in RPMI for 2 h. LGG was fixed with 4% paraformaldehyde overnight at 4°C, permeabilized with 0.5% Triton X-100 for 5 min, and blocked with 5% goat serum for 1 h. LGG was then incubated with rabbit anti-p40 antibody overnight at 4°C and then with Cy3-conjugated goat anti-rabbit IgG secondary antibody (Cell Signaling Technology) for 1 h at room temperature. DNA was counterstained with DAPI. Stained LGG was observed through fluorescence microscopy.
Preparation of cellular and mucosal lysates for Western blot analysis.Cultured cells were washed with ice-cold PBS and solubilized in cell lysis buffer (10 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, and 1 mM EGTA) with protease and phosphatase 1 and 2 inhibitor cocktails (Sigma-Aldrich Co.) and then centrifuged (14,000 × g, 10 min) at 4°C. Colonic mucosal lysates were collected by scraping the colonic mucosa, solubilizing the samples in cell lysis buffer, and then disrupting cells with a TissueLyser. Protein content was determined with a DC protein assay.
Cellular and mucosal lysates were mixed with Laemmli sample buffer, and proteins were separated by SDS-PAGE for Western blot analysis with anti-phospho-Tyr1068 EGFR, anti-Rab27a (Cell Signaling Technology), anti-EGFR (Millipore), and anti-cleaved caspase-3 (Invitrogen Corporation, Carlsbad, CA) primary antibodies and horseradish peroxidase (HRP)-conjugated secondary antibodies against rabbit or mouse IgG. Samples were blotted with an anti-β-actin antibody (Sigma-Aldrich Co.) as a protein loading control.
Rabbit anti-p40 and anti-75 generated by our group (22) were used for Western blot analysis. Sera collected from rabbits before injection of p40 or p75 were used as negative controls for anti-p40 and anti-75 antibodies, respectively. For other primary antibodies, blots probed with only secondary antibodies were used as negative controls for primary antibodies.
Dot blot for localization of HSP90α/β in EVs.Dot blotting of EVs was performed as described previously (57). Different concentrations of EVs collected from CM were dotted onto nitrocellulose membranes and allowed to dry at room temperature for 1 h. The membranes were then blocked with 5% nonfat dried milk in Tris-buffered saline (TBS) in the absence or presence of 0.1% (vol/vol) Tween 20 (TBS-T) at room temperature for 1 h and incubated with anti-HSP90α/β antibody (Santa Cruz Biotechnology, Inc.) in TBS or TBS-T overnight at 4°C and HRP-conjugated secondary antibody for 1 h at room temperature.
Statistical analysis.Statistical significance was determined by one-way analysis of variance (ANOVA) for multiple comparisons and a two-tailed t test for comparing data from two groups using Prism 6.0 (GraphPad Software, Inc., San Diego, CA). A P value of <0.05 was defined as statistically significant. All data are presented as means ± standard deviations (SD).
ACKNOWLEDGMENTS
This work was supported by National Institutes of Health (NIH) grant R01DK081134, the Crohn's & Colitis Foundation Senior Research Award, and The Vanderbilt Microbiome Initiative (F.Y.), NIH grants R01DK56008 and R01DK54993 and the Crohn's & Colitis Foundation Senior Research Award (D.B.P.), NIH grants R01DK58587, R01CA77955, and P01CA116087 (R.M.P.), and core services performed through Vanderbilt University Medical Center's Digestive Disease Research Center, supported by NIH grant P30DK058404. We thank the Vanderbilt Microbiome Initiative for the support for this project.
We thank M. Kay Washington from Vanderbilt University Medical Center for examining intestinal tissue sections from mouse models of colitis.
We have no conflicts of interest to declare.
FOOTNOTES
- Received 5 February 2019.
- Returned for modification 7 March 2019.
- Accepted 9 April 2019.
- Accepted manuscript posted online 22 April 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00113-19.
- Copyright © 2019 American Society for Microbiology.
