Lactobacillus rhamnosus GG supernatant enhance neonatal resistance to systemic Escherichia coli K1 infection by accelerating development of intestinal defense

The objective of this study was to determine whether Lactobacillus rhamnosus GG culture supernatant (LCS) has a preventive effect against gut-derived systemic neonatal Escherichia coli (E. coli) K1 infection. The preventive effects were evaluated in human colonic carcinoma cell line Caco-2 and neonatal rat models. Our in vitro results showed that LCS could block adhesion, invasion and translocation of E. coli K1 to Caco-2 monolayer via up-regulating mucin production and maintaining intestinal integrity. In vivo experiments revealed that pre-treatment with LCS significantly decrease susceptibility of neonatal rats to oral E. coli K1 infection as reflected by reduced bacterial intestinal colonization, translocation, dissemination and systemic infections. Further, we found that LCS treated neonatal rats have higher intestinal expressions of Ki67, MUC2, ZO-1, IgA, mucin and lower barrier permeability than those in untreated rats. These results indicated that LCS could enhance neonatal resistance to systemic E. coli K1 infection via promoting maturation of neonatal intestinal defense. In conclusions, our findings suggested that LCS has a prophylactic effect against systemic E. coli K1 infection in neonates. Future studies aimed at identifying the specific active ingredients in LCS will be helpful in developing effective pharmacological strategies for preventing neonatal E. coli K1 infection.


Pre-treatment with LCS inhibited the adhesion and invasion of E. coli K1 to Caco-2. Because
adherence and invasion to intestinal epithelium are the pivotal steps for intestinal bacterial translocation and to enter the circulation resulting in a systemic infection, we firstly determined whether LCS has the inhibitory effect on adhesion and invasion of E. coli K1. Caco-2 monolayers were pre-incubated with different concentrations of LCS for 3 hours (h) before the bacterial infection. Monolayers treated with cell culture medium or MRS (LGG culture medium, not used, 2% in the cell culture medium without antibiotics) alone served as controls. Numbers of cell-associated bacteria and intracellular bacteria were determined. The data reveal that pre-incubation with LCS could dose-and time-dependently inhibits adhesion and invasion of E. coli K1 (Fig. 1A-D). Interestingly, we could not find any obvious inhibitory effects on adhesion and invasion of E. coli K1 on adding LCS 1 h after E. coli K1 infection (Data not shown). To examine whether LCS has antibacterial activity, we assessed the influence of LCS on the growth of bacteria cultured in vitro with brain heart infusion (BHI) broth. As shown in Fig. 1E, the growth curves of bacteria grown in BHI with or without LCS were similar. This result demonstrates that LCS has no lethal effect on in vitro growth of E. coli K1. Furthermore, the trypan blue stain assay showed that LCS has no detectable cytotoxicity on Caco-2 (data not shown). Overall, these data suggested that LCS can effectively inhibit adhesion and invasion of E. coli K1 but has no impact on in vitro growth of E. coli K1.

Mucin is required for LCS-mediated inhibitory effect on adhesion and invasion of E. coli
K1. Direct killing and competing adhesion sites with pathogen are two major mechanisms by which viable LGG inhibit bacterial adhesion and invasion. Thus, it is puzzling that how LCS exhibit an inhibitory effect on adhesion and invasion of E. coli K1 (Fig. 1A-D), without any steric-hindrance or direct killing effect on E. coli K1 (Fig. 1E). Mucin layer is an important barrier that separates the pathogen from enterocyte. We thus speculated that, mucin layer may play a pivotal role in LCS-mediated inhibitory effect on adhesion and invasion of E. coli K1. To test this hypothesis, we firstly evaluated the influence of LCS or E. coli K1 on production of mucin in Caco-2 monolayer using Periodic Acid Schiff (PAS) assay as described in Methods. As shown in Fig. 2A, infection with E. coli K1 markedly reduced the expression of mucin from multiplicity of infection (MOI) of 50 to 200 compared with the control. In contrast, LCS could significantly elevate mucin production in a concentration-dependent manner. We next examined whether LCS has any protective role against mucin-depletion effect of E. coli K1. Result shown in Fig. 2B suggested that pre-treatment with LCS could significantly alleviate E. coli-induced loss of mucin. Similar results were also observed with morphological alterations of mucin layers. As shown in Fig. 2C, untreated Caco-2 monolayer was covered with a purple, homogeneous and continuous mucin layer. After exposure to E. coli K1, however, the mucin layer became thinner and disrupted. In contrast, pre-incubation with LCS for 3 h could prevent from E. coli K1-induced disruption of the mucin layer. Furthermore, we evaluated the expression levels of MUC2 in each group using western blot analysis. As shown in Fig. 2D, pre-incubation with LCS alleviated E. coli K1-induced depletion of MUC2.
N-acetylcysteine (NAC) is an agent which could remove the mucin layer from intestinal mucosa 34 . Hence, we used NAC to explore whether the mucin layer is required for LCS-mediated inhibitory effect. Caco-2 monolayers were infected as described in the adhesion assay and then the bacteria that were trapped in mucin and adhering to Caco-2 monolayers were separated using NAC and counted as described in Methods. As shown in Fig. 2E   Caco-2 monolayers were incubated with cell culture medium, MRS or different concentrations of LCS for 3 h, or with 1% LCS for different time duration before infection. MRS (LGG's medium) and cell culture medium were used as controls. Then E. coli K1 (MOI = 100) was added and incubated for 3 h. The numbers of associated bacteria (A,B) and intracellular bacteria (C,D) were determined. The results were expressed as a percentage of the control. Error bars indicate standard deviations. (E) Effect of LCS on the growth of E. coli K1 in BHI broth at different concentrations of LCS. Bacterial growth was monitored by measuring the absorbance of liquid cultures at 600 nm. *P < 0.05, **P < 0.01, ***P < 0.001. protective effects of LCS against gut barrier injury caused by chemicals, such as alcohol, dextran sodium sulfate and hydrogen peroxide [35][36][37][38][39] . However, little is known about the roles of LCS on pathogen-induced intestinal barrier injury. Here, we examined whether LCS could abrogate the deleterious effects of E. coli K1 on intestinal integrity. Firstly, we explored whether treatment with MRS or LCS alone could reduce the trans-epithelial electrical resistance (TEER) values of Caco-2 monolayers. As shown in Fig. 3A, stable TEER values were observed in the control, MRS, 1% LCS and 2% LCS treated Caco-2 monolayers, suggesting that these factors have no detrimental role on intestinal integrity. However, infection with E. coli K1 could reduce the TEER values of Caco-2 monolayer in a time-dependent manner. In contrast, when pre-incubated with LCS for 3 h before infection, the reduction of TEER was alleviated (Fig. 3A). In parallel, the numbers of E. coli K1 translocated from the upper chamber to the lower chamber of transwell insert were dramatically reduced in LCS treated groups (Fig. 3B). These results suggest that LCS could reduce E. coli-caused gut barrier injury and prevent from intestinal translocation of E. coli K1.
To further investigate the protective effect of LCS on intestinal barrier function, we examined the zonula occludens-1 (ZO-1) expression in Caco-2 during E. coli K1 infection using western blot and immunofluorescent staining. As shown in Fig. 3C, E. coli K1 challenge resulted in a marked reduction of ZO-1 expression, but this detrimental role was abrogated by pre-incubation with LCS. Similar results were observed in immunofluorescent staining, which were shown in Fig. 3D. Non-infected Caco-2 exhibited continuous expression of ZO-1 between adjacent cells, and showed a typical "chicken-wire" shape with a continuous lining (red arrow). However, infection with E. coli K1 induced different degrees of morphological damage (yellow arrow). Pre-incubation with LCS significantly inhibited E. coli-induced destruction of ZO-1 tight junction morphology. Thus, we concluded that pre-treatment with LCS protected the intestinal barrier integrity against E. coli K1-caused injury.

Effect of LCS on bacterial intestinal colonization, gut barrier injury and systemic infection during E. coli K1 infection in neonatal rats.
To examine the protective effect of LCS against neonatal E. coli K1 bacteremia and meningitis in vivo, we induced systemic E. coli K1 infection in neonatal rats via feeding live E. coli K1 as described previously 11,40 . In this animal model, the course of E. coli K1 infection mimics the natural route of intestinal colonization, translocation, bacteremia, sepsis and meningitis found in the human neonates. At the beginning of the experiments, all rat pups (1-day-postpartum, P1) from four litters were randomly divided into four groups and received PBS, 20% MRS, 10% or 20% LCS by oral gavage (twice a day for three days). Then systemic infection was induced by oral gavage with 5 × 10 9 colony-forming unit (CFU) of E. coli K1 and the protective effects of LCS against bacterial intestinal colonization, gut barrier injury and systemic infection were evaluated as described in Methods. As shown in Fig. 4A, feeding E. coli K1 to PBS-or MRS-treated pups led to almost 100% intestinal colonization within 24 h after infection and persisted throughout the period of observation (6 days). In contrast, only 73% pups in 10% LCS treated group and 47% pups in 20% LCS treated group were colonized with E. coli within 24 h after infection. These results indicated that pre-treatment with LCS delayed the intestinal colonization of E. coli K1. Next, we evaluated the intestinal barrier permeability using fluorescein isothiocyanate (FITC)-dextran. Results showed that the pups which were pre-treated with LCS have lower serum levels of FITC-dextran than that in untreated pups (Fig. 4B), indicating that LCS could prevent E. coli-induced intestinal barrier injury.
The numbers of E. coli K1 in blood, liver, spleen and cerebrospinal fluid (CSF) were detected as described in Methods. Our results showed that E. coli K1 CFU counts in blood, liver, spleen and CSF were significantly decreased in rat pups that were pre-treated with LCS ( Fig. 5A-D). Taken together, these data suggested that LCS could raise the resistance of neonatal rats to gut-derived E. coli K1 infection. Birchenough GM et al. found that 9-day-old (P9) Wistar rat pups were more resistant to gut-derived E. coli K1 infection than two-day-old (P2) pups 40 . In our study, similar result was observed in Sprague Dawley (SD) rat pups. As shown in Fig. 6A-C, P3 pups, which were infected with E. coli K1 showed higher mortality, bacteremia and gut barrier permeability compared with P5, P7 and P9 pups. P9 pups were refractory to E. coli K1 infection with lowest mortality, bacteremia and intestinal barrier permeability. These results suggested that neonatal intestinal defense progress rapidly after birth. Importantly, we could notice that, on pre-treatment with LCS for three days, P3 pups could exhibit resistance to E. coli K1 infection similar with the P7 pups ( Fig. 6A-C). Thus, we speculated that LCS has considerable potential to promote the maturation of neonatal intestinal defense. Based on these observations, we hypothesize that improvement in the maturation of intestinal defense could represent the underlying mechanism to explain the contribution of LCS to neonatal resistance against oral E. coli K1 infection. Further to test this hypothesis, we firstly explored whether administration with LCS promotes intestinal epithelial cell proliferation and differentiation via detecting the expression of Ki67 and MUC2 in the small intestine and colon respectively. As shown in Fig. 7A and B, the expression levels of Ki67 and MUC2 were significantly increased in the intestine tissues of LCS treated pups.

LCS enhances the neonatal resistance to E. coli K1 infection via promoting postnatal maturation of intestinal defense.
Mucin, immunoglobulin A (IgA) and barrier function are key components of the intestinal defense. Thus, we next compared the production of mucin and IgA and formation of intestinal barrier function between LCS treated and untreated pups. Results showed that LCS treated pups have a higher expression level of mucin and IgA levels (Figs 6D,E and 8A). To evaluate the formation of intestinal barrier function, the gut permeability and expression of tight junction protein ZO-1 were detected. Intestinal permeability assay showed that LCS treated pups have lower intestinal permeability than that in untreated pups (Fig. 6F). Meanwhile, immunohistochemical staining showed that pre-treatment with LCS increased ZO-1 expression on the membrane of ileum (Fig. 8B). Overall, these results suggested that LCS could accelerate the development of neonatal intestinal defense.

Discussion
In the present study, we intended to make a strategy shift from treatment to prevention of the neonatal systemic E. coli K1 infection through the use of probiotic culture product to enhance the intestinal defense, which is a unique preventive barrier against natural E. coli K1 infection. Our data showed that pre-treatment with LCS is able to delay E. coli K1 intestinal colonization, inhibit bacterial translocation and dissemination, and reduce bacteremia and meningitis in neonatal rats. These results suggested that administration with LCS enhances the resistance of neonatal rats to E. coli K1 infection. Further study found that LCS promotes intestinal proliferation and differentiation, accelerate intestinal barrier formation and increase mucin and IgA production in the gut of neonatal rat. These data indicated that neonatal administration with LCS accelerates the maturation of intestinal defense and confers a high resistance to intestinal infection. To the best of our knowledge, this is the first study reporting such beneficial effect of probiotic culture product on maturation of neonatal intestinal defense. Collectively, our data not only indicated that LCS has a great potential in preventing neonatal E. coli K1 sepsis and meningitis, but also has a potential to become an effective prophylaxis for immature intestine-associated diseases, such as inflammatory bowel disease, infectious enteritis and necrotizing enterocolitis.
Probiotics secrete many kinds of antimicrobial substances that inhibit pathogenic infection by direct killing effect, including bacteriocins, reuterin, organic acid, hydrogen peroxide and some heat-resistant small peptides 41 . It has been established that the antimicrobial activity of LGG is executed entirely through secreted cell-free supernatants 41 . In this study, we have found the inhibitory effect of LCS on bacterial adhesion and invasion. This effect, however, is unlikely to execute via direct inhibition on the growth of bacteria, because the concentrations of LCS used here, did not show any extracellular antimicrobial activity (Fig. 1E). In contrast, the anti-adhesion mechanism was based on indirect enhancement of mucin production that separated the pathogen from Caco-2 cells (Fig. 2E and F). (B) Pups were treated as described above, intestinal permeability was evaluated using FITC-dextran. The results are expressed as relative content, taking the value for the control (un-infected pups) as 1. *P < 0.05, ***P < 0.001.
Because LCS has no direct antibacterial activity to pathogens, it is unlikely to develop the microbial resistance. Meanwhile, this indirect mechanism makes LCS exert a preventive effect on most of gut-derived pathogenic infections rather than only E. coli. Furthermore, we found that LCS could also promote mucin production in the intestinal tract of neonatal rats ( Fig. 6D and E), which is much significant during the neonatal period. Firstly, during neaonatal period, a protective intestinal mucus barrier always not fully developed 40 , and the up-regulated mucin production will form a physical barrier to protect the underlying epithelium from the attachment of pathogens. Secondly, mucin could provide carbon, nitrogen, and sulfur source for intestinal microbiota, especially Akkermansia muciniphila, which could through a turnover of mucus (degrade mucins and simulate mucin production), maintain the mucus thickness to protect the intestinal barrier 42 .
Intestinal epithelial barrier is formed and maintained by tight junction complexes, which plays a significant role as access point that checks bacterial translocation across the cell barrier. Disruption of the integrity of this barrier occurs in several diseases, such as inflammatory bowel disease 43  showed that live LGG, but not heat-inactivated LGG protected polarized MDCK-I and T84 epithelial cell against Enterohemorrhagic E. coli-caused changes in TEER, dextran permeability, and redistribution of ZO-1 and claudin-1 46 . However, evidence on probiotic-derived product exerting protective effect against pathogen-caused tight junction injury is still lacking. In the present study, we observed that pre-treatment with LCS could alleviate E. coli K1-induced reduction of TEER values and ZO-1 expression in Caco-2 monolayer (Fig. 3A,C and D). These results indicated that LCS could prevent from E. coli K1-induced tight junction damage, as well as convincingly suggested that LGG culture supernatant has the similar protective effect on intestinal epithelial barrier as live LGG.
The immature intestinal defense in the neonatal period may provide opportunities for pathogenic translocation across the gut barrier into the blood stream, leading to systemic inflammation and infection 15,44,47 . In our study, we observed that feeding of E. coli K1 to P3 pups led to about 87% death within 7 days, with the higher intestinal permeability and more serious bacteremia than that in P5, P7 and P9 pups (Fig. 6A-C). The same phenomenon occurs in humans, because human neonates are also most susceptible to E. coli K1 infection during the  Fig. 4. E. coli K1 were recovered 60 hours after infection from the blood (A), liver (B), spleen (C) and CSF (D). All data were expressed as means ± standard deviation. *P < 0.05, **P < 0.01, ***P < 0.001.
Scientific RepoRts | 7:43305 | DOI: 10.1038/srep43305 early neonatal period 48 . Importantly, these results indicated that the intestinal defense is undergoing rapid maturation after birth. Interestingly, when pre-treated with LCS for three days, P3 pups exhibited the same resistance to oral E. coli K1 infection as P7 pups (Fig. 6A-C), indicating that LCS has potential to promote maturation of neonatal intestinal defense. Subsequently, we found that expression levels of Ki67 and MUC2, the maker of cell proliferation and intestinal differentiation respectively 49 , were significantly higher in LCS treated pups than those in untreated pups ( Fig. 7A and B). Moreover, we observed that LCS could modulate important intestinal defense including ZO-1, mucin, IgA and barrier permeability (Figs 6D-F, 8A and B). These data suggested that LCS is able to accelerate the maturation of neonatal intestinal defense, and thereby reduce the susceptibility of neonates to E. coli K1 infection. These results were consisted with a previous study that, neonatal mice which colonization with live LGG had more sophisticated intestinal functions and were less susceptible to dextran sodium sulfate-induced colitis 50 than the mice without LGG colonization. In another study, Ravi M. Patel et al. reported that neonatal colonization of mice with live or heat-killed LGG accelerates maturation of intestinal barrier function by promoting claudin 3 expression 44 . However, the authors found that administration with high-dose live LGG (10 9 CFU/day for 7 days) lead to an increased mortality in 1-week-old mice. This paper highlight the safety concerns about the application of live probiotics in neonates. Indeed, according to the regulations of Food and Drug Administration in the United States, routine use of live organisms to immunocompromised premature infants is prohibited 51 . In this study, we found that LGG culture supernatant has shown a progressive effect on maturation of neonatal intestinal functions similar to live LGG, and thus support the application of probiotic-derived factors to replace live probiotic to avoid the potential risks if necessary.
Gut microbiome is largely subjected to dynamic changes after birth, and temporally take part in maturation of the intestinal immune system 52 52 . Thus, it is interesting to explore the effect of LCS on formation of neonatal intestinal microbiota. We speculated that LCS has a great potential to promote development of neonatal intestinal microbiota. Two rational bases supported this hypothesis. Firstly, it is established that besides protecting the enterocyte from pathogenic assault, mucus also provide glycan-dependent anchoring sites and nutrients to intestinal commensal microbiota 53 . Thus, it is possible that LCS-induced mucin production may help the colonization of commensal bacteria. Secondly, IgA, the important defense in intestinal (F) FITC-dextran assay was performed to evaluate the gut barrier formation in vivo. All data were expressed as means ± standard deviation. *P < 0.05, **P < 0.01. mucosa to prevent the pathogen from attaching to the enterocyte, has been demonstrated that could promote the establishment of microbiota composition and maintains the diversity of microbiota 54 . Moreover, research has shown that IgA is able to maintain the homeostasis of gut flora using an "immune exclusion" strategy to exclude the pathogenic microbes and control the commensal microbes 55 . Thus, induced IgA expression by LCS may play an important role in formation, development and maintenance of intestinal microbiota composition. Further study should focus on exploring the relationship between LCS and formation of neonatal intestinal microbiota.
In the present study, we suggested that LCS has a promising application in preventing neonatal intestinal infection. However, it is unclear which LCS components exert the beneficial effects. LCS is a complex mixture containing lots of substances, including lipids, organic acids, proteins and other small molecules. Due to the complexity and uncertainty, it is necessary to keep an eye on the side effects of LCS. To date, p40 and p75 are two most characterized proteins purified from LGG-derived soluble factors and both of them have beneficial effects on intestinal barrier functions. However, p40 showed more potent effects than p75 56 . More recently, it has been established that p40 can induce IgA and MUC2 production in intestinal tissues through active epidermal growth factor receptor (EGFR) 57,58 . Furthermore, using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, we also detected two bands, ~40 KDa and ~75 KDa, exist in LCS in our study (data not shown). In accordance with our findings, it is reasonable to speculate that p40 may be the major active ingredient in LCS to exhibit the beneficial effects towards infectious pathogens.
In summary, our data suggested that LCS could prevent neonatal gut-derived E. coli K1 infection through promoting the maturation of neonatal intestinal defense. We believed that LCS has a great potential to become an alternative treatment option for preventing neonatal E. coli K1 sepsis and meningitis. Furthermore, our data also confirmed that supernatant from certain probiotic could exert beneficial effects similar with their live probiotic counterparts. Further identification of the active ingredients in LCS is required to develop newer prophylaxis for preventing neonatal sepsis and meningitis.

Methods
Ethics statement. All research involving animals has been approved by the ethics committee and performed strictly according to the guidelines for animal care in Southern Medical University (SMU, Guangzhou, China). Timed-pregnant SD rats were obtained from Animal Experimental Center of SMU and bred in-house. All supplements including food, water, and enrichment were autoclaved, and animals were kept in the animal facility. All surgeries were performed under anesthesia with ketamine and lidocaine, and utmost efforts were taken to minimize suffering.  LGG was grown in De Man, Rogosa, and Sharpe (MRS) broth (Oxoid, Hampshire, UK) at 37 °C for 24 hours. E. coli K1 was grown for 14 h at 37 °C in brain heart infusion (BHI) broth in the presence of rifampicin (100 μ g/ ml). LCS was prepared by centrifugation at 10,000 rpm for 1 minute (m) at 4 °C and dual filtration through 0.22 μ m millipore filters. To avoid the detrimental effect of acidic compounds, LCS was concentrated using 3 kDa spin columns and diluted into cell culture medium or PBS at appropriate pH value (7)(8) for subsequent experiments. To exclude the potential impact of probiotic culture medium, MRS were processed as LCS and served as controls in certain experiments. Caco-2 was purchased from Shanghai Institute of Cell Biology (Shanghai, China) and routinely cultured in Eagle's Minimum Essential Medium with 10% heat inactivated fetal bovine serum and streptomycin (100 mg/ml) and penicillin G (50 mg/ml) at 37 °C in 5% CO 2 .
Adhesion and invasion assays. Caco-2 cells were grown in a 24-well tissue culture plate (10 5 cells per well) at a minimum of 12 days to allow them fully differentiate 11 . Before the adhesion assays, Caco-2 were pre-incubated with cell culture medium, 2% MRS, 0.5% LCS, 1% LCS or 2% LCS (dilution in cell culture medium) for 3 h. After treatment, cell monolayers were washed with phosphate buffered saline (PBS) three times and examined under the microscope. No morphologic changes were observed. Adhesion assays were performed in Caco-2 as previously described 11,59 . In brief, approximately 10 7 CFU E. coli K1 were added to Caco-2 monolayer with a multiplicity of infection (MOI) = 100. The monolayers were incubated for 3 h at 37 °C. After incubation, the monolayers were washed three times with PBS and lysed by 0.5% Triton X-100 for 8 m. The bacteria were collected and enumerated by BHI agar plates with rifampicin (100 μ g/ml). Each assay was performed in triplicate wells and repeated three times. The results were expressed as a percentage of the control (E. coli infection only).
For invasion assays, Caco-2 monolayers were infected with E. coli K1 as described as above. To eliminate any extracellular bacteria, the monolayers were incubated with experimental medium containing gentamicin (100 μ g/ml) for 1 h at 37 °C. Then monolayers were washed three times with PBS and lysed by 0.5% Triton X-100 for 8 m. Intracellular bacteria were determined as mentioned above. Each assay was performed in triplicate wells and repeated three times. The results were expressed as a percentage of the control (E. coli infection only).
We used the NAC (Sigma-Aldrich, St. Louis, USA) to quantify E. coli K1 which was trapped in mucin or adhered to Caco-2 cells 34 . Caco-2 monolayers were infected with E. coli K1 as described in adhesion assays. Then infected monolayers were incubated with NAC (10 mM in PBS containing 0.2 μ M CaCl 2 , 0.5 mM MgCl 2 and 15 mM glucose) to remove the mucin layer (contains E. coli K1). Cells were washed four times with PBS to remove NAC and lysed with 0.5% Triton X-100 for 8 m. The number of associated bacteria was enumerated by BHI agar plate with rifampicin (100 μ g/ml). To evaluate the bacteria trapped into mucus, removed mucin layers were collected and plated on BHI agar plates with rifampin (100 μ g/ml) for counting. Each assay was performed in triplicate wells and repeated three times. Data are expressed as the percentage of trapped E. coli K1 or adhered E. coli K1 among the added bacteria.
Bacterial translocation and TEER measurements. Caco-2 cells were grown in Transwell inserts (6.5 mm diameter, 3 μ m pore size, Corning Costar Corp., USA) for at least 21 days to differentiate and to form tight junctions. This polarized monolayer mimics the pathogen translocation across the gut barrier (upper chamber) into the blood circulation (lower chamber). To determine whether LCS have protective effect on intestinal barrier against E. coli K1 infection, Caco-2 monolayers growth on Transwell insert were pre-incubated with cell culture medium, 2% MRS, 1% LCS, or 2% LCS for 3 h at 37 °C and 5% CO 2 . Then E. coli K1 (MOI = 100) were added to the upper chamber of Transwell inserts. TEER values were measured at 0, 1, 2, 3, 4, and 5 h post infection using a Millicell electrical resistance apparatus (EVOMAX, World Precision Instruments, USA). The bacteria translocated from the upper chamber to the lower chamber at 5 h post infection were quantified by plating on BHI agar with rifampicin (100 μ g/ml) and incubated at 37 °C overnight. PAS assay. Caco-2 monolayers grown in 6-well plate were treated with different levels of E. coli K1 (MOI = 50, 100 or 200) or LCS (1% or 2%) or infected with E. coli K1 as mentioned in the adhesion assay. Then treated monolayers were collected and lysed in lysis buffer to obtain soluble ingredients. Mucin production in soluble fractions was measured as described by Wang LH et al. 58 . In brief, 0.1% periodic acid was added to each sample and incubated for 2 h at 37 °C. Then Schiff reagent was added and incubated for 30 m in dark at room temperature. Optical density of each sample was assessed using a microtitre plate reader set at 550 nm. All samples were analyzed in triplicate.

PAS staining.
Caco-2 cells grown in 24-well plate were infected with E. coli K1 as described in the adhesion assay. Then monolayers were harvested and fixed in 4% paraformaldehyde at room temperature for 20 m. PAS staining was performed according to the manufacturer's instructions of PAS kit (Solarbio Science & Technology Co., Ltd, Beijing, China). The PAS-stained wells were counterstained with hematoxylin and observed using light microscopy. For PAS staining of intestinal tissues, 0035 μ m sections were fixed in 4% formaldehyde and paraffin-embedded. PAS staining was conducted as described above.

Neonatal rat model of hematogenous E. coli K1 systemic infection. Systemic E. coli K1 infection
was induced using specific pathogen free SD rats as described previously with minor modifications 11,40 . Litters were retained with their natural mothers after birth. Four litters (10 pups per litter) were pooled and randomly distributed into four groups. Pups were fed with 100 ul of PBS, 20% MRS, 10% LCS or 20% LCS (dilution in PBS) twice a day for three days. Then all pups were gavage with 100 μ l of E. coli K1 (5 × 10 9 CFU). Sixty hours post infection, blood, liver and spleen were extracted aseptically after anaesthetizing the rats with ketamine and lidocaine. CSF samples were collected as described previously 60 . Bacteria were quantified in homogenized tissues by serial dilution culture on BHI agar plates with rifampicin (100 μ g/ml). Gastrointestinal tract colonization of E. coli K1 was determined at 24 h intervals by culture of perianal swabs on MacConkey's agar as described previously 61 .
Intestinal permeability was determined as previously described 62 . Briefly, pups were given FITC dextran (MW 4000 at 60 mg/100 g, Sigma) by gavage 4 h before sacrifice. Blood was collected from heart puncture and FITC concentrations were measured using a fluorescence spectrophotometer at an excitation wavelength of 485 nm and emission wavelength of 535 nm.
Histopathological examination and immunohistochemical staining. For the immunohistochemical staining, 5 μ m paraffin-embedded sections were deparaffinized and antigen was retrieved. Sections were blocked in 1% normal goat serum and incubated overnight at 4 °C with antibody specific for MUC2, ZO-1, Ki67 and IgA (all from Abcam, Cambridge, UK) respectively. The primary antibodies were visualized using horseradish peroxidase (HRP)-coupled second antibodies with 50 mM Tris-HCl buffer (pH 7.4) containing DAB (3,3′diaminobenzidine) and H 2 O 2 , and the sections were lightly counterstained with hemotoxylin.
Quantification of immunoreactive signal was performed using National Institutes of Health image analysis software Image J 63 . In brief, the red-green-blue (RGB) bitmap images were firstly converted to 8-bit grayscale, then the threshold was modulated to display only positive signals and eliminate the background. The same cutoff value of threshold was used to analyze all the slides stained concurrently. The staining intensity measurements were calculated for the total area (total intensity/mm 2 ). A total of five sections for each group were analyzed. We expressed the results as the relative area, taking the value of the control (untreated P3 pups) as 1.

Western blotting analysis.
To assess the expression of ZO-1 and MUC2 in Caco-2, cells were collected and lysed on ice in radio-immunoprecipitation assay buffer and boiled at 100 °C for 10 min. Equal amounts of proteins were separated on SDS polyacrylamide gels and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). The membranes were incubated with a rabbit anti-ZO-1 antibody (1:200) or rabbit anti-MUC2 antibody (1:2000). Expression of primary antibodies was visualized using HRP-coupled second antibodies and enhanced chemiluminescence reagent kit (Bio-Rad Laboratories, USA). A goat polyclonal anti-β -actin antibody (1:1500) was used as a loading control. Each western blot assay was repeated at least three times.
Scientific RepoRts | 7:43305 | DOI: 10.1038/srep43305 Immunofluorescence analysis. Caco-2 monolayers were fixed in 4% paraformaldehyde for 10 min at room temperature. Cells were blocked in 5% normal goat serum in PBS for 1 h at room temperature. Then monolayers were probed with a rabbit anti-ZO-1 antibody for 12 h at 4 °C followed by incubation with Alexa Fluor 568-coupled goat anti-rabbit secondary antibody (Invitrogen, Carlsbad, CA) at room temperature in dark for 1 h. Slides were mounted and observed using fluorescence microscopy (Nikon Eclipse: TE 2000-E, Japan). Statistical Analysis. Data are shown in mean ± standard deviation and analyzed by one-way analysis of variance (ANOVA) tests. All statistical analyses were carried out using SPSS 13.0 (SPSS Inc., Chicago, IL, USA) and P < 0.05 was considered to be statistically significant.