Animal care procedures were in strict accordance with the Guidelines for Laboratory Animal Use and Care from the Chinese Center for Disease Control and Prevention and the Rules for Medical Laboratory Animals from the Chinese Ministry of Health, under protocol CAU20151001-1, approved by the Animal Ethics Committee of China Agricultural University. Pigs were sacrificed under xylazine hydrochloride anesthesia, and every effort was made to minimize animal suffering.

Twenty-one healthy (Landrace × Large White) piglets of mixed gender, selected from 8 different litters, weaned at 21 days of age, and weighing 6.29 ± 0.08 kg were obtained from Beijing Hog Raising and Breeding Center and used in this study. Each animal was penned separately and had access to an antibiotic-free feed and water ad libitum from day 0 (when newly weaned and transported to the animal experimental facility of the College of Veterinary Medicine, China Agricultural University) until day 18 (when euthanized).

Rectal temperature and other clinical signs were recorded daily. Severity of diarrhea was scored according to previously published criteria, with minimal modifications (Li et al., 2012). Piglets were considered to have mild diarrhea with a score of 4 and severe diarrhea when the score was 5 or 6.

Salmonella enterica serovar Infantis strain (named as S. Infantis CAU1508) was isolated from the intestinal contents of weaned pigs with diarrhea. Briefly, the fresh sample was added to buffered peptone water (Beijing Aoboxing Bio-tech Co., Beijing, China) and incubated at 37°C for 16–18 h with shaking. A 0.3-ml aliquot of the pre-enrichment broth was transferred onto modified semi-solid Rappaport Vassiliads (MSRV; Beijing Land Bridge Technology Co., Beijing, China) agar and incubated for 24 h at 42 ± 1°C. After selective enrichment, an inoculating loop of white or gray colonies grown on MSRV was plated on xylose lysine tergitol 4 (XLT4; Beijing Land Bridge Technology Co., Beijing, China) agar and incubated at 37 ± 1°C overnight. Bacterial colonies showing morphologic characteristics of Salmonella were then confirmed by amplification of the stn gene by PCR and using the API 20E biochemical identification system (BioMérieux, Beijing, China). Serovar was determined by slide agglutination with commercial Salmonella antisera (Statens Serum Institute, Denmark) following the Kauffmann-White scheme. After identification and purification, a single colony of S. Infantis from an XLT4 agar plate was transferred and grown in Luria-Bertani (LB) broth (Oxoid, Basingstoke, England) overnight at 37°C with shaking and then stored at −70°C.

Next, S. Infantis CAU1508 harboring the plasmid pFPV-mCherry/2 (obtained from Olivia Steele-Mortimer [Addgene plasmid #20956]), which constitutively expresses mCherry under control of the Salmonella rpsM promoter, was prepared as previously described (Drecktrah et al., 2008).

We detected virulence factors associated with diarrhea or fever in the inoculated S. Infantis strain CAU1508 via real-time PCR assay as previously described (Zhou et al., 2015). The virulence factors SipA, SipB, SopA, SopB, SopD, SopE, SopE2, FljB, and SseI were found in this strain, whereas FljA was absent. The primers used are listed in Table S1.

Lactobacillus rhamnosus GG (ATCC 53103) freeze-dried powder (Gefilus, Valio Ltd., Helsinki, Finland) was used to prepare LGG solution, as previously described (Zhang et al., 2010). Briefly, LGG was grown in De Man, Rogosa, and Sharpe (MRS) broth (Oxoid, Hampshire, UK) overnight at 37°C under microaerophilic conditions. Bacteria were pelleted by centrifugation at 3,000 × g for 10 min at 4°C and resuspended in physiological saline. The concentration of LGG was adjusted to 10⁹ CFU/ml.

The S. Infantis CAU1508 (mCherry) challenge strain was grown to mid-log phase in fresh LB broth at 37°C with shaking. The Salmonella were then harvested by centrifugation at 3,000 × g for 10 min at 4°C, washed 3 times with sterile physiologic saline, and resuspended in physiologic saline. An inoculum of S. Infantis CAU1508 (mCherry) containing 5.0 × 10¹⁰ CFU/ml was prepared and quantified by determination of CFUs after plating the serial dilutions of bacterial suspensions onto XLT4 agar plates.

On the day of weaning (day 0), pigs were divided into 3 groups (n = 7 per group) according to weight and ancestry: (1) control (CN) group (oral administration of sterile physiological saline only); (2) S. Infantis (SI) group (oral administration of sterile physiological saline and S. Infantis); and (3) LGG + S. Infantis (LS) group (oral administration of LGG and S. Infantis). All animals were administered the test material orally, as previously described (Yang et al., 2016), and were determined to be free of Salmonella by analysis of feces before the study.

At 9:00 a.m. from days 1 to 7, pigs in group LS were administered 10 ml of LGG (10⁹ CFU/ml) intragastrically, whereas pigs in groups CN and SI were administered 10 ml of sterile physiological saline daily. At 9:00 a.m. on day 8, pigs in groups SI and LS were orally inoculated with 10 ml of S. Infantis (5.0 × 10¹⁰ CFU/ml), whereas CN pigs received 10 ml of sterile physiological saline.

To assess bacterial colonization, fresh fecal samples were collected on days 1, 4, 8, 13, and 16 after weaning. Blood samples were collected from the jugular vein of pigs prior to S. Infantis challenge (0 h) and at 6, 12, 24, 48, 96, 144, and 192 h post-challenge. A 3-ml aliquot of peripheral blood collected in Venoject glass tubes containing EDTA (Terumo Europe NV, Leuven, Belgium) at 0, 24, and 192 h following S. Infantis challenge was used for flow cytometry analysis. Blood samples without additives were centrifuged, and the serum was stored at −80°C for the determination of peripheral levels of IL-7 and IL-22.

On day 10 post-infection, pigs were euthanized and tissue samples were immediately collected. The middle jejunum (without Peyer's patch [PP] involvement) and ileum segments collected from each pig were flash frozen in liquid nitrogen and then stored at −80°C for later mRNA and Western blot analyses. For immunostaining studies, the proximal, mid-, and distal segments of the jejunum and ileum (approximately 10 × 15 × 3 mm) were fixed in 4% paraformaldehyde and embedded in paraffin. For immunofluorescence studies, the specimens were embedded in OCT (optimal cutting temperature) compound during the freezing process.

One gram of feces or mucosal tissues from each animal was homogenized in 9 ml of sterile saline solution, and serial dilutions were then plated on XLT4 agar plates for Salmonella culture and MRS agar plates for lactobacilli culture (Beijing Land Bridge Technology Co., Beijing, China). XLT4 agar plates were incubated for 24 h at 37°C under aerobic conditions, whereas MRS agar plates were incubated under anaerobic conditions for 48 h at 37°C. Results are expressed as log₁₀ CFU/g feces or log₁₀ CFU/g mucosal tissues. All counts were performed in triplicate.

Quantitative PCR was performed as previously described, with minor modifications (Yang et al., 2016). The following primer sets were used: “all bacteria,” 5′-CGGTGAATACGTTCCCGG-3′, 5′-TACGGCTACCTTGTTACGACTT-3′; Escherichia coli: 5′-CATGCCGCGTGTATGAAGAA-3′, 5′-CGGGTAACGTCAATGAGCAAA-3′; Bifidobacterium: 5′-CGGGTGAGTAATGCGTGACC-3′, 5′-TGATAGGACGCGACCCCA-3′. The results for all samples were normalized to the 16S rRNA gene level of “all bacteria.”

The serum concentrations of IL-7 and IL-22 were determined by direct high-sensitivity sandwich ELISA specific for porcine IL-7 (LifeSpan Biosciences, Seattle, WA) or porcine IL-22 (LifeSpan Biosciences), respectively. The concentrations of TNF-α and IL-1β in mucosal tissues of the jejunum and ileum were measured by porcine-specific commercially available ELISA kits (Westang Biotechnology, Shanghai, China). The experimental procedure was based on the manufacturer's instructions, with minimal modifications.

Intestinal pathology was evaluated on hematoxylin & eosin-stained jejunal and ileal sections by a single blinded scorer using a validated scoring system (Zhou et al., 2015). Histologic grading was based on observed epithelial integrity, central lacteal expansion, leukocyte infiltration, submucosal edema, mucosal hyperemia, and lymphoid necrosis of PPs. Each segment was given a score ranging from 0 to 3. For epithelial integrity: grade 0, no change; 1, shedding of <10 epithelial cells per lesion; 2, shedding of 11–20 epithelial cells per lesion; and 3, epithelial ulceration. For central lacteal expansion: 0, no change; 1, mild; 2, moderate; and 3, profound. For leukocyte infiltration: 0, <10 leukocytes per field; 1, 11–15 leukocytes per field; 2, 16–20 leukocytes per field; and 3, >20 leukocytes per field. For submucosal edema: 0, no change; 1, mild; 2, moderate; and 3, profound. For mucosal hyperemia: 0, none; 1, mild; 2, moderate; and 3, severe. For lymphoid necrosis of PPs: 0, none; 1, mild; 2, moderate; and 3, severe. Scores for each criterion were added to give an overall inflammation score of 0–15 for jejunal samples or 0–18 for ileal samples.

Attachment of S. Infantis to the intestinal mucosa was determined by an indirect immunofluorescence assay. Tissue sections (6 μm) obtained from the jejunum and ileum were incubated with monoclonal rabbit anti-pig α-tubulin (1:200 dilution, ab52866; Abcam, Cambridge, UK) overnight at 4°C. The secondary antibody was goat anti-rabbit fluorescein isothiocyanate [FITC]-conjugated (ZF-0311; Zhongshan Golden Bridge Biotechnology, Beijing, China), and DAPI (4′,6′-diamidino-2-phenylindole; Sigma-Aldrich) was used for nuclear staining. The slides were visualized and photographed using a Nikon Eclipse Ti-U inverted fluorescence microscope equipped with a Nikon DS cooled camera head (Nikon).

Single-cell suspensions from peripheral blood and the intestine were isolated as previously described (Zhu et al., 2014). Cells were stimulated in vitro by incubation with 50 ng/ml of phorbol myristate acetate, 500 ng/ml ionomycin, and BD GolgiPlug (BD PharMingen) in complete RPMI 1640 medium at 37°C for 4 h. The monoclonal antibodies used were as follows: mouse anti-pig CD3ε (clone P2G10, FITC-conjugated, 559582; BD Biosciences), mouse anti-pig CD4α (clone 74-12-4, PerCP-Cy5.5-conjugated, 561474; BD Biosciences), mouse anti-human T-bet (clone O4-46, phycoerythrin [PE]-conjugated, 561268; BD Biosciences), and mouse anti-pig IFN-γ (clone P2G10, AlexaFluor 647-conjugated, 561480; BD Biosciences). Isotype controls PerCP-Cy5.5-conjugated mouse IgG2b, κ (558304; BD Biosciences), PE-conjugated mouse IgG1, κ (559320; BD Biosciences) and AlexaFluor 647-conjugated mouse IgG1, κ (557732; BD Biosciences) were included. The stained cells were analyzed on a FACScalibur™ flow cytometer (BD Biosciences), and data analysis was performed using FlowJo 9.3 software (Tree Star).

Total RNA was extracted from frozen tissues using an EASYspin plus RNA extraction kit (Aidlab Biotechnologies, Beijing, China) according to the manufacturer's instructions. Reverse transcription and real-time PCR were performed as previously described (Yang et al., 2016). Data were normalized to the geometric mean C_T values of three reference genes: β-actin, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and hypoxanthine phosphoribosyl-transferase. Results are presented as fold change using the 2^−ΔΔCT method, as previously described (Zhou et al., 2015). Primers are listed in Table S2.

Proteins were extracted from the jejunum and ileum using Radio-Immunoprecipitation Assay buffer (RIPA; Sigma-Aldrich, St. Louis, MO), as previously described (Yang et al., 2016). The primary antibodies used were polyclonal rabbit anti-pig IL-22 (1:1,500 dilution, ab193813; Abcam), polyclonal rabbit anti-human STAT3 (1:1,000 dilution, AP0365; Bioworld Technology, Nanjing, China), polyclonal rabbit anti-human p-STAT3 (phospho-S727) (1:1,000, AP0248; Bioworld Technology), monoclonal mouse anti-human IL-22RA2 (1:1,500 dilution, 66190-1-Ig; Proteintech, Chicago, IL), polyclonal rabbit anti-human IL-7Rα (1:5,000 dilution, ab115249; Abcam), and monoclonal mouse anti-GAPDH (1:2,000 dilution, 60004-1-Ig; Proteintech). Horseradish peroxidase-conjugated secondary antibodies used were goat anti-rabbit IgG (H+L) (SA00001-2; Proteintech), goat anti-mouse IgG (H+L) (SA00001-1;Proteintech), and rabbit anti-goat IgG (H+L) (ZB-2306; Golden Bridge Biological Technology). Immobilon Western chemiluminescent HRP substrate (Millipore, Billerica, MA) was used for immunoblot detection. The bands were visualized using a Tanon-5200 gel image system (Tanon, Shanghai, China). The intensity of bands was quantified by densitometric analysis using ImageJ software (National Institutes of Health, Bethesda, MD). Results are presented as the ratio of the intensity of the IL-22, IL-22BP, or IL-7Rα band to that of the GAPDH band or the ratio of the intensity of the p-STAT3 band to that of the STAT3 band.

The jejunal and ileal tissue samples embedded in paraffin were sectioned at 3-μm thickness, and immunohistochemistry was performed using the un-avidin-biotin complex technique, a highly sensitive system (Envision polymer-system, Biogenex, San Ramon, CA). Briefly, tissue sections were boiled in 10 mM citrate buffer (pH 6.0) for antigen retrieval. Slides were quenched with 3% hydrogen peroxide solution for 10 min at room temperature. After a washing procedure with phosphate-buffered saline (PBS), the slides were incubated with goat anti-pig CCL20 polyclonal primary antibody (1:400 dilution, MBS421719; MyBioSource, San Diego, CA) or mouse anti-human IL-22BP monoclonal primary antibody (1:1,000 dilution, 66190-1-Ig; Proteintech) in a humidified chamber overnight at 4°C. After 3 rinses in PBS, the slides were incubated with the secondary horseradish peroxidase-conjugated anti-goat IgG or anti-mouse IgG (unbiotinylated antibodies, Envision System, DakoCytomation) for 20 min at 37°C. Tissue staining was visualized with 3,3′-diaminobenzidine (Zhongshan Golden Bridge Biotechnology). Negative controls in which PBS and goat serum or mouse serum replaced the primary antibody during incubation were included in each batch. Images were captured using an Olympus BX41 microscope (Olympus, Tokyo, Japan) equipped with a Canon EOS 550D camera head (Canon, Tokyo, Japan). The detection of CCL20 was executed through a ranked score of 0–4, which was used to evaluate the number of positive cells per section taken from each block, as previously described (Li et al., 2014). The indications for the scores were as follows: 0 = no positive cells, 1 = 1–10 positive cells, 2 = 11–30 positive cells, 3 = 31–50 positive cells, and 4 = or > 50 positive cells. The number of CCL20-positive cells in the jejunum of each pig was scored from 10 fields at a magnification of ×400.

The SAS statistical software package, version 9.3 (SAS Institute Inc., Cary, NC) and the software's PROC MIXED procedure were used for statistical analyses. The incidence of diarrhea was analyzed using Pearson's chi-squared test. For non-normally distributed and repeated-measure data analysis, the non-parametric Friedman's test using the SAS procedure FREQ was performed to compare diarrhea scores between treatments. Moreover, the non-parametric Wilcoxon-Mann-Whitney U-test was performed to compare differences between treatments in histologic scores for the small intestine. Differences between least-square means were compared using Tukey's test. Data are presented as the mean ± standard error of the mean (SEM). P-values: ^*P < 0.05; ^**P < 0.01; ^***P < 0.001.

The rectal temperature in all pigs was 39.38 ± 0.31°C before S. Infantis challenge. On day 9 (day 2 after S. Infantis challenge), the rectal temperature in S. Infantis-challenged pigs without LGG pretreatment increased to 40.11 ± 0.14°C compared with CONT pigs (P = 0.011; Figure 1A). On days 16 and 17, pigs pretreated with LGG also exhibited a higher temperature than CONT pigs.

The diarrhea scores were monitored for 17 days. During week 1 after weaning, 33.33% (7/21) of piglets experienced naturally acquired severe diarrhea (weaning diarrhea) lasting for 2 days due to stress. However, the mean diarrhea scores were lower in pigs pretreated with LGG than in CONT pigs (P = 0.001; Table S3). Within 24 h following S. Infantis challenge (day 8), 3 pigs in the S. Infantis treated group experienced severe watery diarrhea (fecal scores ≥5). In contrast, pigs pretreated with LGG had lower diarrhea scores compared with pigs infected only with S. Infantis (P = 0.040; Figure 1B). On day 9 post-challenge, S. Infantis treated pigs continued to exhibit more severe diarrhea than CONT pigs (P = 0.012; Figure 1B). During S. Infantis infection, the incidence of diarrhea in S. Infantis treated pigs was higher than that in CONT pigs (P < 0.001; Table S3); however, a similar difference compared with CONT pigs was also observed in pigs pretreated with LGG (P = 0.001).

Pigs pretreated with LGG had fewer Salmonella in the mucosal tissues of the jejunum after S. Infantis infection, than did pigs only challenged with S. Infantis (P = 0.018, Figure 2A). However, no changes in the number of lactobacilli in jejunal and ileal mucosa were observed (Figure 2B). S. Infantis shedding occurred rapidly over a short time period (within approximately 2 days) (Figure 2C). By contrast, the number of lactobacilli in the feces of S. Infantis-challenged pigs was lower than in pigs pretreated with LGG and CONT pigs on day 16 (P = 0.028 and P < 0.001, respectively; Figure 2D). As expected, orally fed LGG increased the abundance of fecal Bifidobacterium compared with CONT and pigs infected only with S. Infantis on day 13 after infection (P = 0.018 and P = 0.034, respectively; Figure 2F). The amount of E. coli shed by pigs after S. Infantis infection was less than that of CONT pigs (P < 0.01; Figure 2G). No changes in the abundance of total bacteria in feces were detected in the present study (Figure 2E).

On day 18 (10 days after S. Infantis challenge), histologic analyses showed that infection with S. Infantis caused epithelial lesions, central lacteal expansion, substantial leukocyte infiltration, submucosal edema, and mucosal hyperemia in the small intestine, as well as lymphoid necrosis of PPs (Figure 3A). Histologic assessment of the jejunum and ileum from pigs infected only with S. Infantis revealed more severe inflammation relative to that of CONT pigs (P = 0.011 and P = 0.037, respectively; Figures 3B,C), whereas LGG pretreatment tempered the severity of Salmonella-associated inflammation.

Colonization by S. Infantis (mCherry) was observed in the intestinal mucosa but not in the liver or spleen (Figure 3D).

The production of TNF-α and IL-1β in the jejunal mucosa of pigs was increased by challenge with S. Infantis compared with the CONT (P = 0.031 and P = 0.026, respectively; Figures 3E,F), but not in that of pigs pretreated with LGG.

To examine the effect of LGG on modulating systemic T cell-mediated immunity to S. Infantis, we examined the percentage of CD3⁺ CD4⁺ IFNγ⁺ T cells in peripheral blood at 0, 24, and 192 h following S. Infantis challenge. At 24 h after S. Infantis challenge, the percentage of peripheral blood CD4⁺ IFNγ⁺ T cells in pigs without LGG pretreatment was higher than that in CONT pigs (P < 0.001; Figure 4A). At 192 h post-challenge with S. Infantis, the percentage of peripheral blood CD4⁺ IFNγ⁺ T cells was higher than CONT, with or without LGG pretreatment (P < 0.01 and P < 0.001, respectively).

To further explore the role of the transcription factor T-bet in regulating IFNγ expression among CD4⁺ T cells, CD3⁺ CD4⁺ T-bet^+/− IFNγ^+/− T cells were also examined, as described above. Representative results at 24 h after S. Infantis challenge are shown in Figure 4B. Remarkably, at 24 h after challenge, the percentage of CD4⁺ T-bet⁻ IFNγ⁺ T cells in the peripheral blood of pigs pretreated with LGG was lower than in pigs without pretreatment (P < 0.001; Figure 4C). Orally fed LGG induced an increase in the proportion of CD4⁺ T-bet⁺ IFNγ⁺ T cells in peripheral blood before and after S. Infantis challenge compared with CONT pigs (P < 0.05; Figure 4C). Interestingly, the percentage of CD4⁺ T-bet⁺ IFNγ⁻ T cells decreased at 24 h after S. Infantis challenge in pigs not fed LGG compared with CONT pigs (P = 0.013; Figure 4C).

We assessed changes in the proportion of CD4⁺ IFNγ⁺ T cells in the intestinal compartments, including the PPs, intraepithelial layer, and lamina propria of the jejunum and ileum. In the jejunum, the percentage of CD4⁺ IFNγ⁺ T cells from the PPs was significantly higher in pigs challenged with S. Infantis than in CONT pigs (P = 0.003), whereas this increase was attenuated by oral feeding of LGG (P = 0.001; Figure 5A). Similarly, the expression of IFNγ mRNA in the jejunum was downregulated in pigs pretreated with LGG compared with pigs challenged only with S. Infantis (P = 0.047; Figure 5C). Among jejunal lamina propria lymphocytes, there was a higher percentage of CD4⁺ IFNγ⁺ T cells in pigs pretreated with LGG than in CONT pigs (P = 0.007; Figure 5A). In the ileum, an expansion of CD4⁺ IFNγ⁺ T cells among intraepithelial lymphocytes was found in pigs infected with S. Infantis compared with CONT pigs (P < 0.05). Intriguingly, the expression of the transcription factor T-bet in the ileum was elevated in pigs pretreated with LGG compared with CONT pigs and pigs challenged only with S. Infantis (P = 0.048 and P = 0.05, respectively; Figure 5B).

IL-7 is essential for immune cell development and homeostasis. The levels of IL-7 in the serum of pigs pretreated with LGG were lower than those of CONT pigs at 96 h after S. Infantis challenge (P = 0.046; Figure 6A). The expression of IL-7 mRNA in the jejunum of pigs pretreated with LGG was lower than that in pigs only challenged with S. Infantis (P = 0.041; Figure 6B). The simultaneous expression of IL-7Rα (the receptor for IL-7) in the jejunum was higher in pigs challenged with S. Infantis compared with CONT pigs (P = 0.002), and this increase was attenuated by oral administration of LGG (P = 0.007; Figure 6C). No changes in the expression of IL-7 and IL-7Rα in the ileum were observed following S. Infantis challenge.

Levels of IL-22 in the serum of pigs infected only with S. Infantis were higher than in CONT pigs at 24 h after S. Infantis challenge (P = 0.040; Figure 7A). To determine whether LGG induces the activation of IL-22 signaling, we examined the expression of IL-22, pSTAT3, and IL-22BP in the small intestine. The expression of both IL-22 and pSTAT3 was elevated in ileal tissues of pigs pretreated with LGG in comparison with CONT pigs (P = 0.038 and P = 0.027, respectively; Figures 7B,C). However, this phenomenon was not observed in the jejunum (data not shown). Immunohistochemical analysis revealed that IL-22BP-expressing cells in the small intestine were scattered throughout the lamina propria (Figure 7D). IL-22BP expression increased in the ileum following S. Infantis challenge, compared with CONT, including those pretreated with LGG (P = 0.042 and P < 0.001, respectively; Figure 7E). Unexpectedly, pretreatment with LGG increased IL-22BP levels relative to S. Infantis alone (P = 0.006; Figure 7E).

CCL20 was localized primarily at the surface epithelium of the villi but was also present in the lamina propria and goblet cells, as evidenced by sporadic positive staining (Figure 8A). The number of CCL20-positive cells in the jejunum was increased after S. Infantis infection compared with the CONT (P = 0.015), and this increase was attenuated by LGG pretreatment (P = 0.032). Concordantly, the expression of CCL20 mRNA in the jejunum was downregulated in pigs pretreated with LGG compared with CONT pigs and pigs only challenged with S. Infantis (P = 0.038 and P = 0.022, respectively; Figure 8B). However, no changes in CCR6 mRNA expression were observed in the small intestine (Figure 8C).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.