The L. paracasei used in this study was isolated from koumiss made by the local herdsmen of Aluke’erqin qi, Inner Mongolia, China. Escherichia coli O₈ (E. coli O₈) isolated from the feces of calves presenting with diarrhea and was used to establish a mouse diarrhea model. All bacterial strains were stored at the College of Veterinary Medicine of Inner Mongolia Agricultural University. The L. paracasei was prepared from a 48-h static culture grown in deMan, Rogosa, and Sharpe (MRS) broth at 37°C. The E. coli O₈ was prepared from a 12-h shake culture raised in nutrient broth at 37°C.

Sixty male specific-pathogen-free Kunming mice (age 3–4 weeks, weight 18–22 g) were purchased from the Inner Mongolia Medical University Experimental Animal Center, Hohhot, China and housed in a sterile environment there (Product Number: SCXK2020-0001). The mice were maintained at an average temperature of 20°C ± 2°C; under a standard 12 h/12 h light/dark cycle (8:00 a.m. to 8:00 p.m.) and at 45% ± 10% relative humidity (RH). All mice were adaptively fed for 1 week before the experiments and had ad libitum access to a sterilized diet and water. Mouse feeding and the experimental procedures were performed in strict accordance with the provisions of the Experimental Care and Ethics Committee of Inner Mongolia Agricultural University (No. [2020]001).

The experimental scheme is shown in Figure 1. The mice had ad libitum food access and there was no difference among treatment groups in terms of food intake. The diarrhea models were established in the negative control (NC), antibiotic (CIP), and treatment (TE) groups by administering 0.2 mL/10 g E. coli O₈ (3 × 10¹³ CFU/mL) for 1 week. The control group (CK) was intragastrically administered 0.9% (w/v) physiological saline on days 1–14. The NC group was administered E. coli O₈ on days 1–7 and 0.9% (w/v) physiological saline on days 8–14. The CIP group was administered E. coli O₈ on days 1–7 and 125 μg/ml ciprofloxacin on days 8–14. The pretreated (PE) mice were administered 0.2 mL/10 g probiotic L. paracasei (3 × 10⁸ CFU/mL) thrice daily for 1 week followed by E. coli O₈ until day 14. The mice in the treatment (TE) group were administered E. coli O₈ on days 1–7 and then 0.2 mL/10 g probiotic L. paracasei thrice daily on days 8–14. The mode of administration is intragastric administration.

The mice were separated in cages and the floors were covered with filter paper to facilitate observation of the fecal smudges. At days 0, 2, 4, 6, 8, 10, 12, and 14, mouse body weights were recorded. On days of 0, 3, 7, 10, and 14, the rates of diarrhea (1) were calculated. The diarrhea indices (2) were calculated based on the stool rates (3) and stool levels (4) and the fecal scores (Table 1).

On days 8 and 15, six mice per group were sacrificed by intraperitoneal injection of 0.1% (w/v) pentobarbital sodium at a dose of 40 mg/kg. Postmortem examinations were then performed. The livers, spleens, lungs, and left and right kidneys were excised and weighed. The ilea were collected and sliced into two sections. One of these was stored in liquid nitrogen until western blot analysis and RT-PCR. The other was fixed in 4% (v/v) paraformaldehyde and used in histological and immunohistochemical (IHC) analyses. Blood samples were collected and the serum were separated and immediately stored at -80°C. The samples used to determine the cecal bacterial flora were collected and stored at -80°C until 16s rRNA sequencing analysis.

The serum DAO and zonulin levels were measured by rat DAO and zonulin enzyme-linked immunosorbent assay (Elabscience, Bethesda, MD, United States). All procedures were performed according to the manufacturers’ instructions. The tests were performed in triplicate.

The ilea were fixed in 4% (v/v) paraformaldehyde for 1 day and embedded in paraffin. Paraffin sections were prepared for 4 μm, dewaxed with water, stained with H&E, examined under a microscope, photographed, and scored according to Table 2.

The paraffin sections were recovered by heating to 98°C in 10 mM citrate buffer (pH 6.0) for 10 min. Endogenous peroxidase was blocked with 10% (v/v) H₂O₂ for 30 min. Non-specific antigens were blocked with serum at room temperature (20–25 °C) for 30 min. The paraffin sections were incubated with rabbit anti-occludin (1:100; DF7504; Affinity Biosciences Ltd., Liyang, China), rabbit anti-claudin-1 (1:100; AF0127; Affinity Biosciences Ltd., Liyang, China), and rabbit anti-ZO-1 (1:100; AF5145; Affinity Biosciences Ltd., Liyang, China) primary antibodies at 4 °C overnight and then treated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG secondary antibodies (1:200; GB23303; Wuhan Servicebio Technology Co., Ltd., Wuhan, China). Photographic images were acquired under an OLYMPUS microscope and analyzed with ImagePro Plus software (Media Cybernetics, Rockville, MD, United States).

Total RNA was extracted from the ilea with RNA extraction kit (AP-MN-MS-RNA-250; Axygen Biosciences; Union City, CA, United States) according to the kit manufacturer’s instructions was used to evaluated the expression levels of IL-6, IL-1β and TNF-α. Adjust the sample concentration to 400 μg/mL, then 500 ng total RNA was reverse-transcribed in 10 μL final volume with a PrimeScript RT Master Mix Kit (TaKaRa, RR036A, Japan), includes 5 × PrimeScript Master Mix 2 μL, total RNA 1.5 μL and RNase Free H₂O 6.5 μL. Reverse transcription proceeded at 37°C for 15 min, 85°C for 5 s, and 4°C for 10 min. RT-qPCR was performed with SYBR Green Master Mix in an Applied Biosystems 7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, United States). Then 100 ng total RNA amount of DNA template in 20 μL final volume with a Real-Time PCR Kit (TaKaRa, RR820A, Japan), the 20μL reaction system includes TB Green Premix Ex Taq II (Tli RNaseH Plus) 10 μL, PCR Forward Primer 0.8 μL, PCR Reverse Primer 0.8 μL, ROX Reference Dye II 0.4 μL, cDNA 2 μL, Nuclease-Free water 6 μL. The PCR program consisted of an initial denaturation step at 95°C for 30 s followed by 40 cycles of 95°C for 3 s and 60°C for 30 s. Relative mRNA expression was analyzed by the 2^{–ΔΔ Ct} method and normalized to β-actin expression levels. The primers used in the RT-PCR are listed in Table 3, all primers were synthesized by Sangon Biotech, ShangHai, China.

The total protein was extracted from ilea and then quantitated using BCA Protein Assay kit. Equal amounts of protein (20 μg) from different samples were separated by 6∼15% SDS-PAGE and transferred to PVDF membranes. The membranes were blocked with 5% skimmed milk for 4 h at room temperature, then probed with ZO-1 (1:1,000, AF5145, Affinity), claudin-1 (1:1,000, AF0127, Affinity), occludin (1:1,000, DF7504, Affinity), p65 (1:1,000, AF5002, Affinity), MLCK (1:1,000, AF5314, Affinity), MLC2 (1:1,000, 36725, CST), and β-actin (1:3,000, ab8226, Abcam) was used as loading control, overnight at 4°C and with secondary antibody at RT for 1 h. The images were captured using ChemiDoc MP imaging system (Bio-Rad).

Cecal contents were collected from 4 mice per group at days 8 and 15, for microbiome analysis. Microbial genomic DNA was extracted from fecal samples using the E.Z.N.A. soil DNA Kit (Omega Bio-Tek, Norcross, GA, United States) according to the manufacturer’s instructions. The DNA extract was checked on 1% agarose gel, and DNA concentration and purity were determined with NanoDrop 2000 UV-vis spectrophotometer (Thermo Fisher Scientific, Wilmington, United States). The hypervariable region V3-V4 of the bacterial 16S rRNA gene were amplified with primer pairs 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) by an ABI GeneAmp 9700 PCR thermocycler (ABI, CA, United States). The PCR amplification of 16S rRNA gene was performed as follows: initial denaturation at 95°C for 3 min, followed by 27 cycles of denaturing at 95°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 45 s, and single extension at 72°C for 10 min, and end at 4°C. The PCR mixtures contain 5 × TransStart FastPfu buffer 4 μL, 2.5 mM dNTPs 2 μL, forward primer (5 μM) 0.8 μL, reverse primer (5 μM) 0.8 μL, TransStart FastPfu DNA Polymerase 0.4 μL, template DNA 10 ng, and finally ddH₂O up to 20 μL. PCR reactions were performed in triplicate. The PCR product was extracted from 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, United States) according to manufacturer’s instructions and quantified using Quantus Fluorometer (Promega, United States). Purified amplicons were pooled in equimolar and paired-end sequenced on an Illumina MiSeq PE300 platform (Illumina, San Diego, United States) according to the standard protocols by Majorbio Bio-Pharm Technology (Shanghai, China). The raw 16S rRNA gene sequencing reads were demultiplexed, quality-filtered by fastp version 0.20.0 and merged by FLASH version 1.2.7 with the following criteria: (i) the 300 bp reads were truncated at any site receiving an average quality score of < 20 over a 50 bp sliding window, and the truncated reads shorter than 50 bp were discarded, reads containing ambiguous characters were also discarded; (ii) only overlapping sequences longer than 10 bp were assembled according to their overlapped sequence. The maximum mismatch ratio of overlap region is 0.2. Reads that could not be assembled were discarded; (iii) Samples were distinguished according to the barcode and primers, and the sequence direction was adjusted, exact barcode matching, 2 nucleotide mismatches in primer matching. Operational taxonomic units (OTUs) with 97% similarity cutoff were clustered using UPARSE version 7.1, and chimeric sequences were identified and removed. The taxonomy of each OTU representative sequence was analyzed by RDP Classifier version 2.2 against the 16S rRNA database (Silva v138) using confidence threshold of 0.7.

Statistical significance analyzed using the software GraphPad Prism. Student’s t-test to determine the levels of significance for comparison between two groups and statistical significance among more than two groups was calculated using one-way ANOVA with Tukey’s test. P-values of less than 0.05 were considered significant (*P < 0.05; ^**P < 0.01; ^***P < 0.001; ^****P < 0.0001; “ns,” no significant). The results were reported as mean ± SD. Spearman’s correlation between the bacteria and the diarrhea-related parameters was conducted using the programming language R (version 4.0.2).

Compared with CK group, the groups administered E. coli O₈ presented with significant reductions in body weight (Figure 2A), and increased diarrhea rates (Figure 2B) and indices (Figure 2C). Hence, the mouse diarrhea model was successfully established. The E. coli administrate caused diarrhea (Figure 2D) followed by intestinal congestion and edema (Figure 2E). Compared with the NC group, the PE group showed significant increase in body weight and decreases in the diarrhea rates and indices. Thus, the L. paracasei treatment helped prevent E. coli O₈-induced diarrhea. Compared with the NC group, the TE group presented with lower diarrhea rates and indices but no significant change in body weight after L. paracasei administration. On day 8, the NC, CIP, and TE groups presented with significantly lower spleen and lung indices than the CK and PE groups (Figure 2F). By day 15, there were no significant differences among groups in terms of their liver, spleen, lung, left kidney, and right kidney indices (Figure 2G). Therefore, L. paracasei efficiently treated E. coli O₈-induced diarrhea.

DAO is a highly active structural enzyme in the intestinal epithelium (31) and its blood levels reflect intestinal barrier integrity. Figure 2H shows that on day 8, the DAO level was significantly higher in the E. coli-treated group than the CK group (P < 0.01) but there were no significant differences between the PE and CK groups (P > 0.05) in terms of DAO level. On day 15, in the TE group DAO significantly downregulated (P < 0.01). The blood DAO levels were higher for the CIP than the NC group (P < 0.01). After E. coli intervention, the DAO content of the PE group was significantly lower than those of the TE and NC groups (P < 0.01) but higher than that of the CK group (P < 0.05).

Zonulin is the main regulator of the epithelial TJ and, by extension, the intestinal barrier. Figure 2I shows that the E. coli treatment significantly upregulated the serum zonulin level (P < 0.01). On day 15, the zonulin content was significantly lower in the PE and TE groups than the NC group after L. paracasei intervention (P < 0.01). However, the zonulin content was still higher in the TE group than the CK group (P < 0.05). The zonulin content was still significantly higher in the CIP group than the CK group (P < 0.01).

In addition, we also found the levels of DAO and Zonulin in the NC group on day 15 were still higher than those in the PE group, although this value was lower than the value after 7 days of E. coli administrate, it also seemed to indicate that E. coli would affect intestinal permeability for a longer time, while the content of DAO and Zonulin in L. paracasei increased after the intervention was changed to E. coli challenge, but it was not significant compared with the CK group, which indicated that L. paracasei had a certain protective effect on intestinal damage caused by E. coli, and long-term intervention of L. paracasei may have more profound benefits on the body’s intestinal barrier.

H&E staining of the ileal tissue (Figure 3) disclosed that E. coli O₈ administration damaged the intestinal villi and increased the numbers of lymphocytes in the lamina propria and epithelium. In contrast, prophylactic or therapeutic L. paracasei administration restored the intestinal villi. Hence, L. paracasei administration alleviated the E. coli O₈-induced diarrhea pathogenesis. Although L. paracasei could prevent diarrhea, on day 15, the intestinal villi were nonetheless slightly damaged after E. coli administration. This finding suggests that L. paracasei may require long-term administration for optimal efficacy (Figure 4).

In histological morphology observation, after 15 days of intervention in each group, the ileal length of the CK group was 17.5 cm, NC group was 15.6 cm, CIP group was 15 cm, PE group was 20.5 cm, and TE group was 15 cm. There was no significant growth after L. paracasei treatment, but comparable to the NC and CIP groups, the intestinal tissue was not swollen. In the HE staining observation, we observed a very small amount of inflammatory infiltration in the TE group and observed intact intestinal villi with more goblet cells, which was consistent with the ileal morphology of the TE group in Figure 2E.

Immunohistochemistry (IHC) (Figure 5) and western blotting (WB) (Figure 6) were used to investigate the roles of L. paracasei in diarrhea and determine whether its probiotic action depends on TJ protein expression. Figures 5A,C,E, shows IHC staining of the ileum tissue indicated that E. coli treatment groups NC, CIP, and TE presented with occludin, claudin-1 and ZO-1 protein downregulation. These effects were largely reversed in response to prophylactic L. paracasei administration. This treatment kept the TJ proteins and the ileal architecture intact (Figures 5B,D,F). WB confirmed that E. coli downregulated the TJ protein whereas L. paracasei alleviated the damage E. coli caused to the TJ (Figure 6A). The antibiotic therapy also upregulated the TJ Proteins (Figure 6B). Therefore, L. paracasei administration reinforced the TJ which, in turn, played an important role in the antidiarrheal efficacy of the probiotic.

To explore the effects of L. paracasei on diarrhea, we orally administered L. paracasei or saline to the mice thrice daily for 1 week before or after they were subjected to E. coli O₈-induced injury. Using RT-PCR, we identified the relative differences in ileal IL-6, IL-1β, and TNF-α expression. As shown in Figure 7 E. coli administrated upregulated IL-6, IL-1β, and TNF-α in the NC, CIP, and TE group (Figures 7A–C), L. paracasei treatment before E. coli O₈ administration did not upregulate IL-6, IL-1β, or TNF-α compared with the CK group (P > 0.05), however, L. paracasei treatment after E. coli O₈ administration significantly downregulated IL-6, IL-1β, and TNF-α in the TE group compared with the NC group (P < 0.01). There were no significant differences between the CIP and NC groups in terms of their IL-6, IL-1β, and TNF-α expression levels (P > 0.05) (Figures 7D–F).

We used WB to analyze the NF-κB-MLCK pathway and determine whether the observed changes in IL-6, IL-1β, and TNF-α were associated with the NF-κB pathway (Figures 8A,B). The E. coli administration activated the NF-κB-MLCK pathway and upregulated NF-κB, p65, MLC2, and MLCK. The L. paracasei administration did not upregulate p65, MLC2, or MLCK (Figure 8A). After L. paracasei therapy, though, p65, MLC2, and MLCK were downregulated but there were no significant decreases in their expression levels in either the TE or the CK group (Figure 8B). Thus, L. paracasei administration was more efficacious prophylactically than therapeutically.

A 16S rRNA analysis was performed on the mouse cecal contents to investigate the intestinal microbiota after L. paracasei administration. Figures 9A,C,D shows that the Chao1 and Shannon indices indicated significant differences between the E. coli O8-administered groups and the CK and PE groups in terms of their intestinal microbiota α-diversity (Figures 9A,C). However, the PE group had a significantly lower Simpson index than the other groups (Figure 9E). The Shannon and Simpson indices were significantly increased in response to L. paracasei administration (Figures 9D,F), but there were no significant differences among the CK, NC, PE, and TE groups in terms of these parameters. The CIP group presented with significantly lower Chao1 and Shannon indices than the other groups (Figures 9B,D). The Simpson indices significantly increased following ciprofloxacin administration (Figure 9F).

We performed a principal coordinate analysis (PCoA) based on a weighted UniFrac distance to analyze the β-diversity of the gut microbial composition. We found significant differences between the E. coli-administered and CK groups in terms of their gut microbiome β-diversity. The diarrhea model mice pretreated with L. paracasei and the CK group differed in terms of their β-diversity (Figure 9G). On day 15, the β-diversity of the groups administered L. paracasei was similar to that of the CK group. β-diversity was similarly elevated for both the PE and TE groups. β-diversity was significantly different between the group administered ciprofloxacin and the CK group (Figure 9H).

The potential compositions of the gut microbiota differed among the five groups. At the phylum level, all of them presented with similar taxonomic communities and relatively high abundances of Firmicutes and Bacteroidota. Compared with CK and PE groups, the NC, CIP, and PE group exhibited significantly reduced Bacteroidota and increased Proteobacteria and Firmicutes/Bacteroidetes ratios (Figure 10A). These changes were reversed in the E. coli O₈-administered groups treated with L. paracasei and ciprofloxacin (Figure 10B). Ciprofloxacin administration lowered relative gut microbiota diversity and increased the abundances of Firmicutes, Bacteroidota, Verrucomicrobiota, and Campylobacteriota (Figure 10C).

We also analyzed the taxonomic communities at the genus level (Figures 11A,B). Enterobacter significantly increased after E. coli O₈ administration. Oral L. paracasei administration decreased Enterobacter and increased Lactobacillus. The gut microbiota comprised Lactobacillus, Dubosiella, Akkermansia, and Anaerosporobacter after ciprofloxacin administration (Figure 11C).

A heatmap was plotted for Spearman’s correlations among gut microbiota abundance and the diarrhea-related indices (body weights, diarrhea rates and indices, liver, spleen, lung, and left and right kidney indices, DAO, zonulin, occludin, claudin-1, ZO-1, IL-6, IL-1β, and TNF-α).

The clustering analysis in Figure 12 shows that the bacterial phyla were clearly divided into three groups (Figure 12A). The phyla in group A were negatively correlated with body weight and the TJ proteins but positively correlated with the diarrhea rates and indices, the organ indices, DAO, zonulin, and the proinflammatory factors. Bacterial abundance was significantly increased in the E. coli treatment group (Figure 11A). Therefore, an increase in any of these phyla could promote diarrhea. Conversely, the phyla in group B were positively correlated with body weight and the TJ proteins but negatively correlated with the diarrhea rates and indices, the organ indices, DAO, zonulin, and the proinflammatory factors (Figures 11A,B). Therefore, an increase in any of these phyla could inhibit diarrhea.

Similar phenomena were also observed at the genus level. Figure 12C shows that the genera were divided into five groups. Those in group A were positively correlated with body weight and TJ proteins but negatively correlated with the diarrhea rate and indices, DAO, zonulin, and the proinflammatory factors. The genera in group B were not significantly correlated with the diarrhea-related indices. The genera in group C were positively correlated with the diarrhea rates and proinflammatory factors but not significantly correlated with the diarrhea-related indices. The abundances of these genera were correlated with IL-6, IL-1β, and TNF-α. The genera in group D were negatively correlated with the TJ proteins and the proinflammatory factors but positively correlated with the liver and spleen indices. The genera in group E were positively correlated with the diarrhea rates and indices, DAO, zonulin, and the proinflammatory factors but negatively correlated with body weight, the liver index, and the TJ proteins. Thus, any increase in these genera has therapeutic efficacy against diarrhea.