Evaluation of Potential Probiotic Properties of Limosilactobacillus fermentum Derived from Piglet Feces and Influence on the Healthy and E. coli-Challenged Porcine Intestine

In this work, we evaluated the probiotic properties of Limosilactobacillus fermentum strains (FL1, FL2, FL3, FL4) isolated from feces of healthy piglets. The in vitro auto-aggregation, hydrophobicity, biofilm-forming capacity, survival in the gastrointestinal tract, antimicrobial activity and anti-oxidation capacity were evaluated. Four strains were resistant to simulated gastrointestinal conditions, including low pH, pepsin, trypsin and bile salts. They also maintained strong self-aggregation and cell surface hydrophobicity. Limosilactobacillus fermentum FL4, which had the strongest adhesion ability and antimicrobial effect on Enterotoxigenic Escherichia coli K88 (ETEC K88), was then tested in porcine intestinal organoid models. The in vitro experiments in basal-out and apical-out organoids demonstrated that L. fermentum FL4 adhered to the apical surfaces more efficiently than basolateral surfaces, had the ability to activate the Wnt/β-catenin pathway to protect the mucosal barrier integrity, stimulated the proliferation and differentiation of the intestinal epithelium, and repaired ETEC K88-induced damage. Moreover, L. fermentum FL4 inhibited inflammatory responses induced by ETEC K88 through the reduced expression of pro-inflammatory cytokines (TNF-α, IL-1β and IFN-γ) and higher levels of anti-inflammatory cytokines (TGF-β and IL-10). These results show that L. fermentum FL4 isolated from feces of healthy Tunchang piglets has the potential to be used as an anti-inflammatory probiotic and for mitigation of intestinal damage in piglets.


Introduction
Many microorganisms are considered to be probiotics that have beneficial effects on their hosts. Lactic acid bacteria (LAB) have been shown to be effective probiotics in swine production [1][2][3]. LAB consist of Aerococcus, Carnobacterium, Enterococcus, Lactobacillus, Lactococcus, Leuconostoc, Oenococcus, Pediococcus, Streptococcus, Tetragenococcus, Vagococcus, and Weissella [4]. Huang et al. [5] showed that the LAB count in native pig feces was much higher than in wild pig feces. Additionally, Wang et al. [6] indicated the strains of Lactobacillus plantarum ZLP001 isolated from the gastrointestinal tracts of healthy pigs modulated their gut microbiota. Since the natural microbiota stabilized in the intestine easily and propagated to a specific species rapidly [7], the isolates from homologous hosts themselves would be more effective probiotics than isolates from other sources. Limosilactobacillus fermentum is the one widely presented in nature, often isolated from natural fermented plants [8], bread [9], dairy products [10], breast milk [11], saliva [12], or human and animal feces [13][14][15]. The inherent characteristics of L. fermentum accelerate the improvement and development of probiotic preparations. However, L. fermentum is understudied compared with other species in the genus Lactobacillus. Existing research has Ten wells in a sterile 96-well plate were filled with 200 µL bacterial suspension [32]. Negative control wells had only medium added. The 96-well plate was incubated in an anaerobic chamber at 37 • C for 24 h. Then, suspensions were removed from all wells and the wells were washed three times. Fixation, 99% methanol, 15 min; Staining, 2% Hucker crystal violet, 2 min; Dissolving, 33% (v/v) glacial acetic acid. The absorbance of every well was measured at 630 nm (A).
The 24 h-cultures were inoculated into MRS medium (pH = 2.0, pH = 3.0) at 3% (v/v) inoculation, then incubated at 37 • C for 3 h. Samples were taken at 0 h and 3 h and counted on MRS agar.
Acid tolerance (%) = [(Log10 CFU/mL (3 h))/(Log10 CFU/mL (0 h))] × 100 The cultures were centrifuged at 6000 rpm for 10 min and washed with sterile PBS (pH = 6.5) at least twice. Then, the bacterial precipitations were resuspended in sterile saline with similar absorbance (A 600 = 0.60 ± 0.02). The suspensions (1 mL) were added into artificial gastric juice (9 mL). The artificial gastric juice was prepared by adding pepsin (Macklin, Shanghai, China) in 0.2% (w/v) sterile saline at a concentration of 3.5 g/L, and the pH was adjusted to 2.0 with HCl. Samples were taken at 0 h and 3 h and counted on MRS agar. The suspensions above (1 mL) were mixed with artificial intestinal juice (9 mL). The artificial intestinal juice was prepared by using 22 g/L NaHCO 3 , 2 g/L trypsin (Macklin, China), 3 g/L bile salt (Solarbio, Beijing, China), and 4.4 g/L NaCl, and the pH was adjusted to 8.0 with NaOH. Artificial gastric and intestinal juices were filtered through a 0.22 µm sterile filter membrane. Samples were taken at 0 h and 3 h and counted on MRS agar. The bacterial survival rate was calculated as follows: Gastric survival rate (%) = [Log10 CFU/mL (3 h))/(Log10 CFU/mL (0 h))] × 100 Intestinal survival rate (%) = [(Log10 CFU/mL (3 h))/(Log10 CFU/mL (0 h))] × 100 Agar diffusion method was used to detect the antimicrobial activity of L. fermentum strains against ETEC K88 strain. 10 mL LB medium containing 20 µL ETEC K88 was poured into sterile petri dish and dried in the clean bench. The sterilized oxford cup was evenly placed on the solidified medium with sterile tweezers. The ends of the tweezers were tapped slightly to ensure the oxford cup was in close contact with the medium. In the oxford cup, 50 µL L. fermentum was added, with PBS as negative control and neomycin sulfate as positive control. After the sample was added, all plates were transferred to an incubator and cultured at 37 • C for about 18 h. The inhibition zone diameters were measured with calipers.
Total antioxidant capacity (TAC) of supernatants was measured using the total antioxidant capacity assay kit-FRAP method (Jiancheng, China). Total antioxidant capacity was represented by the concentration of FeSO 4 standard solution (mmol/gprot).

Porcine Intestinal Organoid Cultivation
The porcine intestinal organoids were kindly provided by Professor Xiaoliang Li at Zhejiang University. The cryopreserved organoids were placed in a 37 • C water bath for 2 min and then contents were transferred into DMEM/F-12 (ThermoFisher, Waltham, MA, USA) with 1% BSA (Biofroxx, Hesse, Germany). The suspensions were centrifuged at 200× g for 5 min to obtain the cell pellets, which were further resuspended in 100 µL complete organoid culture medium (STEMCELL Technologies, Vancouver, BC, Canada) and seeded in 100 µL Matrigel (BD Bioscience, Franklin, NJ, USA) on a prewarmed 24-well culture plate (Corning, Coring, NY, USA). Matrigel was incubated for 15 min at 37 • C to polymerize. Finally, 500 µL complete organoid culture medium supplemented with 1% antibiotics-antimycotic (10,000 IU/mL Penicillin + 10,000 µg/mL Streptomycin + 25 µg/mL Amphotericin B) (Beyotime, Shanghai, China) was added to every well. Intestinal organoids were cultured at 37 • C in a 5% CO 2 chamber. Culture medium was replaced every 2 days and the organoids were passaged every 5 or 7 days based on the growth of organoids. By dissociating the organoids with gentle cell dissociation reagent (GCDR) (STEMCELL Technologies, Vancouver, BC, Canada) on a shaker at 37 • C for 15 min, cells were obtained for seeding in fresh Matrigel. All the experiments were started after at least two passages.

Establishment of Intestinal Apical-Out Organoids
Suspension culture of intestinal organoids was established as described in [33] with modifications. After passaging, organoids were cultured in Matrigel for 24 h, then the medium containing approximately 200 organoids was aspirated from the wells. The Matrigel in each well was broken up using 500 µL of ice-cold 5 mM EDTA/PBS (Gibco, Waltham, MA, USA; Servicebio, Wuhan, China). The solution from every 3 wells was transferred into a 15 mL sterile tube containing 10 mL of ice-cold 5 mM EDTA/PBS, which further dissolved the Matrigel on a shaker for 1 h at 4 • C. The organoid precipitates were collected by centrifugation at 200× g for 5 min and then washed with 5 mL DMEM (Gibco, USA). They were centrifuged again as above, the pellets were resuspended in complete organoid culture medium and cultured in a 24-well ultra-low-attachment plate. The apicalout organoids were cultured in suspension in a 5% CO 2 incubator at 37 • C. Culture medium was replaced every 3 days.

Treatment of Intestinal Organoids
Before the experiment, the bacterial concentration was determined by measuring the optical density at 600 nm. The cells were harvested by centrifugation (1200× g, 10 min), washed twice with 1× sterile PBS, resuspended in DMEM/F12 without antibiotics (final concentration 2 × 10 6 CFU/mL). 500 µL DMEM/F12 was added to each well to meet the requirement of 10 6 CFU bacteria per well. The organoids were cultured for 5 days. To detect whether the L. fermentum selected from in vitro probiotic property tests had a repairing effect on damage organoids, organoids and L. fermentum (10 6 CFU) were co-cultured in advance for 24 h. Then, the organoids were treated with ETEC K88 (10 6 CFU) for 1 h to induce injury. The samples were collected for further experiments at mRNA and protein levels. To verify that the L. fermentum had no negative effects on organoids, organoid samples were collected separately after the 24 h co-culture of organoids and L. fermentum (10 6 CFU). To observe the damage caused by ETEC K88 to organoids, organoids were co-cultured with ETEC K88 (10 6 CFU) for 1 h [33,34].
The growth conditions of organoids were observed under the microscope and images were obtained from different experimental treatments.

Epithelial Barrier Integrity
Apical-out organoids were pelleted in a 2 mL tube and resuspended in a solution of 4 kDa FITC-Dextran (diluted in DMEM-phenol red-free). The suspension was cultured at room temperature for 5 min, then the 4 kDa FITC-Dextran (Sigma-Aldrich, USA) was aspirated, leaving only 10 µL at the bottom. A circle with a diameter of 2cm was drawn on the slide using a Hydrophobic Barrier Pen (PAP pen) to create a well-like region, preventing the samples from flowing randomly. Then, the 10 µL solution was transferred to the center of the circle. To build a 3D-space for keeping the organoid shape, vacuum grease was applied to the four corners of the coverslip with gentle pressing and ultimately played a role in fixation. The slides were immediately observed with a laser confocal microscope (IX81-FV1000, Olympus, Shinjuku, Japan) and the fluorescence images were obtained using FV10-ASW software (version 4.2).

Immunofluorescence Staining
All samples were fixed in 4% paraformaldehyde for 30 min at room temperature, then centrifuged and washed with D-PBS (Servicebio, China) three times (5 min/time). Organoids were permeabilized with 0.5% Triton X-100 (Servicebio, China) and sealed with 3% BSA solution. The primary antibody solution (1:100-1:1000 dilution) (anti-Wnt3a mouse antibody, Thermo Fisher Scientific, Waltham, MA, USA; anti-villin mouse antibody, anti-Ki67 mouse antibody, anti-β-catenin mouse antibody, anti-ZO-1 mouse antibody, Servicebio, China; anti-Lgr5 mouse antibody, Origene technologies, Rockville, MD, USA) were added and the mixture was incubated overnight at 4 • C, then centrifuged and washed with D-PBS. The organoid pellets were resuspended in secondary antibody solution and phalloidin staining solution (1:500 dilution) (iFluor TM 594 Conjugated Goat anti-mouse IgG Antibody, Huabio, Hangzhou, China), and incubated for 2 h at room temperature without light. 10 µL DAPI (2 µg/mL, Servicebio, China) staining solution was added and they were incubated for 10 min at room temperature without light. After the centrifuging and washing, organoids were resuspended in 10 µL Vectashield mounting medium (Servicebio, China). The 10 µL solution was transferred to the center of the PAP pen circle and vacuum grease was applied to corners of the coverslip with gentle pressing. The slides were immediately observed under a laser confocal microscope and the fluorescence images were obtained using FV10-ASW software.

qRT-PCR
Basal-out organoids and apical-out organoids were collected after experiments. Total RNA was extracted from intestinal organoids and L. fermentum (FL4) strain by Trizol assay. RNA was further reverse-transcribed into cDNA using PrimeScript RT Master Mix (Perfect Real Time) (Takara) following the manufacturer's instructions. All cDNA were analyzed on LightCycler 480 Real-time PCR System with the TB green Premix Ex Taq (Tli RNaseH Plus, Takara) and specific primer (Table 1). Reaction conditions: 95 • C, 5 s; 40 cycles of 95 • C, 5 s and 60 • C, 30 s; 95 • C, 5 s; 60 • C, 1 min; 95 • C, 5 s; 50 • C, 30 s. GAPDH was used as endogenous gene for porcine intestinal organoids, while phenyl alanyl synthase (pheS) was used as control gene for L. fermentum. Assays were performed in triplicate, and the relative expression of target gene was analyzed based on the 2 −∆∆Ct method.

Metagenomic Sequencing Analysis
Total genomic DNA was extracted from feces samples using the E.Z.N.A. ® Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's instructions. Concentration and purity of extracted DNA was determined with TBS-380 and NanoDrop2000, respectively. DNA extract quality was checked on 1% agarose gel.
DNA extract was fragmented to an average size of about 400 bp using Covaris M220 (Gene Company Limited, Hong Kong, China) for paired-end library construction. Pairedend library was constructed using NEXTFLEX Rapid DNA-Seq (Bioo Scientific, Austin, TX, USA). Adapters containing the full complement of sequencing primer hybridization sites were ligated to the blunt end of fragments. Paired-end sequencing was performed on Illumina Novaseq 6000 (Illumina Inc., San Diego, CA, USA) at Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) using NovaSeq Reagent Kits according to the manufacturer's instructions (www.illumina.com, accessed on 13 November 2020). Sequence data associated with this project are still being analyzed and are not yet public.
At the outset of analyzing metagenomic sequencing data, thorough splitting and cutting was needed to ensure the raw data was of high quality. The next step was assembly, which involved merging overlapping metagenomic reads into longer contiguous DNA contigs. The assembly software MEGAHIT (version 11), based on the principle of the De-Brujin graph, was used. According to the overlapping relationship between kmers, the De-Brujin graph was constructed to obtain contigs. Contigs of more than 800 bp were screened for statistics and subsequent analysis.
Open reading frames (ORFs) from each assembled contig were predicted using Prodigal/MetaGene (http://metagene.cb.k.u-tokyo.ac.jp/, accessed on 5 October 2006, version 1.15.1). The predicted ORFs with a length ≥ 100 bp were retrieved and translated into amino acid sequences using the NCBI translation table. A non-redundant gene catalog was constructed using CD-HIT (http://www.bioinformatics.org/cd-hit/, accessed on 1 September 2009, version 4.6.1) with 90% sequence identity and 90% coverage. High-quality reads were aligned to the non-redundant gene catalogs to calculate gene abundance with 95% identity using SOAPaligner (http://soap.genomics.org.cn/, accessed on 1 March 2008, version 2.21).

Statistical Analysis
The data from experiments were expressed as the mean ± SD. One-way ANOVA was evaluated to determine significant differences among multiple groups, followed by Bonferroni multiple comparisons or Dunnett's T3 multiple comparisons (SPSS 26.0). Significant differences between two groups were analyzed by the t-test (SPSS 26.0) and data among three groups were evaluated by one-way ANOVA. p < 0.05 and p < 0.001 were considered statistically significant.

Isolation and Identification of Limosilactobacillus fermentum Strains
Through metagenomic sequencing (Supplementary Figure S1), it was found that the abundance of Lactobacillus and Limosilactobacillus fermentum was significantly higher in Tunchang pigs (TC) than in Duroc × Landrace × Yorkshire (DLY). We isolated and purified four LAB species from healthy Tunchang piglet feces after incubating under anaerobic conditions at 37 • C for 48 h. The strains were Gram-positive and rod-shaped, which was consistent with the morphological characteristics of Lactobacillus (Supplementary Figure S1). As shown in Figure 1, all four isolates (FL1, FL2, FL3, FL4) were identified as Limosilactobacillus fermentum by genotype identification and showed 99% homology to the same strains. The conservation and homology of the isolated strains were observed by multiple sequence alignment. MEGA11 (molecular evolutionary genetics analysis) software was used to perform unambiguous alignment of the 16S rRNA gene sequences of the four bacteria identified in this study to construct a phylogenetic tree. Different color codes are used for each nucleotide to show the homology and differences in the obtained sequences. considered statistically significant.

Isolation and Identification of Limosilactobacillus fermentum Strains
Through metagenomic sequencing (Supplementary Figure S1), it was found that the abundance of Lactobacillus and Limosilactobacillus fermentum was significantly higher in Tunchang pigs (TC) than in Duroc × Landrace × Yorkshire (DLY). We isolated and purified four LAB species from healthy Tunchang piglet feces after incubating under anaerobic conditions at 37 °C for 48 h. The strains were Gram-positive and rod-shaped, which was consistent with the morphological characteristics of Lactobacillus (Supplementary Figure  S1). As shown in Figure 1, all four isolates (FL1, FL2, FL3, FL4) were identified as Limosilactobacillus fermentum by genotype identification and showed 99% homology to the same strains. The conservation and homology of the isolated strains were observed by multiple sequence alignment. MEGA11 (molecular evolutionary genetics analysis) software was used to perform unambiguous alignment of the 16S rRNA gene sequences of the four bacteria identified in this study to construct a phylogenetic tree. Different color codes are used for each nucleotide to show the homology and differences in the obtained sequences.

Tolerance of Limosilactobacillus fermentum to Gastrointestinal Tract Conditions
Following the FAO/WHO guidelines (Guidelines for the Evaluation of Probiotics in Food (2002) World Health Organization), probiotic bacteria must be safe and tolerant of gastrointestinal tract (GIT) conditions (including low pH, pepsin, pancreatin, and bile salt) The situations at pH 2 and 3 were examined in this study. The tolerance of L. fermentum to low pH is illustrated in Table 2. All strains showed higher viability when incubated within 3 h in MRS medium at pH 2 and 3, although the survival at pH 2 decreased after 3 h. Four L. fermentum strains except L. fermentum FL2 had a survival rate over 90% when cultured at pH 3, while FL2 had good viability at pH 2, second only to L. fermentum FL4. The experiment to simulate gastric conditions included the acid resistance tests in order to investigate the combined effects of low pH and pepsin. Similarly, four L. fermentum all survived in the simulated gastric juice (pH 2.0) after 3 h, among whom FL4 had the highest survival rate of 59.95%, followed by FL2. Data were expressed as the mean ± SD (n = 3). Different lowercase letters (a, b, c) in the same column indicated significant differences, p < 0.05.
Simulation of intestinal conditions investigated the effects of bile salt and pancreatin on the growth of bacteria, and the bacterial viability under the influence of the simulated intestinal juice is presented in Table 2. After being processed by artificial gastric juice for 3 h, L. fermentum was also tolerant of the simulated intestinal juice and showed good survival rates. After 3 h incubation, FL4 showed the highest survival (73.57%), followed by L. fermentum FL3 (70.83%). According to the results above, the strain FL4 had good resistance to low pH and the gastrointestinal tract environment.

Adhesion Ability between Different Limosilactobacillus fermentum Strains
Can a probiotic bacterium effectively exert its probiotic effect based on the capacity to adhere to the intestinal mucosal surface? The cell surface hydrophobicity, auto-aggregation and biofilm-forming capacity are positively correlated with adhesion ability, and can be considered to be indicators of the adhesion level of probiotics [35]. The four L. fermentum strains showed different degrees of hydrophobicity, varying from 53.90% to 90.16%. The L. fermentum FL4 strain showed the highest hydrophobicity, followed by L. fermentum FL2, showing hydrophobicity of 86.67% (Figure 2A). De souza et al. reported cell surface hydrophobicity as an important physicochemical characteristic that could facilitate the first contact between probiotics and intestinal epithelial cells [9,42]. Furthermore, the hydrophobicity also affected the auto-aggregation of bacteria [43]. Corresponding to the hydrophobicity results, the auto-aggregation test presented similar results, where FL4 was at the highest level and FL2 was second. In addition, there was an upward trend in the degree of auto-aggregation of L. fermentum over time. FL4 showed the fastest speed of auto-aggregation (13.07%) in just one hour, but FL2 (58.74) and FL3 (58.71%) presented higher degrees of auto-aggregation after 4 h culture ( Figure 2B,C). Aggregation is an important feature for biofilm formation, so we further quantified the biofilm-forming ability of L. fermentum. From Figure 2D, the biofilm-forming ability of FL4 was 2.34 at the middle level but the highest of the four L. fermentum strains, as the other three strains were at a weak level.

Antimicrobial and Anti-Oxidation Assays
Antimicrobial activity is an important characteristic for probiotics because the antimicrobial substances they produce can antagonize the invasion of harmful bacteria. The antimicrobial effect of L. fermentum on ETEC K88 was explored using its supernatant and suspension. The results in Figure 3A reveal that the supernatant of L. fermentum had inhibitory effects on ETEC K88; among all the strains, FL4 showed the largest-diameter zone of inhibition (13.91 mm). The presence of organic acids, hydrogen peroxide and bacteriocins in the supernatant lead to the inhibitory effects [35]. We measured the pH of four bacterial solutions in advance, and they were all around 4, creating a low-pH environment. The quantitative analysis of biofilm-forming capacity. Quantitative detection of biofilms: ODc = Negative Control (average) + 3 × SD. A < ODc, do not have the ability to form biofilms; ODc < A ≤ 2ODc, weak biofilmforming ability; 2ODc < A ≤ 4ODc, moderate biofilm-forming ability; A ≥ 4ODc, strong biofilm-forming ability. Data were expressed as the mean ± SD (n = 10). * p < 0.05, ** p < 0.01. The quantitative analysis of biofilm-forming capacity. Quantitative detection of biofilms: ODc = Negative Control (average) + 3 × SD. A < ODc, do not have the ability to form biofilms; ODc < A ≤ 2ODc, weak biofilm-forming ability; 2ODc < A ≤ 4ODc, moderate biofilm-forming ability; A ≥ 4ODc, strong biofilm-forming ability. Data were expressed as the mean ± SD (n = 10). * p < 0.05, ** p < 0.01.
Antimicrobial activity is an important characteristic for probiotics because the antimicrobial substances they produce can antagonize the invasion of harmful bacteria. The antimicrobial effect of L. fermentum on ETEC K88 was explored using its supernatant and suspension. The results in Figure 3A reveal that the supernatant of L. fermentum had inhibitory effects on ETEC K88; among all the strains, FL4 showed the largest-diameter zone of inhibition (13.91 mm). The presence of organic acids, hydrogen peroxide and bacteriocins in the supernatant lead to the inhibitory effects [35]. We measured the pH of four bacterial solutions in advance, and they were all around 4, creating a low-pH environment. Previous studies indicated that LAB had the antioxidant capacity to clean the reactive oxygen species (ROS) in the intestine and further keep the stable state of ROS in vivo [44]. L. fermentum produced a large amount of glutathiones that were natural antioxidants; in this study, total antioxidant capacity was represented by the concentration of FeSO4 standard solution. The standard curve and the results shown in Figure 3B show that antioxidant capacity was mainly in the supernatant rather than the bacterial suspension. The antioxidant capacity of the bacterial suspension ranged from only 0.14 to 0.24 mmol/gprot, while the supernatant of FL4 had the strongest antioxidant capacity (3.45 mmol/gprot), followed by FL1 (3.21 mmol/gprot) and FL2 (3.04 mmol/gprot). In general, the four L. fermentum Previous studies indicated that LAB had the antioxidant capacity to clean the reactive oxygen species (ROS) in the intestine and further keep the stable state of ROS in vivo [44]. L. fermentum produced a large amount of glutathiones that were natural antioxidants; in this study, total antioxidant capacity was represented by the concentration of FeSO 4 standard solution. The standard curve and the results shown in Figure 3B show that antioxidant capacity was mainly in the supernatant rather than the bacterial suspension. The antioxidant capacity of the bacterial suspension ranged from only 0.14 to 0.24 mmol/gprot, while the supernatant of FL4 had the strongest antioxidant capacity (3.45 mmol/gprot), followed by FL1 (3.21 mmol/gprot) and FL2 (3.04 mmol/gprot). In general, the four L. fermentum strains all demonstrated good anti-oxidation performance and have the potential to be antioxidants.

L. fermentum FL4 Influenced the Proliferation of Basal-Out Intestinal Organoids
The porcine intestinal organoids were successfully recovered and passaged two times. Shown in Supplementary Figure S2, the organoids grew slowly and asynchronously when they were just resuscitated; after two passages, the organoids matured within five days with good budding and the growth rate was significantly accelerated ( Figure 4A). To assess the positive effects of L. fermentum FL4 on the intestinal epithelium, the organoids and L. fermentum FL4 (10 6 CFU) were co-cultured on the third day for 24 h. The surface area of intestinal basal-out organoids increased significantly (p = 0.049) when treated with L. fermentum FL4 ( Figure 4B). The significant increase (p = 0.042) in Lgr5 mRNA expression indicated the L. fermentum FL4 treatment might stimulate intestinal epithelial proliferation ( Figure 4D). Considering Wnt/β-catenin signals are essential in intestinal epithelium homeostasis, correlation genes were further examined. L. fermentum FL4 significantly upregulated the mRNA expression of Wnt3a (p = 0.047), and the protein level of β-catenin was significantly increased compared with the control ( Figure 4F).

L. fermentum FL4 Alleviated the Intestinal Epithelial Damage Induced by ETEC K88 in Basal-Out Intestinal Organoids
In addition to evaluating the stimulatory effects of L. fermentum under normal conditions, we also determined whether L. fermentum had a repair function on the damage caused by ETEC K88 in intestinal organoids ( Figure 5A). Significant reduction (p = 0.012) of ZO-1 was observed in Figure 5C that indicated the mucosal barrier damage. Moreover, compared to the control, the villin and ZO-1 protein levels were decreased in organoids treated by ETEC K88 (Figure 5C,D). However, L. fermentum ameliorated the loss of ZO-1 to maintain the mucosal integrity ( Figure 5F). Goblet cells and Paneth cells are secretory cells differentiated from stem cells in the intestine, and Mucin-2 (Muc2) and lysozyme (Lyz) are their respective markers. ETEC K88 stimulation significantly downregulated the mRNA expression of Muc2 (p = 0.015) and Lyz (p = 0.011) ( Figure 5G). Moreover, the Ki67 mRNA levels (p = 0.011) and Ki67 + cell numbers were decreased (p = 0.036). Furthermore, L. fermentum reversed the reduced expression of Ki67 after ETEC K88 treatment ( Figure 5B,E). Compared to the ETEC K88 alone-treated group, L. fermentum significantly enhanced mRNA and protein expression of β-catenin (p = 0.048). However, expression of wnt3a remained unchanged among three treatments ( Figure 6A,B,D). The maintenance of ISCs was also further verified by the increase in the Lgr5 + cell number (p = 0.025) and Lgr5 mRNA expression levels (p = 0.012) in organoids treated with L. fermentum FL4 ( Figure 6C,E).
We also analyzed the expression of adhesion genes such as fbp, mub and sor in L. fermentum; ETEC K88 stimulation significantly decreased fbp (p = 0.003), mub (p = 0.012) and sor (p = 0.026) expression ( Figure 6F). The downregulation of adhesion genes may affect mitigation by L. fermentum of the intestinal epithelial damage in basal-out organoids because of less direct contact between L. fermentum and epithelial tips.

Establishment of Apical-Out Intestinal Organoids and Effects of L. fermentum FL4
It is challenging to study the interaction between the epithelial apical surface and microbes in the intestinal lumen in conventional organoid models. The 3D spheroids were a closed system where outside substances had contact with the epithelial base, such as the nutrients and microbes added in the culture medium [29]. Julia et al. attempted to reverse the polarity of the organoids so that the apical surface turned outward, establishing the apical-out model [45]. After the suspension culture, observation under the microscope ( Figure 7A) showed the morphology of organoids had changed. Organoids embedded in Matrigel had a clear lumen, while in the suspended organoids the edge of the columnar epithelial cells could be seen, often without a lumen ( Figure 7B). Confocal immunofluorescence imaging of F-actin revealed that the epithelial apical surface turned outward successfully ( Figure 7B).   We also analyzed the expression of adhesion genes such as fbp, mub and sor in L. fermentum; ETEC K88 stimulation significantly decreased fbp (p = 0.003), mub (p = 0.012) and sor (p = 0.026) expression ( Figure 6F). The downregulation of adhesion genes may affect mitigation by L. fermentum of the intestinal epithelial damage in basal-out organoids because of less direct contact between L. fermentum and epithelial tips.

Establishment of Apical-Out Intestinal Organoids and Effects of L. fermentum FL4
It is challenging to study the interaction between the epithelial apical surface and microbes in the intestinal lumen in conventional organoid models. The 3D spheroids were a closed system where outside substances had contact with the epithelial base, such as the nutrients and microbes added in the culture medium [29]. Julia et al. attempted to reverse the polarity of the organoids so that the apical surface turned outward, establishing the apical-out model [45]. After the suspension culture, observation under the microscope ( Figure 7A) showed the morphology of organoids had changed. Organoids embedded in We hypothesized that L. fermentum would stimulate the proliferation of the intestinal epithelium more efficiently in the apical-out model than in the basal-out model. The organoids and L. fermentum FL4 (10 6 CFU) were co-cultured on the third day for 24 h, and the surface area of intestinal apical-out organoids increased significantly in L. fermentum FL4 treatment ( Figure 7C). As expected, the treatment of L. fermentum resulted in the significant upregulation of mRNA expression of Ki67 (p = 0.004) and Lgr5 (p = 0.045) ( Figure 7D). There was also an increase of villin (p = 0.025) and Muc2 (p = 0.044) at mRNA levels ( Figure 7E,G). The Wnt/β-catenin pathway was activated in organoids treated by L. fermentum with significantly increasing mRNA expression of Wnt3a (p = 0.006), which was also confirmed by the obvious increase of β-catenin (p = 0.006) mRNA expression ( Figure 7F).
Matrigel had a clear lumen, while in the suspended organoids the edge of the columnar epithelial cells could be seen, often without a lumen ( Figure 7B). Confocal immunofluorescence imaging of F-actin revealed that the epithelial apical surface turned outward successfully ( Figure 7B).

L. fermentum FL4 Repaired the ETEC K88-Induced Damage to the Intestinal Epithelium in Apical-Out Intestinal Organoids
To evaluate whether apical-out intestinal organoids can be applied as effective models for probing infection, ETEC K88 is used as a bacterial pathogen to invade intestinal epithelial cells ( Figure 8A). The dextran diffusion assay was used, and we found that infection of apical-out organoids with ETEC K88 resulted in compromised barrier integrity. Imaged under the fluorescence confocal microscope, the organoids in the control completely excluded the FITC-dextrans, while in those treated with ETEC K88 the FITC-dextrans diffused into the inside of the 3D spheroids ( Figure 8B). The decreased mRNA expression of ZO-1 (p = 0.001) and protein expression of villin further verified tight junctions disruption of barrier integrity ( Figure 8D,F). Additionally, we explored whether the L. fermentum could more effectively alleviate the ETEC K88-induced damage to the intestinal epithelium in apical-out intestinal organoids and maintain the integrity of the intestinal epithelial barrier. Pretreated by L. fermentum, the expression levels of ZO-1 (p = 0.001) and villin (p = 0.022) were significantly increased compared to the control and the ETEC K88 alone-treatment ( Figure 8F). Moreover, L. fermentum prominently increased Ki67 (p = 0.001) as well as Ki67 + cells ( Figure 8C,E). Compared to the organoids treated with ETEC K88 alone, L. fermentum significantly upregulated Wnt3a expression (p = 0.003), which was consistent with the significant increase (p = 0.005) in β-catenin ( Figure 9A,B,D). These results indicated that L. fermentum activated the Wnt/β-catenin pathways to maintain inhibition caused by ETEC K88 and then enhanced the proliferative and differentiative ability of ISCs to repair the damaged epithelia. The significant increase in Lgr5 expression (p = 0.001) and Lgr5 + cell number ( Figure 9C,D) in organoids treated with L. fermentum FL4 also proved this. ETEC K88 markedly decreased lysozyme expression (p = 0.041) compared with the control, indicating Paneth cell dysfunction. However, L. fermentum FL4 reversed the reduced expression of lysozyme ( Figure 8C).  Expression of adhesion genes such as fbp (p = 0.004), mub (p = 0.004) and sor (p = 0.026) was significantly increased both in groups of L. fermentum alone-treated and L. fermentum pretreated with ETEC K88 attack ( Figure 9F), indicating that the adhesion of L. fermentum was less affected by ETEC K88 invasion. We speculated that the particular structure of apical-out intestinal organoids provides the potential to enhance adhesion of probiotics to intestinal epithelial cells. Expression of adhesion genes such as fbp (p = 0.004), mub (p = 0.004) and sor (p = 0.026) was significantly increased both in groups of L. fermentum alone-treated and L. fermentum pretreated with ETEC K88 attack ( Figure 9F), indicating that the adhesion of L. fermentum was less affected by ETEC K88 invasion. We speculated that the particular structure of apical-out intestinal organoids provides the potential to enhance adhesion of probiotics to intestinal epithelial cells.

L. fermentum (FL4) Modulated ETEC K88-Induced Inflammatory Responses in the Intestinal Epithelium
To verify the hypothesis that L. fermentum mitigates pathogen-induced inflammatory responses by modulating cytokines, intestinal epithelial cells were challenged with ETEC K88. Exposed to ETEC K88, the basal-out organoids dramatically enhanced the transcription of pro-inflammatory cytokines TNF-α, IL-6, IL-1β and IFN-γ, individually compared to the control as showed in Figure 8A. Preprocessing with L. fermentum alleviated the ETEC K88induced inflammatory responses by the significant reduction of TNF-α, IL-1β and IFN-γ expression. Additionally, the transcriptional expression of anti-inflammatory markers IL-10 and TGF-β were significantly reduced after attack by ETEC K88. Furthermore, organoid interaction with L. fermentum tremendously increased the mRNA expression of regulatory cytokines (IL-10 and TGF-β) in comparison to ETEC K88-attacked organoids ( Figure 10B). It is worth mentioning that the treatment with L. fermentum significantly increased IL-10 expression. Similar results were observed in experiments with apical-out organoids. ETEC K88-mediated higher pro-inflammatory cytokine (TNF-α, IL-6, IL-1β and IFN-γ) levels were significantly reduced upon incubation with L. fermentum ( Figure 10C). Lower transcription of IL-10 and TGF-β was noticed when ETEC K88 stimulated the intestinal epithelial cells but markedly enhanced when L. fermentum was co-cultured in advance ( Figure 10D).

Discussion
The intestinal microbiota of native pigs is rich and diverse; through multi-omics technology, it has been found that the abundance of Firmicutes is higher in native pigs than in introduced pigs [46]. Therefore, dominant strains such as Lactobacillus and Bacillus could be specifically excavated to study their effects on pigs. Zhang et al. [47] obtained Lactobacillus from the intestine of Rongchang pigs, which could be used to develop probiotics and promote stress resistance. Wang [48] isolated Bacillus subtilis MZ-01 from Min pig feces. It has a good effect on reducing the diarrhea of piglets and inhibiting the expression of inflammatory factors. We found that the abundance of Lactobacillus and Limosilactobacillus fermentum was significantly higher in Tunchang pigs (TC) than in Duroc × Landrace × Yorkshire (DLY) by metagenomic sequencing. In this study, four L. fermentum strains were isolated from Tunchang piglet feces, then the probiotic properties were tested under the FAO/WHO criteria for probiotics (2002) [49].
One desirable property of potential probiotics is surviving under GIT conditions (tolerance to low pH, pepsin, pancreatin and bile salts). The low pH in the stomach is a challenge for bacterial survival, especially since the pH at 2.0 and below often significantly decreases their viability. Our study showed that four L. fermentum strains survived in low pH, artificial gastrointestinal juice within 3 h. Different results in survival rates of L. fermentum were relative to different sources [9,50].
Another essential property in probiotics is the ability to adhere to intestinal mucosa. Hydrophobicity and auto-aggregation are used as primary markers for selecting probiotics, which affect adhesion mechanisms and pathogen colonization [51]. Four strains of L. fermentum showed high auto-aggregation, ranging from 93.12% to 94.47%. Aggregation appeared to be specific to each strain and may vary in the same species [52]. Additionally, the results of cell surface hydrophobicity ranged from 53.90% to 90.16% when the non-polar solvent was toluene.
Antioxidant capacity is one of the main characteristics of LAB. The antioxidant capacity of the fermentation supernatant was always much stronger than the strain itself, which was consistent with Shen et al. [53]. This may be due to the metabolic components of the fermented strain, such as extracellular polysaccharides [54] and organic acids [55], having strong antioxidant capacity [56]. Additionally, the fermentation supernatant of FL4 showed the best inhibitory effect on ETEC K88. Research has revealed that the mechanism of killing Gram-negative bacteria is based on the production of organic acids, hydrogen peroxide and hydroxy fatty acids [57]. Low pH could help the passage of organic acids through the membrane to the intracellular environment.
Organoids have been considered to be effective models for studying the interplay between host and bacteria due to their close relationship to tissues/organs. The pathogenesis of porcine epidemic diarrhea virus (PEDV) [58], Salmonella [59] has been explored in intestinal organoid models. Also, studies of the regulatory effect of microbiota (L. reuteri [60], Akkermansia muciniphila [61]) on ISCs and the interaction with them have been reported. ETEC K88 is the principal microorganism responsible for bacterial diarrhea in pigs, which commonly colonizes in the jejunum and releases enterotoxins to impair the intestinal barrier function and trigger an inflammatory reaction [62]. FL4 reduced the damage induced by ETEC K88; The increase of villin and ZO-1 mRNA expression as well as the number of Ki67 + and Lgr5 + cells [60] indicated FL4 alleviated intestinal epithelial damage. Wu et al. [60] reported that L. reuteri D8 maintained the activation of the Wnt/βcatenin pathway to guarantee the proliferation of intestinal epithelia. From results in our study, FL4 significantly enhanced expression of β-catenin. We speculate that epithelial cell organoids with basolateral surfaces out reduced the adhesion of L. fermentum. Several cell surface-associated proteins associated with mucus and enterocytes have been identified by genomic approaches, namely mub, fbp, S-layer proteins, LTA and exopolysaccharides [63]. ETEC K88 stimulation significantly reduced the expression of fbp, mub and sor.
Co et al. [45] established apical-out polarity of organoids that preserve epithelial integrity, maintain secretory and absorptive functions and allow regulated differentiation.
After co-culture, the mRNA level of villin, Muc2, Ki67 and Lgr5 significantly increased, and the Wnt/β-catenin pathway was activated. Moreover, FL4 reduced damages induced by ETEC K88 to the intestinal barrier, reflected by the significant increase in ZO-1, villin, Lyz, Ki67, Lgr5, Ki67 + and Lgr5 + cells. The Wnt/β-catenin pathway is crucial for the maintenance of intestinal crypt proliferation [64,65]. L. fermentum FL4 activated the Wnt/βcatenin pathway with increased expression of Wnt3 and β-catenin to support intestinal epithelial proliferation and differentiation. The upregulation of adhesion genes fbp, mub and sor confirm the enhanced adhesion of L. fermentum to apical-out organoids. In summary, L. fermentum FL4 adhered to the apical surfaces more efficiently than to basolateral surfaces, which accelerated the Wnt/β-catenin signal activation to protect the mucosal barrier integrity and repair the ETEC K88-induced damage.
Cytokine production is considered to be an important indicator in response to ETEC K88. ETEC K88 significantly increased the expression of TNF-α, IL-1β and IFN-γ, while FL4 remarkably reduced these effects. Duary et al. [66] also reported lower levels of TNF-α, IL-8 and IFN-γ in L. plantarum+LPS treatment. These results suggest that FL4 intervention could control ETEC K88-induced inflammatory responses. Exposed to L. fermentum, FL4 enhanced the expression of TGF-β and IL-10 in organoids challenged by ETEC K88. Lactobacillus amplified TGF-β production through NF-κB signaling and inhibited inflammatory responses [67], which confirmed the anti-inflammatory properties of Lactobacillus.

Conclusions
In conclusion, we obtained four L. fermentum strains isolated from feces of healthy Tunchang piglets, and they exhibited probiotic properties, including resistance to gastrointestinal conditions, adhesion ability, antimicrobial activity against ETEC K88 and antioxidant capacity. Further in vitro experiments in basal-out and apical-out organoids demonstrated that L. fermentum adhered to the apical surfaces more efficiently than to the basolateral surfaces, had the ability to activate the Wnt/β-catenin pathway to protect the mucosal barrier integrity, stimulated the proliferation and differentiation of the intestinal epithelium, and repaired the ETEC K88-induced damage. Moreover, L. fermentum inhibited inflammatory responses induced by ETEC K88 through the reduced expression of proinflammatory cytokines (TNF-α, IL-1β and IFN-γ) and higher levels of anti-inflammatory cytokines (TGF-β and IL-10). Further in vivo experiments and clinical studies of pigs are required to ascertain the safety and efficacy of Tunchang pig-derived probiotics.

Data Availability Statement:
The whole genome sequence of the four tested strains has been deposited at GeneBank ID SUB12926851, with accession numbers OQ552786 to OQ552789.