H. pylori P12 and H. pylori P12-GFP strains were provided by the Max Planck Institute for Infection Biology and H. pylori SS1 (ATCC 43504) (Lee et al., 1997) was provided by the University of Western Australia (UWA), Australia. H. pylori strains were cultured on Columbia agar containing 7% sterile defibrinated sheep blood (Bianzhen, Nanjing, China), 20μg/ml vancomycin (Meilunbio, Dalian, China), 10μg/ml polymyxin (Meilunbio), 10μg/ml amphotericin B (Meilunbio), 10μg/ml trimethoprim (Sigma, St. Louis, USA), then placed in an incubator containing 5% O₂ and 10% CO₂, cultured at 37°C, subcultured once every three days, and used for the experiment after subculture. The tested lactobacilli were provided by Danisco China (Shanghai, China): Lactobacillus acidophilus NCFM (ATCC 7003969), L. acidophilus La-14 (ATCC SD5212), Lactiplantibacillus plantarum Lp-115 (ATCC SD5209), Lacticaseibacillus paracasei Lpc-37 (ATCC SD5275), Lacticaseibacillus rhamnosus Lr-32 (ATCC SD5217) and L. rhamnosus GG (ATCC 53103). After the Gram Staining Kit (Solarbio, Beijing, China) was used to identify the bacterial morphology, the lactobacilli were cultured anaerobically in MRS broth (Solarbio) at 37 °C 48 hours and then subcultured ( Figure S1 ).

The human gastric adenocarcinoma cell line (AGS) was purchased from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences (Shanghai, China) and cultured in RPMI 1640 medium (Thermo Fisher, Waltham, MA, USA) supplemented with 10% fetal bovine serum at 37°C and 5% CO₂ in a humidified incubator.

Fifty male C57BL/6 mice, Specific Pathogen Free (SPF), four weeks old, were purchased from Zhejiang Vital River Laboratory Animal Technology Co. Ltd. (Zhejiang, China). All the animals were housed under standard conditions (SPF grade animal room with individually ventilated cages; temperature range from 23°C to 25°C, humidity range from 50% to 60%, 12/12 hours light/dark cycle, food and water were provided ad libitum). The experimental steps and ethics were approved by the ethics committee of the Fifth Affiliated Hospital of Zhengzhou University (KY2022002).

After infection, the culture medium was collected by centrifugation at 12000x rpm for 5 min to remove cell debris and bacteria and collect the supernatant. The concentration of interleukin (IL)-8 and tumor necrosis factor (TNF)-α were detected by human IL-8 ELISA Kit (Elabscience, Wuhan, China) and human TNF-α ELISA Kit (Elabscience, Houston, TX, USA) respectively, by following the guidelines of the manufacturer. All experiments were performed in triplicate.

H. pylori P12 and lactobacilli were cultured overnight in BHI (Thermo Fisher) and MRS broth. After centrifugation at 5000 x rpm for 8 min and 4000 x rpm for 5 min, the supernatant was discarded, and bacteria were harvested and resuspended in 1 ml serum-free RPMI 1640 medium.

Overnight cultured AGS cells were co-incubated with lactobacilli (multiplicity of infection (MOI) = 100) and H. pylori P12 (MOI = 100) for 6h. After the incubation period, the supernatant was harvested for ELISA and cells were harvested for RNA isolation.

AGS cells were cultured in two 12 well plates and incubated with lactobacilli (MOI = 100) and H. pylori P12-GFP (MOI = 100) for 6h. At the end of incubation, one plate of the cells was washed thrice with PBS and photographed by fluorescence and white light. A urease detection reagent was prepared by adding concentrated hydrochloric acid into the PBS (pH=7.4) to adjust the solution to pH=6.8. Urea (Solarbio) and phenol red (Solarbio) were weighed and added to reach concentrations of 110mmol/L and 10mg/L, respectively, and then dissolved by vigorously shaking. To the other 12well plates, 1 ml urease detection reagent was added. After 1-7 hours of reaction, 80μl medium was withdrawn to record the absorbance value at 540nm(Shmuely et al., 2004; Rokka et al., 2008; Tharmalingam et al., 2014; Yang et al., 2020).

AGS cells and mouse gastric mucosal tissue were lysed with RNAiso plus (TaKaRa, Kyoto, Japan), and gastric mucosal tissue needed to be assisted by an ultrasonic crusher. According to the manufacturer’s instructions, cDNA was converted using the ReverTra Ace qPCR RT Kit (TOYOBO, Shanghai, China). qRT-PCR was performed using 2×ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) in a Roche Lightcycler480II system based on the manufacturer’s recommendations. Primer sequences were designed in NCBI Primer-BLAST ( Table 1 ). The relative gene expression was determined using the 2^-ΔΔCt method. All experiments were repeated thrice.

SPF C57BL/6 mice (male 4 weeks old) were obtained from Zhejiang Vital River Laboratory Animal Technology Co., Ltd. (Zhejiang, China) and fed on Laboratory Rodent Diet 5001. The mice were randomly divided into 5 groups (Uninfected, H. pylori SS1, H. pylori SS1+L. acidophilus NCFM, H. pylori SS1+L. plantarum Lp-115, H. pylori SS1+L. acidophilus NCFM and L. plantarum Lp-115) of 10 individuals each after one week of adaptation. The control group was gavaged with PBS, and the experimental groups were gavaged with H. pylori SS1 (1*10⁹ CFU, 0.2ml/piece) only or combined with the corresponding lactobacilli (1*10⁹ CFU, 0.2ml/piece) (1:1) or together (1:1:1) once every other day for 6 weeks ( Figure 1A ).

Mouse gastric mucosa was analyzed for urease activity using the rapid urease Kit (Sanqiang, Fujian, China) as instructed by the manufacturer. The color change was determined, when the solution turns pink or red, the urease test is positive; when it remains yellow, the urease test is negative.

Gastric tissues were fixed with 4% paraformaldehyde, embedded in paraffin, cut into 4mm sections, and stained by HE and immunohistochemical methods. The Chronic gastritis histological grade scale was used to evaluate the samples (Fang et al., 2018). Five histological changes were graded: H. pylori, chronic inflammation, activity, atrophy, and intestinal metaplasia. Each histological change was classified into one of four grades: none, mild, moderate, and severe. According to the new Sydney system, the degree of inflammation and lymphocyte infiltration of gastric tissue after H. pylori infection were evaluated (Kim et al., 2020; Maluf et al., 2020). Immunohistochemistry for H. pylori was performed using a Rabbit anti-H. pylori polyclonal antibody (Cell Marque) (ZSGB, Beijing, China), and the colonies of H. pylori in mouse gastric mucosa were identified.

All statistical analyses were performed using GraphPad Prism 9 software. During the processing of experimental data, the values whose deviation from the average value of the same group of data exceeded three times the standard deviation were considered outliers and eliminated. We performed one-way ANOVA on raw and lg-converted data to compare the multi groups. The Bonferroni and Tukey tests were conducted to calculate the statistical significance among the groups. P-value < 0.05 was considered significant for all statistical analyses.

To compare the effect of the six selected lactobacilli interfering with H. pylori adhesion, AGS cells were co-infected with H. pylori P12-GFP and the six lactobacilli strains (MOI=100) respectively for 6 hours. After infection, cells and GFP-positive H. pylori were measured under fluorescence microscopy. Upon comparative observation under fluoroscopy, the amount of GFP-positive H. pylori in the lactobacilli groups was less than in the H. pylori only infected group. Furthermore, the morphology of AGS cells was relatively normal in the L. acidophilus NCFM and L. plantarum Lp-115 treated cells compared with the other lactobacilli treated cells, indicating less cell stress on the AGS cells compared with other lactobacilli strains ( Figure 2A ).

To further quantify the inhibition effect between different lactobacilli, we used a urease activity assay for the co-infection plate and recorded the absorbance at 540nm. The absorbance values of the six intervention groups were decreased to different degrees from 1 to 7 hours compared with that of the H. pylori infected group. For further quantification analysis, the absorbance values at each group’s 7h time point were taken into a bar chart for statistical analysis. The absorbance values of the seventh hour were statistically analyzed. The data suggest that samples of L. acidophilus NCFM, L. acidophilus La-14, L. plantarum Lp-115 and L. rhamnosus GG had significant differences compared with samples of H. pylori (P<0.05), indicating that these four lactobacilli strains can effectively reduce the colonization of H. pylori ( Figure 2B ).

To compare the inhibitory effect between the six probiotic strains, we performed a co-infection model of H. pylori and the strains (MOI=100) in AGS cell culture for 6 hours. We analyzed the mRNA and protein levels of IL-8 (Cxcl8) and TNF-α (TNF) to determine whether these probiotic strains transcriptionally regulate the inflammatory markers. The expression of these markers is commonly altered upon H. pylori infection. Expression of Cxcl8 and TNF mRNA with all six probiotic strains was significantly reduced compared with the H. pylori infected group (P<0.001). In Cxcl8, Lp-115 showed significant difference than other probiotics except NCFM, while pattern of NCFM was significantly lower than La-14 and Lpc-37 (P<0.01). The comparison of other probiotics including La-14, Lpc-37, Lr32 and GG did not show any significant difference (P>0.01). In TNF, only Lpc-37 was found to have significant difference than other probiotics in the group (P<0.01) ( Figure 3A ).

Consistent with the mRNA results, IL-8 and TNF-α in the cell supernatant were significantly decreased in the probiotic treatments compared with the H. pylori-infected group (P<0.001). In IL-8, NCFM was significantly lower than La-14, while Lp-115 showed significant difference than the La-14 and Lpc-37 (P<0.01). In TNF-a, the inhibitory effect of the six selected probiotic strains did not show significant differences ( Figure 3B ). Combined with the anti-adhesion results, L. acidophilus NCFM and L. plantarum Lp-115 had a good effect and less cell stress. Therefore, L. acidophilus NCFM and L. plantarum Lp-115 were further selected for validation in the H. pylori infected mouse model.

To further validate whether L. acidophilus NCFM and L. plantarum Lp-115 alone or in combination can counteract H. pylori colonization and attenuate gastric inflammation in vivo, C57BL/6 mice were infected with H. pylori SS1 and co-administered L. acidophilus NCFM and L. plantarum Lp-115 with for 6 weeks ( Figure 1A ). After co-administration the mice were euthanized, and the gastric tissues were assessed for H. pylori infection by the rapid urease test ( Figure S2 ). The H. pylori adhesion on gastric tissues was then analyzed by immunohistochemistry (IHC) assays ( Figure 1B ), which showed that L. acidophilus NCFM and/or L. plantarum Lp-115 intervention groups had comparably less H. pylori adhesion than the H. pylori infected group. To further validate the H. pylori colonization in different groups, the mRNA expression of ureA and ureB was tested from gastric tissues of all groups ( Figure 1C ), which showed that L. acidophilus NCFM and/or L. plantarum Lp-115 intervention groups had less expression of ureA and ureB compared with H. pylori infected group (P<0.001). These results show that L. acidophilus NCFM and L. plantarum Lp-115 alone or combined can reduce H. pylori colonization on gastric mucosa in mice.

To further investigate whether L. acidophilus NCFM and L. plantarum Lp-115 can attenuate H. pylori colonization and gastric inflammation in vivo, gastric tissues from the five study groups were analyzed by Hematoxylin-Eosin staining (HE stain). Compared with the uninfected group, the gastric lamina propria of mice in the H. pylori infected group showed more lymphocyte, plasma cell, and neutrophil infiltration in the active phase. Incidentally, the mice also had local thinning of the mucosal layer, reduction of the glands propria, and thickening of the muscularis mucosae ( Figure 4A ). Compared with the H. pylori group, the inflammatory cells infiltrating the lamina propria of the gastric mucosa in the H. pylori + NCFM group, H. pylori + Lp-115 group, and H. pylori+NCFM+Lp-115 group were reduced to different degrees, which indicated that L. acidophilus NCFM and L. plantarum Lp-115 could ameliorate the H. pylori induced gastric inflammation ( Figure 4A ).

To further verify if NCFM and/or Lp-115 can inhibit H. pylori induced Th1 type inflammation, the mRNA expression of Il4 (IL-4), Cxcl15 (CXCL15), Il10 (IL-10), and Ifng (IFN-γ) was measured by qRT-PCR from mouse gastric mucosal tissue. The results showed that NCFM and Lp-115 reduced the expression of Ifng and promoted the expression of Il4 induced by H. pylori in C57BL/6 mice. In Il4 expression, the difference between the Lp-115 and H. pylori groups was the most significant (P<0.001). In Cxcl15, only the Lp-115 group was significantly different from the H. pylori group (P<0.01); therefore, the overall anti-inflammatory effect of Lp-115 was more pronounced than that of NCFM. The expression of Cxcl15 in the group receiving the combination of Lp-115 and NCFM was not significantly different from that in the H. pylori group. The expression of Il4 was lower than that in the Lp-115 group ( Figure 4B ), indicating that the two LAB strains had no compound effect in improving inflammation caused by H. pylori infection. In conclusion, NCFM and/or Lp-115 can reduce H. pylori induced Th1 type inflammation (Ifng expression) in C57BL/6 mice and tend to transform Th2 type inflammation (Il4 expression).