Strain Lp05 was provided by Wecare Probiotics Co., Ltd. (Suzhou, China). The strain was cultured in De Man, Rogosa, and Sharpe (MRS) medium at 37°C for 18 h (20). Bacterial cells were harvested by centrifugation at 6000 × g for 8 min and resuspended in sterile water to achieve a final concentration of 1×10⁹ CFU/mL.

H. pylori SS1, obtained from BeNa Culture Collection, was revived from -80°C storage on Columbia agar plates supplemented with 5% sterile defibrinated sheep blood. After being revived for two generations, it was cultured on Columbia blood agar plates for 48 h, washed with BHI broth, and the bacterial suspension adjusted to 1×10¹⁰ CFU/mL (21).

Cultured H. pylori SS1 was centrifuged and washed twice with artificial gastric juice (22), then resuspended to adjust the OD₆₀₀ to approximately 0.50. After staining with CFDA SE dye and incubating at 37°C for 30 min, excess dye was removed by centrifuging the bacteria in PBS. The pre-stained H. pylori SS1 was then mixed with Lp05 in equal volumes, vortexed for at least 10 s, and incubated at 37°C for 2 h. A sample of this mixture was placed on a slide and examined under a confocal laser scanning microscope (E_x = 494 nm, E_m = 521 nm) (23).

Strain Lp05 was centrifuged, washed twice with PBS, and resuspended in sterile PBS. Similarly, H. pylori SS1 was washed and resuspended in artificial gastric juice. Both suspensions were adjusted to an OD₆₀₀ of approximately 0.50 and mixed equally to induce co-aggregation. After 2 h at 25°C, the aggregates were fixed overnight in 2.5% glutaraldehyde at 4°C, dehydrated in a graded ethanol series, and gold-sputter coated. The samples were then scanned under a scanning electron microscope (24).

All procedures involving mice were approved by the Experimental Animal Ethics Committee of Zhengzhou University (License No: SYXK(Yu)2021-0011) and followed the ARRIVE guidelines 2.0 (25). To establish the H. pylori infection model, 60 eight-week-old female SPF-grade C57BL/6 mice (SPF Biotechnology Co., Ltd, Beijing, China) were acclimatized for a week and then divided into six groups of ten each for a six-week experiment (26) ( Figure 1 ). The control (CTL) group received daily gavage of 0.2 mL saline throughout the experiment. From the second week, the five treatment groups started modeling H. pylori SS1 infection with bi-daily gavages of a 0.2 mL 0.9% NaCl suspension containing 1×10¹⁰ CFU/mL H. pylori for a total of five times. The model control group (MC) followed the same saline regimen and only H. pylori. The quadruple therapy group (PG) received a normal diet in the first week and post-infection, daily gavages of a mix containing amoxicillin (410 mg/kg), bismuth pectin (123 mg/kg), furazolidone (41 mg/kg), and pantoprazole (16.4 mg/kg) during the second and third weeks, followed by 0.2 mL 0.9% NaCl until week six. The Lp05 low, medium, and high dose groups (Lp05_L, Lp05_M, Lp05_H) received daily gavages of 0.2 mL with Lp05 at 1×10⁸, 1×10⁹, and 1×10¹⁰ CFU/mL respectively, from the first week and began H. pylori infection the same as MC group from the second week. All mice were kept at 25 ± 2°C with bedding changed two to three times weekly and weighed regularly.

After the final gavage, mice fasted for 12 hours were anesthetized using a 4% chloral hydrate solution intraperitoneally, followed by blood collection via orbital extraction. Blood samples were centrifuged at 3500 rpm for 20 min at 4°C, with serum stored at -80°C for analysis. Post-experiment, mice were euthanized and stomach, duodenum, and cecum tissues were extracted and fixed for histological examination.

H. pylori presence in stomach tissues was detected using a modified Warthin-Starry (W-S) silver stain, effective in revealing H. pylori as black deposits in tissues (27). For structural and pathological assessments, tissues underwent H&E staining after fixing, dehydration, and paraffin embedding. Tissue sections of 4 μm were stained using Dong et al.’s (2022) methods to detail the tissues’ structural and pathological states (28).

Gastric mucosa samples were incubated in a solution containing 1 g/mL urea and 850 µg/mL phenol red at 37°C for 30 min, observing color changes (29). Additionally, homogenized samples in PBS were diluted, plated on selective media, and cultured under microaerophilic conditions at 37°C for 2-5 days to identify H. pylori based on typical colony characteristics.

Post-experiment, anti-H. pylori IgG antibodies in mouse serum were detected using an ELISA (Yuanju Biotechnology Center, Shanghai, China). Serum samples, diluted 1:50, were added to H. pylori antigen-coated wells, incubated at 37°C for 30 min, washed thrice, and treated with HRP-labeled rabbit anti-mouse IgG for another 30 min. The reaction developed with Tetramethylbenzidine (TMB) was stopped using sulfuric acid and read at 450 nm. Serum was analyzed for gastric enzymes (α-amylase, pepsin, cellulase), inflammatory markers (Interleukin-17A (IL-17A), Interleukin-23 (IL-23), Interleukin-10 (IL-10), Interferon beta (IFN-β), Interferon gamma (IFN-γ), Forkhead Box P3 (FoxP3), endotoxin (ET), interferon–γ–inducible protein 10 (IP-10), transforming growth factor beta-1 (TGF-β1)), and activities of catalase (CAT), malondialdehyde (MDA), superoxide dismutase (SOD), and myeloperoxidase (MPO). All assays were performed according to the manufacturer’s instructions provided with each kit (BIOSCO Biotechnology Co., Ltd, Jiangsu, China).

We analyzed the microbial communities in stomach and cecum tissues via 16S rRNA gene sequencing. DNA was extracted, and the V3-V4 regions of the 16S rRNA gene were amplified using 341F and 805R primers. The PCR process involved an initial denaturation at 95°C, followed by cycles at 94°C for denaturation, 58°C for annealing, and 72°C for extension, concluding with a final extension at 72°C. Amplified products were indexed, pooled, and sequenced on an Illumina MiSeq platform in a paired-end 2x300 bp format. Raw data were refined using USEARCH to remove low-quality reads and chimeras. Amplicon sequence variants (ASVs) were clustered at 97% similarity with UPARSE (version 7.1) and classified via the RDP classifier (30). Alpha and Beta diversity were analyzed using QIIME, with significant differences between communities determined by linear discriminant analysis effect size (LEfSe), PICRUSt, and statistical analysis of metagenomic profiles (STAMP) analyses (31).

All data were presented as mean ± standard deviation (SD). Statistical analyses were performed using R Studio software. The data were first tested for normal distribution using the Kolmogorov-Smirnov test. Multiple group comparisons were made using the non-parametric Kruskal-Wallis test, while differences between two groups were assessed using the Mann-Whitney U test, with a significance level set at p < 0.05. Microbial community analysis was conducted using QIIME software and principal coordinate analysis (PCoA) based on unweighted UniFrac distances (32). Community phylogenetics were investigated using PICRUSt and LEfSe analyses, while Spearman’s rank correlation test was used to explore correlations between bacterial genera in stomach and cecum tissues. All statistical analyses were performed in the R environment version 4.3.2 (33).

This study utilized CLSM to confirm the co-aggregation phenomenon between Lp05 and SS1. CFDA SE fluorescent staining was employed to label SS1, facilitating the identification of its presence within the co-aggregates. Observations under the microscope are presented in Figure 2A (a-c) , which shows the distribution of SS1 cultured alone under bright field, dark field, and overlay conditions, with bacteria evenly dispersed and no aggregates formed. In contrast, Figure 2A (d-f) displays the situation after 2 h of co-culturing Lp05 with SS1, where large co-aggregates (indicated by arrows) are visible in all three fields. These aggregates clearly demonstrate SS1 being encapsulated by Lp05, clustering together, which may facilitate its expulsion from the stomach, thus potentially reducing the gastric colonization of H. pylori.

The formation of co-aggregates between Lp05 and SS1 was also observed using SEM. After freeze-drying under vacuum, the aggregates were examined under a cold field emission scanning electron microscope to confirm the co-aggregation. Figure 2B (a, b) shows SS1 and Lp05 cultured separately, magnified 5000 times; SS1 appears as curved rods while Lp05 is depicted as short rods, with significant morphological differences between the two. Figure 2B (c, d) display the co-aggregates at magnifications of 5000 and 10000 times, respectively, with arrows pointing to multiple SS1 bodies bonded with Lp05, forming larger aggregates. Notably, the morphology of the Lp05 strain remains stable, whereas the co-cultured SS1 shows signs of edge dispersion, cellular deformation, collapse, and breakage, indicating significant impact on its cell structure. These observations confirm the effective formation of co-aggregates by Lp05 with H. pylori, reducing its gastric colonization.

The urease activity in gastric mucosa samples was assessed as shown in Figure 3A . Notably, the MC group exhibited significantly increased urease activity, leading to a rise in pH that changed the phenol red indicator from yellow to pink or red. Quantitatively, urease activity in the MC group was markedly higher, with an optical density (OD) measurement at 562 nm reaching 2.0. In comparison, groups treated with various doses of Lp05 and the PG group demonstrated lower urease activities, with minimal change in the color of the phenol red indicator or maintenance of its yellow color. Particularly, the Lp05_H group showed urease activities comparable to those of the PG group (OD_562nm = 0.5) and significantly lower than those in the MC group. This suggests that Lp05 can reduce gastric colonization by H. pylori, thereby lowering urease activity induced by H. pylori infection and alleviating tissue inflammation caused by H. pylori colonization and infection. These findings are substantiated by the OD measurements, indicating effective suppression of urease by Lp05_H treatment. Additionally, serum levels of H. pylori-IgG antibodies (as depicted in Figure 3B ) were significantly lower (p < 0.01) in mice treated with Lp05 compared to the MC group, approaching levels seen in the CTL group and comparable to those in the PG group. This indicates that Lp05 can effectively reduce the systemic immune response to H. pylori.

Figure 3C presents W-S silver-stained sections of mouse gastric tissue. Compared to the CTL group, the MC group displayed dark brown to black rod-shaped or slightly helical bacteria (indicated by arrows) on the mucosal layer, epithelial surface, and intercellular spaces. H. pylori predominantly colonized the gastric pits and gland lumina at the junction of the gastric body and antrum, generally appearing arc-shaped, S-shaped, and occasionally as short rods. After homogenizing gastric tissues and culturing on Brucella agar plates for three days, Gram staining of the colonies ( Supplementary Figure S1 ) revealed typical red S-shaped, slightly helical, or curved rod-shaped morphology of H. pylori, confirming its Gram-negative nature and successful establishment of the H. pylori infection model in the MC group with observable colonization of the gastric mucosa. By contrast, in Lp05_L group, occasional H. pylori colonization was still visible. However, in Lp05_M and Lp05_H groups, the colonization of H. pylori was comparable to that in the CTL and PG groups, indicating that medium and high doses of Lp05 significantly inhibit the colonization of H. pylori on the gastric mucosa.

HE staining revealed the effects of Lp05 on gastric and duodenal tissues in H. pylori-infected mice. Figures 4A–L show that in the CTL group, gastric structures were intact with well-defined gastric pits, orderly epithelial cells, and no inflammation. Conversely, the MC group exhibited H. pylori-induced pathologies such as epithelial necrosis, disorganized glands, and inflammation, with fewer parietal cells. The PG group had mild inflammation and minor ulcers. The Lp05_L group showed mild to moderate inflammation, whereas the Lp05_M and Lp05_H groups demonstrated notable tissue improvements, especially the Lp05_H group, where the gastric mucosa was clear, intact, and exhibited minimal inflammation, similar to the CTL group. For the duodenum, Figures 4M–R indicates regular mucosal structures in the CTL group, while the MC group showed extensive cellular necrosis and inflammation, disrupting the structure significantly. Despite treatment, the PG group had disordered epithelial cells. Post Lp05 treatment, especially in the Lp05_M and Lp05_H groups, the duodenal villi were intact with regular epithelial cell arrangements, suggesting that medium and high doses of Lp05 effectively alleviate H. pylori-induced duodenal damage.

This test assessed the impact of Lp05 on digestive enzymes in the gastric tissue of mice infected with H. pylori, focusing on α-amylase, cellulase, and pepsin. Specifically, in the CTL group, the mean α-amylase activity was 27.5 mg/min/g. Post-infection in the MC group, activity decreased significantly to 17.2 mg/min/g (p < 0.01), indicating the detrimental effects of H. pylori infection. Following treatment with Lp05, α-amylase activity in the Lp05_H group almost completely recovered to 28 mg/min/g, slightly surpassing the CTL group. The baseline pepsin activity in the CTL group was 1501 nmol/min/g. After H. pylori infection (MC group), activity dropped to 1260 nmol/min/g (p < 0.01). Remarkably, pepsin activity increased to 1614.9 nmol/min/g in the Lp05_H group, significantly higher than the CTL group, which underscores the effectiveness of Lp05 in not only reversing the infection impact but also enhancing enzyme activity. For cellulase, the activity was 454.9 μg/min/g in the CTL group, reduced to 329.5 μg/min/g in the MC group (p < 0.01) due to H. pylori infection. Post-Lp05 treatment, activity in the Lp05_H group reached 481.6 μg/min/g, again exceeding the CTL, indicating significant recovery and enhancement of function. These results ( Figures 5A–C ) suggest that Lp05 intervention, particularly at higher doses, supports overall gastric function recovery by normalizing and enhancing digestive enzyme activities, which are otherwise compromised by H. pylori infection.

As shown in Figures 5D–L , in the MC group, ET levels were measured at 420.1 pg/mL, elevated compared to the CTL group, which had levels of 331 pg/mL. The Lp05_H group effectively reduced ET to 332.24 pg/mL (p < 0.05), closely aligning with the CTL group. IL-17A was higher in MC at 236.15 pg/mL compared to 186.3 pg/mL in CTL, and was reduced to 204 pg/mL in Lp05_H (p < 0.05). IL-23 levels rose to 117.5 pg/mL in MC from 83.9 pg/mL in CTL, but were moderated to 75.70 pg/mL with Lp05 treatment (p < 0.05). IP-10 levels increased to 495.5 pg/mL in MC from 387.1 pg/mL in CTL, with a reduction to 348.7 pg/mL observed in Lp05_H (p < 0.05). TGF-β1 also showed an increase in MC to 11.5 pg/mL from 9.4 pg/mL in CTL (p < 0.05), and was reduced to 8.7 pg/mL in Lp05_H (p < 0.05), indicating a substantial reduction. Conversely, IFN-β and IFN-γ, which decreased to 58.9 pg/mL and 391 pg/mL in MC from CTL levels of 67.7 pg/mL and 646.5 pg/mL respectively, were restored in Lp05_H to 67.8 pg/mL and 672.2 pg/mL (both p < 0.05). Additionally, FoxP3 levels were reduced in MC to 206.76 pg/mL from 343 pg/mL in CTL and were restored in Lp05_H to 341.0 pg/mL (p < 0.05). The effects of Lp05 were similar to those of PG group, particularly in Lp05_H group, where the modulation of inflammatory cytokines was especially pronounced, indicating that Lp05 effectively suppresses the systemic inflammatory response caused by H. pylori infection.

As depicted in Figures 5M–P , oxidative stress markers were quantitatively analyzed to evaluate the impact of H. pylori infection and subsequent treatments. In the CTL group, CAT activity was measured at 38.1 nmol/min/mL, and SOD activity at 40.4 U/mL. Following H. pylori infection, the MC group exhibited significant reductions in these antioxidant enzymes, with CAT decreasing to 21.3 nmol/min/mL and SOD to 23.3 U/mL (all p < 0.05). This decline coincided with increases in MDA and MPO levels, from 0.72 nmol/mL to 1.03 nmol/mL and from 0.75 U/mL to 1.24 U/mL respectively, indicating a marked increase in oxidative stress. Post-treatment analyses showed significant reversals in these oxidative stress markers across all intervention groups (all p < 0.05). After treatment with varying doses of Lp05, CAT levels were increased to 42.2 nmol/min/mL in the Lp05_H group, surpassing CTL levels. Similarly, SOD levels improved to 38.2 U/mL in Lp05_H, nearly reaching CTL values. MDA and MPO levels were significantly reduced to 0.86 nmol/mL and 1.09 U/mL, respectively, in the Lp05_H group. These results suggest that high-dose Lp05 possesses potent antioxidant capabilities, effectively alleviating the oxidative damage caused by H. pylori infection.

The 16S rRNA gene sequencing assessed Lp05 effect on the gastric microbiome of H. pylori-infected mice. Figure 6A identifies 13,888 ASVs, with 420 common across groups. Unique ASVs were 2366 in the CTL group, 1815 in MC, 3046 in PG, and 2041, 2002, and 2198 in the low, medium, and high dose Lp05 groups respectively, indicating significant microbial diversity changes due to treatment. Observed richness, Chao1, Shannon, and Simpson indices ( Figures 6B–E ) evaluated microbial abundance, diversity, and evenness, showing a decrease in α-diversity in the MC group compared to CTL. Post-BQT and Lp05 interventions, richness and Chao1 indices increased, albeit not significantly versus the MC group. β-diversity analysis ( Figure 6F ) showed distinct microbial structures, with significant composition shifts in Lp05, MC, and PG groups compared to CTL. Dominant species and their abundances are detailed in Figures 6G, H . Firmicutes and Bacteroidota were predominant, with significant variations in relative abundance, for instance, 71.18% in MG versus 26.67% in CTL. Lp05 significantly adjusted these proportions, particularly in the low-dose group, aligning more closely with CTL. Dominant genera included Muribaculaceae_unclassified, Lactobacillus, and Muribaculum, with Muribaculaceae_unclassified showing increased abundance in PG and Lp05 groups versus MC, suggesting effective microbial modulation. LEfSe analysis ( Figure 7A ) highlighted marker species with significant differences across groups, showing beneficial genera like Rikenella were enhanced by Lp05, indicating its role in optimizing the gastric microbiome. STAMP analysis ( Figure 7B ) noted a significant recovery in Muribaculaceae_unclassified in Lp05-treated groups compared to MC, particularly in low doses. PICRUSt2 analysis ( Figure 7C ) predicted upregulated metabolic pathways such as amino acid metabolism and polysaccharide biosynthesis in Lp05 groups, essential for microbial community stability and nearing levels in the CTL group.

Venn diagram analysis ( Figure 8A ) showed 14,342 ASVs, with 340 shared across all six groups. Unique ASVs were 2760 in CTL, 2518 in MC, 1627 in PG, and 2595, 1894, and 2268 in the Lp05 low, medium, and high-dose groups, respectively, indicating significant impacts on intestinal microbiome structure. Alpha diversity metrics, including observed richness, Chao1, Shannon, and Simpson indices ( Figures 8B–E ), demonstrated a decline in microbial richness and diversity in the MC group compared to CTL. Post-Lp05 treatment, particularly at high doses, significantly enhanced these indices (p < 0.05), suggesting Lp05’s effectiveness in restoring diversity lost due to H. pylori infection. PCoA analysis ( Figure 8F ) revealed significant microbial composition shifts in the Lp05, MC, and PG groups relative to CTL, indicating effective microbiome rebalancing. Figures 8G, H detailed dominant species at various taxonomic levels, with Firmicutes and Bacteroidota being prevalent. For instance, Firmicutes comprised 60.85% in MC versus 51.91% in CTL. Genus level analysis showed prominence of Lachnospiraceae_NK4A136_group and Muribaculaceae_unclassified, with Lp05 notably enhancing Lachnospiraceae_NK4A136_group, highlighting its potential for supporting intestinal health. LEfSe analysis ( Figure 9A ) identified significant biomarkers, with dominant species in CTL including Muribaculum and Paramuribaculum. MC showed enrichment in Atopobiaceae and Olsenella, while PG’s biomarkers were primarily Firmicutes and Bifidobacterium. Lp05’s intervention highlighted beneficial biomarkers such as Bacilli and Ligilactobacillus, suggesting a reduction in potential pathogens. STAMP analysis ( Figure 9B ) at the genus level confirmed significant abundance changes in Ligilactobacillus, Bifidobacterium, and Olsenella, indicating their roles in H. pylori-associated pathologies. PICRUSt2 analysis ( Figure 9C ) revealed significant enrichment in metabolic pathways, particularly carbohydrate and nucleotide metabolism, essential for cellular growth and repair. This suggests that Lp05 intervention regulates metabolic activities in the gastric microbiome, potentially reducing metabolic disorders and long-term health issues associated with H. pylori infection.

Using Spearman’s correlation, we analyzed the relationship between the relative abundances of microbial genera in the gastric and intestinal microbiomes, as shown in Figure 10 . In the gastric microbiome, genera such as Raoultella, Streptomyces, Pseudoxanthomonas, Gemmata, Sporosarcina, and Steroidobacter were found to have significant positive correlations (p < 0.05) with intestinal genera including Actinobacillus, Ureaplasma, Anaerococcus, Oceanivirga, Fusobacterium, Atopostipes, Succinivibrio, Rikenella, Campylobacter, and Anaerovibrio. Conversely, these gastric genera showed negative correlations with Christensenellaceae_R-7_group, Eubacterium_nodatum_group, Bifidobacterium, Mitochondria, Brevundimonas, and Subdoligranulum. Additionally, gastric genera such as Ruminococcus_gnavus_group, Rothia, UBA1819, Subdoligranulum, Blautia, and Phascolarctobacterium demonstrated positive correlations with Christensenellaceae_R-7_group, Eubacterium_nodatum_group, Bifidobacterium, Mitochondria, Brevundimonas, and Subdoligranulum. However, these were significantly negatively correlated (p < 0.05) with the intestinal genera including Actinobacillus, Ureaplasma, Anaerococcus, Oceanivirga, Fusobacterium, Atopostipes, Succinivibrio, Rikenella, Campylobacter, and Anaerovibrio. This complex interplay of correlations between gastric and intestinal microbiomes indicates that Lp05 intervention not only affects specific genera but also influences the broader microbial ecosystem relationships within the host.

Using Spearman’s correlation, the study evaluated the relationships between gastric and intestinal microbial genera in H. pylori-infected mice, depicted in Figure 10 . Gastric genera like Raoultella, Streptomyces, and Pseudoxanthomonas showed significant positive correlations (p < 0.05) with several intestinal genera such as Actinobacillus and Ureaplasma, indicating potential shared roles in gastrointestinal processes. Conversely, these gastric genera negatively correlated with Christensenellaceae_R-7_group and Bifidobacterium, suggesting differing physiological roles or adaptations to environmental conditions. Additionally, genera like Ruminococcus_gnavus_group displayed both positive and negative correlations with various intestinal genera, reflecting their complex interactions within the gastrointestinal tract. These findings highlight how Lp05 intervention impacts not only specific microbial genera but also the broader inter-genus relationships, affecting overall microbial ecosystem dynamics within the host.