From April 2018 to January 2020, among the endoscopic screening population of upper gastrointestinal cancer in Yangzhong City, 193 individuals diagnosed as chronic gastritis and RUT-positive were brought into this study. According to the consensus opinion on chronic gastritis in China formulated by the Digestive Diseases Branch of the Chinese Medical Association (Shanghai, 2017) (Fang, 2023), “Atrophy and/or intestinal metaplasia” in pathological diagnosis was defined as CAG, and others were defined as nCAG. There were 124 CAG patients and 69 nCAG patients in the 193 individuals, and the clinical data were collected using a unified questionnaire. Exclusion criteria: 1) patients with a history of major organic lesions, such as stroke, malignancy, organ transplantation, gastrointestinal surgery, etc.; 2) history of antibiotics, probiotics, proton pump inhibition and other drugs in the last month; 3) age below 40 or above 70 because most asymptomatic patients screened out in the range of 40 - 70 years old; 4) neoplasia, cancer and other suspected gastric mucosal cancer in pathological diagnosis; 5) residence in Yangzhong area for less than 5 years; 6) unwillingness to cooperate. The study was approved by the Ethics Committee of People’s Hospital of Yangzhong City (No.2018-039) and the patients gave informed consent.

Gastric sinus mucosal tissues were clamped during early morning fasting and electronic gastroscopy, quickly frozen in liquid nitrogen and stored at -80°C. Gastric microbiota were profiled according to the previous literature (Xu et al., 2023). Total DNA of gastric mucosa was extracted by QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA), and the quality of DNA was detected by 1% agarose gel electrophoresis. PCR amplification was performed using universal primers in V3-V4 region of 16S rRNA gene, and Miseq PE library was prepared for high-throughput sequencing. QIIME (Version 1.17) was used to optimize sequencing data, and the operational taxonomic unit (OTU) was established in UP-ARSE35 software according to 97% similarity, and taxonomic annotations of microbiota species were obtained from the RDP (Ribosomal Database Project II database).

Fasting venous blood 2–3 ml with a procoagulant tube was collected, the serum was separated by 3000 rpm centrifugation for 10min. According to the previous literature (Liu et al., 2021), Hp antibody typing classification assay kit (western blotting) (Shenzhen Blot Biological Products Co., Ltd., No. 201905016) was used to detect Hp-related IgG against four antigens including CagA (116 kD), VacA (95 kD, 91 kD), UreB (66 kD) and UreA (30 kD). The experimental process was briefly described as following: 1) Place the blotting membrane strips into the assay tank, add 1000 μL dilution solution and 10 μL serum, and shake (40 times/min) for 30 minutes at room temperature (22-23°C); 2) Wash the strips with 1000 μL washing buffer for 4 times, add 500 μL working solution and 10 μL enzyme-linked antibodies, and shake (40 times/min) for 30 minutes at room temperature (22-23°C); 3) Wash the strips with 1000 μL washing buffer for 3 times, add 500 μL chromogenic reagent solution, stand for 5 mins, and terminate the reaction by washing 3 times with 1000 μL distilled water.

According to Hp antibody typing classification, Hp infection was divided into Hp-I infection and Hp-II infection. If the bands of CagA and/or VacA were positive, it could be diagnosed as Hp-I infection; if the bands of both CagA and VacA were negative, but the bands of urease A and/or urease B were positive, it could be diagnosed as Hp-II infection; if all of the four bands were negative, it could be diagnosed as non-Hp infection. The band of quality control should appear normally, otherwise, the test should be invalid.

SPSS software (Version 25.0, SPSS, inc.) was used for clinical parameter analysis. Visual Genomics software (Release 1, Shanghai Infinity Biotechnology Co., Ltd.) was conducted for alpha diversity and community structure analysis. Spearman correlation analysis was used to construct the symbiotic relationship of the microbiota, and R software (Version 2022.02.0-443, https://www.r-project.org/) was used for volcano plot, Spearman correlation analysis, and graphical representation. Two-tailed P < 0.05 indicated significant statistical significance.

The present 193 individuals included 124 CAG patients and 69 nCAG patients, who were well matched by the gender, body mass index (BMI), and four lifestyle habits (eating fried foods, eating fresh fruits, smoking and drinking) (P > 0.05). However, the CAG patients were significantly elder (P < 0.05), and had higher proportion of bile reflux and inflammatory activity than nCAG patients (P < 0.01). It is worth noting that the CAG patients had a higher proportion of Hp-I infection than nCAG patients (P < 0.001). Additionally, there were statistically significant differences in the distribution of eating speed and salt preference (P < 0.05) ( Table S1 ).

Alpha diversity showed that the richness indices (Ace, Chao and Observed OUTs) of gastric microbiota were significantly higher in CAG patients than in nCAG patients (P < 0.05) in the patients with Hp-I infection; however, among the patients with Hp-II infection, no differences of the richness indices were observed between CAG patients and nCAG patients (P > 0.05). Meanwhile, no significant difference of Shannon was observed between CAG patients and nCAG patients (P > 0.05) ( Figure 1 ). Interestingly, among nCAG patients, the Hp-I infected patients had significantly lower Chao and Shannon indices than the Hp-II infected patients (P < 0.05). The results suggested that Hp-I infection significantly decreased the richness and diversity of gastric microbiota in nCAG patients, while significantly increased the richness of gastric microbiota in CAG patients.

Here investigated the effect of Hp-infection typing on the gastric microbiota in patients with chronic gastritis ( Figure 2 ). At the phylum level, the dominant gastric phyla in gastritis patients included Proteobacteria, Epsilonbacteraeota, Actinobacteria, Bacteroidetes, and Firmicutes, especially Proteobacteria was the most prevalent phylum (relative abundance > 50%) ( Figure 2A ). Among dominant phyla, non-parametric analysis showed that Hp-I infected patients had a higher Epsilonbacteraeota and lower Actinobacteria than Hp-II infected patients (P < 0.05). The 17 prevalent bacterial genera (prevalent rate 100%) included Ralstonia (23.94%), Helicobacter (20.28%), Pseudonocardia (9.99%), Mesorhizobium (9.21%), Bradyrhizobium (5.05%), Labrys (4.75%), Hydrobacter (3.56%), Variovorax (2.63%), Pseudolabrys (2.25%), Reyranella (1.51%), Acidovorax (1.11%), Methylobacterium (1.05%), Aliidiomarina (0.83%), Nesterenkonia (0.68%), Pseudomonas (0.62%), Cupriavidus (0.43%) and Pelomonas (0.25%). At OUT level, different analysis showed that the Hp-infection typing exerted significant effect on the gastric microbial community in patients with chronic gastritis ( Figure 2B ). After calculating the fold of change (FC) in each OUT between CAG and nCAG patients, the different OUTs were collected based on Log₂(FC) greater than 0.5 or less than -0.5 with a correction of P < 0.05. The present study observed 40 significantly up-regulated OTUs and 98 significantly down-regulated OTUs in the gastric mucosa of CAG patients. Among Hp-I infected patients, 29 up-regulated OTUs and 81 down-regulated OTUs were observed in the stomach of CAG patients compared with nCAG patients. Among Hp-II infected patients, 29 up-regulated OTUs and 17 down-regulated OTUs were collected in the stomach of CAG patients compared with nCAG patients.

On the one side, here observed the effect of CAG on gastric microbiota in the three respective groups (the total population, the Hp-I infected patients, and the Hp-II infected patients) ( Figure 3 ). Venn intersection analysis obtained 197 differential OTUs between CAG and nCAG. Hp-I infected patients had much more differential OTUs than Hp-II infected patients, and 15 shared OTUs may be the different bacteria to distinguish CAG from nCAG ( Figure 3A ). Here presented the abundant status of 69 differential OTUs with a prevalence rate greater than 10%. Among them, fifteen sharing differential OTUs were considered to be CAG-related bacteria independent of Hp-infection, and they belonged to phylum Proteobacteria, including genus Labrys, Rhodopseudomonas, Pseudolabrys, Methylobacterium, Acidovorax, Bradyrhizobium, Variovorax, Pelomonas, and Herbaspirillum ( Figure 3B ). In Hp-I infected patients, 37 unique differential OTUs were regarded as CAG-related bacteria, among them, four abundant OTUs belonged to phylum Proteobacteria and Actinobacteria, including genus Actinomyces, Neisseria, Actinobacillus and Ralstonia ( Figure 3C ). In Hp-II infected patients, 22 unique differential OTUs were regarded as CAG-related bacteria, among them, 12 abundant OTUs belonged to the phylum Proteobacteria, Fusobacteria, Firmicutes and Bacteroidetes, including genus Labrys, Mesorhizobium, Fusobacterium, Bacteroides, Escherichia-Shigella, Alkalicoccus, Ralstonia, Bradyrhizobium and Variovorax ( Figure 3D ). In addition, the other shared OTUs in the Venn were considered as environment-related bacteria, all of which belonged to the phylum Proteobacteria ( Figure S1 ). The results suggested that the gastric microbiota differ significantly during the progression from nCAG to CAG along with Hp-infection typing, which again confirmed that gastric non-Helicobacter bacteria play a certain role in the evolution of gastric mucosal inflammation and carcinoma, and Hp typing might be the key role in shaping the gastric microbiome.

On the other hand, here also explored the effect of Hp-infection typing on the gastric microbiota in CAG and nCAG patients, and observed 334 and 309 Hp-infection related OTUs in CAG and nCAG patients, respectively ( Figure S2 ). Venn analysis showed that seven OTUs were associated with Hp-infection types, 28 OTUs were related to Hp-infection typing in CAG patients, and 86 OTUs linked with Hp-infection typing in nCAG patients ( Figure S2A ). Among the seven shared OTUs, Hp-I infected patients had a higher abundance of OTU1 (Helicobacter pylori 26695-1) than Hp-II infected patients, but had lower abundances of six non-Helicobacter OTUs ( Figure S2B ). Furthermore, OTU1 was a typical Hp-I strain with CagA and VacA genes, whose prevalence rate was 95.9% (185/193) in total population, suggesting that virulent Hp-infection inhibit the proliferation of non-Helicobacter bacteria in the stomach of gastritis patient. Compared with Hp-II infected patients, regardless of CAG ( Figure S2C ) or nCAG ( Figure S2D ), most of the differential OTUs were significantly decreased in the stomach of Hp-I infected patients. In the CAG patients, the significantly decreased OTUs were mainly oral dominant microbiota, such as genus Neisseria, Staphylococcus, Haemophilus. However, in the nCAG patients, the significantly reduced OTUs were mainly the gastric dominant microbiota, such as Aliidiomarina, Reyranella, Halomonas, Pseudomonas, and Acidovorax. The results suggest that virulent Hp infection inhibit the proliferation of normal gastric microbiota at the early stage of chronic gastritis, while inhibit the colonization of oral microbiota in the stomach at the late stage of chronic gastritis.

Based on the previous literature (Yang et al., 2022), Spearman correlation analysis was performed at OUT level, and a significant correlation (P < 0.001) with absolute values of correlation coefficients |r| > 0.6 was selected to construct gastric symbiosis network in CAG and nCAG patients, respectively ( Figure 4 , Table S2 ). The gastric symbiosis network of nCAG patients included 290 nodes, and eight nodes belonging to Helicobacter genus significantly linked with sparse OTUs of non-Helicobacter bacteria. On the contrary, in the stomach of the CAG patients, the symbiosis network included 342 nodes, and the Helicobacter OTUs linked with numerous non-Helicobacter bacteria ( Figure 4A ).

Here focused on 127 differential OTUs in the network, the OTUs in the symbiotic network of CAG patients linking more than 30 nodes were defined as core OTUs, and 40 core OTUs were identified in the symbiotic network of CAG patients. Among the 40 core OTUs, ten OTUs (OTU195, OTU315, OTU374, OTU186, OTU312, OTU213, OTU41, OTU833, OTU174, OTU35) belonged to genus Ralstonia, nine OTUs (OTU753, OTU560, OTU109, OTU228, OTU188, OTU118 OTU126, OTU707, OTU748) belonged to genus Pseudolabrys, phylum Proteobacteria. Similarly, five core OTUs were identified in the symbiotic network of nCAG patients, and they were OTU629 (Nesterenkonia), OTU257 (Herbaspirillum), OTU1010 (Halomonas), OTU420 (Cupriavidus), and OTU102 (Acidovorax) ( Figure 4B ). The eight Helicobacter OTUs (OTU194, OTU606, OTU250, OTU323, OTU1, OTU60, OTU114, OTU1173) were extracted from the commensal network, and found that they had much more connections in the symbiotic network of CAG patients than that of nCAG patients ( Figure 4C ). Furthermore, in the stomach of nCAG patients, only OTU1 was significantly negatively associated with a non-Helicobacter OTU53 (Reyranella), and the symbiotic network was mainly composed of Helicobacter OTUs ( Figure 4D ). Meanwhile, in CAG patients, Helicobacter OTU194 and OTU1173 were associated with OTU126 (Pseudolabrys), and Helicobacter OTU194 linked with 35 non-Helicobacter OTUs including eight Pseudolabs OTUs, six Ralstonia OTUs, five Bradyrhizobium OTUs, five Mesorhizobium OTUs, and four Variovorax OTUs ( Figure 4E ). Overall, in the stomach of nCAG patients, Helicobacter bacteria might synergistically proliferate, and inhibit the proliferation of non-Helicobacter bacteria; in the stomach of CAG patients, both synergistic proliferation and antagonistic competition were observed between the Helicobacter bacteria and non-Helicobacter bacteria. The results provided new clues for probiotic application in preventing and treating Hp infection in patients with chronic gastritis.

Based on the COG prediction function of gastric microbiota, here explored the potential signaling pathways in the development of CAG ( Figure 5 ). With the criterion of Log₂ (FC) > 0.5 or < - 0.5, and P <0.05, there were 605 up-regulated and 38 down-regulated COG predictive functions in CAG patients compared to nCAG patients. Among Hp-I infected patients, CAG patients had 251 up-regulated and 24 down-regulated COG predictive functions compared to nCAG patients; and in Hp-II infected patients, 1126 COG predictive functions were up-regulated in CAG patients compared to nCAG patients ( Figure 5A ). Venn analysis showed that 82 sharing differential COG prediction functions were related to the occurrence of CAG independent of Hp-infected type ( Figure 5B ). Among the 82 sharing functions, seven enhanced COG predictive functions included ATPase component, cell wall-associated hydrolases, membrane-associated phospholipid phosphatase, copper chaperone, response regulator of the LytR/AlgR family, bacteriophytochrome, and ABC-type Na⁺ efflux pump, permease component ( Figure 5C ). Since the seven functions links with microbial physiological metabolism, the gastric non-Helicobacter bacteria should also play certain role in the evolution of the Hp-induced nCAG to CAG, even to gastric carcinoma. For KEGG prediction function, there were 23 up-regulated and one down-regulated KEGG prediction functions. Twelve up-regulated and one down-regulated KEGG prediction functions were related to CAG with Hp-I infection, while 155 up-regulated KEGG prediction functions were related to CAG with Hp-II infection ( Figure S3A ). Venn analysis showed that only two sharing differential prediction functions linked with Hp-infection typing ( Figure S3B ), namely, sphingolipid metabolism and butirosin and neomycin biosynthesis were significantly enhanced in CAG patients ( Figure S3C ). These data provided new clues for understanding the development of Hp-infection inducing CAG.

Smoking and drinking are the two famous living habits for GC risk, and bile reflux and inflammatory activity are the two pathologic factors driving the progression of nCAG to CAG/GC (Poorolajal et al., 2020). Several studies proved that smoking and drinking could alter the gastrointestinal microbiota. Leite et al. (2022) found that smoking distinctly altered the duodenal microbiota, but quitters and non-smokers occupied similar duodenal microbiota, suggesting that living habits-induced microbial alteration is plastic and reversible. Bai et al. (2022) found that the exposure of cigarette smoking could increase the incidence of colorectal cancer in mice, accompanied with the absence of gut Parabacteroides distasonis and Lactobacillus spp., and the increased Eggerthella lenta was positively related to the fecal taurine deoxycholic acid (TDCA), which could in turn activate the intestinal epithelial MAPK/ERK pathway. Meanwhile, cigarette smoking could reduce the expressive levels of intestinal mucosal junction protein-3 and ZO-1, but increase the serum LPS level, leading to intestinal barrier damage and colorectal cancer. Zhao et al. (2022) found that the occurrence of alcoholic fatty liver was associated with altered gut microbiota, while Nie et al. (2022) found that the long-chain saturated fatty acids (palmitic acid, stearic acid) in Jinhua ham could prevent chronic alcohol-associated liver damage (ALD), and the preventive effect was associated with increased gut Akkermansia muciniphila and Lactobacillus. Chen et al. (2022) found that Bacteroides xylanisolvens could degrade the cumulative nicotine in the intestine, and prevent the development of nonalcoholic fatty liver disease (NAFLD). However, the relationship between smoking/drinking and gastric microbiota is still unclear in the occurrence of CAG.

Thus, this study investigated relationship between smoking/drinking and gastric microbiota in CAG patients ( Figure 6 ). Among drinkers, the CAG patients occupied 26 significantly up-regulated and 94 down-regulated bacterial OTUs compared to nCAG patients; among non-drinkers, there were 24 significantly up-regulated and 33 down-regulated bacterial OTUs in CAG patients compared to nCAG patients. Similarly, compared to smoking nCAG patients, 53 down-regulated and 20 up-regulated bacterial OTUs were observed in the smoking CAG patients; among the non-smokers, 85 down-regulated and 29 up-regulated bacterial OTUs were found in CAG patients compared to nCAG patients ( Figure 6A ). Venn analysis showed that 20 sharing bacterial OTUs might be related to CAG development. The drinking patients had 45 unique OTUs, while the non-drinking patients had four unique OTUs; the smoking patients had ten unique OTUs, while the non-smoking patients had 26 unique OTUs ( Figure 6B ). Here demonstrated the dominant OTUs in the stomach of chronic gastritis patients with different lifestyle status. Among the 20 sharing OTUs, the ten up-regulated OTUs belonged to genus Labrys, Rhodopseudomonas and Methylobacterium, while the ten down-regulated OTUs mainly belonged to Labrys and Pseudolabrys, suggesting that gastric Labrys play a key role in the development of CAG ( Figure 6C ). In drinking CAG patients, among the 45 unique OTUs, most of them were significant down-regulated, and belonged to genus Bradyrhizobium, Aliidiomarina and Halomonas ( Figure S4A ). In non-drinking CAG patients, among the four unique OTUs, the increasing gastric OTU341 belonged to phylum Actinobacteria ( Figure S4B ). In smoking CAG patients, among the ten unique OTUs, increasing OTU439 (genus Ralstonia) and decreased OTU1067 (genus Labrys) were observed ( Figure S4C ). Among the 26 unique OTUs in the non-smoking patients, CAG patients occupied an elevated OTU323 (genus Helicobacter) with decreasing OTU21 (genus Prevotella 7) and OTU704 (genus Acidovorax) ( Figure S4D ). Compared to nCAG patients, most non-Helicobacter bacteria were significantly decreased in CAG patients, which was consistent with the decreased richness of gastric microbiota in CAG patients, suggesting that Hp-infection may inhibit the proliferation of non-Helicobacter bacteria in the stomach of CAG patients.

Moreover, the present study explored the effects of smoking, alcohol consumption, bile reflux, and inflammatory activity on the gastric microbiota in CAG and nCAG patients, respectively ( Figure S5 ). Comparative analysis was conducted between smoking and non-smoking individuals, only one different OUT was observed in the stomach of both nCAG and CAG patients ( Figure S5A ). Similarly, comparative analysis was conducted between drinking and non-drinking individuals, even no different OUT was observed in the stomach of both nCAG and CAG patients ( Figure S5B ). These results suggested that gastritis development stage, but not smoking/drinking, play a key role in shaping the gastric microbiota. Compared to the patients without inflammatory activity, there were 12 down-regulated and 24 up-regulated OTUs in the stomach of nCAG patients, and 56 down-regulated and six up-regulated OTUs in the stomach of CAG patients. Three sharing differential OTUs (OTU1, OTU60, OTU194) belonged to genus Helicobacter, and OTU1969 belonged to genus Streptococcus ( Figure S5C ), suggesting that Helicobacter bacteria are the key bacteria inducing gastric inflammatory activity. Due to the low prevalence rate of bile reflux in nCAG patients, here just observed the effect of bile reflux on gastric microbiota in CAG patients. Compared with CAG patients without bile reflux, 307 up-regulated OTUs were observed in the stomach of CAG patients with bile reflux, and most of the up-regulated OTUs were oral dominant bacteria, such as Veillonella, Prevotella 7 and Rothia ( Figure S5D ). These results suggested that bile reflux could promote the colonization of oral microbiota in the gastric mucosa of CAG patients.

In summary, several interrelated Helicobacter bacteria were observed in the stomach of gastritis patients with Hp-infection. At the phase of nCAG, Helicobacter bacteria were almost unrelated to the non-Helicobacter bacteria, and virulent Hp-I infection decreased gastric dominant bacteria including Aliidiomarina, Reyranella, Halomonas, Pseudomonas and Acidovorax. In case of CAG development, bile reflux increased oral microbiota (Rothia, Prevotella, Veillonella) to colonize in the gastric mucosa, and in the virulent Hp-infected patients, Helicobacter bacteria linked closely with non-Helicobacter bacteria (Bradyrhizobium, Psedolabrys, Ralstonia, Mesorhizobium and Variovorax), and inhibited oral microbiota from colonizing in the gastric mucosa ( Figure 7 ). Thus, the results bring new understanding of etiology in the development of chronic gastritis.