Gastric tumorigenesis induced either by Helicobacter pylori infection or chronic alcohol consumption through IL-10 inhibition

Background Alcohol is class 1 carcinogen and results in 3.3 million deaths every year. H. pylori is also an important factor for gastric carcinogen. Alcohol consumption is emerging as an important contributor to gastric cancer, but there is no direct or experimental evidence of alcohol and H. pylori infection produce gastric cancer in human and animal model alone. Here, we provide insight into the molecular mechanisms driving gastric carcinogenesis. Results Alcohol consumption, together with H. pylori infection, causes gastric cancer; interleukin-10 (IL-10) downregulation and mitochondrial metabolic dysfunction in CD8+ cells are also involved. IL-10 knockout accelerates tumor development in mice with either H. pylori infection or alcohol induced gastric cancer or both. IL-10 downregulation and CD8+ cell dysfunction stimulates IL-1β secretion. Specifically, we show IL-10 inhibits glucose uptake and glycolysis and promotes oxidative phosphorylation with lactate inhibition. Consequently, In the absence of IL-10 signaling, CD8+ cells accumulate damaged mitochondria in a mouse model of gastric cancer induced with the combination of alcohol plus H. pylori infection, and this results in mitochondrial dysfunction and production of IL-1 β. IL-1β promotes H. pylori infection and reduces NKX6.3 gene expression, resulting in increased cancer cell survival and proliferation. Conclusions Overall, the molecular mechanisms of gastric carcinogenesis include IL-10 inhibition resulting in lowered host immunity via mitochondrial dysfunction of CD8+ lymphocytes; gastric inflammation due to H. pylori infection, alcohol intake, and IL-1β production; and disruption of gastric-specific tumor suppressor NKX6.3 expression, which increases cancer cell survival and proliferation.

1. Introduction 32 weeks) as the initiating process, which led to tumor development (at 40 weeks) in the stomach of mice with H. pylori infection (Fig. 1b). Macroscopic analysis of tumor area revealed development of tumors only in mice treated with both H. pylori and alcohol (52.88 ± 9.085 mm 2 ; p < 0.001; f = 33.88; Fig. 1c). Solid tumor growth in mice treated with both alcohol and H. pylori confirmed the combinational role of alcohol and H. pylori in gastric tumorigenesis (Fig. 1c). Overall, these results indicated that H. pylori and alcohol together caused a progressive shift from gastritis to gastric cancer. These results Histological analysis also showed multifocal elongation of the gastric pits, glandular atrophy, and a significant reduction in the glandular zone in atrophic foci. These changes were significantly less in mice treated with alcohol and infected with H. pylori exhibiting tumors, compared to mice treated with alcohol only (2.20 ± 0.2) or infected with H. pylori only (1.6 ± 0.245; p < 0.001; f = 27.50; Supplementary Fig. 1a-d). Dual treatment resulted in reduction of parietal cell components, especially parietal and chief cells, which were replaced by gastric mucus-producing cells. Pepsinogen I and II, gastrin, somatostatin, and H + /K + ATPase showed substantially lowered expression in mice exhibiting gastric tumorigenesis compared with untreated mice or mice treated singularly (p < 0.001; f = 315.9; Supplementary Fig. 2a-g). Gastrin, somatostatin, and H + /K + ATPase are required for normal gastric mucosal development and parietal cell activation [15,16].
In mice exhibiting carcinogenesis, these biomolecules were almost completely ablated. As a result, a combination of alcohol and H. pylori directly accelerates loss of differentiated epithelial cell types, leading to chronic atrophic gastritis and eventually to gastric tumorigenesis. The combinatorial treatment shifts gastric tissue to a tumor phenotype with increased numbers of inflammatory cells found in the tumor gastric mucosa. This observation suggests that H. pylori or alcohol alone is not enough to cause gastric tumorigenesis; a combination is required to trigger the formation of submucosal glands and invasion of tumor cells.
This effect is associated with increased tumor progression due to substantially reduced expression of several gastric tumor suppressor genes (TSGs), including Tff1 (0.329 ± 0.087), Tff2 (0.218 ± 0.041), Gkn1 (0.246 ± 0.080) and Gkn2 (0.372 ± 0.135) in the mouse group that developed tumors compared to other groups (p < 0.001;. Loss of these gastric-specific TSGs promotes tumorigenesis [17,18,19]. Tff1 and Tff2 genes are upstream regulators of gastrokine (Gkn) gene expression. Therefore, loss of Tff1 or Tff2 expression could lead to tumor growth [20]. These genes were substantially downregulated in mice exhibiting gastric tumors, and the reduction was associated with the combined treatment of alcohol and H. pylori infection. Overall, inhibition of Tff1, Tff2, Gkn1, and Gkn2 is known to be associated with tumorigenesis [20,21].
The extent of gastric tissue damage was evaluated as well. We found that mice treated simultaneously with alcohol and H. pylori exhibited tumor development. These mice displayed significantly higher activities of malondialdehyde (MDA; 364.479 ± 34.583 mg/mL), myeloperoxidases (MPO; 0.548 ± 0.060 U/L), catalase (28.369 ± 2.392 U/mL), carbonyl (102.558 ± 19.085 nmol/mg), and lipoperoxidase (LPO; 173.853 ± 27.849 µmol/L) compared to mice treated with either H. pylori or alcohol alone (p < 0.001; Supplementary   Fig. 4a-e). A combination of alcohol and H. pylori infection increases peroxidation processes and leads to the production of superoxide radicals and aldehydes, such as MDA, MPO, and hydrogen peroxide, that can form adducts with DNA and proteins. Subsequently, we proposed that H. pylori infection and alcohol metabolites, including acetaldehydes, aldehydes, and peroxides, damage DNA, DNA repair enzymes, and immune cell protein cytokines, resulting in gastric tissue damage [22,23,24].

Alcohol dose in mice is relevant to human alcohol consumption and stagedependent gastric cancer development
To apply our observations to humans, we optimized the mice's alcohol intake to reflect human intake to determine the dose-dependent (0.5-5 g/kg/day) effect of alcohol consumption. Mice treated with a combination of H. pylori infection and alcohol developed gastritis, leading to cancer in a dose-dependent manner (Fig. 2a,also 2g). A dose greater than 3.5 g/kg/day showed significant gastric carcinogenesis. Consequently, we treated mice with 5 g alcohol/kg/day to induce gastric carcinogenesis. This dose is non-toxic, and more importantly, it is comparable to the alcohol consumption of human gastric cancer patients (0.405 g/kg/day) [8]. These results suggest that alcohol dosage contributes substantially to the development of gastric inflammation and carcinogenesis.
We also examined alcohol intake, metabolism, and excretion and found a high blood assimilation of alcohol in mice with gastritis (0.344 ± 0.024 g/kg/day) or gastric tumors (0.654 ± 0.041 g/kg/day) compared to mice treated only with alcohol (0.074 ± 0.027 g/kg/day; p < 0.001). This observation corresponded with low urine dissimilation of alcohol in the gastric tumor group (0.103 ± 0.012 g/kg/day) compared to the gastritis group (0.181 ± 0.011 g/kg/day) or the group treated with only alcohol (0.620 ± 0.066 g/kg/day; p < 0.001). Similar results were found in human samples; gastric cancer patients (0.025 ± 0.000 g/kg/day) and chronic gastritis patients (0.022 ± 0.001g/kg/day) showed high levels of alcohol blood assimilation compared to healthy subjects (0.005 ± 0.001g/kg/day; p < 0.001; Fig. 2b, c). Alcohol dehydrogenase (i.e., an alcohol metabolizing enzyme) activity was significantly higher in mice exhibiting gastric tumorigenesis (93.259 ± 15.213 U/mL) compared to mice with the cohort treated with only alcohol (34.829 ±10.132 U/mL;p < 0.001;Fig. 2d). Moreover, mice with gastritis and gastric tumorigenesis exhibited a dark brown urine color with alkaline pH (Fig. 2e, f), suggesting the presence of a high level of alcohol dehydrogenase (ADH). These results also confirmed the role of alcohol consumption in the development of physiological abnormalities and gastric cancer.
High levels of ADH metabolites in urine following alcohol consumption may be a sign of gastric tumor development as well. The risk for alcohol-related cancer is influenced by the expression of alcohol metabolizing enzymes, such as ADH. Many Chinese, Korean, and Japanese individuals carry a version of the gene for ADH that codes for an overactive form of the enzyme [25]. This overactive ADH enzyme speeds the conversion of alcohol (ethanol) to toxic acetaldehydes [26,27] and increases the risk for cancers like pancreatic, esophageal, liver, colorectal, and female breast cancers [28]. As a result, epidemiologists have regarded alcohol and acetaldehydes as potential carcinogens [10,29]. Ultimately, we developed a mouse model that utilizes an alcohol dosage relevant to human alcohol intake. With this model, gastric tumorigenesis is induced without toxicity in peripheral organs (Fig. 2g). As expected, activities of major alcohol metabolites (Acetaldehyde) were increased after combinational treatment of ethanol and H .pylori infection (45.382± 2.264 µM; p < 0.0001; Fig. 2h). As a result, this mouse model was used to study the effect of alcohol and its metabolites on the development of gastric cancer.

IL-10 knockout accelerates tumor growth in alcohol-treated and pyloriinfected mice
Gastric inflammation leads to gastric tumor development. In our mice model, the number of pro-inflammatory cell types was significantly increased (i.e., leukocyte infiltration), as Similarly, IL-10 levels showed reduction in human gastric cancer and gastritis serum samples compared to healthy subjects (p < 0.001; Fig. 3a-g). Our data suggested that alcohol intake combined with H. pylori infection causes a marked imbalance in IL-10 cytokine levels (p < 0.001; Fig. 3a-g). We evaluated the association of gastric cancer with the combination of H. pylori and alcohol consumption. Figure 4H shows the ROC-curve for the discrimination IL-10 and H. pylori expression status with alcohol consumption data.
The area under the curve (AUC) was found to be HP (area: 0.927, CI: 0.860-0.994), Cancer stage (area: 0.836, CI: 0.740-0.931) and IL-10 (area: 0.939, CI: 0.882-0.996) were associated with alcohol consumption (Supplementary Tables 1). Overall, the H. pylori infection and alcohol consumption showed great sensitivity and specificity towards IL-10; especially this combination could be used to confirmation of gastric cancer development.
We further examined the functional impact of IL-10 loss on gastric inflammation leading to gastric tumorigenesis induced by the combination of H. pylori infection and alcohol consumption. IL-10 ablation accelerated the development of gastric tumorigenesis in mice carrying IL-10-null alleles (Fig. 4a). Macroscopic morphometric analysis revealed that loss of the IL-10 gene accelerated the process of tumor development in mice treated with a combination of alcohol and H. pylori infection up to 32 weeks (p < 0.0001; Fig. 4b). These data support the view that IL-10 might act locally to resist chronic gastritis and subsequent gastric tumor development. We analyzed IL-10 -/mice for changes in acetaldehyde, CD8 + cells and IL-1β levels, after alcohol or H. pylori stimulation. IL-10 -/mice treated with alcohol and H. pylori showed high acetaldehyde levels compared to mice treated alcohol or H. pylori alone or wild-type (WT) (p < 0.0001; Fig. 4c-e). Furthermore, the immune and inflammatory profile of IL-10 KO mice was analyzed by quantifying CD8 + cells and IL-1β secretion in IL-10 knockout mice treated with alcohol and H. pylori infection. We found significantly diminished numbers of CD8 + immune cells (p < 0.0001;

IL-10-deficient CD8 + cells exhibit altered metabolic profiles after pylori infection and acetaldehyde stimulation
Functional mitochondria are crucial for tumor maintenance [30,31]. However, the role of mitochondria in immune cell function is largely unknown. Consequently, we examined the mitochondrial respiratory capacity of immune cells. We analyzed the effect of alcohol and H. pylori in several immune cell types: macrophages (CD-11b), CD-4+, CD-8+, NK (CD-49b) Fig. 8a-f). For instance, IL-10-/-mice showed greatly diminished CD8 + percentages compared to WT mice upon alcohol metabolite and/or H. pylori addition (p < 0.0001; Fig. 4f-h).

and B-cells (CD-19) and dendritic cells (CD-11c) (Supplementary
Furthermore, we analyzed the effect of alcohol consumption and H. pylori infection on mitochondrial protein by mRNA sequencing. RNA sequencing (RNA-seq) data showed that phosphoglycerate mutase (PGM2) predominantly expressed in WT group compared to mice treated with alone alcohol, H. pylori, or both alcohol and H. pylori) as well as 6phosphofructo-2-kinase/fructose-2, 6-biphosphatase 3 (Pfkfb3). However, PGM2 expression was affected by IL-10 production (p < 0.001; Fig. 5A). PGM2 is a glycolytic enzyme, whereas pfkfb3 gene has the highest kinase and phosphatase activity ratio, which in turn sustains high glycolytic rates. As a result, we investigated whether ablation of IL-10 signaling inhibits PGM2 expression in IL10 -/cells. To test this, we analyzed PGM2expression at the mRNA level in IL-10 -/mice. PGM2 expression was downregulated by IL-10 knockdown after alcohol stimulation and H. pylori infection (p < 0.0002; Fig. 5bd). These data illustrate that IL-10 critically play an important role in regulate the glycolitic flux through glucose metabolism.
We next asked whether the inhibition of glycolysis by IL-10 is due to suppression of glycolytic flux. RNA-sequencing data showed that mice treated with both alcohol and H. pylori predominantly expressed Glut1 and Glut10 compared to mice treated with alcohol or H. pylori alone (p < 0.0001; Supplementary Fig. 10a-c). Specifically, Glut-10 expression was perturbed in IL-10 -/-CD8 + cells with stimulation of acetaldehyde and H. pylori infection. However, in the absence of exogenous IL-10, glucose uptake was maintained at higher levels in IL-10 -/-CD8 + cells. In contrast, in the presence of exogenous IL-10, glucose uptake was reduced in IL-10 -/-CD8 + cells (p < 0.0001; Supplementary Fig. 10d-e).
Additionally, PKM2 production was elevated in alcohol + H. pylori infected IL-10 -/mice ( Fig. 5e). Lactate dehydrogenase levels were elevated as well (LDH-a & LDH-b), (p < 0.0001; Supplementary Fig. 11a-b). Higher glucose uptake was likely utilized for lactic acid fermentation by H. pylori, since only alcohol-treated mice showed low glucose uptake and low levels of lactate dehydrogenase production (p < 0.0001; Supplementary Fig. 10ac). High glucose uptake shifts mitochondrial metabolism from oxidative phosphorylation to glycolysis in IL-10 -/-CD8 cells, which accounts for the decrease in oxidative phosphorylation and CD8+ cell population discussed previously ( Fig. 5e-f). Previous studies also demonstrate that IL-10 inhibits glucose translocation via downregulation of glycolytic gene expression [30]. Our RNA-seq analysis in IL-10 KO mice revealed that IL-10 inhibits gene expression of mitochondrial glycolytic pathway enzymes, including Aco, Pgm2, Mot2, Fabp, Mdh1, Glut, Pdk2 and Pfkp (Fig. 5a). Together, these data illustrate that IL-10 inhibits glycolytic flux by regulating GLUT10, a glucose transporter and gene expression of glycolytic enzymes.
IL-10 -/-T-lymphocytes (CD8 + ) were also analyzed for changes in mitochondrial content, total reactive oxygen species (ROS), and mitochondrial ROS production. IL-10 -/-Tlymphocytes exhibited decreased mitochondrial mass after acetaldehyde (500 µM) plus H. pylori (1x10 3 cells/mL) stimulation compared to WT T-lymphocytes or cells treated with acetaldehyde or H. pylori stimulation alone (p < 0.001; Supplementary Fig. 12a-c). The reduction in mitochondrial content could be due to apoptosis and clearance of dysfunctional mitochondria (MitoTracker Green+high) with loss of Δᴪm under acetaldehyde + H. pylori stimulation (Fig. 5i). Loss of Δᴪm is associated with accumulation of mitochondrial reactive oxygen species. We therefore examined whether accumulation of Δᴪm low mitochondria in IL-10 -/-T-lymphocytes was associated with ROS production.
Mitochondrial membrane potentials (MitoSox assay) were elevated in T-lymphocytes cells treated with acetaldehyde and H. pylori in the absence of IL-10 (p < 0.01; Supplementary   Fig. 13a-c). We evaluated ROS levels through activation of the electron transport chain (ETC) in mitochondria. Hydrogen peroxide (H 2 O 2 ) and mitochondrial superoxide levels were elevated in T-lymphocytes cells with combinatorial treatment compared to single treatment or WT (p < 0.001; Supplementary Fig. 14a-c). To assess ROS levels, we used the total ROS and mitochondria-specific ROS indicator DCFDA and MitoSox staining, respectively, to selectively detect mitochondrial superoxide. Total ROS and mitochondrial ROS production was enhanced in the absence of exogenous IL-10 and inhibited in the presence of exogenous IL-10. This observation correlated with total mitochondrial mass.
The accumulation of ROS-producing mitochondria in IL-10 -/cells was also visualized via live-cell imaging with fluorescent dyes, MitoTracker Green and MitoSox. Under alcohol and H. pylori stimulation, the absence of IL-10 resulted in lowered maximal respiratory capacity (MRC) compared with untreated CD8 cells. As a result, loss of mitochondrial fitness could be responsible for reduced IL-10 levels in CD8 + cells after alcohol treatment and H. pylori infection. Consistent with this idea, basal cellular MitoSox levels were also reduced in CD8 + cells after acetaldehyde and H. pylori infection. These data demonstrate metabolic reprogramming and altered ROS production as a consequence of combinational treatment of acetaldehyde (alcohol metabolite) and H. pylori. This combinatorial treatment contributes to metabolic dysfunction in CD8 + cells, leading to lowered host immunity and increased inflammation. We also assessed the functional profile of mitochondria based on cytokine levels in IL-10 -/-CD8 + (Fig. 5g) and IL-10 -/-CD11b + cells (p < 0.001; Supplementary Fig. 15a) by qPCR. IL-1β levels were elevated in combinatorial treatment of acetaldehyde and H. pylori compared to WT or IL-10 -/-CD8 + cells treated singularly(p < 0.0001; Fig. 5h). IL-1β secretion was caspase-1-dependent. Exogenous IL-10 reverse the mito-dysfunction and switch from glycolysis to oxidative phosphorylation and leading to survival of CD8 + cells with low gastric inflammation. As a result, we hypothesized that enhanced mitochondrial ROS production in IL-10 -/cells serves as an endogenous signal for inflammasome activation [30].

Ablation of IL-10 signaling lowers host immunity (CD8 + ) and promotes pylori infection
We quantified the colony forming units of H. pylori in IL-10 -/mice treated with alcohol; there was no significance difference between treatment groups and WT (p < 0.001; Fig.   6a-c). These results suggest that IL-10 expression has no significant effect on H. pylori growth or cell number. We also examined the rate of H. pylori infection by measuring H. Urease and CagA toxin secretion was enhanced in alcohol plus H. pylori-stimulated IL-10 knockout mice compared to those treated with H. pylori alone or WT.
We investigated the dose dependent effect of alcohol consumption on IL-10 and CD8 + inhibition and H. pylori CagA infection. We found gradual decrease of CD8 + and IL-10 levels in gastric cancer mouse model treated with continuing increase of alcohol dose (0.5-5g/kg/day) (p < 0.001; Fig. 6g-h). In contrast, H. pylori CagA infection was stimulated at high doses of alcohol (p < 0.001; Fig. 6i). These data suggest that alcohol consumption is strongly associated with the loss of IL-10 expression and lowered CD8 + cell count, resulting in low host immunity. As a result, we investigated whether ablation of IL-10 signaling promotes H. pylori infection in low immune environments in gastric mucosae [32].
To test this, we analyzed the expression of CagA mRNA in IL-10 -/-AGS gastric cancer cells stimulated with acetaldehyde, H. pylori, and IL-1β. CagA levels were dramatically increased in IL-10 -/cells compared to wild-type (WT) AGS gastric cancer cells (p < 0.001; Fig. 6j-k). Overall, these data illustrate that the tumor microenvironment is influenced by acetaldehyde (alcohol metabolites) and H. pylori infection with IL-10 ablation and IL-1β overexpression, ultimately leading to reduced CD8 + cell count (low immunity) and stimulation of H. pylori infection (Fig. 6l).

Discussion
Gastric cancer begins with inflammation and progresses to gastritis and later to gastric carcinogenesis [33]. Our mouse model confirms the contribution of alcohol in the induction of gastric inflammation leading to gastric cancer in the presence of H. pylori (Fig: 1 & 2).
In addition, the effects of interleukin-10 inhibition suggest that pre-existing gastric immunopathology accelerates the loss of differentiated epithelial cell types leading to profound glandular atrophy and gastric tumorigenesis. In sum, IL-10 depletion may promote the shift from gastric inflammation to gastric carcinogenesis in mice treated with a combination of alcohol and H. pylori infection (Fig: 3 & 4).
Using mouse models, this study is the first to provide experimental evidence for the epidemiological link between high gastric cancer incidence and regions with high alcohol consumption and H. pylori infection [8]. This study also reveals the critical role of interleukin-10 in lowered host immunity via CD8+ cell downregulation and IL-1β upregulation. Defects in IL-10 regulation can result in mitochondrial dysfunction, observed as a metabolic shift from oxidative phosphorylation to glycolysis in activated Tlymphocytes (CD8+). In these pro-inflammatory conditions, gastric tumor suppressor gene (NKX6.3) expression is downregulated as well, resulting in increased gastric cancer cell survival and proliferation. As a result, stimulation of anti-inflammatory cytokine IL-10 can serve as a therapy against gastric cancer, in addition to lowered alcohol intake and pathogen clearance via antibiotic treatment. Gastritis patients with low IL-10 levels could also benefit from IL-10 stimulation as a possible preventive measure against gastric carcinogenesis (Fig: 5).
Beyond the clinical applications of this work, this study provides insight into the molecular mechanisms of gastric tumorigenesis. In mice, alcohol metabolites, such as peroxides or acetaldehydes, bind to immune cell GABAA receptors that in turn dysregulate and degrade immune cell populations (e.g., CD8 and CD4) and anti-inflammatory cytokines (e.g., IL-10) by metabolic reprogramming. Lowered host immunity promotes H. pylori infection of gastric cells, which further inflames gastric tissue (Fig. 5a). Here we provide evidence for IL-10-dependent regulation of glucose metabolism, mitochondrial metabolism and inflammatory responses in activated CD8+ cells. For example, upon alcohol activation, IL-10-deficient T-lymphocytes (CD8+) exhibited reduced oxidative phosphorylation (OXPHOS). The reduction in OXPHOS also coincided with elevated glucose uptake (e.g. GLUT10) for lactic acid fermentation under H. pylori infection with alcohol stimulation (e.g. LDHA). This metabolic shift leads to lactate production via PKM2 and LDH-A activation in IL-10-/-deficient T-lymphocytes [34,35,36]. Exogenous IL-10 blocks GLUT-10, PKM2 and LDH-A, decreasing glycolysis by lactate production and restoring oxidative phosphorylation activity. These data combined suggested that IL-10 uses an alternative metabolic pathway to sustain CD8+ cells mitochondrial activity, like glutamine As a result; ablation of IL-10 signaling perturbs glucose metabolism by dysregulating the PGM2, PFKFB3, and PKM2 expression and affects mitochondrial metabolism since glucose is not more an available metabolic substrate. Mitochondrial dysfunction also explains CD8+ cell death in IL-10-deficient mice (Fig. 5). As a result, H. pylori infection is exacerbated by low host immunity. Pathogen-induced inflammation is coupled with increased IL-1β secretion in IL-10-/-CD8+ cells with mitochondrial dysfunction. This chronic inflammation predisposes gastric cells towards carcinogenesis (Fig: 6). downregulation, gastric cancer cell survival and proliferation is increased. We conclude that NKX6.3 acts as a tumor suppressor gene involved in maintenance of gastric homeostasis and prevention of gastric tumorigenesis (Fig: 7).

Conclusions
In sum, many molecular mechanisms contribute to gastric carcinogenesis. Here we outline the synergistic effects of cytokine production (i.e. IL-10 and IL-1β), T-lymphocyte dysfunction, H. pylori infection and alcohol consumption on gastric cancer development and progression. In the future, these mechanisms and their molecular targets offer treatment opportunities for alcohol-associated gastric cancer.
After one week of acclimation, mice were administered ethanol (5 g/kg/day) in drinking water for two weeks prior to infection with H. pylori, and alcohol administration was continued throughout the experiment. To facilitate H. pylori colonization, pantoprazole (20 mg/kg) was administered by gavage 3 times to lower gastric acidity. Each mouse was administered a suspension of H. pylori SS1 strain containing 10 8 CFUs/mL by gavage 3 times per week. Mice (n = 5) were euthanized on consecutive weeks of 4, 8, 12, 16, 20, 24, 28, 32, 36, and 40 weeks. A control group was maintained without any treatment.
Doses of alcohol ranging from 0.5-5.0 g/kg/day were administered to mice (n = 5) to determine the optimal alcohol dose to induce gastric tumorigenesis.

Patients and gastroendoscopy
All gastric patients were subjected to gastroendoscopy and examination in the Second Affiliated Zhengzhou University Hospital and Henan Cancer Hospital (Zhengzhou, Henan, China). Blood samples and tissue biopsies were obtained from consenting patients from the antral and corpus portions of the stomach during gastrointestinal endoscopy and gastric surgery. The patients who donated the primary tumors were completely informed and provided written consent (Access No: CUHCI2015009).
The signal was visualized with peroxidase-labeled streptavidin complexes by DAB, and the sections were briefly counterstained with hematoxylin. The immunohistochemical localization pattern was also recorded by digital imaging (Nikon Ti-DS, Japan). ImageScope (11.1.1.752) software program was used, and the labeling index was calculated as a percentage of positive cells relative to the total number of counted cells.

Real-time RT-PCR
Total RNA was extracted using the commercial RNA extraction kit (Ambion by Life Technologies, Van Allen Way Carsbad, USA, 92008) and cDNA synthesized by amfiRivert cDNA synthesis platinum master mix (GenDEPOT, Katy, TX, USA, Cat number R5600-200).
Real-time PCR (qRT-PCR) was conducted using a 7500 FastDX (Applied Biosystems, MA, USA) and the power SYBR green PCR master mix (Applied biosystem, Warrington WA1 4SR, UK, Cat number 4367659). Primer IDs and sequences are shown in Supplementary Tables 2, 3 and 4.

Cytokine and chemokine protein measurement using a multiplex bead array.
Cytokine and chemokine protein levels in mouse (Cat number 740446) and human (Cat number 740118) serum were measured using a multiplex magnetic bead array kit (customized by BioLegend LEGENDplex, San Diego, CA, USA). The multiplex bead arrays were performed according to the manufacturer's instructions with a minimum detectable concentration varying from 0.96-11.27 pg/mL. Legendplex (version: 7.0) software program was used to analyze the FACS data and the cytokine concentrations were calculated in pg/mL against standard values.

Urine collection and measurement (metabolic cage experiments)
Alcohol metabolism and urinary flow rate were determined by placing mice individually in metabolic cages. Mice were allowed a 3-day habituation period to adapt to the environment. Later, food and water intake, urinary flow rate, and body weight were recorded every day. Subsequently, a 12 h collection (9 p.m. to 9 a.m.) of urine was performed to obtain the urinary parameters, including volume, pH, and color. Data were analyzed for alcohol intake, metabolism, and excretion. Alcohol content in the serum and urine samples from mice and humans was measured according to the Enzychrom ethanol commercial assay kit's instructions (BioAssay System, ECET-048, Hayward, CA, USA, Cat number ECET-048).

Mouse-derived xenograft model
A mouse-derived xenograft model was developed from gastric tumor tissue of mice treated with alcohol (Access No: CUHCI2016011). Then additional mice were subcutaneously implanted with tissues weighing 0.10-0.12 g and measuring ~3 mm. Animals were monitored periodically for their weight and tumor growth. A second passage was performed and the same protocol was followed as described above.

RNA-seq analysis
Total RNA of stomach tissues was prepared by using an RNeasy kit (QIAGEN) with an RNase free DNase set (QIAGEN). Sample preparation and sequencing were performed in the BGI (Beijing Genomics Institute, Shenzhen Beishan Industrial Zone, Shenzhen 518083, China) using an Illimina HiSeq sequencing system at 50 bp read length.

Tissue harvesting for cell sorting and FACS analysis
Mice stomach tissue was minced into 3-4 pieces and pressed firmly to force the fragments apart; this allowed cells to pass through a wire mesh (

Measurement of mitochondrial activity and ROS production
Jurkat T-lymphocytes cells seeded in non-tissue culture plates were stimulated as

Statistical analysis
The experiments were randomized, and investigators were blinded to histological examination during all experiments. All statistical analyses were performed using Graphpad Prism 5.0 software (San Diego, CA, USA), with differences between groups considered significant with a p value <0.001. Data are presented as mean values ± S.E.M.
Histopathological scores and all other experimental data were compared using a t-test    Alcohol consumption in the development of gastric tumorigenesis. a) Dosedependent response of alcohol intake on progression from gastritis to cancerous cell growth. The lowest dose (0.5 g/kg/day) of alcohol had no effect on gastric pathology compared to an intermediate dose (1.0-3.5 g/kg/day) or a higher dose (4.0-5.0 g/kg/day). Scale bars represent 100 μm (10x); the scale bar in the inset images represents 20 μm (40x magnification), (n = 5/cohort). Arrow and A= atrophy, H= hyperplasia, MT= metaplasia, I= inflammatory cells, IN= muscular mucosae invasion. Association of alcohol intake, metabolism, and excretion in mouse and human gastric disease samples. Alcohol concentration in b) human (n = 10) and c) mouse serum and urine samples was used to calculate alcohol intake, alcohol absorption (serum), and excretion (urine). d) Alcohol dehydrogenase (ADH) levels in mouse serum were measured. Mice exhibiting gastritis and tumorigenesis showed higher ADH biochemical activity compared to alcohol-only-treated or untreated mice. e) Analysis of urine samples from the gastric disease mouse models. Physical characteristics of mouse urine, including volume, color, and pH, were recorded for 35 days. f) Urine samples were collected and color differences were recorded from the different groups of mice. g) An equivalent dose of alcohol in mice and humans was determined by using a Km value of 3 (0.810). h) Acetaldehyde (ACE) levels in mouse urine were measured.
Mice exhibiting gastritis and tumorigenesis showed higher acetaldehyde compared to alcohol-only-treated or untreated mice  IL-10 knockdown accelerates progression from gastritis to tumorigenesis in H.

Figure 6
Loss of IL-10 enhances H. pylori CagA infection due to lowered immunity upon