CagA⁺ H. pylori in the present study was isolated from the gastric specimens of a gastric ulcer patient during gastroscopy at the Third Xiangya Hospital of Central South University (Changsha, Hunan, China), and its identity confirmed using the complete sequence data of H. pylori 16s rRNA gene from GenBank data and positive biochemical tests as described (10). CagA^– H. pylori (ATCC 51932) were purchased from American Type Culture Collection (ATCC, Manassas, VA, United States). Both strains were cultured for 3–4 days on Columbia blood agar plates supplemented with antibiotics and 10% sheep blood (Fisher Scientific 50863755, Waltham, MA, United States) under a microaerophilic milieu (5% O2, 10% CO2, and 85% N2) at 37°C. The concentration of H. pylori was determined by measuring the optical density at OD 600 nm, where 1 unit of OD 600 nm corresponds to about 2 × 108 colony-forming unit (CFU)/ml (10).

Human gastric epithelial cell line (GES-1) was obtained from Professor Canxia Xu in Department of Gastroenterology, the Third Xiangya Hospital, Central South University; Changsha, Hunan, China. Human umbilical vein endothelial cells (HUVECs) and mouse brain microvascular endothelial cells (bEnd.3) were purchased from ATCC and cultured in RPMI-1640 (Gibco, Grand Island, NY, United States), endothelial cell medium (Sciencell Research Laboratories, Carlsbad, CA, United States), and DMEM (Gibco, Grand Island, NY, United States), respectively, supplemented with 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY, United States), 100 U/ml penicillin, and 100 mg/ml streptomycin in a controlled humidified incubator with 5% CO2 and 95% room air. To evaluate the effect of exosomes on endothelial cells, exosomes (100 ug/ml) from conditioned media of GES-1 co-cultured with CagA⁺ H. pylori, CagA^– H. pylori or PBS without H. pylori, or exosomes from the serum of mice with CagA⁺ H. pylori, CagA^– H. pylori infection or without H. pylori infection were cultured with HUVECs or bEND.3. After being cultured for 4 h, HUVECs or bEND.3 were tested for ROS production using the fluorescent dye 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA, Invitrogen D399, Waltham, MA, United States) as described (14). GES-1 was cultured with H. pylori at a MOI (multiplicity of infection) of 100 for 12 h as described (10).

All animal experiments were performed in accordance with the “Guide for the Care and Use of Laboratory Animals of the US National Institutes of Health.” The experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Missouri School of Medicine, Columbia, MO, United States. Specific-pathogen-free 4–6 week-old male C57BL/6 wild-type mice (WT) and LDLR knockout mice (LDLR^–/–) were from Jackson Lab (ME, United States), and were fed rodent diet and water ad libitum.

After fasting overnight, mice were given 0.2 ml of PBS, CagA^– H. pylori or CagA⁺ H. pylori inoculums (approximate 4 × 109 CFU/ml) by intragastric gavage as described (10). The presence of H. pylori infection was assessed at 1-week post-infection or at the time when mice were sacrificed, using Rapid Urease Test (RUT) and Giemsa staining of gastric mucosa. To evaluate the effect of H. pylori infection on the development of early atherosclerosis, LDLR^–/– mice were fed a high fat diet (HFD) (Envigo TD.88137, Indianapolis, IN, United States) beginning 1 week after H. pylori gavage. WT mice were sacrificed 1 week after the last gavage to collect thoracic aorta to determine endothelium-dependent and -independent vascular relaxation responses and ROS production (see details below). LDLR^–/– mice were sacrificed after week 3, 5, and 12 of HFD feeding to collect whole aorta and aortic root for atherosclerotic lesion analysis (see details below).

Thoracic aorta was isolated from mice to evaluate endothelium-dependent and -independent vascular relaxation responses as described (10). The thoracic aorta was cut into 2–3 mm segments and mounted onto a four-channel Wire Myograph System (610M; DMT, Aarhus, Denmark). Aortic segments were equilibrated with a resting tension of 4.9 mN for 45–60 min in a temperature-controlled tissue bath filled with 5 ml of Krebs’ solution and bubbled continuously with 95% O2 and 5% CO2, at 37°C. The aortic preparations were then tested for maximal contraction with 50 mM KCl, and concentration-dependent vasocontractile responses to phenylephrine (PE). After adequate washout with Krebs solution and equilibration, the aortic tissues were examined for endothelium-dependent relaxation to cumulative doses of acetylcholine (ACh, 10^–9–10^–5M) and endothelium-independent relaxation to cumulative doses of nitroglycerin (NTG, 10^–9–10^–5M) after submaximal contraction with PE.

To further confirmed the role of oxidative stress on endothelial dysfunction caused by H. pylori infection, the antioxidant NAC was used to treat the mice in vivo and endothelial cells in vitro. NAC is an FDA-approved drug and has been traditionally considered an antioxidant that effectively attenuates ROS production (15).

Mice in the NAC treatment group received NAC (Sigma-Aldrich, MO, United States, 1 mg/ml in drinking water) 3 days before the first gavage until the end of the experiment (16). NAC was changed every other day and covered with aluminum foil to avoid exposure to direct light. For the in vitro study, endothelial cells were incubated with 10 mM NAC as described (17).

Atherosclerotic burden in thoracic aorta and cross sections of aortic root was quantified with Oil Red O staining as described (18). Briefly, the thoracic aorta was collected immediately from euthanized mice and carefully prepared with the removal of periadventitial fat after perfusion with sterile phosphate-buffered saline (PBS). Thoracic aorta was cut open longitudinally after being washed twice with 60% isopropanol and images taken with a digital camera after being stained with Oil Red O dye for 30 min at room temperature. Aortic roots were frozen with optimum cutting temperature O.C.T. Compound (Fisher, Waltham, MA, United States) and serial sections (10-μm thick) were cut through the aortic valve as described (18). The sections were stained with Oil Red O for plaque quantification in the cross-section areas. The total lesion area and lesion area percentage were analyzed and quantified using Image J software.

Dihydroethidium (DHE, Invitrogen D23107, Waltham, MA, United States) was used to assess ROS formation in cryostat sections of aorta using fluorescence microscopy (19). Cryostat sections (5 μm) of mouse aortic rings were incubated with 5 μM DHE in normal physiological saline solution (NPSS, composition in mmol L-1: NaCl 140, KCl 5, CaCl2 1, MgCl2 1, glucose 10, and HEPES 5) for 7 min as described (19). Aortic rings were washed three times to remove DHE after incubation, and then examined with a fluorescence microscope. The fluorescence intensity was evaluated with 518 nm excitation and 606 nm emission, and the images were analyzed using Image J software.

The level of intracellular ROS in HUVECs or bEND.3 was evaluated using the fluorescent dye 2’7’-dichlorodihydrofluorescein diacetate (H2DCFDA; Invitrogen D399) (14) after treatment with exosomes from conditioned media or mouse serum as describe (14). H2DCFDA is a non-polar compound that is converted by cellular esterase to the polar and membrane impermeable derivative H2DCF. H2DCF is non-fluorescent but becomes highly fluorescent 2’,7’-dichlorofluorescein (DCF) when oxidized in the presence of intracellular ROS. After treatment with exosomes for 4 h, endothelial cells were washed twice with pre-warmed PBS, and then incubated with 15 uM H2DCFDA for 30 min at 37°C in the dark. Cells were washed twice with pre-warmed PBS after removing the dye. The fluorimetric signal in the cells was examined and analyzed using a fluorescence microscope. The fluorescence intensity was evaluated using ImageJ software.

To prepare exosomes from conditioned media, human gastric epithelium cells (GES-1) were cultured with PBS, CagA^– H. pylori or CagA⁺ H. pylori at MOI of 100 for 12 h. The conditioned media and mice serum were collected to isolate the exosomes as described (20). Briefly, the conditioned media were successive centrifuged at 4°C (300 × g for 10 min, 2,000 × g for 20 min, and 10,000 × g for 30 min) to eliminate the cells and cell debris. The supernatant was then ultracentrifuged at 100,000 × g at 4°C for 70 min for two times (Beckman Coulter, Indianapolis, IN, United States). Exosome pellets were re-suspended in PBS for further analysis. Serum from mice with CagA^– H. pylori or CagA⁺ H. pylori infection and control mice was collected for exosome isolation similar to conditioned media as described above (20). Exosomes identification of morphologies, size distribution, and biomarkers were assessed using a transmission electron microscopy (TECNAI G2 Spirit; FEI, Hillsboro, OR, United States), dynamic light scattering with a particle and molecular size analyzer (Zetasizer Nano ZS) and Western blotting as described (10).

Data were expressed as mean ± standard error of the mean (SEM) and analyzed with SPSS statistical software (22.0 for Windows; SPSS, Chicago, IL, United States). A two-tailed unpaired t-test was used for the analysis of two groups of data with normal distribution and equal variance, and a two-tailed unpaired t-test with Welch’s correction for analyzing two groups of data with normal distribution and unequal variance. Mann–Whitney U test was used for comparisons between two groups of data with abnormal distributions. One-way analysis of variance (ANOVA) was used for three or more groups of data analysis with normal distributions and equal variance, and the Kruskal–Wallis test with Dunn post-hoc multiple comparison tests was used for three or more groups of data analysis with normal distributions and unequal variance or abnormal distributions. P < 0.05 was considered significant.

As over 90% of Asian H. pylori patients are infected with CagA⁺ H. pylori, we hypothesized that CagA⁺ H. pylori would exhibit greater gastric colonization than CagA^– H. pylori. A total of 110 male C57BL/6 mice were divided into 11 groups (10 mice in each group) and infected with either CagA⁺H. pylori or CagA^– H. pylori using 4 different infection protocols (Table 1). Phospho-buffered saline (PBS) treatment served as negative control. Animals were considered infected when both RUT and Giemsa staining were positive. A 100% infection rate was achieved in mice receiving three daily doses of CagA⁺ H. pylori, whereas six doses of CagA^– H. pylori were required to achieve a 100% infection rate (Table 1). CagA⁺ H. pylori exhibited a higher gastric colonization rate than CagA^– H. pylori under all infection conditions. All control mice were negative for H. pylori infection, confirming that no H. pylori contamination existed in the animal facility.

To determine if a significant difference in endothelial dysfunction existed between mice infected with CagA⁺ H. pylori and CagA^– H. pylori, both acute (1 week) and chronic (12 weeks) infection models were established using C57BL/6 mice with PBS as control. Ex vivo acetylcholine (ACh)-induced endothelium-dependent relaxation of aortic rings was significantly decreased in mice with CagA⁺ H. pylori infection at both 1 week and 12 weeks post-infection compared with control (Figures 1A,B). Endothelium-independent relaxation to nitroglycerin (NTG) was unchanged between the groups (Supplementary Figures 1A,B). In contrast, no significant changes in ACh-induced relaxation (Figures 1C,D) or NTG-induced relaxation (Supplementary Figures 1C,D) were observed in mice with CagA^– H. pylori infection compared with control.

To determine if CagA⁺ H. pylori infection promotes the development of atherosclerosis, LDLR^–/– mice were infected with CagA⁺ H. pylori or CagA^– H. pylori with PBS as control. Infection with CagA⁺ H. pylori significantly increased aortic atherosclerotic lesion areas in mice compared with PBS-treated controls after 3 and 5 weeks of HFD. However, no significant differences in aortic atherosclerotic lesion areas were present between LDLR^–/– mice with CagA^– H. pylori infection and PBS-treated controls. After 12 weeks of HFD, all LDLR^–/– mice developed extensive atherosclerotic lesions with or without CagA⁺ H. pylori or CagA^– H. pylori infection (Figures 1E–N). These data suggest that CagA⁺ H. pylori, but not CagA^– H. pylori infection, accelerates early atherosclerotic development in LDLR^–/– mice on an HFD.

To test the hypothesis that CagA⁺ H. pylori infection impairs endothelial function through increased ROS production, aortic vasodilation and ROS levels were assessed in C57BL/6 mice infected with CagA⁺ H. pylori or CagA^– H. pylori and controls receiving PBS. CagA⁺ H. pylori infection significantly decreased ACh-induced aortic relaxation in association with significantly increased aortic ROS production compared with their controls (Figure 2A). In contrast, CagA^– H. pylori infection had no significant effect on either aortic ACh-induced relaxation or ROS levels compared with controls (Figure 2A). There was no change in NTG-induced aortic relaxation in mice infected with CagA⁺ H. pylori or CagA^– H. pylori (Supplementary Figure 1). N-acetylcysteine (NAC) treatment effectively blocked excessive ROS production in aortic segments (Figure 2B) and preserved ACh-induced relaxation in mice after 1 or 12 weeks of CagA⁺ H. pylori infection (Figures 2C,D), without differences in NTG-induced relaxation (Figures 2E,F).

CagA⁺ H. pylori, not CagA^– H. pylori infection also significantly increased aortic ROS production in LDLR^–/– mice with 3 or 5 weeks of HFD feeding compared with their controls (Supplementary Figures 2A,B). After 12 weeks of HFD feeding, all LDLR^–/– mice exhibited increased ROS production with or without CagA⁺ H. pylori or CagA^– H. pylori infection (Supplementary Figure 2C).

To test the hypothesis that CagA⁺ H. pylori, but not CagA^– H. pylori, infection promotes exosomes production, leading to endothelial dysfunction, serum exosomes were prepared from mice with CagA⁺ H. pylori and CagA^– H. pylori infection as well as non-infected control mice. Exosomes were characterized using transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and western blotting (Figures 3A–C). Although there was no significant difference in serum exosomes level from mice infected with either H. pylori strain or controls (Figure 3D), treatment with serum exosomes from mice infected with CagA⁺ H. pylori significantly inhibited the function of mouse bEND.3 cells in vitro with decreases in migration, tube formation and proliferation compared with control exosomes (Figures 3E–G). Culture with serum exosomes from mice with CagA^– H. pylori infection had no significant effect on endothelial function (Figures 3E–G).

Exosomes from conditioned medium of human gastric epithelial cells (GES-1) cultured with CagA⁺ H. pylori or with CagA^– H. pylori were prepared and similarly characterized by TEM, NTA and western blotting (Figures 4A–C). Co-culture of GES-1 with CagA⁺ H. pylori, not CagA^– H. pylori, significantly increased the exosomes level in conditioned media (Figure 4D). When using the same amount of exosomes (by protein level), exosomes from GES-1 co-cultured with CagA⁺ H. pylori significantly inhibited the functional properties of HUVECs in vitro with decreased proliferation, migration, and tube formation compared with non-infected control, while exosomes from GES-1 co-cultured with CagA^– H. pylori did not significantly impact endothelial cell function (Figures 4E–G).

To test the hypothesis that CagA-containing exosomes impair endothelial function via ROS-mediated mechanisms, mouse bEND.3 cells were treated with serum exosomes from mice infected with CagA⁺ H. pylori, or CagA^– H. pylori or from non-infected control mice. HUVECs were treated with exosomes from conditioned medium of GES-1 co-cultured with CagA⁺ H. pylori, CagA^– H. pylori or PBS. H2DCFDA assay showed that serum exosomes from mice with CagA⁺ H. pylori, not CagA^– H. pylori infection, significantly increased intracellular ROS formation in mouse bEND.3 cells compared to exosomes from non-infected controls (Figure 5A). Similarly, exosomes from conditioned media of GES-1 co-cultured with CagA⁺ H. pylori, not CagA^– H. pylori, significantly increased intracellular ROS levels in HUVECs (Figure 5B). CagA-containing exosomes-induced increase in intracellular ROS formation was associated with decreased endothelial function with reduced proliferation, migration, and tube formation in both bEND.3 (Figures 5C–F) and HUVECs (Figures 5G–J). NAC treatment decreased intracellular ROS production, and preserved function of bEND.3 (Figure 5C) and HUVECs (Figure 5G) in the presence of CagA-containing exosomes (Figures 5D–F,H–J). These data suggest that the effect of CagA-containing exosomes on endothelial function was indeed mediated via ROS formation.

To further test the hypothesis that CagA⁺ H. pylori infection impairs endothelial function through CagA-containing exosomes-mediated ROS formation, C57BL/6 mice were pre-treated with GW4869 to block exosomes release in vivo. As detailed earlier, CagA⁺ H. pylori infection significantly increased aortic ROS level (Figure 2A) in mice in association with impaired ACh-induced relaxation (Figures 1A,C). Treatment with GW4869 significantly reduced serum exosomes levels (Figure 6A), prevented excessive aortic ROS production (Figures 6B,C) and preserved ACh-induced relaxation in mice with CagA⁺ H. pylori infection (Figures 6D,E).

Treatment with GW4869 significantly decreased serum exosomes level in LDLR^–/– mice with H. pylori infection (Figure 6F), and effectively prevented excessive aortic ROS formation (Figures 6G,H). Blocking exosomes release with GW4869 also prevented CagA⁺ H. pylori infection-induced increases in aortic atherosclerotic burden. At three and 5 weeks of HFD feeding, there were no differences in the prevalence of aortic atherosclerotic lesion and total lesion areas between LDLR^–/– mice with CagA⁺ H. pylori infection with GW4869 treatment and non-infected LDLR^–/– control mice (Figures 6I–N). No significant effects of DMSO (vehicle for GW4869) on aortic ROS formation or aortic atherosclerotic burden were observed in hyperlipidemic LDLR^–/– mice with CagA⁺ H. pylori infection (Figures 6A–N). Treatment with GW4869 had no significant effect on atherosclerosis in hyperlipidemic LDLR^–/– mice without CagA⁺ H. pylori infection (Supplementary Figure 3). These data suggest that CagA⁺ H. pylori infection leads to significant endothelial dysfunction and increased atherosclerosis through CagA-containing exosomes-mediated ROS formation.

The limitations in the present study include: (1) only male mice were used; (2) no studies were performed to determine the specific molecules in CagA-containing exosomes that could be primarily responsible for increasing ROS production and endothelial dysfunction and related mechanism(s); and (3) no studies were conducted to define the key pathway(s) that may significantly contribute to increased ROS levels in endothelial cells with CagA⁺ H. pylori infection.