Maintenance of Type IV Secretion Function During Helicobacter pylori Infection in Mice

ABSTRACT The Helicobacter pylori type IV secretion system (T4SS) encoded on the cag pathogenicity island (cagPAI) secretes the CagA oncoprotein and other effectors into the gastric epithelium. During murine infection, T4SS function is lost in an immune-dependent manner, typically as a result of in-frame recombination in the middle repeat region of cagY, though single nucleotide polymorphisms (SNPs) in cagY or in other essential genes may also occur. Loss of T4SS function also occurs in gerbils, nonhuman primates, and humans, suggesting that it is biologically relevant and not simply an artifact of the murine model. Here, we sought to identify physiologically relevant conditions under which T4SS function is maintained in the murine model. We found that loss of H. pylori T4SS function in mice was blunted by systemic Salmonella coinfection and completely eliminated by dietary iron restriction. Both have epidemiologic parallels in humans, since H. pylori strains from individuals in developing countries, where iron deficiency and systemic infections are common, are also more often cagPAI+ than strains from developed countries. These results have implications for our fundamental understanding of the cagPAI and also provide experimental tools that permit the study of T4SS function in the murine model.

ABSTRACT The Helicobacter pylori type IV secretion system (T4SS) encoded on the cag pathogenicity island (cagPAI) secretes the CagA oncoprotein and other effectors into the gastric epithelium. During murine infection, T4SS function is lost in an immune-dependent manner, typically as a result of in-frame recombination in the middle repeat region of cagY, though single nucleotide polymorphisms (SNPs) in cagY or in other essential genes may also occur. Loss of T4SS function also occurs in gerbils, nonhuman primates, and humans, suggesting that it is biologically relevant and not simply an artifact of the murine model. Here, we sought to identify physiologically relevant conditions under which T4SS function is maintained in the murine model. We found that loss of H. pylori T4SS function in mice was blunted by systemic Salmonella coinfection and completely eliminated by dietary iron restriction. Both have epidemiologic parallels in humans, since H. pylori strains from individuals in developing countries, where iron deficiency and systemic infections are common, are also more often cagPAI 1 than strains from developed countries. These results have implications for our fundamental understanding of the cagPAI and also provide experimental tools that permit the study of T4SS function in the murine model. IMPORTANCE The type IV secretion system (T4SS) is the major Helicobacter pylori virulence factor, though its function is lost during murine infection. Loss of function also occurs in gerbils and in humans, suggesting that it is biologically relevant, but the conditions under which T4SS regulation occurs are unknown. Here, we found that systemic coinfection with Salmonella and iron deprivation each promote retention of T4SS function. These results improve our understanding of the cag pathogenicity island (cagPAI) and provide experimental tools that permit the study of T4SS function in the murine model. KEYWORDS Helicobacter pylori, Salmonella, type IV secretion system, cagY, pathogenicity island I nfection with Helicobacter pylori causes chronic gastric inflammation that sometimes progresses to peptic ulcer disease or gastric adenocarcinoma, which is the third most common cause of cancer mortality worldwide (1). The virulence locus most strongly associated with disease is the type IV secretion system (T4SS) encoded on the cytotoxin-associated gene pathogenicity island (cagPAI). The H. pylori T4SS is essential for injection of several known bacterial effectors into host cells, including the CagA previously described (29). Intravenous inoculation is followed rapidly by high bacterial burden in the spleen and expansion of IFN-g CD4 1 T cells, which peak 1 to 2 weeks later and largely dissipate by 6 to 8 weeks. C57BL/6 mice were orally gavaged with H. pylori PMSS1, injected with S. Typhimurium BRD509 intravenously (i.v.) 1 week later, and then sacrificed 3, 5, or 8 weeks postinoculation (p.i.) with H. pylori (Fig. 1A). As expected, S. Typhimurium challenge produced a robust systemic infection, with rapid colonization of the spleen that decreased over the course of the 8-week experiment (Fig. 1B). Salmonella was also recovered from gastric tissue, though in much smaller numbers. Serum IFN-g detected by enzyme-linked immunosorbent assay (ELISA) was markedly elevated 3 weeks after Salmonella challenge in mice infected with H. pylori and declined rapidly (Fig. 1C). Eight weeks p.i., when H. pylori colonization and T4SS were characterized, IFN-g levels were low, and mice had largely recovered from infection with Salmonella, which was no longer detectable in the stomach.
Salmonella coinfection decreases H. pylori colonization and enhances local and systemic inflammation. H. pylori colonization was assessed 8 weeks p.i. (7 weeks after Salmonella challenge), when cagY recombination and loss of T4SS function become apparent. Mice challenged with wild-type (WT) H. pylori harbored 10 5 CFU/g of gastric tissue, which was approximately 10-fold lower in mice that also received S. Typhimurium ( Fig. 2A). Colonization with H. pylori SS1cagY or H. pylori DcagE, which have a nonfunctional T4SS, was greater than with WT H. pylori but showed a similar decrease when coinfected with Salmonella (Fig. 2B). Gastric inflammation was induced by H. pylori infection and was somewhat increased in mice coinfected with S. Typhimurium (Fig. 3A). Since IFN-g is induced by Salmonella challenge (Fig. 1) and is important for control of H. pylori (16), we  3C) compared to that with H. pylori alone, though only the latter was statistically significant. Together, these results show that coinfection with Salmonella induces a systemic and local inflammatory response that is associated with reduced H. pylori colonization, independent of T4SS function.
Gastric overexpression of IFN-g is sufficient to decrease H. pylori colonization. To further characterize the role of IFN-g in controlling H. pylori colonization, we challenged heterozygous mice overexpressing mouse IFN-g under the control of the stomach-specific H 1 /K 1 ATPase b promoter (tgIFN-g). Gastric IFN-g transcript levels were increased by H. pylori infection and were markedly greater in tgIFN-g than in WT mice (Fig. 4A). Gastric inflammation was also increased in tgIFN-g compared to that in WT mice, both uninfected and 4 weeks p.i. (Fig. 4B), and was accompanied by decreased H. pylori bacterial load at 4 and at 8 weeks p.i., when most animals had cleared the infection (Fig. 4C). Functional T cells are required to control H. pylori infection (16), and they are also the major source of IFN-g. To determine if IFN-g is sufficient to reduce H. pylori colonization in the absence of T cells, we compared infection in TCR 2/2 mice and TCR 2/2 mice bearing the IFN-g transgene. As in mice with functional T cells (Fig. 4A), IFN-g expression was increased in T cell receptor-deficient (TCR 2/2 ) mice expressing the IFN-g transgene (see Fig. S1A in the supplemental material), while H. pylori colonization was decreased (Fig. S1B). Similar to previous results (16), cagY recombination  was eliminated in TCR 2/2 mice but increased somewhat by overexpression of IFN-g, even in the absence of functional T cells (Fig. S1C). These results suggest that reduction in H. pylori colonization in the Salmonella coinfection model may be explained, at least in part, by increases in IFN-g-mediated inflammation.
Salmonella coinfection promotes retention of H. pylori T4SS function. We previously found that increased gastritis in IL-10 2/2 mice is associated with decreased H. pylori colonization (16), similar to our findings here in tgIFN-g mice. However, some IL-10 2/2 mice were colonized at levels similar to those of WT mice, and H. pylori recovered from them had typically recombined cagY, suggesting that loss of T4SS function permits increased colonization in the face of a robust immune response. Consistent with this observation, we also found that overexpression of IFN-g caused increased cagY recombination (Fig. 4D), even in the absence of functional T cells (Fig. S1C). Since Salmonella coinfection enhances the systemic and local inflammatory responses and decreases H. pylori colonization, we hypothesized that it too might select for cagYmediated loss of T4SS function and enable H. pylori to partially escape increased immune pressure from IFN-g and other inflammatory cytokines. However, we observed the opposite result. cagY recombination was reduced (Fig. 5A) and T4SS function was increased ( Fig. 5B) in H. pylori colonies recovered from coinfected mice compared to those in mice infected with H. pylori alone. H. pylori colonization was inversely correlated with T4SS function in coinfected mice ( Fig. 5C) but not in mice infected with H. pylori alone, where colonization was overall greater and less variable (Fig. 5D).
H. pylori T4SS function is retained in iron-deficient mice. In vitro experiments suggest that CagA translocation via the T4SS serves to increase iron acquisition from host cells (28). These results are consistent with in vivo studies demonstrating that deletion of cagA impairs H. pylori colonization in iron-deficient, but not iron-replete,  (28). Moreover, H. pylori strains recovered from iron-deficient gerbils, or grown in vitro under iron-deficient conditions, demonstrate greater numbers of T4SS pili and show enhanced T4SS function (30). Since Salmonella induces anemia (31) and inflammation-mediated iron sequestration (32), we hypothesized that Salmonella coinfection might select for a functional T4SS by competing with H. pylori for iron. To first test the effects of anemia on colonization and T4SS function, we gavaged H. pylori into EPO 2/2 mice bearing a homozygous disruption in the 59 untranslated region of the erythropoietin gene (Epo-Tag h ), which reduces whole-body erythropoietin expression (33). EPO 2/2 mice are severely anemic, with a hematocrit level approximately half that of WT mice (see Fig. S2A), though they may not be iron deficient. Hematocrit was further reduced by H. pylori infection in both WT and EPO 2/2 mice, but H. pylori colonization and T4SS function were unaffected ( Fig. S2B and C).
Anemia can occur with or without iron deficiency. To test specifically for the role of iron deficiency on H. pylori colonization and T4SS function, we compared H. pylori infections in mice fed for 5 weeks with an iron-deficient diet (2 to 6 ppm iron) to infections in mice fed a standard diet (200 ppm iron). H. pylori infection was followed by a decrease in food intake, which was restored in mice fed an iron-replete diet but not in those fed an iron-deficient diet (see Fig. S3A). The iron-deficient mice also weighed slightly less than control mice at the end of the experiment (Fig. S3B). Serum iron levels were highly variable in mice fed an iron-deficient diet, though anemia was severe and uniform ( Fig. S3C and D). Similar to anemia in EPO 2/2 mice, iron deficiency anemia did not affect H. pylori colonization (Fig. 6A). However, the effect on T4SS function was dramatic, with no cagY recombination and no loss of IL-8 induction in iron-deficient mice 8 weeks p.i. (Fig. 6B and C). Gastric IFN-g levels were also increased in H. pylori-infected iron-deficient mice, likely as a result of the increased T4SS activity (Fig. 6D). Role of iron deficiency in Salmonella coinfection. Since iron deficiency maintains T4SS function in mice, we hypothesized that retention of T4SS function during Salmonella coinfection may be a result of iron starvation. To determine if Salmonella coinfection induces iron deficiency, we analyzed gastric expression of lipocalin-2, hepcidin, and fur, which are known to be regulated by iron and therefore serve as a functional readout of iron deficiency. Salmonella induces lipocalin-2 in the gut, which binds siderophores from Enterobacteriaceae and enables Salmonella to overcome colonization resistance (32). Although Salmonella coinfection increased gastric lipocalin-2 expression somewhat more than H. pylori alone (Fig. 7A), it seems unlikely that this is relevant for iron sequestration from H. pylori in the stomach, since H. pylori is not known to produce siderophores. Hepcidin is also upregulated in Salmonella infection, causing iron retention in macrophages (34). We confirmed previous observations (35) that H. pylori decreases hepcidin expression in mice, but this was unaffected by Salmonella coinfection (Fig. 7A). Finally, we investigated expression of the H. pylori ferric uptake regulator, fur, which is activated by iron restriction (36). Since we could not readily detect low levels of H. pylori gene expression from infected gastric tissue, we examined fur expression in H. pylori cocultured with AGS cells to mimic interactions with the gastric epithelia. Low-passage-number H. pylori isolated from iron-deficient mice showed greater expression of fur than H. pylori from iron-replete mice, which validated that the assay served as a bacterial readout of iron deprivation (Fig. 7B). However, fur expression in H. pylori was unaffected by coinfection with Salmonella (Fig. 7C). Together, these results suggest that neither host nor bacterial expression

DISCUSSION
After the initial demonstration in mice (11), loss of T4SS function has been demonstrated in gerbils (21), rhesus macaques (12), and humans (16,37), suggesting that modulation of T4SS function is an important feature of the biology of H. pylori. The mechanism is typically in-frame recombination in the middle repeat region of cagY, which encodes an essential T4SS protein, though single nucleotide polymorphisms (SNPs) in cagY or in other genes that are essential for T4SS function may also occur (13). While we know how T4SS function is lost during experimental infection-or even gained (12)-we do not know why. The T4SS is typically functional during chronic infection in humans (38,39) and in rhesus monkeys (26), unless essential genes are absent or present as pseudogenes. Yet, the extraordinary repetitive genetic structure of cagY and the demonstration that recombination in these repeat regions can alter T4SS function strongly suggest that there must be conditions under which T4SS-dependent inflammation is advantageous for the bacterium and others under which it is not. We know from studies with genetically modified mice that cagY recombination does not occur in the absence of CD4 T cells expressing IFN-g, and so loss of T4SS function is immune sensitive (16). But these are not physiological experiments. Here, we sought to use the murine model to identify conditions relevant to human infection under which H. pylori modulates T4SS function.
Just as loss of T4SS function and cagY recombination do not occur in immunodeficient mice, they occur more commonly when the immune response is increased in IL-10 knockout mice, which have more severe gastritis and lower bacterial load when challenged with H. pylori (16). Since deletion of the cagPAI reduces inflammation and increases colonization (23,24), loss of T4SS function may serve to maintain colonization in the face of increased inflammation, such as might occur during a severe, transient systemic infection. The notion that one infection can alter the outcome of another has been repeatedly demonstrated and is thought to occur via enhancement of nonspecific innate immunity (40)(41)(42)(43). To test this hypothesis, we coinfected mice with H. pylori and Salmonella, predicting that, like in IL-10 knockout mice, we would find increased inflammation, lower bacterial load, and reduced T4SS function compared to that in mice infected with H. pylori alone. As expected, Salmonella coinfection decreased H. pylori colonization and was accompanied by increased gastritis as well as local and systemic levels of IFN-g. The importance of IFN-g was supported by the observation that its overexpression was itself sufficient to increase gastritis and reduce H. pylori colonization (Fig. 4). Similar to that in IL-10 knockout mice (16), H. pylori colonization in coinfected mice was inversely correlated with T4SS function, which we initially interpreted as consistent with immune escape leading to increased bacterial load.
However, in contrast to our prediction, T4SS function was increased in Salmonellacoinfected mice, and cagY recombination was reduced (Fig. 5). This suggested the possibility that Salmonella coinfection selected strains with a functional T4SS, which would be expected to be associated with more inflammation and therefore a lower bacterial load. Decreased H. pylori colonization during coinfection might then be a result of increased inflammation, not just from Salmonella infection but also from retention of T4SS function. Since the T4SS enables H. pylori to acquire iron from the host (28) and systemic infection causes iron sequestration (44), we asked whether iron deficiency might cause retention of T4SS function and explain why it is enhanced in Salmonella coinfection. While anemia per se had no effect, anemia induced by dietary iron restriction completely eliminated cagY recombination and loss of T4SS function (Fig. 6).
Together, these studies have identified two physiologically relevant variables that affect T4SS function in the murine model, Salmonella coinfection and especially dietary iron restriction, which completely eliminated loss of T4SS function. Both have epidemiologic parallels in humans, since H. pylori strains from individuals in developing countries, where iron deficiency and systemic infections are common, are also more often cagPAI 1 than strains from developed countries. However, there are important caveats to our findings. For example, while the results of Salmonella coinfection prompted the iron restriction studies and both promote retention of T4SS function, it appears that they are not mechanistically linked, though we did not directly measure gastric iron levels. We currently do not have an explanation for why H. pylori in coinfected mice more often retains T4SS function. While iron deficiency maintains T4SS function and increases gastric IFN-g, it does not affect colonization, which might be expected to decrease. It seems likely that the effects of Salmonella coinfection, and perhaps iron restriction, are pleomorphic and will require additional studies to understand their impact on T4SS function. One approach to separate the effects of Salmonella infection from the inflammation it causes might be to pharmacologically induce innate immunity with, for example, lipopolysaccharide (LPS) administration. Nonetheless, here we have identified two physiological variable factors relevant to human infection that affect T4SS function. The results have implications for our fundamental understanding of the cagPAI and also provide experimental tools that permit the study of T4SS function in the murine model.

MATERIALS AND METHODS
Ethics statement. Experiments were carried out at the University of California, Davis, under protocols approved by the U.C. Davis Institutional Animal Care and Use Committee, which has been accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. All animal experiments were performed in accordance with NIH guidelines, the Animal Welfare Act, and U. S. federal law.
Animals. Female C57BL/6J WT and TCR b/d 2/2 mice were purchased from the Jackson Laboratory (Sacramento, CA). A male mouse with an extra copy of the mouse IFN-g gene under the control of the H/ K ATPase b promoter (H/K-IFN-g line 944 mice) was provided by Andrzej Dlugosz (45). These tgIFN-g mice were bred with WT and TCR 2/2 mice to obtain heterozygous H/K-IFN-g mice and H/K-IFN-g 1/2 TCR 2/2 mice as well as littermate controls without the H/K-IFN-g gene. Heterozygous erythropoietin knockout mice obtained from Nicolas Voituron (46) were bred in-house to obtain EPO 2/2 mice and littermate controls. Mice were housed in microisolator cages and provided with irradiated food and autoclaved water ad libitum. Iron deficiency was induced by providing mice with food containing traces of iron in the range of 2 to 6 ppm (TD 10210; Envigo Teklad Diets, Madison, WI) starting from 3 weeks of age. Control mice were given an equivalent food but with 200 ppm iron (TD 150282; Envigo Teklad Diets). Food intake for each cage of 4 mice was monitored, and iron-deficient and -replete mice were weighed before and after H. pylori infection.
H. pylori and Salmonella challenge. At 8 to 9 weeks of age, mice were challenged with 1 Â 10 9 CFU of H. pylori suspended in 0.25 ml of brucella broth administered by oral gavage with a ball-end feeding needle. For coinfection experiments, mice were infected with Salmonella 1 week after H. pylori challenge by intravenous (i.v.) injection in the lateral tail vein with 5 Â 10 5 CFU of bacteria diluted in 0.2 ml phosphate-buffered saline (PBS). Bacterial concentrations were estimated by optical density at 600 nm and confirmed by plating serial dilutions. At the endpoint, mice were euthanized with an overdose of pentobarbital sodium injection (50 mg/ml intraperitoneally [i.p.]). Blood was collected in microcapillary tubes for hematocrit measurement and into serum separator tubes (BD Microtainer). The forestomach was removed, and the glandular stomach was cut longitudinally along the lesser curvature. Half of the stomach was placed in 300 ml brucella broth, weighed, ground with a sterile glass rod until the mucosal cells were homogenized, and then plated in serial dilution to determine CFU per gram. The limit of detection (LOD) for H. pylori was 20 CFU/g. Salmonella CFU were counted from homogenized stomach (LOD = 7.5 CFU/organ) or spleen (LOD = 24 CFU/organ). Absence of H. pylori or Salmonella colonies was represented as 0 CFU.
Histology. A quarter of the stomach was fixed in 10% formalin, sectioned, and stained with hematoxylin and eosin. The percentage of fields containing neutrophil infiltration (polymorphonuclear leukocytes), gastritis (mononuclear cells), and metaplasia was identified by a veterinary pathologist blinded to experimental condition, using a scoring system previously validated in mice (47). The results for the three histological criteria were averaged and defined as percent gastric inflammation.
Gene expression and iron analysis. Serum IFN-g levels were analyzed with the mouse IFN-g uncoated ELISA kit according to the manufacturer's protocol (Invitrogen). Murine gene expression of IFN-g, hepcidin, and lipocalin-2 was analyzed from gastric tissue homogenized in TRIzol reagent (Ambion). Bacterial expression of fur was analyzed from sweeps isolated from gastric tissue and cocultured for 20 h with AGS cells as described below. RNA was purified by phenol-chloroform phase separation as described by the manufacturer or by the Direct-zol RNA miniprep kit (Zymogen). RNA was transcribed to cDNA by Superscript III (Qiagen) and added to quantitative PCR (qPCR) mixtures with TB green premix Ex Taq (TaKaRa) using primers shown in Table S1 in the supplemental material. Gene expression was normalized to murine glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or bacterial 16S rRNA expression. Amplification was performed using a QuantStudio 6 Flex real-time PCR system (Applied Biosystems). Total serum iron was analyzed on a chemistry analyzer by the Comparative Pathology Laboratory at UC Davis.
IL-8 ELISA. IL-8 was measured as described previously (48). Briefly, human AGS gastric adenocarcinoma cells (ATCC, Manassas, VA) were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ml streptomycin at 5% CO 2 and 37°C. The cells were seeded in six-well plates at a density of approximately 5 Â 10 4 cells/cm 2 with 1.8 ml antibioticfree RPMI medium-10% fetal bovine serum, incubated overnight, and then cocultured with H. pylori sweeps from an individual mouse diluted in 200 ml brucella broth at a multiplicity of infection (MOI) of 100:1. Supernatants were harvested after 20 to 22 h of culture (37°C, 5% CO2) and diluted 1:8 prior to IL-8 assay by ELISA (Invitrogen) performed according to the manufacturer's protocol. WT H. pylori PMSS1 and the isogenic cagY deletion mutant were included on every plate as positive and negative controls, respectively. IL-8 expression was normalized to that for WT PMSS1.
cagY PCR restriction fragment length polymorphism. cagY genotyping was performed on isolated single colonies by PCR-restriction fragment length polymorphism (RFLP) essentially as previously described (12). cagY was amplified with Herculase II fusion DNA polymerase (Agilent Technologies), digested with DdeI and BfuCI or Sau3AI restriction enzymes (New England BioLabs, Ipswich, MA), visualized by agarose gel electrophoresis, and compared to that for WT H. pylori PMSS1. For convenience, cagY recombination was defined as a change in PCR-RFLP pattern compared to that for the WT, though we did not formally measure recombination and changes could occur by other mechanisms.
Statistics. All statistical analyses were performed using GraphPad Prism 7.00 (GraphPad Software, San Diego, CA). Multiple groups were compared using analysis of variance (ANOVA) with Tukey's or Bonferroni's post hoc tests. Differences in H. pylori colonization (CFU/g) and IL-8 expression between two groups were analyzed using the Mann-Whitney test. Proportions of samples with changed cagY were compared between groups using the Fisher's exact test. Correlation between IL-8 expression and H. pylori CFU was analyzed by linear regression. Data are reported as means 6 standard errors of the means (SEMs), and a P value of ,0.05 was considered statistically significant.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.