Pre-existing Helicobacter pylori serum IgG enhances the vibriocidal antibody response to CVD 103-HgR live oral cholera vaccine in Malian adults

Accumulating evidence indicates that persistent Helicobacter pylori gastric infection influences immune responses to oral enteric vaccines. We studied the association between pre-existing H. pylori serum IgG and serum pepsinogens levels (PGs) as markers of gastric inflammation and the immune response to single-dose live oral cholera vaccine CVD 103-HgR in Malian adults. Baseline sera obtained during a phase 2 safety/immunogenicity clinical trial of cholera vaccine CVD 103-HgR among 93 healthy Malian adults were tested for H. pylori IgG antibodies and PGI and PGII levels using enzyme linked immunosorbent assays. Overall 74/93 (80%) vaccine recipients were H. pylori IgG seropositive at baseline. Vibriocidal antibody seroconversion (≥ fourfold increase 14 days following administration of CVD 103-HgR compared to baseline) among vaccine recipients was 56%. However, vibriocidal antibody seroconversion was markedly higher among H. pylori seropositives than seronegatives 64% vs. 26% (p = 0.004); adjusted relative risk: 2.20 (95% confidence intervals 1.00–4.80; p = 0.049). Among H. pylori seropositive vaccine recipients, there were no significant associations between PGI, PGII and PGI:PGII levels and vibriocidal seroconversion. The enhanced seroconversion to oral cholera vaccine CVD 103-HgR among H. pylori seropositive African adults provides further evidence of the immunomodulating impact of H. pylori on oral vaccine immunogenicity.

bacterial overgrowth were some risk factors associated with diminished vibriocidal responses 14 . Individuals with blood group O were shown to develop higher vibriocidal titers 15 , and the pre-immunization elimination of helminths was shown to significantly raise the vibriocidal antibody response in non-blood group O children 16 .
Achieving the desired immune response to a live oral vaccine requires successful passage of viable vaccine organisms through the hostile acidic gastric environment to reach the immunologic inductive sites of the gutassociated lymphoid tissue in the small intestine. Helicobacter pylori chronically colonizes the stomach of persons in developing countries, sometimes resulting in the development of peptic ulcers 17 and gastric cancer in older individuals 18 (reviewed in 19,20 ). Given the high prevalence of chronic H. pylori infection in developing countries, the role such chronic infection in modulating the immune response to oral enteric vaccines is of particular interest. Chronic H. pylori infection can markedly alter gastric physiology, including gastric acid secretion 19,20 , and as a consequence can indirectly influence immune responses to a variety of antigens delivered orally, including cholera vaccine.
We previously reported a notable inverse association between baseline H. pylori seropositivity and vibriocidal antibody seroconversion following oral vaccination with CVD 103-HgR in Chilean children < 5 years of age 21 . In contrast, in older children 5-9 years of age, a positive association was found in relation to gastric inflammation, as measured by serum pepsinogen I (PGI) 21 . In a somewhat analogous study of a live oral typhoid vaccine conducted among U.S. adults, superior immune responses were recorded following immunization with oral attenuated Salmonella Typhi vaccine strain CVD 908-htrA in adults who were seropositive for H. pylori and whose serum PGs indicated more advanced gastritis 22 . Thus, the association between chronic H. pylori infection and the immune response to oral vaccines is complex and multifactorial, and differences according to age and socioeconomic level should be assessed.
A study of the immunogenicity of CVD 103-HgR in Malian adults offered the opportunity to examine in another geographic site (West Africa) the association of pre-existing H. pylori immunoglobulin G (IgG) serum antibodies and serum PGs as markers of gastric inflammation with vibriocidal antibody seroconversion following immunization of adults with this live oral cholera vaccine. Based on our previous data in different populations, our hypothesis was that H. pylori seropositivity would be associated with higher vibriocidal antibody seroconversion in Malian adults.

Materials and methods
Study design and population. We used anonymized specimens of stored sera obtained from participants in a phase II clinical trial with CVD 103-HgR vaccine in Mali, West Africa 23 . Healthy adults aged 18-45 years from Bamako received a single oral dose of ≥ 2 × 10 8 CFU (low-dose) or ≥ 2 × 10 9 CFU (high-dose) of CVD 103-HgR administered in a buffer cocktail. The original trial included in its design the tertiary objective of assessing the association between H. pylori IgG sero-prevalence and the immune response to CVD 103-HgR.
Definition of the study variables. The dependent variable, vibriocidal seroconversion 23 , was defined as a ≥ fourfold increase in serum vibriocidal antibody titer 14 days post-vaccination compared to the baseline titer.
The main independent variable was H. pylori seropositivity at baseline defined as the presence of H. pylorispecific IgG antibodies in the pre-vaccination sera (for cutoff value please see the laboratory methods section). H. pylori-positive sera were further tested for the presence of IgG antibodies to the cytotoxin-associated gene A (CagA) protein, a marker for virulent H. pylori strains. Additional independent variables included serum levels of PGs as a measure of gastric inflammation 24,25 . Serum PGI and PGI are proenzymes of the enzyme pepsin. The PGs are secreted into the gastric lumen and can be found in the serum 26,27 . With increasing severity of H. pylori gastritis, serum PGI and PGII levels become elevated, but when atrophic changes occur in the corpus, the PGI level and the PGI:PGII ratio decrease. These markers can predict atrophic gastritis, intestinal metaplasia and gastric cancer 24,25,28,29 . A cutoff value of serum PGI of < 30 μg/L or of a PGI:PGII ratio < 3.0, is used to determine the presence of atrophic gastritis/ gastric atrophy, as recommended by the manufacturer. However, there is no consensus in the literature regarding the best cut-off values 24,25,[28][29][30][31][32] . A PGI:PGII ratio of 4 to 6 was proposed to enhance the sensitivity of serum PGs in identifying atrophic gastritis 24,[30][31][32] . The occurrence of atrophic gastritis is low, and increases after age 45 years 33 . Thus, the expected number of persons with atrophic gastritis in this small sample of healthy young adult (ages 18-45, median 22 years) is low. Accordingly, we used the lowest quartiles as cut-off values (73.6 μg/L for PGI and 5.35 for PGI:PGII); these are equivalent to clinically relevant values 24,30-32 .
The following variables were considered as covariates and potential confounders: age in years (analyzed as a continuous variable), and sex. Pre-vaccination vibriocidal antibody titer was also assessed, given our previous observations of an inverse association between baseline levels and the immune response to CVD 103-HgR 6,13,15 .
Since H. pylori infection might be a marker of broad exposure to enteric pathogens, we assessed the level of serum IgG antibodies to Shigella flexneri 2a lipopolysaccharide as an additional independent variable. This enhances the understanding of the specificity of the association between H. pylori infection and oral enteric vaccines.
The categories of S. flexneri 2a serum IgG antibody were defined according to the knowledge that Shigella is highly endemic in developing countries, already at a very young age [34][35][36] . Thus, Malian adults were presumed to have been exposed to Shigella infections during their life course. Therefore, we compared participants with the highest levels of S. flexneri 2a serum IgG antibody (highest quartile [Q4]) to those with lower levels (Q1-Q3).
Laboratory methods. Sera obtained in the course of the above-mentioned clinical trial 23 were shipped from Mali to the Center for Vaccine Development and Global Health, University of Maryland School of Medicine, Baltimore, USA and stored at − 20 °C until testing. Vibriocidal antibody levels before and after vaccination were examined as described 23  www.nature.com/scientificreports/ antibody and PGs levels testing. H. pylori IgG antibodies were measured using a commercial enzyme linked immunosorbent assay (ELISA) kit (Enzygnost Anti-Helicobacter pylori II/IgG kit, Siemens Diagnostics Product GmbH, Marburg, Germany) following the manufacturer's instructions. An optical density (OD) of 0.250 or greater were considered positive. The presence of IgG antibody against recombinant CagA protein was measured in H. pylori-positive sera according to an in-house ELISA protocol 18,33 . Concentrations of serum PGI and PGII were quantified by ELISA (Biohit Inc., Helsinki, Finland) following the manufacturer's instructions. Serum IgG antibody titers to S. flexneri 2a lipopolysaccharide was measured using an in-house ELISA protocol at the laboratory of Prof. Daniel Cohen, Tel Aviv University, Israel 37 and results were expressed in ELISA units.

Statistical analysis.
Descriptive statistics (medians and interquartile ranges (IQR) for continuous variables, and frequencies and percentages for categorical variables) were used to describe the study sample. Chi square and Fisher Exact tests, where appropriate, were used to examine differences between H. pylori IgG seropositive and seronegative vaccinees in categorical variables and Mann-Whitney test for continuous variables (e.g., serum PGs). The Chi square test was also used to examine differences in the percentage of vaccinees who developed vibriocidal antibody seroconversion 14 days following vaccination with CVD 103-HgR, according to H. pylori seropositivity and other categorical variables. The difference in the median age between participants with and without seroconversion at day 14 was assessed using the Mann-Whitney test.
A log-binomial model of the generalized linear models [38][39][40] was performed to examine the association between H. pylori IgG seropositivity and vibriocidal seroconversion, while controlling for potential confounders. The use of the log-binomial model is recommended 38,40 if the dependent variable is common, as in our study, in which 52/93 (55.9%) of the participants had vibriocidal antibody seroconversion 14 days following immunization with CVD 103-HgR.
The selection of variables to be included in the multivariable analysis was based on our pre-specified main hypothesis of the relation between H. pylori seropositivity (the main independent variable) and vibriocidal seroconversion following immunization with CVD 103-HgR, while relying on microbiological, immunological and epidemiological knowledge in the field 21,22 . We adjusted for age given the evidence that this variable might affect the immune response to oral enteric vaccines 21,22 , and that H. pylori gastric inflammation and pathology might increase with age [41][42][43][44] . Lastly, we considered S. flexneri IgG serum antibody as a proxy for exposure enteric infections 21 . Accordingly, we assessed the specificity of the association between H. pylori sero-prevalence and vibriocidal seroconversion following immunization with CVD 103 HgR. Relative risk (RR) and 95% confidence intervals (CI) were obtained from this model.
The Mann-Whitney test and box plot were used to assess difference in vibriocidal antibody seroconversion level (with natural logarithm transmutation) according to H. pylori seropositivity. In additional analyses we assessed difference in the median baseline level of vibriocidal antibody titer (with natural logarithm transmutation) between participants with and without seroconversion at day 14, using the Mann-Whitney test and box plot. We also performed a multivariable model that included both H. pylori seropositivity and baseline vibriocidal antibody level. Statistical significance was set at p < 0.05. Data were analysed using IBM SPSS version 25 (Armonk, New York, USA). ethical aspects. The Institutional Review Board (IRB) of the University of Maryland, Baltimore and the local Malian IRB approved the original trial with CVD 103-HgR including the objective of this study that was defined as a tertiary objective. Volunteers were enrolled after providing an informed consent 23 . The ethics committee of Tel Aviv University also approved the protocol of the current study.
We confirm that all methods were performed in accordance with the relevant guidelines and regulations.

Results
The  Table 1). The fold-increase in vibriocidal antibody titer at day 14 compared to the baseline level ranged from 0.02 to 1,024, the value of the 25th percentile was 0.5, the median was 4, and the 75th percentile value was 16. The distribution of the fold increase is presented in Supplementary Table S1. The levels of fold-increase of vibriocidal antibody titer was higher among H. pylori positive vs. negative vaccinees (p = 0.034 by the Mann-Whitney test, see Supplementary Fig.S1 online).
Among H. pylori seropositive individuals, there was no significant difference between vaccinees with and without vibriocidal antibody seroconversion with respect to CagA IgG seropositivity (70% vs. 82%, p = 0.2) or serum PGI, PGII or PGI:PGII levels (

Discussion
We found a positive association between baseline H. pylori IgG antibody seropositivity and vibriocidal antibody seroconversion 14-days following immunization with a single dose of CVD 103-HgR live oral cholera vaccine among Malian adults. This association is reported for the first time in persons from a very low-income country, www.nature.com/scientificreports/  www.nature.com/scientificreports/ a setting in which many oral enteric vaccines have shown lower immunogenicity and efficacy compared to populations in high-income countries. The positive association between H. pylori infection and vibriocidal antibody seroconversion was independent of age among these adults and of a high level of S. flexneri serum IgG anti-LPS antibody, a proxy for exposure to enteric pathogens. Interestingly, the association between H. pylori seroprevalence and the immune response to CVD 103-HgR was not affected by baseline vibriocidal antibody level, serum PGs levels, markers of gastric inflammation or infections with H. pylori CagA-positive (putatively more virulent) strains. Age and S. flexneri IgG antibodies were related to vibriocidal antibody serconversion in the bivariate analysis only. There was no significant correlation between age and S. flexneri IgG antibodies, likely since exposure to Shigella infection occurs during childhood in Mali [34][35][36] .
Findings from the current study confirm our previous observations regarding the involvement of H. pylori infection in the immune response to oral enteric vaccines 21,22 . The positive association between baseline H. pylori seropositivity and vibriocidal antibody seroconversion following vaccination with CVD 103-HgR among Malian adults is in agreement with findings documented among school-age (5-9 years) Chilean children 21 . In younger Chilean children < 5 years of age, on the other hand, H. pylori seroprevalence was inversely related to vibriocidal antibody seroconversion after immunization with CVD 103-HgR 21 . These observations can be explained by the age-related impact of H. pylori infection on gastric physiology. H. pylori infection is acquired in childhood 45 and typically most people do not develop symptoms. However, the histopathological severity of gastritis increases with age, even in the absence of symptoms 44 . H. pylori antral-predominant chronic gastritis might increase the secretion of gastric acid, resulting in an enhanced gastric acid barrier 20,46 , while H. pylori infection and corpuspredominant gastritis might induce gastric hypochlorhydria, which usually occurs in adulthood. Thus, the transit of attenuated oral vaccine strains from the stomach into the small intestine might be enhanced in the presence of hypochlorhydria, and the immune responses to oral vaccines might actually increase. This likely explains our observations in Malian adults.
The finding of a positive association between H. pylori seropositivity at baseline and vibriocidal antibody seroconversion following vaccination with CVD 103-HgR in Malian adults corroborates our findings among U.S. adults who were immunized with a live oral typhoid vaccine. In the US adults we observed stronger humoral immune responses following immunization with oral attenuated Salmonella Typhi vaccine strain CVD 908-htrA 22 in persons who were H. pylori seropositive at baseline and exhibited evidence of advanced gastritis 22 .
In the current study, we did not find an association between serum pepsinogen levels and the immune response to the cholera vaccine CVD 103-HgR among Malian adults. This contrasts with the finding of a positive association between serum pepsinogen level and advanced gastritis among US volunteers 22 . A number of possibilities could explain this discrepancy between the studies, in the associations of serum pepsinogens and the immune response with oral enteric vaccines. One explanation relates to the variation in the epidemiology of H. pylori infection and its related diseases; another relates to differences in public health between the US and the sub-Saharan Africa populations 47,48 .
Salmonella and V. cholerae are both acid sensitive bacteria 49,50 , thus supporting the notion that hypochlorhydria can be the mechanism for the association between H. pylori infection and the enhanced immune responses to these vaccines. Further support stems from early studies on cholera disease, showing that hypochlorhydria increases the risk of cholera [49][50][51][52][53] , as well as studies following the discovery of H. pylori showing a positive association between H. pylori infection and cholera [54][55][56] and typhoid fever 57,58 .
The longstanding systemic and local immune responses induced by H. pylori infection cannot clear the infection 19,20 . The consequences of persistent H. pylori infection may include having an immunomodulatory influence on the immune response to oral enteric vaccines, in particular to cholera and typhoid vaccines. The findings of Mattsson et al. 59 who assessed the production of antibody responses in the stomach as well as serum antibodies following immunization with an oral B subunit whole-cell cholera vaccine in Swedish volunteers 59 , are consistent with this assumption. Most of their study participants had high levels of vaccine-specific antibodysecreting cells (ASCs) in the duodenum and high serum antibody titers. At the gastric mucosa level, a high frequency of vaccine-specific ASC in the antrum was observed among most H. pylori-infected vaccine recipients but not among those who were not infected with H. pylori. The H. pylori-infected group had substantially higher frequencies of total immunoglobulin A (IgA) ASC in the antrum as compared to the non-infected group 59 . This report with oral killed cholera vaccine is consistent with our findings with live vaccine that pre-existing H. pylori infection enhances the immune response to oral cholera vaccines. These observations add evidence to support the hypothesis that although there are deleterious effects of chronic H. pylori infection (duodenal ulcers and www.nature.com/scientificreports/ gastric adenocarcinoma), there may be beneficial effects such as reduced risk of asthma and allergies 46,60 and less diarrheal disease in some populations 61,62 . Recent reports showed associations between the gut microbiota and the immune response to oral rotavirus vaccines in Ghana and Pakistan [63][64][65] , which suggest that modulation of the gut microbiome might yield better immune responses 65 . Collectively, these and our findings support the notion that pre-existing colonization of the gastrointestinal tract with certain bacterial microflora might affect the immunogenicity of oral enteric vaccines.
The relatively small sample size is a potential limitation of our study. Nonetheless, the positive association between H. pylori seropositivity and the immune response to CVD 103-HgR was strong and clear, despite the high prevalence of H. pylori infection in this cohort, which corroborates prevalence estimates from low-income countries 19,66 . Further studies with larger sample sizes are needed to confirm our findings. Moreover, larger studies with sufficient statistical power are needed to assess possible interactions between H. pylori infection and the vaccine dosage on the immune response to the vaccine. The discrepancy between findings from the current study and our previous findings among Chilean children on the association between H. pylori sero-prevalence and the immune response to CVD 103-HgR, may stimulate the need for multi-center international studies employing the same protocol to explore differences across populations in this association. Although H. pylori seropositivity was measured in baseline sera, prior to immunization, the observational design of our study does not enable causal inference. Since the serological assay that measures H. pylori IgG antibody level does not differentiate between past and recent infection, our findings reflect past exposure to H. pylori and not necessarily recent infection. However, it is well documented that H. pylori infection is mostly acquired in childhood and persists throughout adulthood 19,20 . Moreover, use of a serological assay in this setting in Mali might be the best tool in terms of specificity given reports of potential misclassification due to non-H. pylori urease-producing bacteria identified by urea breath test 67 .
Our study has several strengths stemming from both the study design and the novelty of the findings. First, H. pylori IgG seropositivity was correlated with serum PGII and PGI:PGII ratio markers of gastric inflammation, rather than just exposure to the infection. Second, the infection was examined in baseline sera, thus establishing a temporal association. Third, the association was specific to H. pylori and not related to exposure to other enteric infection such as S. flexneri, as demonstrated in the multivariable model. Lastly, the unique positive association between H. pylori seroprevalence and the immune response to an oral enteric vaccine (i.e., V. cholerae oral vaccine strain CVD 103-HgR) is reported for the first time in sub-Saharan Africa, a setting with frequent cholera outbreaks and a high prevalence of H. pylori infection. This increases the generalizability of our findings to other populations in low-income countries.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.