Probiotics-Containing Yogurt Ingestion and H. pylori Eradication Can Restore Fecal Faecalibacterium prausnitzii Dysbiosis in H. pylori-Infected Children

This study investigated the compositional differences in fecal microbiota between children with and without H. pylori infection and tested whether probiotics-containing yogurt and bacterial eradication improve H. pylori-related dysbiosis. Ten H. pylori-infected children and 10 controls ingested probiotics-containing yogurt for 4 weeks. Ten-day triple therapy plus yogurt was given to the infected children on the 4th week. Fecal samples were collected at enrollment, after yogurt ingestion, and 4 weeks after successful H. pylori eradication for cytokines and microbiota analysis using ELISA and metagenomic sequencing of the V4 region of the 16S rRNA gene, respectively. The results showed H. pylori-infected children had significantly higher levels of fecal TGF-β1 than those who were not infected. Eight of 295 significantly altered OTUs in the H. pylori-infected children were identified. Among them, the abundance of F. prausnitzii was significantly lower in the H. pylori-infected children, and then increased after yogurt ingestion and successful bacterial eradication. We further confirmed probiotics promoted F. prausnitzii growth in vitro and in ex vivo using real-time PCR. Moreover, F. prausnitzii supernatant significantly ameliorated lipopolysaccharide-induced IL-8 in HT-29 cells. In conclusions, Probiotics-containing yogurt ingestion and H. pylori eradication can restore the decrease of fecal F. prausnitzii in H. pylori-infected children.


Introduction
Helicobacter pylori (H. pylori) infection primarily starts in childhood [1]. This microorganism has been reported to cause peptic ulcers, gastric lymphoma, and adenocarcinoma mainly in adults, and also iron deficiency anemia and growth retardation in children [2][3][4][5]. A case-control study of a Japanese cohort showed that acquisition of H. pylori in early life increased the risk of developing

Prevalence of H. pylori Infection
A total of 179 children from primary schools were enrolled. The mean age was 10.7 years with a male to female ratio of 1.1. The seroprevalence of H. pylori infection was 12.8%, and there was no significant difference between the male and female children (p = 0.36). In addition, there were no significant differences in body weight (45.0 vs. 41.4 kg, p = 0.19) and height (148.0 vs. 145.1 cm, p = 0.22) between the children with and without H. pylori infection.

Numbers of Study Cases and Fecal Samples
The numbers of enrolled subjects and fecal samples are shown in Figure 1. A total of 37 stool samples including H. pylori-infected (HPS1, n = 10), H. pylori-infected with yogurt (HPS2, n = 6), H. pylori-infected with yogurt and eradication (HPS3, n = 7), non-H. pylori infected controls (CS1, n = 9), and controls with yogurt (CS3, n = 5) were processed for DNA purification, 16S-rRNA gene amplification, and amplicon sequencing. A stool sample from a control subject (CS1) was excluded from the analysis as the child had consumed yogurt twice a week prior to sample collection. Another fecal sample from HPS2 was not analyzed owing to a low yield of DNA.

Fecal Inflammatory Parameters between H. pylori-Infected Children and Controls
The fecal samples were tested for calprotectin, lactoferrin, IL-6, TGF-β1, and sIgA (Supplementary Table S1). The results showed that the H. pylori-infected children had a significantly higher fecal TGF-β1 level (12.0 vs. 7.0 ng/mL, p = 0.02) than the non-infected controls. In addition, the fecal calprotectin (13.2 vs. 2.5 µg/g, p = 0.13) and lactoferrin (15.4 vs. 6.8 µg/g, p = 0.18) levels were also higher in the H. pylori-infected children than in the non-infected children, although with only marginal significance. The sIgA level was lower in the H. pylori-infected children than in the controls (240.3 vs. 505.0 µg/mL, p = 0.07), however there was no significant difference in IL-6 level (3.4 vs. 1.3 pg/mL, p = 0.27) between the two groups.

Sequencing Results
16S rRNA sequencing reads were classified to 295 OTUs from the 37 fecal samples. In phylogenetic analysis, the fecal bacterial community mostly belonged to five major phyla, including Bacteroidetes (54%), Firmicutes (32.8%), Proteobacteria (7%), Actinobacteria (3.3%), and Fusobacteria (0.1%). H. pylori could only be identified in 60% of the samples from the infected patients by 16S rRNA sequencing. The dendrogram of hierarchical clustering and proportional changes in bacterial OTU abundance at the genus level is shown in Figure 2A. The most abundant genus in the 37 samples was Bacteroides spp., followed by Prevotella spp. and Faecalibacterium spp. However, there were no significant differences in the microbiota composition between the five groups in AMOVA (p = 0.331) and HOMOVA (p = 0.522) analysis.

Library Coverage and Sequence Diversity
The rarefaction curve, a comparison of α diversity between subjects categorized by HPS1, HPS2, HPS3, CS1, and CS3 showed high biodiversity between the H. pylori-infected and control groups ( Figure 2B). The slope of each curve reached a plateau with increasing sequence depth. The inverse Simpson index was significantly different between CS1 and CS3, but not between HPS1, HPS2, and HPS3 ( Figure 2C). Weighted UniFrac phylogenetic distance matrices were used to calculate the β diversity, which was shown in PCoA plots. In comparisons of group diversity between CS1 and HPS1, CS1 and CS3, and HPS1, HPS2, and HPS3, and all groups both AMOVA and HOMOVA showed no significant difference (p > 0.05) in microbial diversity between groups (Supplementary Figure S1).

Significant Genus Difference between H. pylori-Infected Children and Controls
In comparisons of significant microbial changes in the phylum (Supplementary Figure S2) and genus (Table 1) levels between the H. pylori-infected (HPS1) and control (CS1) subjects, the mean abundance ratio of Proteobacteria was significantly higher in HPS1 than in CS1 at the phylum level (0.044 vs. 0.026, p = 0.02).
In addition, we identified eight OTUs with significant changes in relative abundance at the genus level between the two groups by Metastats. The H. pylori-infected children had a significantly lower mean proportional abundance of the fecal microbes   Figure 3A shows the serial changes in the abundance of F. prausnitzii in the five groups. In addition to a significant reduction in the H. pylori-infected children (HPS1) compared to the controls (CS1), probiotics-containing yogurt ingestion significantly increased the proportion of F. prausnitzii in the non-H. pylori infected children (0.11 vs. 0.05, p = 0.003). However, this increase was diminished in the H. pylori-infected group after yogurt ingestion (0.04 vs. 0.02, p = 0.11). Moreover, the mean proportional abundance of F. prausnitzii in the H. pylori-infected children with successful eradication was significantly increased compared to that before yogurt ingestion and eradication (0.10 vs. 0.02, p = 0.0005). The abundance level of F. prausnitzii after H. pylori eradication was comparable to that of the non-infected level. Real-time PCR quantification confirmed the increase in the abundance of fecal F. prausnitzii in the samples ( Figure 3B).

Probiotics Facilitate F. prausnitzii Growth in Vitro and In Ex Vivo
To identify the probiotics directly or indirectly facilitate F. prausnitzii growth, we tested the F. prausnitzii growth condition with relative DNA level by real-time PCR in vitro and in ex vivo. Figure 4A showed that culture supernatants of L. acidophilus and B. lactis directly promoted F. prausnitzii growth in a co-culture system. The enhanced effect was disappeared, if the cecum fluid collecting from C57B/6 mice was selected instead of Brain Heart Infusion Supplement (BHIS) media ( Figure 4B). Furthermore, the cecum fluids of mice had positive effect on promotion of F. prausnitzii growth when compared to BHIS culture media ( Figure 4C). Moreover, the supernatant of L. acidophilus or/and B. lactis cultured with cecum tissue (ex vivo) of mice induced a higher relative ratio of F. prausnitzii growth than culture fluid of extracorporeal cecum tissue ( Figure 4D). The equal-sized cecum tissue was co-cultured with 3 mL RPMI medium with or without L. acidophilus and B. lactis for 24 h. The 100 µL supernatants was added in 30 mL YCFA broth with F. prausnitzii co-cultured for 4, 8, and 12 h. The relative abundance of F. prausnitzii were measured as the percentage of total bacteria DNA copy numbers by real-time PCR. All tests were triplet.

F. prausnitzii Ameliorated LPS-Induced IL-8 Expression in HT-29 Cells
To evaluate the functional benefits of probiotics in restoring the proportion of fecal F. prausnitzii associated with H. pylori infection, AGS and HT-29 cells were treated with supernatants of F. prausnitzii culture for 4 h before H. pylori or LPS inoculation. Figure 5 shows that F. prausnitzii supernatant treatment significantly abolished LPS-induced IL-8 production in the HT-29 cells at 6 h ( Figure 5A). However, F. prausnitzii supernatant did not affect the H. pylori-induced IL-8 expression in the AGS cells ( Figure 5B).

Discussion
This study demonstrated a significant reduction of fecal F. prausnitzii in H. pylori-infected children. Moreover, both probiotics-containing yogurt ingestion and H. pylori eradication with yogurt ingestion improved the abundance of fecal F. prausnitzii. Based on in vitro and ex vivo assessment, supplement of L. acidophilus and B. lactis can facilitate F. prausnitzii growth, which can offer benefit to decrease the LPS-induced gut inflammation.
Gut sIgA is a first-line barrier to play an important role in the regulation of host-microbiota homeostasis [22,23]. Our previous study showed H. pylori infection in children significantly reduced serum IgA levels [17]. In this study, we again found that the H. pylori-infected children had an obvious reduction in sIgA than the non-infected controls. The TGF-β and TGF-β receptor signaling are important for systemic and mucosal IgA production [24][25][26]. Our data was compatible to show a significant increase in fecal TGF-β of the H. pylori-infected children due to negative feedback to sIgA depletion. Concerning the fecal calprotectin concentration in H. pylori-infected children, our results being compatible with others showed gastric H. pylori infection did not increase fecal inflammatory parameters [9,27]. The insignificant correlation may cause by wild variation of inflammatory markers between individuals. Further study need to more strict control confounding factors, which may affect microbiota and gut inflammation in study subjects.
Previous reports have shown that H. pylori colonization of the stomach alters the diversity and richness of other gastric microbiomes [28,29]. This study is the first to investigate differences in fecal microbiota between H. pylori-infected and non-infected children using next generation sequencing. We identified eight various genus/species which were significantly different in abundance between the two groups. Among them, only F. prausnitzii was relatively abundant in the gut. This bacterium is one of the most abundant commensal bacteria, and it has been shown to produce a large amount of butyrate in the guts of humans and other animals [30,31]. Clinically, the depletion of fecal F. prausnitzii may serve as a biological marker in patients with inflammatory bowel diseases. This is the first study to report a close association between pediatric H. pylori infection and depletion of gut F. prausnitzii. These results highlight the importance of H. pylori and F. prausnitzii in gut inflammatory diseases. Further studies are needed to clarify how H. pylori colonization in the stomach influences colonic microorganisms.
The ingestion of probiotics or probiotics-containing yogurt is beneficial to human health. The positive effects may be through modulation of the gut microbiota, immune function, and metabolomics [17,32]. Although this study has disclosed that fecal inflammatory parameters (calprotectin, lactoferrin, and IL-6) increased in H. pylori-infected children than controls, it is lack for the follow-up data in children after eradication and yogurt ingestion. Few studies have reported that probiotics-containing yogurt ingestion can increase the levels of certain commensal organisms, which are beneficial to gut health. Because the study yogurt containing Lactobacillus and Bifidobacterium, we are interested in whether the yogurt ingestion can increase the abundance of fecal Lactobacillus and Bifidobacterium in children. As shown in the Supplementary Figure S3, the relative ratio of Bifidobacterium increased in non-H. pylori infected children with yogurt ingestion. However, supplement of yogurt did not increase the fecal Bifidobacterium colonization in H. pylori-infected children. Unfortunately, the abundance of Lactobacillus was very rare even children having yogurt ingestion. Our study has shown yogurt ingestion really promotes gut F. prausnitzii growth in children with and without H. pylori infection. Moreover, our in vitro and ex vivo study also confirmed that L. acidophilus and B. lactis can enhance F. prausnitzii growth via direct (co-culture) and indirect (probiotics-intestine culture supernatant) manners. Finally, in agreement with other studies, we confirmed that F. prausnitzii supernatant can ameliorate LPS-induced inflammatory cytokines in HT-29 cells [33][34][35]. Further studies are needed to elucidate the relationship between H. pylori infection and reduction in gut F. prausnitzii abundance and to investigate the functional proteomics and metabolomics of F. prausnitzii and the effects on gut inflammatory diseases.
The major strength of the study is that the in vivo experiments utilizing the clinical samples identify the microbiome prominently in H. pylori-infected children. Moreover, to demonstrate the probiotics facilitate F. prausnitzii growth by clinical trial, ex vivo, and in vitro experiments. However, there are several limitations in this study, including the small number of participants limits the strength of this finding. Second, how H. pylori colonization in the stomach influences gut sIgA and cytokines that alters the abundance of colonic microorganisms is not investigated. Third, we do not study the other significantly changed microorganisms with tiny amount after H. pylori infection and yogurt ingestion. Whether these bacteria play positive or negative roles desire futher investigations. Finally, what components or species of probiotics can regulate gut homeostasis is unclear.
In conclusions, H. pylori-infected children have a dysbiosis in gut F. prausnitzii, which can be restored by H. pylori eradication and probiotics-containing yogurt. Probiotics-containing yogurt can also increase the abundance of F. prausnitzii in the non-infected children. The effect of the restored abundance of F. prausnitzii is anticipating to the control of gut inflammation.

Subject Inclusion and Exclusion Criteria
This study enrolled students from four elementary schools in Tainan City, Taiwan. Their age ranged from 10 to 12 years. After obtaining consent (A-BR-102-105 approval from the Institutional Review Board, National Cheng Kung University Hospital on 18 February 2014) from each individual and their parents, information on underlying diseases, H. pylori infection status, antibiotics intake, yogurt (probiotics) consumption, and H 2 -blocker or proton pump inhibitor use was recorded. Children who had known major organic diseases such as immunodeficiency disorders, malignancy, and diseases treated with chemotherapy, steroids and antibiotics (within 1 month), predisposed to the influence of gastrointestinal microbial colonization and immunological function were excluded. The children who ingested probiotics or probiotics-containing yogurt with a frequency of more than twice per week 1 month prior to enrollment were dropped out from fecal examinations.

Serum Collection and Diagnosis of H. pylori Infection
Overnight fasting blood samples (6-8 mL) were drawn from each participant after obtaining consent from the children and their parents. The serum was then tested for anti-H. pylori IgG antibodies (HEL-p TEST TM II; AMRAD Biotech, Perth, Western Australia) using an enzyme-linked immunosorbent assay (ELISA) with a sensitivity and specificity > 90% [36]. The seropositive and control children were further confirmed by 13 C-UBT to diagnose ongoing H. pylori infection. The cutoff value of positive 13 C-UBT was defined as an excess 13 CO2 ⁄ 12 CO2 ratio of more than 4.0% [37].

Probiotics-Containing Yogurt Ingestion, H. pylori Eradication and Follow-Up
The 13 C-UBT-confirmed H. pylori-infected children and age-and sex-matched controls ingested one bottle of AB-yogurt twice daily for 4 weeks. The yogurts (200 mL per bottle) containing at least 5 × 10 9 live organisms (Lactobacillus acidophilus, Bifidobacterium lactis, Lactobacillus bulgaricus, and Streptococcus thermophilus). The yogurt ingestion procedure followed our previous study [17]. To improve the compliance of yogurt consumption, school nurses directly watched and recorded its drinking at 07:30 a.m. and 03:30 p.m. on school days. On the weekend, the parents were requested to record those details on the same paper. After the course was completed, returned records were analyzed for compliance. Good compliance was defined as at least 80% consumption of total designed amount. Thereafter, the H. pylori-infected children received triple therapy for 10 days (pantoprazole: 2 mg/kg/day, max. 40 mg bid, amoxicillin: 50 mg/kg per day, max. 1 g bid, and clarithromycin: 15 mg/kg per day, max. 500 mg bid) with yogurt. The controls took yogurt only for 10 days. The success or failure of eradication therapy was tested by 13 C-UBT 4 weeks after completing treatment and yogurt ingestion in the H. pylori-infected children.

Stool Collection and Preparation
Two thumb-sized fresh stool samples each weighing approximately 0.2-0.4 g (HPS1 and CS1) were collected from each participant in the morning. A second stool sample (HPS2) was collected from the H. pylori-infected group 4 weeks after yogurt ingestion. Third fecal samples (HPS3 and CS3) were collected at the completion of the 4 weeks of treatment and/or yogurt ingestion in both groups. Total fecal DNA was extracted using a Qiagen stool kit (Qiagen, Chatworth, CA, USA) and then stored as −20 • C. The other sample was diluted in PBS containing 2.5 mg/mL leupeptin, 11 mg/mL aprotinin, and 0.5 mM 4-(2-aminoethyl) benzenesulfonyl fluoride (Sigma, St. Louis, MO, USA). After thorough mixing and centrifugation for 10 min at 10,000 g, the supernatant was stored as −80 • C for further ELISA tests.

Amplicon Sequencing (16S-rRNA) for Fecal Microbiota Diversity
Multiplex bar-coded indexes were used for paired-end sequencing of the 16S rRNA variable region 4 (V4). We used the universal bacterial primer pairs 515F (5 -TCGTCGGCAGCGTCAGATGT GTATAAGAGACAGGTGCCAGCMGCCGCGGTAA-3 ) and 806R (5 -GTCTCGTGGGCTCGGAGA TGTGTATAAGAGACAGGGACTACHVGGGTWTC TAAT-3 ) for maximal coverage of bacterial phylogeny [38]. Additional appropriate barcode and linker oligonucleotides were appended to the primer pairs. Illumina amplicon library generation was performed as described previously, except for the additional steps of purification of the PCR products with AMPure (Beckman Coulter, Brea, CA, USA) and quantification using a Qubit fluorometer (Invitrogen Life Technologies, London, UK) and quantitative PCR (qPCR) (Kapa Biosystems, Wilmington, MA, USA). The amplified bar-coded DNA from 119 samples was then pooled. The samples were diluted to a final dilution of 10 pM, combined at a 90:10 ratio with 10 pM of balancing library, and run with a 2 × 250 cycle reaction on an illumina MiSeq platform (Illumina, San Diego, CA. USA). Sequences of 16S rRNA were aligned and analyzed with standard operation protocol by mothur software v.1.39 (https://www.mothur.org/ wiki/MiSeq_SOP).

Probiotics Facilitate F. prausnitzii Growth In Vitro and Ex Vivo
F. prausnitzii (APC 918/95b) was purchased from the Institute of Food Science and Technology, National Taiwan University. Bacteria were cultured using YCFA GSC medium and broth in absolute anaerobic conditions [39]. Lactobacillus acidophilus and Bifidobacterium lactis (LA5 ® and Bb12, originated from the Chr. Hansen, Denmark, provided by the President Corp., Tainan, Taiwan) was used. They were maintained on a Brucella agar, incubated in anaerobic conditions.The cecum fluid and tissue were obtained form 8-week-old C57B/6 mice. The cecum tissue was cut flat using sterile scissors, and the equal-sized tissue was co-cultured with 3 mL RPMI medium with or without L. acidophilus and B. lactis in dishes for 24 h. The supernatant was filtered through a 0.22 µL filter, and then added to quantitative F. prausnitzii and YCFA for co-culture. After the 4th, 8th, and 12th hour of co-culture, the relative abundance of F. prausnitzii were measured as the percentage of total bacteria DNA copy numbers by real-time PCR using the specific primers [40].
Gastric AGS and colonic HT-29 cells (3 × 10 6 cells/well) were prepared and pretreated with supernatants for 4 h. Clinically isolated H. pylori (HP238, MOI = 100) and LPS (100 ng/mL) were then added to the dishes for serial time periods. The supernatants were collected at various time points and centrifuged for IL-8 analysis by ELISA (R&D, Minneapolis, MN, USA).

Statistical Analysis
All analyses were performed with built-in commands in mothur. Alpha diversity was determined by rarefaction curves describing the number of operational taxonomic unit (OTUs) with inverse Simpson diversity estimates. Differences in alpha diversity were assessed using the t-test and a repeated measure ANOVA. Beta diversity was assessed using unweighted and weighted UniFrac distance matrices and visualized using principal coordinate analysis (PCoA) [41]. Statistically significant differences in the relative abundance of OTUs between variable groups of patients were analyzed using Metastats (https://www.mothur.org/wiki/metastats/).

Conflicts of Interest:
The authors declare no conflict of interest.

TGF-β1
Transforming growth factor-β1 sIgA Secretory immunoglobulin A UBT Urea breath test PCoA Principal coordinate analysis OUT Operational taxonomic unit. ELISA Enzyme-linked immunosorbent assay