Alteration in the gastric microbiota of Erosive Reflux Disease and Non-erosive Reflux Disease patients: An indicator for disease progression

Introduction and Aim : Erosive reflux disease (ERD) and Non-erosive reflux disease (NERD) are common gastrointestinal (GI) diseases in primary healthcare facilities that might result in alterations in gastric microflora. The aim was to characterize the altered composition of the gastric microflora in both groups and their effects on normal metabolic pathways. Materials and Methods : A total of 27 individuals, consisting of 18 patients of ERD, 4 patients of NERD, and 5 controls based on the questionnaire data and clinical diagnosis. The gastric microbiome was sequenced using 16S rRNA next-generation sequencing. Gastric microbial diversity and compositions were analyzed using MicrobiomeAnalyst. Functional analysis was performed using PICRUST. Results : Dysbiosis was observed in both ERD and NERD groups when compared to the control. Alpha diversity was found to be significant at the ACE index, at the order (p = 0.0239) and class (p = 0.019) taxonomic levels. The gastric microbiome composition at the genus level of the study groups represented a significant decrease of Gram-positive bacterial genera such as Streptococcus , Corynebacterium , and Granulicatella along with an increase of Gram-negative genera such as Helicobacter and Veillonella . Significant alterations in the metabolic pathways due to this dysbiosis were also predicted for both groups. Conclusion : The alterations in the gastric microflora are caused by both ERD and NERD, these alterations are further associated with a shift in microbial consortia towards the abundance of gram-negative bacterial genera that might lead to the progression of GI diseases to a chronic state.


INTRODUCTION
astroesophageal reflux disease (GERD), a common gastrointestinal (GI) condition for which the cause is still unknown, has been defined as a condition that develops when there is a retrograde flow of stomach contents back into the stomach.One of the most common complications of GERD is oesophageal inflammation.When Gastric juices are refluxed into the lower oesophagus, it causes predisposition and transformation of squamous epithelial mucosa to the metaplastic columnar epithelium.Based on the findings of conventional endoscopy and histopathological examination, GERD is categorized into non-erosive reflux disease (NERD) and erosive reflux disease (ERD) (1).ERD is described as GERD patients having endoscopic evidence of esophagitis whereas NERD is described as GERD patients with no endoscopic evidence of esophagitis.Thus, ERD and NERD are differentiated on the basis of the presence or absence of mucosal lesions (2).The symptoms of GERD include decreased esophageal clearance, delayed gastric emptying, and reduced function of the lower esophageal sphincter.Obesity, high-fat diets, alcohol consumption, and smoking have also been linked to increased acid reflux severity (3,4).A particular composition of the microbiota is maintained in each part of the normal GI tract, for example, the esophagus is predominated by transient Gram-positive bacteria, whereas a microbial consortium with more stability occupies the low acidic gastric environment of healthy individuals.This commensalism affects the metabolism of microorganisms and is significantly related to diseases and phenotypes (5).Any disruption in this balance of the gastric microflora causes dysbiosis, a term that includes any alteration in the composition of the commensal microbiome (6).According to a recent study, gastric bacteria create peptides that modify ghrelin secretion, which in turn modulates the enterochromaffin cells that produce gastric juice and ultimately affect the stomach's acidity level (7).The shifting of gastric microflora from Gram-positive bacterial genera to Gram-negative bacterial genera has been reported in the oesophagus of ERD patients which is associated with GC development (8).Helicobacter pylori is the most prevalent bacterial genera in patients suffering from gastrointestinal diseases which extensively promotes the growth of other Gram-G negative bacteria.The link between H. pylori and GERD, however, remains controversial.A recent metaanalysis in Indonesia reported a high prevalence of GERD, whereas the prevalence of H. pylori infection was comparatively low (9).This phenomenon can further be justified with a detailed microbial perspective.The gastric biopsy specimens were exploited using the 16S rRNA approach for the evaluation of differences in the gastric microflora between ERD or NERD patients.The abundance of H. pylori in the disease groups was inferred and the shift in the microflora was considered for the functional analysis in the disease groups.Our objective was to analyze the alterations in the gastric microbial consortium of ERD and NERD groups which could give insight into the progression of GI diseases.

Collection and screening of samples
This study was conducted to evaluate the microbial dysbiosis in ERD and NERD samples in comparison to control samples.Patients (n = 50) visiting the Department of Gastroenterology, Max Super Specialty Hospital, Vaishali, Ghaziabad, Uttar Pradesh, for their upper gastroesophageal symptoms were included in this study and were evaluated by a gastroenterologist.The patients underwent upper gastrointestinal endoscopic screening for their evaluation and treatment protocol.Three biopsy samples of each patient were collected from their stomach antrum as per inclusion and exclusion criteria.One sample was used for a rapid urease test (RUT) and the second sample was collected in phosphate buffer saline for DNA isolation and sequencing and the third sample was collected in 10% buffered formalin for sequencing.Age 15 to 90 years old, upper gastrointestinal endoscopic consent and clinical indications of gastro-duodenal disease associated with H. pylori were among the inclusion criteria.On the other hand, as an exclusion criterion, the use of proton pump inhibitors and antimicrobial medications was prohibited for three months before the sample collection.Based on a questionnaire, demographic details were collected, consisting of gender, smoking, and alcohol-drinking habits as risk factors for GERD and dyspepsia (Table 1).On the basis of our study design, subjects with incomplete specimens or with GC, gastric or duodenal ulcers, and intestinal metaplasia were excluded.The symptoms and upper-endoscopy evaluations by the gastroenterologist were considered the basis of disease characterization.The patients who were diagnosed with both reflux symptoms and mucosal lesions (esophageal erosions) and were also fulfilling the Los Angeles classification were concluded with ERD, whereas the patients diagnosed with only reflux symptoms, heartburn and chest pain were concluded with NERD.Random sampling was performed to obtain 50 samples of ERD, NERD, and Control for sequencing and post-filtering based on the questionnaire and clinical diagnosis, a total of 27 samples were selected for this study.

Ethics approval
All the procedures performed in the study were in accordance with the ethical standards of the institution.This study was approved by the institutional ethical committee of Amity Institute of Biotechnology, Noida, U.P, dated 2-12-2015, and Max Hospital, Vaishali, U.P, India (RS/MSSH/VSH/CRL/IEC/ GASTRO/17-19).

Consent to participate and publish
Written informed consent was taken from all participants.The authors affirm that human research participants provided informed consent for publication.

Data availability
The datasets generated and analysed during the current study are available in the NCBI Sequence Read Archive under Bio Project accession no PRJNA904926 with the accession numbers SRR22426106 to SRR22426144 and SRR22859444 to SRR22859448 a p-value: calculated using the Fisher Exact test between ERD and control groups.b p-value: calculated using the Fisher Exact test between NERD and control groups.

DNA extraction and metagenomic profiling of 16S rRNA sequences
Genomic DNA from human tissue biopsy samples was extracted using the QIAamp DNA Mini Kit (QIAGEN GmbH, Hilden, Germany).It was quantified and assessed for quality using agarose gel electrophoresis.A specific primer with a sequencing linker was created for the 16S rRNA V3 and V4 region using the KAPA HiFi HotStart PCR) Kit (R&D Cape Town, South Africa) for 26 cycles.Amplified signals were checked on a 1.2 percent agarose gel.Round 1 PCR amplicons were further amplified (10 cycles) to include Illumina sequencing barcoded adaptors.Round-2 PCR amplicons (the sequencing libraries) were examined on a gel.Further for multiplex sequencing, the libraries were additionally normalized and pooled.For 16S rRNA gene amplification, the illumina adapter sequences listed below were utilized.

Statistical analysis
The microbial alpha diversity analysis of human tissue biopsy samples was determined by using QIIME (Version 1.7.0) and represented by various indices, using MicrobiomeAnalyst.The relative abundance of the bacterial diversity among the ERD, NERD, and control samples was determined by analysis of variance (ANOVA) and represented using a stacked bar plot chart.Beta diversity was analysed using the Bray-Curtis dissimilarity index with Permutational Multivariate Analysis of Variance (PERMANOVA), represented by the PCA plot.The correlation analysis was performed using Spearman Rank Correlation as the distance measure.Using the Shotgun Data Profiling feature of MicrobiomeAnalyst, Linear discriminant analysis effective size (LEfSe) was performed to determine the significant Kyoto Encyclopaedia of Genes and Genomes (KEGG) IDs (p-values < 0.05) that explain the involvement of the microbial flora in the functional differences between the ERD, NERD, and control samples.Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) was used to estimate the functional content of a metagenome using marker gene data and a database of reference genomes.STAMP was used to graphically represent the results.The p-value was determined using the Kruskal-Wallis H-test and Welch's t-tests; differences were deemed significant for p-values less than 0.05.

Patients' characteristics
The samples were divided into three study groups: 18 ERD, 4 NERD, and 5 Control.Compared to the control samples, alcohol, and smoking status did not significantly correlate with study groups.The rates of alcohol consumption and smoking were 44.44% in the ERD group (Table 1).

Gastric microbiota diversity comparison among study groups
In this dataset, 1144 taxonomic units were detected and assigned to respective phyla, among them 7 phyla were found to be significantly abundant among the study groups.The relative abundance of the phyla in each sample across the diseased groups is illustrated below (Fig. 1).At the phylum level, the gastric microbiota was majorly composed of Proteobacteria, Firmicutes, Bacteroidetes, Campylobacteria, Actinobacteria, and Fusobacteria, with varied percentages of relative abundance, in both H. pyloripositive and -negative individuals.The gastric microbiome profiles in the three groups of subjects were inhabited by Bacteroidetes (ERD 10.42%, NERD 1.87%, and control 14.7%), Firmicutes (ERD 21.5%, NERD 19.08%, and control 68.5%), Proteobacteria (ERD 5.29%, NERD 5.09%, and control 2.95%) (Fig. 1).The relative abundance between the study groups at the genus level showed an increase in the abundance of Helicobacter, Corynebacterium, Staphylococcus, and Veillonella in the disease groups; whereas the abundance of Streptococcus decreased in the diseased groups as compared to the control group (Fig. 1).
Additionally, the alpha-diversity (within-sample diversity) indices of the bacteriomes of the research groups-Chao1, ACE, Fisher, Shannon, and Simpson-were used to estimate the OTU richness.A significant difference in alpha diversity was found in the ACE index, at the order (p = 0.0192) (Fig. 2a) and class (p = 0.0239) (Fig. 2b) taxonomic levels.Using the Bray-Curtis dissimilarity metric, the beta-diversity reflecting differences in the bacterial community profile between the samples was analyzed and utilized in the principal component analysis (PCA) plot.This showed that there was no significant difference between the bacteriome of either group (F=0.92976,R 2 = 0.071908, p > 0.05) (Fig. 2c).

Core microbiome in disease categories
Despite the differences in the disease outcomes, a diverse core microbiome was observed at the genus level which remains unaltered in its composition across the study groups.Core microbiome analysis was performed with sample prevalence ≥ 20% and relative abundance ≥ 0.01% (Fig. 3).The 21 core bacterial genera were identified amongst which, Streptococcus, Veillonella, Helicobacter, and Prevotella_7, Staphylococcus genera were observed to be prevalent among the disease categories (Fig. 3).

Alteration pattern in study groups
We performed a pattern search at the genus level to identify the correlation between the disease categories using Spearman Rank Correlation as the distance measure.At the genus level, Helicobacter, Ralstonia, Veillonella, and Staphylococcus were positively correlated with the disease categories of patients (Fig. 4a).Since Helicobacter showed maximum positive correlation, another pattern search was performed along with Helicobacter as a feature (Fig. 4b).
Staphylococcus, Micrococcus, Neisseria, Haemophilus, and Veillonella were found to be positively correlated and these genera were found to be prevalent in ERD and NERD samples.According to the heat map generated, we observed that the abundance of these bacterial taxa was maximum in the ERD group, followed by the NERD group and control (Fig. 4b, 4c).Gemella, Paracoccus, Porphyromonas, Alloprevotela, and Enhydrobacter showed a negative correlation with the disease categories.The abundance of these bacterial taxa was found to be decreasing in NERD, followed by a further decrease in ERD patients.

Co-occurrence network of bacterial genera associated with study groups
A co-occurrence plot at the genus level was generated to determine plausible associations within the bacterial communities of the study groups, using MicrobiomeAnalyst software.Overall, it was discovered that there were significant differences in the prevalence of 289 genera throughout the study groups (Fig. 5).Of them, 283 significant positive correlations and 6 negative correlations (p-value < 0.05 and coefficient correlation > 0.3) were found.An interactive correlation network was generated and as Helicobacter showed the highest abundance in ERD and NERD group and control, its correlation was further analysed with other bacterial genera.Helicobacter was found to be positively correlated with Lawsonella (p=0.0374),Methylobacterium methylorubrum (p=0.0012),Micrococcus (p= 0.0245), Ralstonia (p = 0.0024), Staphylococcus (p = 0.0007), and Paracoccus (p = 0.0036) (Fig. 5).

Predictive function of the gastric microbiome in study groups
Using the PICRUSt algorithm, we investigated the probable interactions between the gastric microbiome and the functional pathways based on inferred metagenomes and contrasted the variations of these differences between the ERD, NERD, and control groups.Of the 116 associated KEGG pathways, nine were statistically significant (P< 0.05) and distinct among the ERD and control groups, whereas two pathways were statistically significant (P <0.05) and distinct among the NERD and control groups.Interestingly, pathways related to genetic information (Purine metabolism,Thymine metabolism, and Nucleotide metabolism), pathways related to human diseases (Benign chronic pemphigus), along with other major metabolic pathways (carbon metabolism, methane metabolism, lipid metabolism, and pyruvate metabolism) were found to be enriched in ERD group as compared to the control (Fig. 6a).Three KEGG pathways were statistically significant (P<0.05) and different amongst NERD and control pathways related to human diseases (Benign chronic pemphigus), and metabolism (nitrogen metabolism) were enriched in contrast to the control group 6b).
These findings might correlate the effects of the gastric microbiome on both ERD and NERD groups, leading to plausible changes in the body's disease susceptibility and basal metabolic functions.The role of the gastric microbiome in influencing gastric inflammation can also be concluded from this analysis, which in turn indicates the relation between H. pylori infection and ERD resulting in a progression toward Barette's oesophagus, oesophageal adenocarcinoma (EAC) and GC.

DISCUSSION
A prior found a correlation between specific microbial changes in ERD and NERD, conditions that may not be fatal but can significantly lower patients' quality of life (10).The metaplastic columnar epithelium that results from the esophageal inflammation caused by ERD at the gastroesophageal junction (GEJ) might spread into the esophageal cavity (ECC) by transforming the native squamous epithelium (11).Along with the presence of H. pylori, EAC has been found to host unique microbiomes which could be a side effect of the highly acidic environment created by ERD (12,13).This study was done to understand the dysbiosis of the microbiome in the case of ERD and NERD patients when compared to control patients.We observed a possible increased influence of H. pylori on the microbial diversity in both groups.Our results were found to be in line with the study of Serrano et al., in which they reported that the predominant phyla in the gastric microbiota were Firmicutes, Proteobacteria and Bacteroidetes in the absence of H. pylori infection, whereas in the case of dominance of H. pylori, the gastric microbiota showed a reduced bacterial diversity (14).
In the present study, an evaluation of the relative abundances of different bacterial taxa inhabiting the gastric environment of the three study groups was performed which showed that microbiota composition among ERD and NERD groups differ from the control group.In ERD and NERD groups, the microbial consortium was dominated by: Firmicutes, Proteobacteriota, Actinobacteriota, Campylobacterota, and Bacteroidota.Our results are consistent with research by Nardone et al., that found that the lower esophageal tract microbiome is specific to patients with either ERD or NERD and contains high concentrations of Gram-negative species such as Proteobacteria, Fusobacteria, and Campylobacter (15).The gastric microbiome composition at the genus level of the study groups, as represented in (Fig. 1) showed a decrease in the abundance of Gram-positive bacterial species such as Streptococcus, Corynebacterium, and Granulicatella along with an increase in the abundance of Gram-negative species such as Helicobacter and Veillonella.In another study, Macfarlane et al., reported a gradual shift from Streptococcus-dominated Gram-positive bacterial species in normal patients to Gram-negative colonisation in patients with either ERD or NERD, along with the appearance of genera not usually found in the distal oesophagus, like Campylobacter.Our study is also in line with this study where we found the richness of Campylobacter in both disease groups (16).
To evaluate the species richness, the α-diversity of bacteria present between the ERD, NERD, and control patients was observed, which was more in the ERD as compared to that in the control, whereas was less in the NERD group.Findings in the case of the NERD group were in contrast with the study conducted by Yang et al., wherein they reported that the gastric microflora of the NERD patients had much more genuslevel diversity than the control group (17).Similarly, in the case of ERD, our findings were in contrast to the study conducted by Shi et al., which reported that bacterial diversity is decreased in the ERD groups in contrast to the control group (18).
To evaluate the correlation, a pattern search was conducted, which illustrated an increase in the abundance of different bacterial genera in the case of ERD and NERD groups as compared to the control group.Veillonella, Haemophilus, Ralstonia, Neisseria, and Helicobacter showed increased abundance in the ERD group; whereas the abundance of TM7x, and Neisseria was increased in case of NERD group as compared to control.Our findings were in accordance with Hao et al., which reported that Veillonella and Ralstonia show increased abundances in the ERD group (19).Another study conducted by Yang et al., reported an increase in the abundance of Haemophilus, Neisseria, and Veillonella in the ERD group (16).According to the findings of Sugihartono et al., Helicobacter is a bacterium able to dominate gastric microbiota and induce significant inflammation in the case of the ERD group (20).In a different study, Samolka et al., described a condition called hypochlorhydria, which is elevated stomach pH brought on by gastric atrophy or H. pylori-induced suppression of the gastric proton pump.However, this condition is only seen in a small percentage of patients who also have elevated gastrin levels and antralpredominant gastritis (21).Another study by Pero et al., found that H. pylori infection increases β-Defensin 2, a type of innate antimicrobial factor produced by host epithelia, which likely affects the proliferation of other gastric microorganisms and suggests that the host immune response may affect the gastric microbiota (22).Nevertheless, the relationship between ERD and H. pylori infection still remains controversial, and to draw conclusions needs further research.In accordance with our findings of the NERD group, Zhou et al., reported an increased abundance of TM7x and Neisseria (23).In light of the above findings, the alteration of gastric microflora is evident in both groups, which could in turn promote the alteration of various pathways.Thus, it becomes essential to analyse the possible pathways enriched with the abundance of Gram-negative bacterial genera in both ERD and NERD groups.
To identify the altered KEGG pathways, a metagenome-wide association study was performed which showed that pathways related to purine metabolism, carbon metabolism, and lipid metabolism were enriched in the ERD group, whereas the nitrogen metabolism was enriched in the NERD group.The end product of purine metabolism is uric acid, an enrichment in this metabolism in the ERD group might result in the of hyperuricemia further leading to GOUT; this observation is in accordance with the research by Yuan et al., (24).The association of polymorphisms in the one-carbon metabolism (OCM) pathway with GC has been reported by Zhao et al., (25).Another study by Haber et.al., reported the association between GERD and gastritis, hence it can be concluded that enrichment in the carbon metabolism might result in the progression of ERD to gastritis and further to gastric carcinoma.Bi et al., reported the dependence of Helicobacter on unsaturated fatty acids for the maintenance of its membrane structure and functions (26).Ktsoyan et al., stated a significant elevation in the levels of unsaturated fatty acids in the blood of patients with H. pylori infection-induced ulcerations (27), suggesting an association between H. pylori infection and higher lipid metabolism, our analysis also revealed the enrichment of lipid metabolism in ERD patients.Liu et al., reported that nitrate reduction by bacteria that produce nitrite is thought to be the reason for the damage associated with NERD; in acidic environments, nitrate reduction can result in the production of cancercausing N-nitroso compounds and nitrous oxide, which inhibits DNA repair enzymes thereby enhancing the mutagenesis (28).Our analysis also revealed the enrichment of nitrogen metabolism in the case of the NERD group.These alterations in KEGG pathways suggest that gastric microflora dysbiosis might influence the progression of ERD and NERD, towards gastritis and GC.Consequently, additional analysis of the alteration of the gastric microbiome is required to confirm the actual association between H. pylori, ERD, and NERD.However, the limitations of our study include a small sample size, the nature of the selfdisclosure questionnaire and the existence of partial recall bias due to which the symptoms reported might not be exact.Similar research with a larger sample size might overcome these limitations, yielding helpful insights into clinical therapeutics.

CONCLUSION
Conclusively, this study showed that the gastric microflora in the ERD and NERD groups had compositional and functional dysbiosis, which was further aggravated by H. pylori infection.This study indicates that a combination of certain specific bacterial genera could modulate the gastric microbiota, reversing the shift of the microbiome to improve the clinical outcome in ERD and NERD patients.The findings of this study might pave the way for the initiation of larger-cohort clinical validations and the development of guidance for therapeutic strategies with probiotics.

Fig. 1 :
Fig. 1: To compare the relative taxa abundance among control, ERD, and NERD samples at the (a) phylum level and (b) genus level

Fig. 2 :Fig. 3 :
Fig. 2: Alpha diversity between the study groups was found statistically significant at the ACE index.The p-value was determined using T-test/ANOVA.Control, ERD, and NERD are represented by red, green, and blue bars, respectively.(a) Class Level and (b) Order Level.(c) Beta diversity analysis was performed using the Bray-Curtis dissimilarity metric and represented using PCoA for bacterial community comparison in ERD, NERD, and control samples

Fig. 4 :
Fig. 4: To identify the correlation of different taxa along with the sample categories, we performed a pattern search with Spearman rank correlation as a distance measure.Red bars represented a positive correlation and blue bars represented a negative correlation, (a) correlation of top 25 genera with disease categories, (b) correlation of top 25 genera with Helicobacter, and (c) relative change in abundance of important genera

Table 1 :
Demographic characteristics of the samples