MALDI-TOF MS and 16S RNA Identification of Culturable Gastric Microbiota: Variability Associated with the Presence of Helicobacter pylori

Helicobacter pylori is the main bacteria associated with gastroduodenal diseases. Recent studies have reported that gastric microbiota might be modified by the H. pylori colonization, favoring gastric lesions′ development. In Chile, the region of La Araucanía concentrates a high risk of gastric cancer associated with Helicobacter pylori colonization, rurality, poverty, and Mapuche ethnicity. Hence, we aimed to identify the culturable gastric microbiota and characterize its variability at different stages of epithelial injury, based on its H. pylori colonization in dyspeptic patients from this Chilean region. Microaerophilic bacteria strains were isolated from antrum biopsies of 155 dyspeptic patients′ biopsies and identified using MALDI-TOF MS or 16sRNA gene sequencing for non-pylori species identification, and UreC gene amplification for H. pylori confirmation. We found 48 species from 18 families, mainly belonging to Neisseriaceae (21.3%), Streptococcaceae (20.0%), Actynomicetaceae (9.0%), Enterobacteriaceae, and Lactobacillaceae (4.5%); however, Streptococcaceae and Actinomycetaceae families showed a significant reduction in samples infected with H. pylori, along with a considerably lower diversity of species. Our results revealed a microbiota modification due to H. pylori colonization associated with the gastric epithelial state, suggesting a potential microbiota role for developing and progressing gastric diseases.


Introduction
The discovery of Helicobacter pylori (H. pylori) in 1983 started the concept of colonization in the gastric epithelium. For decades, this bacterium has been considered the primary agent responsible committee mentioned above also approved this project (code 111_17) by the resolution NO 4183 on 17 October 2017. Participants were interviewed to complete their information, including age, gender, educational level, Mapuche ethnicity, household, place of residence, smoking and drinking habits, metabolic and cardiovascular diseases (diabetes, hypercholesterolemia, and arterial hypertension), family history of gastric cancer, and H. pylori infection history. Every abnormality (gastritis, ulceration, erosion, and others) was recorded during the gastrointestinal endoscopy when examining the esophagus, stomach, and duodenum.
Two biopsy specimens of the stomach antral zone were obtained from each patient. Antrum samples were chosen as recommended by the standardized biopsies method [19], which the national health institutions used for sampling. Additionally, this sample type is reported as a common stomach area for H. pylori colonization and its detection [19]. One biopsy sample was used for the urease test, which was performed in the respective Health Center using the Rapid Urease Test (RUT) CLOtest ® Kimberly Clark, Irving, TX, USA. The second sample was stored in a sterile environment in an Eppendorf tube with 200 µL of Brucella Broth (BD-DIFCO, Berkshire, UK) and transported in containers at 4 • C to the Applied Biology Molecular Laboratory of the Centre of Translational Medicine at the Universidad de La Frontera.

Categorization of State Gastric Epithelial
Based on the endoscopic observation from the gastric mucosa, including the different gastric mucosa (antrum, corpus, fundus) and duodenum, two categories for describing its state were assigned: lesion and non-lesion. Non-lesion indicated tissue without inflammatory or injury signals, while lesion indicated tissue with epithelia showing some gastric or duodenal injuries. Furthermore, the lesion group was subdivided into Non-erosive Lesion (NEL) for the inflammation or non-erosive epithelial injuries, Erosive Lesion (EL) for the tissue with wounds, and Pre-malignant or Malignant Lesion (PML) for gastric atrophic, metaplasia, dysplasia, or gastric cancer [23,24].

Isolation of Culturable Microaerophilic Microbiota
In a period of fewer than two hours after the collection of the samples, the biopsy was macerated and homogenized with sterile polypropylene baguette. A fraction of 100 µL of the homogenate sample was inoculated in Columbia agar (Oxoid CMO 331, UK) enriched with 7% (v/v) equine blood. The other fraction (100 µL) was inoculated in culture media plates (Columbia agar) enriched with 7% (v/v) equine blood and supplemented with the antibiotics Trimethoprim (5 µg mL −1 ), Amphotericin B (5 µg mL −1 ), Cefsulodin (5 µg mL −1 ), and Vancomycin (10 µg mL −1 ) (Supplement DENT 2%, Oxoid TM , Hampshire, UK) to favor the H. pylori growth. After seeding, both plates were incubated in a microaerophilic atmosphere (5% O 2 , 10% CO 2 , 85% N 2 , and 90% humidity) using the CampyGen closed system (Oxoid TM , Hampshire, UK) from three up to seven days at 37 • C. Visible bacterial colonies were separated according to its phenotype, morphology, Gram stain, catalase, and oxidase test. All isolates were stored at −80 • C in Brucella Broth (BD-DIFCO, Berkshire, UK) with 5% (v/v) equine serum and 20% glycerol (Sigma-Aldrich, Gillingham, UK) for future molecular identification.

Identification of Non-Pylori Species (Culturable Microaerophilic Microbiota)
Bacterial isolated strains were identified using the MALDI-TOF (Matrix-Assisted-Laser-Desorption-Ionization Time-of-Flight) method with the MALDI-TOF MS Autoflex Speed equipment (Bruker Daltonics, Bremen, Germany) and the Flex Control 1.4 software (Bruker Daltonics, Bremen, Germany), for the automatic acquisition of the mass spectrum in the positive linear mode, in a range of mz −1 from 2 to 20 kDa and according to the manufacturer s instructions. The raw spectra were compared using the Biotyper Compass 4.1 database (Bruker Daltonics, Bremen, Germany) for isolated identification. The spectra were recorded at 20 Hz laser frequency, and identification was set from 0 to 3.0, with matching scores higher than 1.7 for genus and 2.0 for species [25].
Bacteria not available in the MALDI TOF database were analyzed by sequencing the 16sRNA gene. Genomic DNA was obtained from the pure culture using a microbial DNA extraction kit (DNeasy UltraClean Microbial Kit ® Qiagen, Germany) following the manufacturer s recommendations. Primer 27F 5 -AGAGTTTGATCCTGGCTCAG-3 and 1392R 5 -GGTTACCTTGTTACGACTT-3 were used to amplify the conserved region 16sRNA, as described by Srinivasan et al. [26]. The reaction mixture was initially subjected to a one denaturation cycle at 94 • C for 5 min, followed by 35 cycles: denaturation at 94 • C for 30 s, hybridization at 53 • C for 30 s and extension at 72 • C for 30 s, and finally extension cycle at 72 • C for 5 min in a Labnet digital thermocycler, Multigene Optimax model (LabNet International Inc., Edison, NJ, USA). Subsequently, the amplified DNA was visualized in 1% agarose gel. Then, PCR products were purified using the Gel Band purification Kit (GE Healthcare Life Sciences, Buckinghamshire, UK) and sequenced in Macrogen Korea ® . Sequencing results were analyzed with the MEGA 7.0 program (Molecular Evolutionary Genetic Analysis, Software 64-bit) [27], and compared with the NCBI database (https://www.ncbi.nlm.nih.gov).

Helicobacter pylori Confirmation
Bacteria with suggestive H. pylori phenotype were subjected for confirmation using the Polymerase Chain Reaction (PCR). Flanking sequences of the ureC gene were amplified using the Fw 5 -AGCTATAAAGTGGGCGAGAG-3 and Rv 5 ATTGCACCCGTTAGGCTCAT-3 , as described by Wormwood et al. [28]. A patient was considered infected with H. pylori when confirmed positive for the rapid urease test, matching comparison with the 16SrRNA gene (as established in Section 2.4), and when the ureC gene was detected from the gastric biopsy.

Statistical Analyses
Association studies of the characteristics of the patients were carried out using the Graph Pad Prism 8.4.3. program. The Chi-square test of association was used to establish the different characteristics of participants and to compare the diversity and number of isolated species between H. pylori-infected and non-infected patients (henceforth referred to as HP+ and HP-groups, respectively) with the different states of the epithelial injury. The comparison of age (average ± SD) between the HP+ and HP-groups were analyzed using t-test. A p-value < 0.05 was considered significant for all the statistical tests.
Regarding the age range, younger individuals were more significantly represented in HP+, with a mean (±SD) of 48.2 ± 13.3 years (p = 0.041) ( Table 1). H. pylori infection highest prevalence was observed for the 46-55 years-old group (n = 19, 59.4%), which began to decrease significantly after 56 years of age ( Figure 1). Furthermore, the 36-45 year-old group had a minor H. pylori prevalence than other age ranges (n = 15, 40.5%). It exhibited a more significant population belonging to Mapuche ethnic in HP-. For the gastric epithelial conditions, according to the endoscopic report, 66.4% (n = 103) of participants did not present a gastric injury detectable using this method. However, 52 (33.6%) of the samples presented lesions, mainly for the HP-(n = 32/61.5%), which had a 64.5% (n = 20) of NEL and an 70.0% (n = 7) of PML, 4 (80.0%) specimens suggestive of metaplasia injurie. Less frequent gastric lesions were found on the HP+ (n = 20, 38.5%), where NEL was the most abundant (n = 11, 35.5%) and malignant gastric lesions were found in 30% (n = 3) mainly due to atrophic injurie (n = 2, 40.0%) ( Table 2). Despite this, the distribution of epithelial injuries was similar for both groups.

Bacterial Identification
The microbiological analysis allowed the isolation of bacteria in 126 (81.3%) antrum biopsy samples, corresponding to 48 species, highlighting a wide diversity. Those isolates belonged to the Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria phyla, grouped into 18 families and 24 genera (not considering H. pylori). The most prevalent non-pylori bacteria were the Neisseriaceae (n= 33, 21.3%) family, followed by the Streptococcaceae (n = 31, 20.0%)-with the most extensive diversity of species-and Actinomycetaceae (n = 14, 9.0%), as well as Enterobacteriaceae and Lactobacillaceae families, both registered in seven biopsies (4.5% respectively). The remaining bacterial families were

Distribution of Culturable Bacteria According to the State of the Gastric Epithelium
Different microbiota profiles (diversity and number of species) were observed according to the epithelium state. An impressive decrease of Gram-negative non-fermenting Bacilli (BNF), (Haemophilus and Flavobacterium) was detected in the injured epithelium. Those were replaced with different microbiota mainly comprised by Gram-positive bacteria (Propionibacteriaceae, Carnobacteriaceae, Micrococcaceae, and Prevotellaceae), representing a potential dysbiosis of the culturable microbiota regarding the epithelial damage progress (Figure 3a,b).
A significant decrease in culturable microbiota diversity was observed in tissues colonized by H. pylori (HP+) (Figure 3a). In this group, specifically for the non-erosive lesion, five species belonging to Proteobacteria (Neisseria), Actinobacteria (Actinomyces), and Firmicutes (Streptococcus, Lactobacillus,

Distribution of Culturable Bacteria According to the State of the Gastric Epithelium
Different microbiota profiles (diversity and number of species) were observed according to the epithelium state. An impressive decrease of Gram-negative non-fermenting Bacilli (BNF), (Haemophilus and Flavobacterium) was detected in the injured epithelium. Those were replaced with different microbiota mainly comprised by Gram-positive bacteria (Propionibacteriaceae, Carnobacteriaceae, Micrococcaceae, and Prevotellaceae), representing a potential dysbiosis of the culturable microbiota regarding the epithelial damage progress (Figure 3a,b).
A significant decrease in culturable microbiota diversity was observed in tissues colonized by H. pylori (HP+) (Figure 3a). In this group, specifically for the non-erosive lesion, five species belonging to Proteobacteria (Neisseria), Actinobacteria (Actinomyces), and Firmicutes (Streptococcus, Lactobacillus, Staphylococcus) phyla were observed. When the inflamed epithelium progressed towards ulcerous/erosive states, bacterial diversity was reduced to two genera of different phyla, Proteobacteria (Acinetobacter johnsonii), and Firmicutes (Enterococcus faecium). Only the Proteobacteria phylum remained in malignant states, specifically the species Neisseria perflava and N. flavescens. On the other hand, there was a reduced variability of species diversity in the HP-group. In this group, the Proteobacteria, Actinobacteria, Flavobacterium, and Firmicutes phyla were present in all epithelial states except in the advanced damaged ones, in which members of the Flavobacterium phylum were not detected. However, a microbiota dynamic was detected, since it changed from nine genera present on non-erosive lesions (NEL) to five genera on erosive lesions (EL); moreover, it increased to seven genera in malignant lesions (PML) (Figure 3b). the other hand, there was a reduced variability of species diversity in the HP-group. In this group, the Proteobacteria, Actinobacteria, Flavobacterium, and Firmicutes phyla were present in all epithelial states except in the advanced damaged ones, in which members of the Flavobacterium phylum were not detected. However, a microbiota dynamic was detected, since it changed from nine genera present on non-erosive lesions (NEL) to five genera on erosive lesions (EL); moreover, it increased to seven genera in malignant lesions (PML) (Figure 3b).

Discussion
Despite the harsh conditions of the gastric environment, various studies have described a colonizing microbiota in the stomach. Moreover, it has been recognized that chronic transmissible and non-transmissible diseases can modify this microbiota, affecting the intestinal homeostasis and promoting gastric lesions [29,30]. However, their specific conformation and interaction for disease progression are not fully understood yet. According to our results, bacteria belonging to the Proteobacteria, Firmicutes, Bacteroidetes, and Actinobacteria phyla were identified in the HP-group. Conversely, Proteobacteria, Actinobacteria, and Firmicutes phyla were also identified in the HP+ group showing significantly less abundance (number of species identified from each phylum), even when including participants who had taken an eradication treatment in the HP-group (p = 0.031, Tables S1 and S2). The results mentioned above highlight that bacteria diversity decreases in people infected with H. pylori, and might be recovered after eradication treatments, as reported in similar studies [15,[31][32][33].
We found 47 different species distributed in 18 families with a predominance of Neisseriaceae, Streptococcaceae, and Actinomyceteae, consistent with a Chinese microbiota report regarding dyspeptic patients without an H. pylori infection [34]. Moreover, microbiota studies performed in healthy Swedish individuals showed a greater diversity and number of Streptococcus, Actinomyces, Prevotella, and Gemella bacteria, specifically in those without an H. pylori infection [35]. In this study, those

Discussion
Despite the harsh conditions of the gastric environment, various studies have described a colonizing microbiota in the stomach. Moreover, it has been recognized that chronic transmissible and non-transmissible diseases can modify this microbiota, affecting the intestinal homeostasis and promoting gastric lesions [29,30]. However, their specific conformation and interaction for disease progression are not fully understood yet. According to our results, bacteria belonging to the Proteobacteria, Firmicutes, Bacteroidetes, and Actinobacteria phyla were identified in the HP-group. Conversely, Proteobacteria, Actinobacteria, and Firmicutes phyla were also identified in the HP+ group showing significantly less abundance (number of species identified from each phylum), even when including participants who had taken an eradication treatment in the HP-group (p = 0.031, Tables S1 and S2). The results mentioned above highlight that bacteria diversity decreases in people infected with H. pylori, and might be recovered after eradication treatments, as reported in similar studies [15,[31][32][33].
We found 47 different species distributed in 18 families with a predominance of Neisseriaceae, Streptococcaceae, and Actinomyceteae, consistent with a Chinese microbiota report regarding dyspeptic patients without an H. pylori infection [33]. Moreover, microbiota studies performed in healthy Swedish individuals showed a greater diversity and number of Streptococcus, Actinomyces, Prevotella, and Gemella bacteria, specifically in those without an H. pylori infection [34]. In this study, those species were also identified in normal and injured epithelia, in both groups, mainly in the H. pylori non-infected tissue (HP-).
In the samples colonized by H. pylori, less diversity of bacteria and fewer tissues with lesions were detected when compared with the HP-group. This modification in the diversity of bacteria caused by H. pylori is consistent with reports that indicate that H. pylori causes a change in the colonized environment, reducing the microbiota, mainly in advanced lesions [35]. Probably, once the lesion occurred and its damage and size progressed, the H. pylori moved to a healthy epithelium. Besides, due to its mobility, H. pylori colonizes throughout the gastric epithelium, it remains in the antral region causing chronic inflammation and histological changes, where lesions can evolve to atrophic gastritis and achlorhydria [6,36]. In turn, advanced lesions might facilitate the growth of other species unable to colonize the acidic environment, while the competition among bacteria promotes coccoid forms of H. pylori and finally inhibition of its growth [23,33], driving a gastric dysbiosis. In our study, H. pylori was less frequently detected in pre-malignant lesions (atrophic and metaplasia). Based on the aforementioned preliminary evidence, we cannot ignore the possibility of infected patients where H. pilory was localized in other gastric areas but absent in the antral zone. Despite this, it was reported that antral samples yielded similar H. pylori detection sensitivity compared with corpus and antrum+corpus samples regarding the atrophic lesion degree [37,38], which might be necessary for early diagnosis. Therefore, additional studies should be conducted for conclusive results regarding the H. pylori sample-dependent detection in infected patients.
Other reports have shown that H. pylori can reduce microbiota in children and adults [39]. Also, it has been described that co-excluding interactions between H. pylori and Fusobacterium, Neisseria, Prevotella, Veillonella, and Rothia could be involved in the progression of gastric lesions or carcinogenesis [22,35], while the H. pylori eradication treatment allows the microbiota to be restored [22,35,39]. Our results exemplify those observations, showing that Neisseriaceae, Streptococcaceae, Enterobacteriaceae, and Lactobacillaceae families persist after H. pylori infection, while Gram-positive bacteria such as Gamellaceae, Propionibacteriaceae, Granilucatella, Bacillaceae, and Corynebacterium are excluded even in different states of gastric epithelial injury. Moreover, we showed Neisseria species persistence that were only observed in the atrophic lesions and metaplasia. This decrease in bacterial families also had been reported in gastric biopsies isolates obtained from Chinese dyspeptic patients [9,13,29]. Similarly, Zeng et al. identified N. flavescens as a positive urease agent responsible for gastric inflammation and emphasized the importance of the Neisseria interaction in the epithelium state [29]; however, studies on its implications are still insufficient. In our research, Neisseria was present in erosive tissue and pre-malignant lesions in the HP+ group. Muto et al. stated the carcinogenic capacity of Neisseria when cohabiting with H. pylori [40]. In that sense, Neisseria can produce large amounts of alcohol dehydrogenase, which produces carcinogen acetaldehyde and may also contribute to gastric carcinogenesis [40]. Therefore, Neisseria genus presence should be carefully considered since it was found in advanced lesions of both HP-and HP+ groups, even though Pero et al. indicated that this genus is related to a lower risk of cancer [15]. Beyond the factors above, other studies have pointed out that ancestral population conditions can determine an H. pylori infection s benign or malignant course [41]. Recent studies carried out by our research group reported a higher prevalence of H. pylori infection in a population with Mapuche ancestry, with a 2.3 times greater risk of being colonized by H. pylori [42]. Even though we did not find differences regarding ethnicity, a study focused on Mapuche ethnicity (including a larger number of samples) would be important to investigate the microbiota variability of European and African ancestral H. pylori species in the Mapuche population.
The greatest diversity and number of bacteria were observed in tissues non-colonized by H. pylori (HP-). From 18 bacterial families detected, only eight genera were recognized on HP+. This microbiota behavior can be attributed to the interaction between the bacterial groups [35], whose detection is often limited because many of these species are not cultivable or impossible to isolate [7]. It should be noticed that other factors such as age, diet, metabolic disorders, smoking and drinking habits can influence the microbiota behavior and how those factors modify the H. pylori-associated microbiota should be studied, particularly for age and hypercholesterolemia, which exhibited significant differences in this study.
Moreover, a greater proportion of participants with metabolic diseases and smoking/drinking habits were found in the HP-group-which exhibited the most advanced lesions-suggesting the need for future investigation to relate these parameters with non-H. pylori-associated microbiota, as discussed in recent works [43]. Some metabolic disorders can generate a leptin increase and a reduction of vitamins that have been shown to gastric intestinal metaplasia in mice [7] and the modification of human microbiota [8]. Furthermore, these antecedents might favor changes in the gastric epithelium, including the synergistic interaction between the intestinal epithelium, its microbiota, and the immune system, a condition that is considered an essential factor for proper gastrointestinal function [44,45].
Likewise, in our study, patients with a higher prevalence of those factors showed a tendency to more severe lesions of the gastric epithelium.
For erosive tissue colonized by H. pylori, we only detected N. mucosa, N. perflava, Actinomyces johnsonii, and Enterococcus faecium; however, in pre-malignant lesions, only N. flavescens and N. perflava remained. Streptococcus was detected in all damaged and undamaged tissue stages in the absence of H. pylori, and in tissues without lesion and non-erosive lesion, disappearing in the erosive and advanced lesion in the HP+ group. Liu et al., highlighted Streptococcus as a vital agent present in the chronic process including gastric cancer, independent of H. pylori colonization [21]. Also, Eun et al., detected an increased abundance of the Streptococcaceae family in chronic gastritis and intestinal metaplasia in Asian patients [46]. Here we found that L. rhamnosus, Enterobacteria, Rothia, Gemella, and Neisseria species remained in injured tissue for HP-and while others Lactobacillus species decreased, highlighting the necessity to study these groups in gastric lesions. Other authors have also pointed out a link between Lactobacillus/Streptococcus species and the progression of lesions towards malignancy [46,47]. In our study, more lesions were found in the HP-group; hence, increasing the samples for better casuistry might be necessary to understand the effect of non-H. pylori microbiota on the state of gastric tissue.
The microbiota modifications observed in injured epithelia were mainly associated with an increase in Gram-positive bacteria, specifically from the Bacillaceae, Carnobacteriaceae, Enterococcaceae, Micrococcaceae, and Propionibacteriaceae families. These species thrive in the oral and respiratory microbiota, suggesting a possible and continued propagation from these anatomical regions to the stomach [15,16,34]. In this sense, a study carried out to characterize transient or resident species in the intestinal microbiota by high-throughput sequencing showed that H. pylori influences the oral bacterial community composition, and vice versa [48].
Similar results were also reported for studies carried out with American and Swedish dyspeptic individuals [5,34], which confirmed that an H. pylori infection alters the gastric physiology and would be more likely to affect the gastric microbiota [15,31,32,[49][50][51]. Our work confirmed that few culturable species persisted in the malignant epithelium. In gastric cancer of the Swedish population, a decrease of H. pylori and an increase in other species such as Lactobacillus, Veillonella, Haemophilus, Streptococcus, Provetella, and Neisseriawas detected [2]. Similar results were reported by Wang et al., who studied the Chinese population with gastric cancer, showing that certain bacteria can produce nitrites and generate inflammation, consequently altering the gastric mucosa [6,15]. Therefore, the interaction and contribution of different bacteria species-such as those reported here-in the epithelium state in HP+ and HP-groups need to be evaluated in future studies. Moreover, new insights are required regarding the microbiota impact on H. pylori-associated gastric diseases (e.g., risk of cancer), including its response to treatment and lethality.

Conclusions
The culturable gastric microbiota present in the antrum of dyspeptic patients exhibited significant changes associated with an active H. pylori infection. We detected a different diversity and abundance of microorganisms directly associated with the different states of the gastric lesion, suggesting that this infection may contribute to the presence of gastric injuries.
The description and identification of culturable microbiota could be the first step for future studies regarding the interaction of bacterial species and the gastric epithelial state, providing an important breakthrough for the understanding of gastric pathologies and, through the microbiota characteristics, defining a bacterial profile to be used as a biomarker for detecting risks in developing of gastric epithelial alterations.
Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/2076-2607/8/11/1763/ s1, Table S1: Diversity distribution of bacterial Phyla in Pylori and Non-pylori groups including participants without and/or never go on to an H. pylori eradication treatment, Table S2: Diversity distribution of bacterial Phyla in Pylori and Non-pylori groups including participants with H. pylori eradication treatment as part of Non/pylori group.