The Influence of Modernization and Disease on the Gastric Microbiome of Orang Asli, Myanmars and Modern Malaysians

The present study explored the differences in gastric microbiome between three distinct populations of Southeast Asia. These include the isolated Orang Asli population and modern Malaysians, as well as patients from Myanmar, the least developed country in the region. All 79 subjects recruited in this study had Helicobacter pylori infection. Based on alpha diversity analysis, Orang Asli had the richest and most diverse gastric microbiome, followed by Myanmar and modern Malaysian groups. Beta diversity analysis revealed significant separation of samples between different populations. These observations are likely to be associated with the level of modernization of each population. Our data further suggested increased bacterial species richness and diversity of the gastric microbiome in individuals who were less modernized, particularly in the Orang Asli group, could suppress the growth of H. pylori. In addition, there were significant variations in the gastric microbiome between modern Malaysians with different types of gastric diseases. Notably, Cutibacterium acnes was present at significantly greater abundance level in patients with non-ulcerative dyspepsia than those with peptic-ulcer diagnosis. This suggests that C. acnes may also play a role in gastritis besides H. pylori, which merits further investigation.


Introduction
The human body does not contain only our own cells but is also a host to numerous microorganisms. To illustrate this scenario, Joshua Lederberg coined the term "microbiome" to signify "the ecological community of commensal, symbiotic, and pathogenic microorganisms that literally share our body space" [1]. Up to 60% of human-associated microbes could not be cultured in vitro [2][3][4], leading to substantial underestimation of the actual microbiota diversity in different parts of human body.

Collection of Gastric Biopsies
Of the 79 gastric tissue biopsy samples collected for this study, from September 2011 to June 2014, 10 and 32 were from UBT-positive; Orang Asli volunteers and non-Orang Asli Malaysian patients who were symptomatic and previously tested positive for H. pylori infection, respectively. These individuals underwent endoscopic examination followed by the collection of gastric biopsies at UMMC's Endoscopy Unit. The remaining 37 samples were collected from Myanmar patients at the Department of Gastroenterology of Thingangyun Sanpya General Hospital, Yangon, Myanmar. All subjects were further diagnosed with gastric cancer (GC), peptic ulcer disease (PUD) or non-ulcerative dyspepsia (NUD) based on endoscopy evaluation.
Biopsy specimens were obtained from the gastric antrum and fundus for metagenomics purposes, CLOtest rapid urease testing, H. pylori culturing, and histopathological examination. Culturing of H. pylori isolate was not performed on Myanmar patient samples in this study. Histopathological examination was performed independently by Dr. Sri Ganesh Kalimuthu of the CliniPath Malaysia Sdn Bhd.

Illumina 16S rRNA Gene Sequencing
Genomic DNA was extracted from gastric biopsy samples using MasterPure TM DNA purification kit (Epicentre, Madison, WI, USA) according to manufacturer's instructions. The V3-V4 region of 16S rRNA gene was amplified using S-D-Bact-0341-b-S-17 and S-D-Bact-0785-a-A-21 primers designed to include the Illumina-compatible adaptors [17,18]. The 16S amplicon libraries were prepared according to Illumina 16S library preparation protocol [19]. Initial PCR amplification was performed using NEBNext High-Fidelity Master Mix (New England Biolabs, Ipswich, MA, USA) with the following conditions: An initial denaturation at 98 • C for 30 s, followed by 30 cycles consisting of denaturation (98 • C for 10 s), annealing (60 • C for 2 min) and extension (72 • C for 20 s), and a final extension step at 72 • C for 1 min. Automated cluster generation and a 2 × 250 bp paired-end sequencing (MiSeq 500-cycle reagent kit V2) was carried out on the MiSeq platform (Illumina, San Diego, CA, USA) at the Monash University Malaysia Genomics Facility. Metagenomics data reported in this paper are available for public access through the MG-RAST server. The accession number and demographic characteristics of each sample are listed in Table S1.

Data Analysis
The 16S sequencing data were quality-trimmed using Sickle (https://github.com/najoshi/sickle) using the following parameters: -q 20 -l 200. Merging of overlapping paired-end sequences was performed using MeFit with default parameters [20]. Filtering of chimeric sequences, de novo greedy clustering of 16S rRNA sequences into Operational Taxonomic Units (OTUs) at 97% similarity threshold, and removal of singleton and chimeric OTUs were conducted using Micca (version 1.7.0) [21]. Taxonomic assignment of the representative OTUs was performed using the Bayesian LCA-based taxonomic classification method with a 1 −100 cut-off e-value and 100 bootstrap replications, against NCBI 16S microbial database [22]. Taxonomic assignment at each level was accepted with a minimum confidence score of 80. The OTU table including taxonomic information is available in Table S2. Unclassified OTUs were retained and included in all following analyses.
Microbial diversity analysis was performed on rarefied OTU abundance matrix (depth value of 17358) using QIIME (version 1.9.1) [23]. Alpha diversity was evaluated based on the following metrics: Observed species and Shannon diversity indices. A non-parametric two-sample t-test with Bonferroni correction was used to compare the alpha diversity metrics between different groups. Principle coordinate analysis (PcoA) using Bray-Curtis and Jaccard distance metrics was performed to visualize separation of samples. Non-parametric statistical analysis of the distance metric was performed using ANOSIM with 1000 permutations. Plots were generated using PhyloToAST software [24]. Prediction of KEGG functional pathways was conducted using Piphillin and the results are available in Table S3 [25,26]. Pearson correlations between KEGG pathways and bacterial genera were examined using microbiomeSeq R package (https://github.com/umerijaz/microbiomeSeq.git). Only negative and positive correlations of ≤−0.8 and ≥0.8, respectively, and with p-values less than 0.01, were considered.

Statistical Analysis
Bacterial genera and KEGG functional pathways with minimum average relative abundances of 0.1% were compared between groups by Wilcoxon rank-sum testing, followed by Bonferroni correction. Differences with adjusted p-values of less than or equal to 0.01 were considered significant.

H. pylori Prevalence among Orang Asli
Among 177 Orang Asli subjects who were screened for H. pylori infection, only 18 (10.2%) and 53 (30.3%) were tested positive and negative, respectively. The rest, interestingly, had indeterminate UBT readings. The demographics and UBT results of the Orang Asli subjects in this study are summarized in Table 1. Ten UBT-positive Orang Asli individuals consented for further endoscopy examination and provided gastric biopsies for this study. The general characteristics and clinical diagnosis of study subjects who underwent endoscopic examination are summarized in Table 2. A total of 22,526,214 reads were generated for all 79 samples. Following quality trimming and merging of overlapping paired-end sequences, 9,912,950 sequences with an average length of 454.3 ± 8.1 bp were retained. The number of sequences per sample ranged from 20,042 to 233,069 (median 115,971). Of 2233 OTUs acquired via de novo clustering algorithm, 1466 and 1105 were taxonomically assigned down to and phylum and genus level, respectively, with an 80% confidence threshold, which were subsequently aggregated to 20 bacterial phyla and 443 bacterial genera, respectively, for further analysis. As our study subjects differed not only by population but also by disease status, a comparison of those with only NUD was therefore performed to exclude the influence of PUD and GC when examining the microbial differences between three major populations including Orang Asli, modern Malaysians comprising Chinese, Indian and Malay ethnic groups, and Myanmar subjects. Overall, Proteobacteria, Firmicute and Actinobacteria constituted the three most predominant phyla, at 71.9%, 10.4% and 9.1% of relative abundances ( Figure 1A), respectively, in the gastric microbiome of modern Malaysian patients. Similarly, in the Orang Asli samples, while having both Proteobacteria and Actinobacteria as the first and third most abundant phyla, at 57.3% and 5.6%, respectively, Bacteroidetes, not Firmicute, was the second most common bacterial phylum (13.4%) identified. In Myanmar patients, the top three most abundant phyla were Firmicutes, Proteobacteria and Bacteroidetes, at 39%, 28.1% and 15.5%, respectively.
At the genus level, Helicobacter was the most abundant bacterial genus identified in the modern Malaysians, with a relative abundance of 61% ( Figure 1B). In both Myanmar and Orang Asli patients, however, the Helicobacter bacterial loads were only 9.8% and 0.25%, respectively. While a high abundance of Lactobacillus (7.2%) was detected in Myanmar patients, Orang Asli group contained mostly environmental bacterial genera including Flavobacterium, Sulfuritalea, Desulfatiglans, Brevibacillus and Massilia.
We next analyzed the diversity of bacterial populations amongst Orang Asli, Myanmar and modern Malaysian samples. Alpha diversity based on species richness and Shannon diversity indices was estimated. Both modern Malaysian and Myanmar groups had significantly reduced microbial richness ( Figure 2A) and diversity ( Figure 2B) relative to Orang Asli. It is worth mentioning that the gastric microbial community of modern Malaysians was the least diverse among all. To assess bacterial community differences between groups, principal coordinate analysis (PCoA) based on Jaccard metric was performed. The analysis revealed significant dissimilarities in microbial composition among the groups ( Figure 3, R = 0.75, p < 0.001).
To identify distinguishing bacterial genera between Orang Asli, Myanmar and modern Malaysian NUD samples, 76 genera with minimum average abundances of 0.1% were subjected to Wilcoxon rank-sum testing followed by Bonferroni correction. Of 28 significant genera identified, as summarized in Table 3, 17 were profoundly enriched in the Orang Asli gastric microbiome compared to that of Myanmar and modern Malaysian samples. While Nocardioides was nearly absent in the Myanmar samples, modern Malaysians had the highest Helicobacter bacterial load, significantly greater than that of Orang Asli and Myanmar cohorts.         Functional prediction using Piphillin revealed 306 KEGG pathways (Table S3). Following rarefaction and filtering non-NUD samples, we then compared between different populations the pathways with total average abundances of at least 0.1% and considered only those with a minimum relative fold change of 1.25 and a Bonferroni-adjusted p-value of less than or equal to 0.01 as statistically significant. Seventy-six pathways were significant (Table S4). Notably more than one third of pathways (28/76) were significantly up-or down-regulated in Orang Asli samples in comparison to both Myanmar and modern Malaysian cohorts. Of 28 significantly differential pathways identified in Orang Asli, half were involved in metabolism of amino acids and organic compounds like ascorbate and aldarate, histidine, tryptophan, inositol phosphate, arginine and proline, phenylalanine, tyrosine and beta-alanine, which had positive significant association with Bosea, Brevundimonas, Emticicia, Sphingorhabdus, Tardiphaga and Variovorax; six environmental bacterial genera that were found present almost exclusively in the Orang Asli population (Table S5). Compared to both Orang Asli and Myanmar cohorts, significant up-regulation of FoxO signaling was observed in the modern Malaysian samples and was shown to be positively and negatively associated with Helicobacter and Streptococcus, respectively.

Gastric Diseases Altered the Microbial Community Structure in Modern Malaysian Patients
We attempted additional diversity analysis to determine whether different gastric disease outcomes would affect the structural composition of gastric microbial community. The analysis was performed separately for each population to eliminate the confounding effect of population factor on disease status. Orang Asli group was excluded as NUD was the only gastric disorder diagnosed in this population. While clear separations between NUD, PUD and GC samples were observed in modern Malaysian samples (Figure 4, Jaccard index, R = 0.4, p = 0.002), interestingly, no evident grouping of Myanmar samples according to disease types could be observed.

237
Functional prediction using Piphillin revealed 306 KEGG pathways (Table S3). Following 238 rarefaction and filtering non-NUD samples, we then compared between different populations the 239 pathways with total average abundances of at least 0.1% and considered only those with a minimum 240 relative fold change of 1.25 and a Bonferroni-adjusted p-value of less than or equal to 0.01 as 241 statistically significant. Seventy-six pathways were significant (Table S4). Notably more than one  (Table S5).   Both NUD and PUD samples shared similar major bacterial phyla compositions with Proteobacteria and Firmicutes being the two most predominant phyla, at 72.1% and 10.1% in the former and 85.6% and 7.4% in the latter ( Figure 5). The GC samples, however, had comparable abundances of both Proteobacteria and Firmicutes, at 33.2% and 36.3%, respectively. At genus level, while Helicobacter bacteria constituted 60.5% and 74.4% of gastric microbiome in both NUD and PUD groups, respectively, in GC group the total abundance of top five most predominant genera including Helicobacter was 37.8%, indicating that GC samples had a more diverse and complex gastric bacterial population.
former and 85.6% and 7.4% in the latter ( Figure 5). The GC samples, however, had comparable abundances of both Proteobacteria and Firmicutes, at 33.2% and 36.3%, respectively. At genus level, while Helicobacter bacteria constituted 60.5% and 74.4% of gastric microbiome in both NUD and PUD 268 groups, respectively, in GC group the total abundance of top five most predominant genera including 269 Helicobacter was 37.8%, indicating that GC samples had a more diverse and complex gastric bacterial   Due to small GC sample size (n = 4), comparison of bacterial genera abundances by Wilcoxon ranked-sum testing was performed only between NUD and PUD groups. Following Bonferroni correction, of 55 genera tested, only one appeared significant, which is Cutibacterium (adjusted p = 0.005, median abundances of 0.159% and 0.002% for NUD and PUD, respectively).

Discussion
In this study, the gastric microbial compositions of three distinct populations, Orang Asli, Myanmars and modern Malaysians, who all had H. pylori infection based on positive UBT result, were investigated. It is important to note that the median abundance of Helicobacter bacteria in Orang Asli was merely 0.25%, which is several orders of magnitude less than that of modern Malaysian and Myanmar individuals. This is likely due to the significantly more diverse bacterial makeup in the gastric environment of Orang Asli compared to that of modern Malaysians and Myanmars. Although the human stomach serves as the primary reservoir for H. pylori colonization, the findings above indicate that the degree of colonization can be affected by the changes in the gastric microenvironment. It is evident that the presence of environmental bacterial species in the stomach of Orang Asli, as well as in the Myanmar individuals to a certain extent but less, could suppress the growth of H. pylori, while the lack of these foreign environmental isolates in modern Malaysians allows H. pylori to flourish. This finding could possibly explain why most Orang Asli individuals had a borderline UBT outcome, presuming most of them are likely to be H. pylori-positive.
In both Orang Asli and modern Malaysian groups, Proteobacteria was found to be the predominant bacterial phylum. While Proteobacteria in modern Malaysians contained primarily Helicobacter bacteria belonging to the class Epsilon Proteobacteria, Orang Asli harbored mostly environmental microbes of Alpha-, Beta-, Delta-and Gamma-proteobacteria classes. Notably, the Orang Asli gastric microbiome had a Firmicutes/Bacteroidetes ratio of 0.41, which is at least 6-fold less than that of Myanmar and modern Malaysian groups, at 2.51 and 2.65, respectively. It was previously shown that a high-fat/high-sugar diet promotes the growth of Firmicutes while reducing the abundance of Bacteroidetes in the distal gut of rodents [27]. Additionally, when an obese individual shifted to a fat-restricted diet or a carbohydrate-restricted low-calorie diet, the abundance levels of Bacteroidetes and Firmicutes were increased and decreased, respectively [28]. Together, these observations suggest that both modern Malaysian and Myanmar populations are likely to consume a diet high in carbohydrate and/or fat content possibly attributed to influx of Western culture as part of modernization process, while Orang Asli are still maintaining traditional plant-based diets low in fat and carbohydrate corresponding to their hunting/planting lifestyle [29].
Of the several bacterial genera that were significantly enriched in Orang Asli, Bosea, Sphingorhabdus, and Variovorax, they also displayed significant positive correlations to arginine and proline metabolism, and β-alanine metabolism, respectively. While a diet supplemented with arginine and proline was shown to improve the wound healing process in diabetic rats [30], β-alanine supplementation as reported in a recent meta-analysis was able to improve exercise performance [31]. It is thought that these bacteria may facilitate the metabolism and uptake of these amino acids that are widely available in vegetables and grains consumed by Orang Asli, which is then likely to increase the wound healing and physical capacities in Orang Asli to enhance their survival fitness in an outback setting.
While bacteria of the genus Nocardioides was found in nearly all modern Malaysian (95%, 19/20) and Orang Asli (100%, 10/10) samples, it was detected in only 23.1% (3/13) of Myanmar samples. Members of the genus Nocardioides are mostly environmental bacteria found in soil or plants and only one Nocardioides species has thus far been isolated from a human faecal sample [32]. The high prevalence of Nocardioides species could be a distinct feature of gastric microbiome in peninsular Malaysians including Orang Asli, albeit the influence of this Nocardioides species on human health is uncertain.
Contrary to modern Malaysian samples, the Myanmar population did not show distinct clustering of samples by gastric disease conditions. It is thought that lifestyle and dietary habits, geographical and socioeconomic factors, and/or host genetics may have greater influence on the gastric microbiome than disease conditions in the Myanmar population. In modern Malaysians, PUD samples harbored the most H. pylori, which is expected as this bacterium is an important driving factor for peptic ulcer disease outcome [33]. Notably, NUD samples had significantly more Cutibacterium bacteria than in the PUD samples. At the species level, it was further identified as C. acnes, formerly known as Propionibacterium acnes [34]. Interestingly, while H. pylori has been known for its dominant role in causing different gastric disorders, C. acnes was identified as a possible causative agent for lymphocytic gastritis [35]. It is important to clarify that NUD, also known as functional dyspepsia, is the preferred clinical diagnosis term assigned to patients with gastritis, on the basis of endoscopic and histological results [36]. To examine the potential role of C. acnes in the development of gastric diseases alongside H. pylori, further investigation is required. This would help to expand our current knowledge of bacterium-driven gastric diseases.

Conclusions
This study highlighted the taxonomic differences between the gastric microbiome of three distinct populations from Southeast Asia including Orang Asli, Myanmars and modern Malaysia, and such differences are possibly due to different levels of modernization. Distinguishing gastric microbial community structure could also be seen between different gastric disease conditions in modern Malaysians, albeit not in Myanmar patients. Importantly, the gastric microbiome of Orang Asli, who lived in isolation from urbanization, and a near primitive lifestyle with no use of any modern antibiotics and few modern medicines, may represent one of the most "original" snapshots of human gastric microbiome to-date.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-2607/7/6/174/s1, Table S1, Demographics data and accession numbers; Table S2, OTU table with taxonomic information; Table S3, KEGG pathway abundances predicted based on the gastric microbiome of Orang Asli, Myanmar and modern Malaysian groups; Table S4, Distinguishing KEGG pathways between Orang Asli, Myanmar and modern Malaysians groups; Table S5, Pearson correlation analysis between KEGG pathways and bacterial genera of Orang Asli, Myanmar and modern Malaysian groups.