A Human and Animal Based Study Reveals That a Traditionally Fermented Rice Beverage Alters Gut Microbiota and Fecal Metabolites for Better Gut Health

Abstract

Fermented rice beverages are consumed globally, especially in Southeast Asia. In India, such beverages are consumed by a substantial population of ethnic communities. In this study, the gut bacterial diversity of rice beverage drinkers from Assam, India (n = 27) was compared with that of nondrinkers (n = 21) with the next-generation sequencing (NGS) of fecal metagenomic 16S rDNA, which indicated changes in 20 bacterial genera. Further, mice (n = 6, per treatment group) were gavaged daily for 30 days with different fractions of the beverage, which included rice beverage (RB), soluble (SF), and insoluble fractions (IF) to determine the effects of different components of the beverage. A comparison of gut bacteria at two time points, 0 and 30 days of treatments, suggested changes in 48 bacterial genera across the different treatment groups in mice. Major bacterial changes were suggestive of functional components associated with gut health, as observed in both humans and mice. Next, the Gas Chromatography–Mass Spectrometry (GC–MS) of mice stool after 30 days of treatments showed a total of 68 metabolites, of which hexadecanoic acid, a flavor component of this beverage, was present in the feces of all mouse treatment groups except controls. These metabolites showed treatmentwise clustering in groups in a partial least-squares discriminant analysis (PLS–DA) plot. Blood endotoxin levels were lower in all treatment groups in the mice compared to those of the controls. The findings of the study are suggestive of the gut modulatory effects of the beverage on the basis of the observed features of the bacterial changes.

1. Introduction

Microbes residing in the intestine, collectively called the ‘gut microbiota’, influence the health and well being of the host. Some important functions of the gut microbiota include metabolic functions and energy salvage, the production of short-chain fatty acids (SCFAs) and vitamins, strengthening the immune system, and preventing colonization from pathogens [1]. Microbes colonize the host gradually after birth through a complex process of host–microbe interactions [2]. The disturbance of microbiota composition leads to a condition called dysbiosis that implicates diseased conditions such as allergies, inflammatory bowel disease, obesity, diabetes, and cancer [3]. Various factors, such as diet, genetics, ethnicity, and environmental conditions, affect the microbiota and its functions, of which diet plays the dominant role [4]. In recent years, extensive studies have unveiled an understanding of the relationships between the gut microbiota and diet [5]. There are studies reporting the modulation of the human gut microbiota and immunity via dietary interventions that included fermented foods [6]. Interestingly, fermented foods and beverages are gaining global popularity not only because of their organoleptic properties, but also for their health benefits. These foods confer several benefits to human health and modulate the composition of the host’s gut microbiota [7]. Derrien and van Hylckama (2015) described three possible mechanisms by which the health-beneficial microbes, when ingested, affect the host: (1) trophic interactions (e.g., production of SCFA), (2) the direct inhibition of pathogens (by producing bacteriocins, lowering the pH, etc.), and (3) indirectly, via the production of host-derived molecules (IgA, mucins, etc.) [8]. The preparation and consumption of different fermented foods are practised as a culinary art by many ethnic communities of Northeast India [9]. Tiwa is an eminent community of Northeast India residing mainly in the Morigaon, Kamrup, and Karbi Anglong districts of Assam. The diet of Tiwas mostly includes rice, along with several indigenous herbs and meat products. The edible plants and wild vegetation are often utilised for the preparation of several food products, and preserved by drying or fermentation. The preparation of the rice beverage by the Tiwa community utilises their traditional knowledge of the use of medicinal plants and the fermentation process. The fermentation process is carried out using a traditional starter cake that is prepared using different plants and serves as an inoculum. The starter consists of amylolytic and alcohol-fermenting microbes, and lactic acid bacteria that act synchronously upon the rice substrate, resulting in an alcoholic beverage. During the fermentation process, rice starch is broken down into sugars that are further converted into alcohol, acids, and other metabolites through the action of different microbes. Biochemical analyses of the beverage detected the presence of microbe-originated organic acids such as lactic acid, amino acids, and prebiotic sugars such as raffinose and trehalose [10]. Antimicrobial activities against common pathogens and antioxidant properties based on free radical scavenging were also studied by different groups [11,12]. Besides regular consumption, it is traditionally acclaimed for having health-modulating effects for insomnia, headaches, body aches, inflammation, diarrhea, and urinary problems [13]. Previously, we had reported the presence of lactic acid bacteria and prebiotics in this beverage [14,15]. A very recent review summarised the microbial diversity of the beverage, including probiotics along with metabolites conferring prebiotic attributes. The beverage has been reported as “symbiotic” along with its immunity-enhancing properties [16]. The present study was conducted to understand the effect of the beverage consumed by the Tiwa community on the gut microbes of humans, followed by a study of other parameters such as metabolites and intestinal morphology in mice using the high-throughput NGS technique and GC–MS analyses. The use of the mouse model allowed for controlled interventions that were difficult in humans because of confounding factors. It was hypothesised that the consumption of the rice beverage might be associated with changes in the gut microbiota and other health parameters. Accordingly, gut bacteria in humans were compared between drinker and nondrinker volunteers of the Tiwa community. Mice were treated with different fractions of the rice beverage to determine the distinct effects. To the best of our knowledge, the present study is the first to report on effects of the traditionally fermented rice beverage of Northeast India on the gut microbiota through studies on human and mouse models.

2. Materials and Methods

2.1. Recruitment of Human Subjects and Collection of Samples

Before the study, a survey was conducted in villages where the Tiwa community is dominant and for their willingness to participate in the study. The volunteers were recruited using written informed consent and a bilingual (English and Assamese) questionnaire, from the Morigaon district of Assam, India. Inclusion criteria for the study were ethnicity, age (18–60 years), and the consumption of the beverage; those taking antibiotics or other drugs were excluded. A total of 48 participants, both males and females, belonging to the Tiwa community were selected on the basis of their rice beverage drinking habits and classified as drinkers or nondrinkers. Height, weight, pulse rate, blood group, and blood pressure were determined during their visit to the sample collection site. The stool samples were collected in sterile containers with a spoon attachment (HIMEDIA, Mumbai, India). Then, 3 mL of blood was withdrawn in vacutainer collection tubes (BD, New Jersey, USA). The samples were brought to the laboratory in frozen conditions. The biochemical parameters of blood, namely, gamma-glutamyltransferase GGT), albumin, globulin, total protein, and blood sugar, were determined using standard kits (Tulip Diagnostics, India) following the manufacturer’s instructions. The human study was approved by the institutional ethics committee (approval no. IEC(HS)/IASST/1082/2014-15/6).

2.2. Rice Fermentation

Rice was fermented following the traditional knowledge of the Tiwa community of Assam, India. The starter cake was prepared in the laboratory as described in the patent filed earlier [17]. One kg of rice was boiled to cook and allowed to cool-down to room temperature. A 30 g of the starter cake was crushed to produce a homogeneous powder, mixed with the rice, and kept for fermentation at 30 ${}^{\circ }$C. After fermentation, the mass was filtered, and the liquid (RB) was harvested and centrifuged at 5000 rpm for 10 min to separate the supernatant containing the soluble fraction (SF); the pellet was resuspended in sterile water to the initial volume, denoted as the insoluble fraction (IF). The biochemical parameters of the beverage were discussed in our earlier report [14].

2.3. Mouse Experiment

Five-week-old male Swiss albino mice (n = 6 per group) were used for this study. The animals were housed in polypropylene cages (cat. no. 041000 Tarsons, India), and a standard rodent chow diet, drinking water, and a 12 h day/night cycle were maintained throughout the experiment. Dietary intakes per cage and body weight per animal were monitored during the experimental period. After a 7-day acclimatization period, the animals were divided randomly into four treatments: control (CT), rice beverage (RB), insoluble fraction (IF), and soluble fraction (SF). A total of 200–230 $\mathsf{\mu }$L of RB, IF, and SF were administered, whereas the controls (CT) received sterile water using gavage. The dosage was equivalent to the reported human intake of the beverage based on their body weight to produce a consistent gavage volume [10]. The treatments were performed once daily in the morning. fecal samples of individual animals were collected on the 0th and 30th days of treatments, and immediately stored at −80 ${}^{\circ }$C. On the 30th day, animals from each group were sacrificed, and the intestines were collected for histology. All experiments were conducted following the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Ministry of the Environment, Forests, and Climate Change, Government of India, and approved by the institutional animal ethics committee (approval no. IASST/IAEC/2016-17/07).

2.4. Isolation of Fecal Metagenomic DNA and Next-Generation Sequencing (NGS) Analyses

Metagenomic DNA was extracted from the fecal samples of humans and three animals from each treatment group using a QIAGEN DNA Stool Mini-Kit (QIAGEN, Hilden, Germany) following the instructions, and subjected to NGS analysis with sequencing service provider Macrogen Inc. (Seoul, Republic of Korea). The PCR amplification of metagenomic DNA was performed using 341F-805R primer pairs spanning the V3–V4 (Bakt 341F, CCTACGGGNGGCWGCAG Bakt 805R, GACTACHVGGGTATCTAATCC ) region of the 16S rDNA gene. The NGS library was prepared using the Nextera XT library preparation kit according to the Illumina MiSeq protocol (Illumina Inc., California, USA, 2017). Sequencing was carried out on an Illumina MiSeq machine (MiSEq 2500) following 2 × 300 bp paired-end chemistry with the multiplexed pooled samples. The NGS data were analysed using the phyloseq (version 1.30.0) package in RStudio with R (version 3.6.3). Analyses of read quality, the filtering and trimming of sequences, the calculation of error rates, sample inference, and the merging and removal of chimera were performed with reference-free divisive amplicon denoising algorithm 2 (Dada2, version 1.16.0) [18]. Sequences having a minimal Phred score of 25 were considered for analysis, and forward and reverse sequences were truncated at 280 and 250 bp, respectively, for a proper merging. A distinct ASV table was constructed, and bacterial taxonomy was assigned from the Silva database (release 138) at 99% homology [19]. The diversity of the bacterial communities was established and statistical analyses were performed using the vegan package (2.6-2); results were visualised using the ggplot2 (version 3.3.0) and ggpubr (version 0.4.0) packages [20,21]. Alpha diversity, a measure of species richness and evenness within a single sample, was used to compare the taxon diversity between the fecal samples. Beta diversity was estimated to measure changes in the compositions of the bacterial communities. Jaccard and Bray–Curtis beta diversity indices were used for qualitative and quantitative measurements, respectively, of community differences.

2.5. Nontargeted Fecal Metabolite Profiling of Mice Feces via Gas Chromatography–Mass Spectrometry (GC–MS)

Fecal sample preparations and the GC–MS run program were followed as per the protocol described elsewhere, with slight modifications [22]. Briefly, 10 mg of fecal sample from each animal was extracted with 50 $\mathsf{\mu }$L of 3:2:2 isopropanol:acetonitrile:water and centrifuged at 14,000 rpm for 5 min. A 25 $\mathsf{\mu }$L of methoxyamine hydrochloride, dissolved in pyridine (15 mg/mL), was added to the extract and incubated at 60 ${}^{\circ }$C for 45 min. A 50 $\mathsf{\mu }$L of N-methyl-N-trimethylsilyltrifluoroacetamide with 1% trimethylchlorosilane was added, and samples were further incubated at 60 ${}^{\circ }$C for 30 min, centrifuged at 3000 rpm for 5 min, and brought to room temperature before injecting. Analysis was carried out on a Shimadzu GC 2010 Plus-triple Quadrupole (TP-8030) GC–MS/MS system fitted with an EB-5MS column (length 30 m, thickness 0.25 $\mathsf{\mu }$m and ID, 0.25 mm). The programme consisted of 80 ${}^{\circ }$C for 30 s, a ramp of 15 ${}^{\circ }$C/min to 330 ${}^{\circ }$C, and an 8 min hold. Masses in the range of 50–650 m/z were scanned at 5 scans/s after electron impact ionisation. The mass spectra were matched in the National Institute of Standards and Technology library, USA. Noisy peaks and column bleeds were removed before data analysis. Partial least-squares discriminant analysis of the fecal metabolites on the 30th day of treatment was performed using Metaboanalyst [23].

2.6. Assay of Plasma Endotoxin

Endotoxin concentration in the mouse plasma was determined with Pierce LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fisher Scientific) following the manufacturer’s instructions. A standard linear graph was prepared using Escherichia coli endotoxin within the range of 0.1–1.0 EU/mL. A 100 $\mathsf{\mu }$L of the plasma sample was used for assay, and the endotoxin concentration is expressed as EU/mL.

2.7. Histology of Ileum

About 2–3 cm of the distal part of the ileum was isolated and fixed in Carnoy’s solution (60% ethanol, 30% chloroform, and 10% glacial acetic acid). After 2 h of fixation, tissues were dehydrated in graded alcohol (50–100%) and lastly cleared in xylene. The clear tissues were embedded in paraffin and cut into 3–5 $\mathsf{\mu }$m thickness using a microtome (Leica RM 2235, Wetzlar, Germany). The tissue sections were stained with haematoxylin and eosin (HE), the villus length was measured from the base to the top of the villus, and the crypt depth was measured between the crypt–villus junction and the base of the crypt [24]. The histological sections were examined using a light microscope (ZEISS Axio Lab.A1, Carl Zeiss Microscopy GmbH, Oberkochen Germany) and ZEISS Zen Microscopy software for imaging without knowledge of the origin of the tissue samples.

2.8. Statistical Analyses

The relative abundance data of the selected bacterial genera at 0th and 30th days of treatment were compared using the Mann–Whitney U test within SPSS (IBM SPSS 20, SPSS Inc., Chicago, IL, USA). The Kruskal–Wallis H test was performed to determine the differences in diversity indices. One-way ANOVA was performed to determine the differences in plasma endotoxin levels within SigmaPlot (Systat Software Inc., San Jose, CA, USA).

3. Results

3.1. Dietary and Anthropometric Features of the Study Subjects

A total of 48 human volunteers belonging to the Tiwa ethnic community of Assam participated in the study. All of them followed a nonvegetarian diet. In total, 27 participants consumed the beverage, referred to as drinkers (D), and 21 did not consume it, referred to as nondrinkers (ND). The details of health and blood parameters are listed in Supplementary Table S1. In mice, the intake of different fractions of the rice beverage during the treatment period did not interfere with the diet consumption. Therefore, no significant changes were observed in food consumption (Supplementary Figure S1).

3.2. Gut Bacterial Diversity of Humans and Mice via NGS

A total of 234 and 239 individual amplicon sequence variants (ASVs) were observed depicting the bacterial communities at different taxonomic levels in human and mouse samples, respectively (Supplementary Figure S2). In both subject groups, Prevotella was an abundant genus (Figure 1). The Shannon and Simpson indices of alpha diversity showed significant differences (p ≤ 0.05) on the 0th and 30th days of treatment (Figure 2). However, no significant differences were observed in human (Supplementary Figure S3) and mouse samples (Supplementary Figure S4), compared on the basis of treatments. For the human samples, principal coordinates analysis on the beta diversity indices showed subtle dissimilarities based on drinking habit with a variance of 5.1% (Axix 1) and 5.2% (Axis 2), for Jaccard, and 8.6% (Axix 1) and 7.6% (Axis 2) Bray–Curtis. Similarly, in mouse samples, the distance matrices showed a variance of 7.1% (Axix 1) and 7.6% (Axis 2), for Jaccard, and 9.1% (Axix 1) and 7.2% (Axis 2) Bray–Curtis (Figure 3).

A comparison between drinkers and nondrinkers revealed changes in 20 bacterial genera in humans (

Table 1. Comparison between gut bacteria in drinkers and nondrinkers of the rice beverage in the Tiwa community. Upward arrows indicate an increase in drinkers vs. nondrinkers and vice versa.

Genus	Inference	p Values
Odoribacter	↑	0.001
Christensenellaceae R-7	↓	0.001
Coprococcus	↑	0.002
Defluviitaleaceae UCG-011	↑	0.004
Oscillospiraceae UCG-005	↑	0.004
Rikenellaceae RC9	↑	0.006
Succinivibrio	↑	0.007
Aeromonas	↓	0.007
[Eubacterium] siraeum	↑	0.01
Oscillospiraceae UCG-002	↑	0.013
Ruminococcus	↑	0.016
Lachnospiraceae NK4A136	↑	0.017
Lachnospiraceae NC2004	↑	0.042
Butyrivibrio	↑	0.023
Butyricimonas	↑	0.027
Lachnoclostridium	↓	0.03
Bilophila	↑	0.038
Rothia	↓	0.04
Senegalimassilia	↑	0.045
Anaerovoracaceae Family XIII UCG-001	↑	0.047

). An increased abundance of bacterial genera: Odoribacter, Coprococcus, Defluviitaleaceae UCG-011, Oscillospiraceae UCG-005, Rikenellaceae RC9, Succinivibrio, Eubacterium siraeum, Oscillospiraceae UCG-002, Ruminococcus, Lachnospiraceae NK4A136, Lachnospiraceae NC2004, Butyrivibrio, Butyricimonas, Bilophila, Senegalimassilia and Anaerovoracaceae. Family XIII UCG-001 increased, whereas a decrease in Christensenellaceae R-7, Aeromonas, Lachnoclostridium and Rothia was observed.

In mice, a maximal shift was observed in the abundance of 16 and 15 bacterial genera in the RB and SF groups, followed by 12 in the CT group, and IF showed a significant difference in 4 bacterial genera between the 0th and 30th days of treatments with different fractions of the beverage, as determined with the Mann–Whitney U test (p ≤ 0.05) (

Table 2. Changes in gut bacteria in mice after 30 days of treatment with different fractions of the beverage. Upward arrows indicate increase after 30 days of treatment and vice versa.

Treatment	Bacteria	Inference	p Values
CT	Eubacterium ventriosum	↑	0.037
	Lachnospiraceae A2	↓	0.037
	Cuneatibacter	↓	0.037
	Leptotrichia	↓	0.046
	Negativibacillus	↓	0.046
	Mesocricetibacter	↓	0.046
	Rodentibacter	↓	0.046
	Ruminococcus	↑	0.046
	Candidatus Arthromitus	↓	0.050
	Intestinimonas	↓	0.050
	Lachnospiraceae NK4A136	↑	0.050
	Odoribacter	↓	0.050
RB	Eubacterium siraeum	↑	0.037
	Anaerovoracaceae Family XIII UCG-001	↓	0.037
	Ruminococcaceae UBA1819	↑	0.037
	Eubacterium nodatum	↓	0.046
	Lachnospiraceae ASF356	↓	0.046
	Rodentibacter	↓	0.046
	Anaerovoracaceae Family XIII AD3011	↓	0.046
	Streptococcus	↓	0.046
	Negativibacillus	↓	0.046
	Roseburia	↓	0.050
	Acetatifactor	↑	0.050
	Bacteroides	↓	0.050
	Ligilactobacillus	↓	0.050
	Anaerotruncus	↓	0.050
	Erysipelatoclostridium	↓	0.050
	Prevotellaceae UCG-001	↑	0.050
IF	Lachnospiraceae ASF356	↑	0.000
	Paludicola	↑	0.000
	Anaerovoracaceae Family XIII AD3011	↓	0.001
	Parasporobacterium	↓	0.001
SF	Wautersiella	↑	0.037
	Leptotrichia	↓	0.046
	Candidatus Stoquefichus	↓	0.046
	Lachnospiraceae UCG-004	↓	0.050
	Lachnospiraceae ASF356	↑	0.050
	Rikenella	↑	0.050
	Desulfovibrio	↑	0.050
	Lactobacillaceae HT002	↓	0.050
	Intestinimonas	↑	0.050
	Bacteroides	↓	0.050
	Lactobacillus	↓	0.050
	Ligilactobacillus	↑	0.050
	Eubacterium xylanophilum	↓	0.050
	Prevotellaceae UCG-001	↑	0.050
	Ruminococcus	↑	0.050

). In CT, 3 bacterial genera, namely, Eubacterium ventriosum, Ruminococcus and Lachnospiraceae NK4A136 increased, whereas others including Lachnospiraceae A2, Cuneatibacter, Leptotrichia, Negativibacillus, Mesocricetibacter, Rodentibacter, Candidatus Arthromitus, Intestinimonas and Odoribacter decreased significantly. In RB, Eubacterium siraeum, Ruminococcaceae UBA1819, Acetatifactor, and Prevotellaceae UCG-001 increased significantly, whereas a decrease in Anaerovoracaceae Family XIII UCG-001, Eubacterium nodatum, Lachnospiraceae ASF356, Rodentibacter, Anaerovoracaceae Family XIII AD3011, Streptococcus, Negativibacillus, Roseburia, Acetatifactor, Bacteroides, Ligilactobacillus, Anaerotruncus, Erysipelatoclostridium, was observed. In the IF group, an increase in 2 genera, namely, Lachnospiraceae ASF356, and Paludicola and a decrease in Anaerovoracaceae Family XIII AD3011 and Parasporobacterium, were observed. Next, in the SF, Wautersiella, Ruminococcaceae ASF356, Rikenella, Desulfovibrio, Intestinimonas, Ligilactobacillus, Ruminococcus increased, and Leptotrichia, Candidatus Stoquefichus, Lachnospiraceae UCG-004, Lactobacillaceae HT002, Bacteroides, Lactobacillus and Eubacterium xylanophilum decreased significantly after 30 days of treatment (p ≤ 0.05).

Considering the microbes that changed exclusively due to RB and its two fractions, IF and SF, Lachnospiraceae genus ASF356 was common in all the treatment groups except the controls. The abundance of this genus increased in RB, but decreased in IF and SF. In a similar trend, two other genera showed such differences in RB and SF treatments. Ligilactobacillus decreased in RB and increased in SF, and Prevotellaceae UCG-001 increased in RB and decreased in SF. We speculate that these variations might have arisen due to complex microbial interactions or other factors. Interestingly, the abundances of Anaerovoracaceae Family XIII AD3011 decreased in RB and IF, whereas Bacteroides decreased in the RB and SF treatments. Further, comparing the RB treatment with humans, Eubacterium siraeum was increased. Another genus, Anaerovoracaceae Family XIII UCG-001, was common, but showed a varied difference. Ruminococcus, on the other hand, was commonly increased in drinkers and in the SF treatment in mice.

3.3. Fecal Metabolites after 30 Days of Rice Beverage Treatment

A total of 68 compounds were detected with a peak area greater than 0.1%. Of these, 25 compounds were majorly categorised as organic acids, 8 sugar alcohols, 7 amino acids, and 6 short-chain fatty acids. Others included inorganic compounds and acids sugars, and amino acid derivatives (

Table 3. List of mouse fecal metabolites (peak area percentage) after 30 days of treatment with different fractions of the beverage.(CT = control, RB = rice beverage, IF = insoluble fraction, and SF = soluble fraction).

Metabolites	CT	RB	IF	SF
alpha Linolenic acid	2.54	0.955	0.505	0.13
17-Octadecynoic acid	0	0.075	0	0
1-Monolinoleoylglycerol	0.75	2.59	0.205	3.075
1-O-Heptadecylglycerol	0	0	0	0.03
1-O-hexadecylglycerol	0	0.2	0.255	0.595
2-alpha -Mannobiose	0	0.13	0	0
2-Desoxy-pentos-3-ulose	0.045	0.11	0.095	0.175
2-Pyrrolidone-5-carboxylic acid	0.525	0.07	0.43	0.455
3-Hydroxydodecanedioic acid	0.17	0	0	0
4-Hydroxyphenylbutyric acid	0	0.155	0.235	0
4-Nitrophenyl-.beta.-D-galacturonide	0	0.065	0	0
5-Methyluridine	0	0	0.275	0
9,12,15-Octadecatrienoic acid	0	0	0.1	0
9-Octadecenoic acid	0	2.1	2.645	1.485
Acetic acid	0.14	0.44	0.11	0.215
alpha-L-Mannopyranose	0	0.025	0	0
Arachidonic acid	0	0	0	0.32
Benzenepropanoic acid	0.16	0	0.11	0
Benzoic acid	0	0	0	0.17
Butanedioic acid	1.91	0.725	1.07	3.045
Butanoic acid	0	0	1.225	0.11
Cinnamate	0	0.025	0	0.03
D-(-)-Erythrose	0	0.015	0	0
D-(-)-Rhamnose	0	1.04	1.905	2.625
D-(-)-Tagatose	0	0.11	0.345	0.195
D-(+)-Cellobiose	0	0.885	0.76	0.595
D-(+)-Glucuronic acid	0	0.07	0.075	0
D-(+)-Xylose	8.415	5.38	9.36	5.79
D-Arabinonic acid	0	0	0	0.02
D-Galactose	16.96	12.04	11.11	13.67
D-Glucitol	0	0	0	1.315
D-Glucose	0	16.4	0	0
D-Lactose	0	2.015	0	0
D-Mannose	0.34	0.215	0	0.895
Dodecanedioic acid	0	0.12	0.285	0
D-Xylose	0	0	0	0.05
Eicosanoic acid	0	0	0	0.095
Erythro-Pentonic acid	0	0.04	0.21	0.05
Ethanedioic acid	0.4	0	0.955	0.845
Ethanol	0	0	0	0.01
Glycerol	4.41	4.84	4.31	5.295
Glycine	0	0	0	0.035
Hexadecanoic acid	0	7.495	6.225	5.85
L-(-)-Sorbose	0	0	0	0.02
L-Asparagine	0	0	0	0.155
L-Isoleucine	0.07	0.03	0.125	0.35
L-Threonine	0	0	0.25	0.79
L-Valine	0.135	0.1	0.2	0.355
Maltose	0	0.155	0	0
Monoamidoethylmalonic acid	0	0	0.01	0
N-Acetyl-D-galactosaminitol	0	0	0	1.07
N-Acetyl-D-glucosamine	2.955	1.415	4.525	4.78
n-Pentadecanoic acid	0	0	0	0.27
N-phthalimide	1.415	1.355	0	2.44
Octadecanoic acid	0	4.4	0	0
Oleic acid	7.62	5.125	4.19	4.84
Pentanedioic acid	0.045	0	0.015	0.015
Pentanoic acid	0	0	0	1.22
Phenyl alanine	0	0	0	0.045
Phosphoric acid	0	0	0.04	0
Propanedioic acid	0	0	0	0.12
Propanetriol	0	0.015	0.03	0.04
Propanoic acid	9.89	7.06	6.075	13.01
Ribitol	2.38	0.03	0	0
Serine	0.1	0.04	0.19	0.265
Tetracosanoic acid	0	0.18	0.28	0.42
Tetradecanoic acid	0.065	0	0.395	0.2
Uridine	0	0	0.105	0.04

). Multivariate statistical analysis was performed for the fecal metabolites using partial least-squares discriminant analysis among the treatment groups. The PLS-DA plot indicates the clustering of the treatment groups on the basis of the metabolite profiles (Figure 4).

3.4. Plasma Endotoxin and Histological Study of Ileum

The level of endotoxin in the CT group (0.515 EU/mL) was higher than those in the RB (0.123 EU/mL), IF (0.073 EU/mL) and SF (0.215 EU/mL) groups (p ≤ 0.05) (Figure 5). HE-stained ileum revealed changes in enterocytes, villus length and crypt depth, lymphocytic infiltration, and goblet cell distribution of the ileum in the treatment groups. Villus length was 169.85 $\mathsf{\mu }$m in the CT group with intact epithelia. The length was reduced to 155.89 and 151.84 $\mathsf{\mu }$m in the IF and RB-treated groups, respectively, and increased in the SF-treated group (176.91 $\mathsf{\mu }$m). The length of the villus in all the treatment groups differed significantly compared to that of the controls. Crypt depth in the CT group (84.02 $\mathsf{\mu }$m) had no significant difference from other treatment groups. The width of the lamina propia and the depth of the submucosal layer were not altered among the treatment groups (

Table 4. Morphological features of ileum tissues ( mean of 5 replicates ± standard error, CT = control, RB = rice beverage, IF = insoluble fraction, and SF = soluble fraction).

	CT	RB	IF	SF
Villus length ($\mathsf{\mu }$m)	169.852 ± 5.92	151.848 ± 3.59	155.89 ± 10.73	176.914 ± 9.92
Crypt depth ($\mathsf{\mu }$m)	84.022 ± 1.96	82.892 ± 2.98	74.65 ± 2.21	93.246 ± 2.39
Width of lamina propria ($\mathsf{\mu }$m)	26.94 ± 2.64	27.236 ± 1.80	30.108 ± 2.86	31.902 ± 1.33
Depth of submucosa ($\mathsf{\mu }$m)	22.51 ± 1.44	25.376 ± 3.00	21.23 ± 1.64	29.43 ± 2.35

). A width of lamina propria with much more lymphocytic infiltration was also observed in the RB-treated group, followed by the SF-treated group. The number of goblet cells was comparatively higher in the SF-treated group (Figure 6).

4. Discussion

The gut microbiota regulates the health and diseases of the host. Various factors that shape the population and composition of gut microbiota include the diet, genetics, mode of delivery, age, and geography. Notably, diet remains the most important factor influencing the gut microbiota. Further, dietary components such as carbohydrates, proteins, fats, and other micronutrients play distinct roles [5]. Fermented foods contain nutrients, metabolites, bioactive peptides, and several components produced by microbial action that affect the gut microbiota. Very often, microbes participating in the fermentation process directly pass the intestinal barrier, colonise the gastrointestinal tract of the host, and exert health benefits. Hence, the multiple food components and microbes, and their interactions differently influence the microbiota. A recent review by Precup and Vodner (2019) demonstrated the role of Prevotella in the gut in both human and rodents, and suggested it as a “possible biomarker of diet” [25]. Further, a comparative study on the gut microbiota of urban and traditional rural diets suggested an association of this bacterium with the rural diets [26]. Another study, in a similar line, suggested a reduced abundance of Prevotella in humans consuming a Western diet [27]. In addition, Dehingia et al. (2015) showed Prevotella as a major bacterium in different tribes of Northeast India. Collectively, these findings support our observation of the dominance of Prevotalla in the human gut. Moreover, in this study, alpha diversity demonstrating the richness of bacterial communities remained unchanged in the human cohorts, in line with earlier findings with different ethnic communities of Northeast India [4]. Succinovibrio is yet another bacterium associated with Western and non-Western diets that was found to be abundant in our findings, associated with humans drinking the beverage [27].

The IF was devoid of alcohol, and contained microbes and nutrients present in the fermented residues, whereas the SF was devoid of microbes, but soluble constituents were present. The overall changes in bacteria observed in the treatment groups suggest the effect of different components of the beverage. For example, alcohol consumption changes microbial diversity in both rodents and humans [28]. In SF, the change in Bacteroides as a result of alcohol is in line with earlier reports [28,29]. Alcohol-induced reduction in LAB, as reported earlier, was also evident in our findings [30]. However, in RB, Ligilactobacillus was reported to have increased antiobesity and anti-inflammatory effects in mice [31,32]. This bacterium, often considered as an immunobiotic, has exhibited synergistic effects with wine prebiotics against colitis in mice [33,34]. Prevotellaceae UCG-001 is a Gram-negative bacillus that ferments carbohydrates to generate SCFAs that exert anti-inflammatory effects, inhibiting pathogens [35,36]. An increased abundance of this bacterium in mice was associated earlier with prebiotic and other functional components [37,38]. In our findings, the increase in this bacterium in the RB and SF treatments indicated such components, in line with our earlier report [14].

Varied differences in unclassified bacteria belonging to the Lachnospiraceae family were observed in different treatments in both mouse and human samples. For instance, Lachnospiraceae NK4A136 is an anaerobic, spore-forming bacterium that ferments plant polysaccharides into SCFAs [39]. It was associated with bile acid metabolism and cadaverine leading to an improvement in insulin levels [40]. The abundance of this bacterium was higher among drinkers in the human subjects, similar to a study on alcohol-induced changes in the microbiota of Rhesus macaques [41]. Further, a decreased abundance of Christensenellaceae among drinkers is also supported by this study.

Interestingly, another study reported increased abundances of Lachnospiraceae and Bacteroides in stressed mice compared with in the normal group, whereas these two contrarily decreased significantly in the RB and SF treatments [42]. However, in the IF treatment, we observed an increase in Lachnospiraceae, though Bacteriodes did not change. These variations in the observations might have resulted due to the complex interactions between different bacteria and the differences in the treatments. In the SF group, Paludicola increased, which had earlier been reported to increase by treating with an antioxidant component of fermented rice [43,44]. Ruminococcus was previously reported to (i) enhance immunity by promoting the gut barrier function, (ii) produce butyrate in the gut, and (iii) be associated inversely with intestinal permeability [45,46]. An increase in Ruminococcaceae due to prebiotic treatment was also reported [47]. Interestingly, in human samples, SCFA producers Coprococcus, Oscillospiraceae UCG-005, Butyrivibrio, and Butyricimonas were increased in drinkers [48,49,50,51].

Dietary differences have been associated with changes in the morphology of the small intestine, leading to functional alterations in its structure. The goblet cells present in the intestinal epithelium produce several factors, including trefoil factors, and different mucins that form the mucus layer. Mucin 2 is the most abundant mucin and is associated with the organisation of the intestinal mucus layers at the epithelial surface. This layer serves as a barrier to pathogens, restricting their entry towards epithelial cells [52]. An increased goblet cell number and hyperplasia in the SF-treated group could have been an indication of escalated mucin production due to the impact of alcohol, as supported by an earlier study [53]. The absorption and depletion of fatty acids and carbohydrates take place during the transition through the small intestine; hence, bacteria residing in the colon are sustained by the fermentation of the complex sugars [54]. The SF treatment contained alcohol along with soluble sugars that could elevate the triglyceride levels [55]. This might have increased the villus length in the SF treatment. Undigested fibres serve as the energy source for the gut bacteria, leading to the production of SCFAs that contribute to anti-inflammatory effects. We had earlier reported the presence of hexadecanoic acid, a long-chain fatty acid, in a rice-based beverage that was fermented using a similar method in this study [14]. It is one of the flavour components and is abundant in rice wine from China [56]. Interestingly, it was present in the feces of all treatments, and absent in the control, depicting its origin to be from the components of the beverage. There have been reports suggesting the endotoxin-reducing activity of fermented foods [57]. As expected, plasma endotoxins were lowest in the IF, followed by the RB and SF treatments, compared to the controls. Contrarily, the inflammation-inducing effects, as observed in the histology of SF-treated mouse intestine, also had a higher endotoxin level. The IF treatment, having no alcohol, showed the lowest endotoxin levels and an RB having a combination of IF and SF had intermediate endotoxin levels in the blood, but less than that in the controls.

5. Conclusions

Fermented foods and beverages produced traditionally with indigenous practices have gained popularity due to the associated health benefits. These foods have a long history of consumption, and are strongly associated with the culture and traditions of many ethnic communities. There are traditional claims from these communities of these foods having health modulatory effects. However, due to a lack of scientific justification to these claims, these foods often remain confined to local markets. Rice beverage, a popular fermented drink, is prepared by different tribes of Northeast India and traditionally believed to have several health benefits. This study, in an attempt to validate the claims, was conducted using human and mouse subjects. Overall, our associative findings in humans and mice collectively pave the way to understanding the changes in the microbiota due to the consumption of this beverage. Altogether, the observed bacterial changes due to the consumption of the beverage in both subject groups may have resulted due to the functional components of the beverage. Changes in SCFA-producing bacteria and those affected by prebiotics indicated the gut health modulatory aspects of the beverage. A low level of endotoxin in the treatments compared to that in the controls further strengthens this attribute. In connection with the emergent reports on the health benefits of moderate alcohol consumption, this study further opens new dimensions for studying other aspects of the rice beverage [58,59].