Abstract

   The short- and long-term consumption of monosodium glutamate (MSG)
   increases urinary pH but the effects on the metabolic pathways in the
   liver, kidney and the gut microbiota remain unknown. To address this
   issue, we investigated adult male Wistar rats allocated to receive
   drinking water with or without 1 g% MSG for 2 weeks (n = 10, each). We
   performed a Nuclear Magnetic Resonance (NMR) spectroscopy-based
   metabolomic study of the jejunum, liver, and kidneys, while faecal
   samples were collected for bacterial DNA extraction to investigate the
   gut microbiota using 16S rRNA gene sequencing. We observed significant
   changes in the liver of MSG-treated rats compared to controls in the
   levels of glucose, pyridoxine, leucine, isoleucine, valine, alanine,
   kynurenate, and nicotinamide. Among kidney metabolites, the level of
   trimethylamine (TMA) was increased, and pyridoxine was decreased after
   MSG-treatment. Sequencing of the 16S rRNA gene revealed that
   MSG-treated rats had increased Firmicutes, the gut bacteria associated
   with TMA metabolism, along with decreased Bifidobacterium species. Our
   data support the impact of MSG consumption on liver and kidney
   metabolism. Based on the gut microbiome changes, we speculate that TMA
   and its metabolites such as trimethylamine-N-oxide (TMAO) may be
   mediators of the effects of MSG on the kidney health.

   Keywords: monosodium glutamate, gut microbiota, metabolic pathway,
   metabolomics, microbiome, trimethylamine

1. Introduction

   Monosodium glutamate (MSG) is commonly added to foods to increase
   palatability, especially in Asian cuisine and industrially processed
   foods [[45]1]. MSG is considered a safe ingredient for human
   consumption regardless of the amount by the Food and Drug
   Administration (FDA) [[46]2] despite recent data showing that the
   average daily MSG intake is 3–4 g and that every additional gram of MSG
   increases the risk of developing metabolic syndrome [[47]3]. Elevated
   amounts of dietary MSG are associated with overweight [[48]4,[49]5] and
   hypertension [[50]6], but the results are inconsistent [[51]7,[52]8].

   Attempts have been made to reveal the effects of MSG on metabolic
   organs of animals by either parenteral [[53]9,[54]10] or oral intake
   [[55]11,[56]12]. In one of such efforts to investigate the effects of
   MSG in rat kidneys, the results showed that both short- [[57]13] and
   long-term [[58]11] MSG consumption caused alkaline urine in rats,
   although the mechanism of urine alkalinization is not established.
   Metabolomics is a commonly used approach to define the changes in
   metabolites from various conditions [[59]14,[60]15]. The short-term MSG
   consumption caused specific changes in urinary metabolites including
   dimethylamine (DMA) and methylamine (MA), which are gut-derived
   metabolites from trimethylamine (TMA). TMA is a volatile short-chain
   aliphatic amine that provides the characteristic fish odour. In fact,
   the gut bacteria produce TMA from dietary fish or it can be generated
   from other nutrients including choline and carnitine, which are
   abundant in eggs and red meat [[61]16]. The major part of TMA is
   enzymatically converted to the odourless trimethylamine-N-oxide (TMAO),
   and high TMAO plasma levels are associated with cardiovascular disease
   (CVD) [[62]17], chronic kidney disease (CKD) [[63]18] and diabetes
   [[64]19].

   We herein took advantage of metabolomics tools to investigate the
   effects of short-term MSG consumption on the metabolome of plasma,
   liver, kidney and gut, and analysed the correlation between the
   metabolomic results and the changes of the gut microbiota. Our data may
   provide new mechanistic insights into the effects of dietary MSG while
   also demonstrating biomarkers of MSG-induced organ damage.

2. Materials and Methods

2.1. Animals

   Twenty 6-week-old male Wistar rats (weight approximately 200 g) were
   obtained from the National Laboratory Animal Center (Mahidol
   University, Salaya, Bangkok, Thailand), acclimatized in individual
   stainless-steel metabolic cages for 2 weeks, and then maintained under
   a standard condition of temperature (23 ± 2 °C), humidity (30–60%) and
   brightness (350–400 Lux) of 12 h dark/12 h light cycle. Rats were
   provided with a commercial pellet diet No. CP 082 (Perfect Companion
   Group, Bangkok, Thailand) during the study. All experiments were
   performed following the guidelines of the Northeast Laboratory Animal
   Center (NELAC), Khon Kaen University, Thailand. Moreover, the study was
   approved by the Animal Ethics Committee of Khon Kaen University,
   Thailand (AEKKU-NELAC 5/2558).

2.2. Reagents

   Pure food-grade (99%) MSG (Ajinomoto, Tokyo, Japan) was used to feed
   the animals. Commercial grade methanol and chloroform were purchased
   from RCI LABSCAN LIMITED (Bangkok, Thailand) for tissue extraction,
   LC-MS-grade water was purchased from Merck (Darmstadt, Germany),
   potassium dihydrogen phosphate (KH[2]PO[4]) and deuterium oxide (D[2]O)
   were purchased from Merck (Darmstadt, Switzerland), sodium azide
   (NaN[3]) was produced by the European Chemicals Agency (ECHA, Helsinki,
   Finland), sodium trimethylsilyl-[2,2,3,3-^2H[4]]-propionate (TSP), an
   internal standard for NMR analysis, was obtained from Santa Cruz
   Biotechnology (Santa Cruz CA, USA). For mass spectrometry,
   commercial-grade isopropanol (C[3]H[8]O) and formic acid (CH[2]O[2])
   were purchased from RCI LABSCAN LIMITED (Bangkok, Thailand).

2.3. Experimental Design

   Rats were randomly assigned to receive drinking water either with (n =
   10) or without (n = 10) 1 g% MSG for 2 weeks. Reverse osmosis (RO)
   water was used (chlorine concentration of 3–4 ppm) for the animal study
   when all the rats had free access to food. Daily intake of water (ml)
   and food (g) were recorded, as well as the weekly body weight (g). The
   24 h urinary excretion (ml/rat/day) was recorded throughout the study
   period. Faeces and tail blood samples (100 µL of plasma) were collected
   at 2 weeks prior to the sacrifice of animals using carbon dioxide
   (CO[2]) after a 12-h fasting. Tissue samples, i.e., jejunum, liver and
   kidney, were collected and snap-frozen in liquid nitrogen before being
   stored at −80 °C until used for analysis.

2.4. Sample Preparation and Analysis

   Each tissue (100 mg wet mass) was used for metabolite extraction
   according to the previously published protocols [[65]20]. Before NMR
   acquisition, the polar phase of tissue extracts was re-suspended in an
   NMR buffer of 580 µL (100 mM sodium phosphate buffer, pH 7.4, in D[2]O,
   containing 0.1 mM TSP and 0.2% NaN[3]), vortexed briefly, and
   centrifuged at 12,000× g for 5 min at 4 °C. Subsequently, 550 µL of the
   mixture was transferred to an NMR glass tube (Duran Group, Mainz,
   Germany) with an outer diameter of 5 mm prior to NMR analysis.
   Approximately 250 mg of faeces was mixed with 500 µL HPLC-grade water
   (Merck, Darmstadt, Germany) and homogenized using a vortex mixer with
   the speed 2500 rpm for 15 min at room temperature. The faecal
   suspension was then centrifuged at 12,000× g for 15 min at 4 °C and 540
   µL of supernatant was transferred into clean 1.5 mL microtubes, and 60
   µL of NMR buffer (1.5 M KH[2]PO[4] buffer, pH 7.4, in D[2]O, containing
   2 mM TSP and 1% NaN[3]) was added. Next, the tubes were vortexed
   briefly, centrifuged at 12,000× g for 10 min, and finally, 580 µL of
   the mixture was transferred to a 5 mm outer diameter NMR glass tube for
   analysis.

   Tissue extracts and faecal samples were analysed using a 400 MHz NMR
   spectrometer (Bruker Biospin, Rheinstetten, USA) at a temperature of
   298.15 K. The spectra were referenced to the TSP peak (δ^1H 0.00),
   phased and baseline corrected using MATLAB (Mathworks, Natrick, MA
   USA). The signal of TSP peak (δ^1H −1.000–0.005) from all tissues and
   TSP peak (δ^1H −1.20–0.157) from faeces were removed. In addition, the
   water peak was removed from the jejunum (δ^1H 4.50 and 5.20), liver
   (δ^1H 4.68 and 5.00), kidney (δ^1H 4.31 and 5.74) and faeces (δ^1H 4.18
   and 5.23). Moreover, raw spectra were subjected to the peak alignment
   and normalization [[66]21]. The spectral data of all samples were
   analysed using unsupervised principal component analysis (PCA) and
   supervised orthogonal signal correction-projection to latent
   structures-discriminant analysis (O-PLS-DA). The data were mean-centred
   and scaled with unit variance (UV). The O-PLS-DA models are evaluated
   by the R^2X, R^2Y and Q^2Y values, representing the fitness, fraction
   of variances of Y matrix, and forecasting ability of the model,
   respectively [[67]22]. To avoid over-fitting, 7-fold cross-validation
   was performed for 500 repeats. Permutation p-value was used to indicate
   the model validity. Significant variables of each valid model were
   selected through O-PLS-DA correlation coefficients with
   Benjamini–Hochberg false-discovery rate correction (p < 0.05).
   Statistical total correlation spectroscopy (STOCSY) [[68]23], in-house
   chemical shift databases and the Human Metabolome Database (HMDB
   version 4, USA) [[69]24] were used for metabolite identification.

   Pathway enrichment analysis was also conducted using MetaboAnalyst
   ([70]http://www.metaboanalyst.ca/ (accessed on 14 July 2020)) [[71]25].
   The metabolic pathway illustration was generated by Cytoscape [[72]26].
   Changes in metabolites were identified by STOCSY and univariate
   analysis was performed by investigating the relative concentration of
   significantly differential metabolites, calculating the integration of
   spectra peaks. According to the p-value calculated by the Student’s
   t-test using GraphPad Prism 7 (Ver. 7, GraphPad Software, Inc., La
   Jolla, CA, USA).

2.5. Plasma Sample Preparation for UHPLC-ESI-QTOF-MS Analysis

   Plasma (50 µL) was mixed with 150 µL of isopropanol (IPA), followed by
   24-h incubation at 20 °C to precipitate the protein. All samples were
   centrifuged twice at 13,000× g at 4 °C for 10 min. A total of 50 µL was
   taken from each sample and pooled in a 1.5 mL microcentrifuge tube for
   constructing a quality control (QC) sample and 120 µL of each sample
   mixture was transferred to an HPLC glass insert.

2.6. UHPLC-ESI-QTOF-MS Analysis

   UHPLC-ESI-QTOF-MS analysis was employed at Khon Kaen University
   International Phenome Laboratory (KKUIPL). The aqueous phase extracts
   of samples were analysed on a reverse-phase platform. The separation
   part was performed using a UHPLC system (Bruker, Darmstadt, Germany)
   when the Bruker intensity solo HPLC C18 2.1 × 100 mm, 2 µm column
   (Bruker, Darmstadt, Germany) was used. The column temperature was set
   at 55 °C and the autosampler temperature was set at 4 °C. Mobile phase
   A was 100% HPLC grade water with 0.1% formic acid (FA), and mobile
   phase B was 100% methanol with 0.1% FA. The flow rate was set at 0.4
   mL/min and the elution gradient was set as—99.9% A (0.0–2.0 min, 0.25
   mL/min), 99.9–75% A (2.0–10.0 min, 0.4 mL/min), 20% A (10.0–12.0 min,
   0.4 mL/min), 10% A (12.0–21.0 min, 0.4 mL/min), 0.1% A (21.0–23.0 min,
   0.4 mL/min), 99.9% A (24.0–26.0 min, 0.4 mL/min). A sample injection
   volume of 4 µL was applied for both positive and negative ionization
   polarity modes.

   Mass spectrometry analyses were performed using a compact ESI-Q-TOF
   system (Bruker, Darmstadt, Germany). Sodium formate (HCOONa) containing
   2 mM sodium hydroxide, 0.1% FA and 50% IPA was directly injected as an
   external calibrant with flow rate 0.5 µL/min. The condition in positive
   ionization polarity mode—mass range 50–1300 m/z, cone voltage 31 V,
   capillary voltage 4500 V, source temperature 220 °C, desolvation
   temperature 220 °C, desolvation gas flow 8 L/min. The conditions of
   negative ionization polarity mode—m/z range: 50–900 m/z, cone voltage
   31 V, capillary voltage 4500 V, source temperature 220 °C, desolvation
   temperature 220 °C, desolvation gas flow 8 L/min.

2.7. Gut Microbiome Analysis

   DNA was extracted using the QiAamp^® PowerFecal^® Pro DNA kit (Qiagen,
   Hilden, Germany). Extracted DNA was measured at OD 260/280 using a
   Nanodrop2000c spectrophotometer (Thermo scientific, Waltham, MA, USA).
   All of the extracted DNA were preserved at −20 °C until analyses. The
   V3-V4 region of the 16S ribosomal RNA (rRNA) gene for each sample was
   amplified using the universal forward primer V3
   (5′-CCTACGGGNGGCWGCAG-3′) and the reverse primer V4
   (5′-GACTACHVGGGTATCTAATCC-3′). PCR reaction consisted of 2x KAPA HiFi
   HotStart ReadyMix, 12.5 ng DNA template, and 5 µM of each primer, with
   initial denaturation at 95 °C for 3 min followed by 25 cycles of
   denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, extension
   at 72 °C for 30 s, and final extension at 72 °C for 10 min. An Agilent
   DNA 1000 Chip and Agilent 2100 Bioanalyzer (Agilent Technologies, Palo
   Alto, CA, USA) were combined to quantify the amplified product. PCR
   amplicons were purified using AMPure XP beads to purify. The partial
   16S rRNA gene was sequenced using Illumina MiSeq platform (Illumina
   Inc., San Diego, CA, USA).

   The 16S rRNA sequencing library was prepared using genomic DNA (gDNA),
   and matching paired-end sequences were merged using FLASH
   ([73]http://ccb.jhu.edu/software/FLASH/ (accessed on 29 June 2020))
   [[74]27]. Quality filtering removed sequences containing low-quality
   reads and was performed using CD-HIT-OTU
   ([75]http://weizhong-lab.ucsd.edu/cd-hit-otu (accessed on 29 June
   2020)) [[76]28] and the rDnaTools package. OTUs (Operational taxonomic
   units) were classified with 97% threshold identity using UCLUST
   algorithm based on 100% similarity. Sequences sharing ≥97% similarity
   were assigned to the same OTU. Representative OTU sequences were
   aligned and were taxonomically classified using Ribosomal Database
   Project (RDP) and were compared with the reference NCBI databases. The
   classification is based on the Green genes database
   ([77]http://greengenes.lbl.gov/ (accessed on 29 June 2020)). The output
   of a classification of reads at several taxonomic levels—kingdom,
   phylum, class, order, family, genus, and species using Quantitative
   Insights Into Microbial Ecology (QIIME) [[78]29].

2.8. Statistical Analysis

   The statistical analysis of the food and water intake, urine volume
   output and body weight were reported as mean ±SD per group of animals
   and the differences between groups were compared for statistical
   significance by Student’s t-test with p-value < 0.05 considered as
   statistically significant.

3. Results

3.1. IMPACT of MSG on Food and Water Intake, Body Weight and Urine Volume
Output

   MSG-treated and control rats had a similar amount of food intake (17.44
   ± 1.94 and 18.46 ± 1.37 g/rat/day, respectively) ([79]Figure 1A) and
   body weight (333.58 ± 17.23 and 333.80 ± 15.50 g/rat, respectively)
   ([80]Figure 1B). However, water intake was significantly higher in
   MSG-treated rats (52.38 ± 18.36 mL/day) compared to controls (38.38 ±
   8.39 mL/day) ([81]Figure 1C). In both MSG-treated and control groups,
   urine output tended to increase with time, although a significant
   increase was seen only in MSG-treated rats (15.42 ± 3.92 mL/rat/day at
   pre-treatment and 29.09 ± 8.78 mL/rat/day post-treatment). Also,
   although statistically not significant, urine output of MSG-treated
   rats (29.09 ± 8.78 mL/rat/day) was apparently higher than that of
   controls (23.15 ± 10.55 mL/rat/day) ([82]Figure 1D).

Figure 1.

   [83]Figure 1
   [84]Open in a new tab

   (A) Food intake (B) Body weight (C) Water intake and (D) Urine output
   of male Wistar rats supplemented with 1 g% MSG and control animals (n =
   10 per group) at baseline and at 2 weeks. Data are shown as mean ± SD
   and p-values calculated by Student’s t-test, ** p < 0.01, *** p <
   0.001).

3.2. Metabolic Changes in Metabolically Important Organs

   The NMR spectra of the collected tissue samples were analysed after two
   weeks of MSG treatment. PCA of NMR spectral data was initially carried
   out to observe any obvious clustering and outliers ([85]Figure 2). The
   clear clustering and complete separation were observed in hepatic
   metabolite profiles between the control and MSG-treated group with a
   permutation p-value of 0.002, R^2X of 58%, Q^2Y of 0.84, and R^2Y of
   0.93 illustrated in PCA and O-PLS-DA score plots ([86]Figure 2A, B). A
   representative OPLS-corresponding coefficient loading plot with
   metabolite assignment is presented in [87]Figure 3 with significant
   liver models indicated in [88]Figure 3A; all significant changes in
   metabolites are summarized in [89]Table 1. In the liver, higher levels
   of six metabolites including glucose, pyridoxine, 2-deoxyuridine,
   inosine, unknown 1 and 2 were found in the MSG-treated group, whereas
   eleven metabolites, namely leucine, isoleucine, valine, alanine,
   N-acetylglycoprotein, acetone, 1,3 dimethylurate, histamine, xanthine,
   kynurenate and nicotinamide, were significantly higher in the control
   group. The PCA score plot based on the renal metabolites ([90]Figure
   2C) indicated no clear clustering between two groups, and the O-PLS-DA
   model was statistically significant with a permutation p-value of
   0.018, R^2X of 56%, Q^2Y of 0.29, and R^2Y of 0.74 ([91]Figure 2D). Two
   altered metabolites were found in kidney tissues with significantly
   higher trimethylamine level and significantly lower pyridoxine in the
   MSG-treated group compared to controls ([92]Figure 3B). In contrast,
   the results of jejunum ([93]Figure 2E, F), faeces and plasma
   ([94]Figure 2 and [95]Figure S1) revealed no clustering between the
   treatment and control groups illustrated in PCA and O-PLS-DA scores
   plots.

Figure 2.

   [96]Figure 2
   [97]Open in a new tab

   Principal component analysis (left panel) and O-PLS-DA score plot
   (right panel) of tissues of MSG-treated (Red) and control groups
   (Blue). (A,B) Liver; (C,D) kidney; and (E,F) jejunum.% PC represents
   variation explained by each principal component and Q^2 represents
   predictive ability of the model.

Figure 3.

   [98]Figure 3
   [99]Open in a new tab

   OPLS-corresponding coefficient loading plots of liver (A) and kidney
   (B). Positive and negative signals denote significant metabolites in
   MSG-treated rats (downwards pointing) and control groups (upwards
   pointing) with intensity indicated by colour (red is highly correlated
   and blue is weakly correlated) and also correlation per metabolite.
   Abbreviations: Ile, isoleucine; NAG, N-acetylglycoprotein; IrI,
   correlation coefficients.

Table 1.

   Relative changes in liver and kidney metabolites of control and
   MSG-treated rats, using the ^1H NMR profiles.
   Metabolites Chemical SHIFT
   (Multiplicity) MSG Induced Metabolic Changes Compared to Control
   (−) Control vs. (+) MSG
   R2X = 58%, Q2Y = 0.84,
   p = 0.002 (liver) (−) Control vs. (+) MSG
   R2X = 56%, Q2Y = 0.29,
   p = 0.018 (kidney)
   Leucine 0.9644 (t), 1.735 (t), 3.787 (m) −0.7823 -
   Isoleucine 0.9644 (t), 1.006 (d), 1.476 (d), 3.833 (d) −0.8539 -
   Valine 0.9644 (t), 1.006 (d), 1.039 (d), 1.476 (d) −0.7768 -
   Alanine 1.494 (d), 4.106 (q) −0.6109 -
   N-Acetylglycoprotein 2.086 (s) −0.8440 -
   Acetone 2.237 (s) −0.8654 -
   Trimethylamine 2.89 (s) - 0.6579
   1,3-dimethylurate 3.306 (s), 3.359 (s) −0.7834 -
   Glucose 3.27 (t), 3.426 (m), 4.642 (d), 5.234 (d) 0.4626 -
   Unknown 1 3.426 (m), 3.521 (m), 4.642 (d), 5.233 (d) 0.5254 -
   Unknown 2 7.089 (s) 0.7485 -
   Histamine 3.095 (m), 3.306 (t), 7.108 (s), 7.906 (s) −0.8240 -
   Pyridoxine 7.684 (s) 0.9173 −0.6069
   2-Deoxyuridine 3.558 (d), 3.719 (m), 3.833 (m), 5.244 (d), 7.859 (d)
   0.8117 -
   Xanthine 7.892 (s) −0.7847 -
   Kynurenate 8.198 (d) −0.6075 -
   Inosine 6.099 (d), 8.241 (s), 8.348 (s) 0.5690 -
   Nicotinamide 8.712 (d), 8.945 (s) −0.6430 -
   [100]Open in a new tab

   R^2X and Q^2Y show the variance explained and predicted by each model
   while p-values for all models were derived from a permutation test (n =
   500). (+) Indicates a higher correlation, whereas (−) indicates a lower
   correlation of urinary metabolites after MSG consumption. The bolded
   chemical shift per metabolite was used as the STOCSY driver peak and
   for deriving the correlation and p-value. Abbreviations: s, singlet; d,
   doublet; t, triplet; m, multiple.

   The correlation pathways of the significant metabolites in the liver
   and kidney tissues were investigated using KEGG IDs and generated by
   Cytoscape software. The metabolites were found to link the metabolic
   network. When these results were combined to draw a metabolic pathway
   map in order to illustrate a more intuitive correlation, the
   metabolites were revealed to be mainly associated with metabolic
   pathways, such as hepatic gluconeogenesis, branched-chain amino acids,
   vitamin B6, and trimethylamine metabolisms ([101]Figure S3).

3.3. MSG Consumption Alters Gut Microbiome

   The results of faecal bacterial composition analysis indicated seven
   phyla, 15 classes, 19 orders, 46 families, 127 genera and 220 species
   from all animal groups. The top 7–10 species were selected and
   percentage relative abundance of bacterial communities at the phylum,
   class, order, family, genus and species levels was generated
   ([102]Figure 4). At the phylum level, the top four relatively abundant
   phyla were in the following order—Firmicutes > Bacteroidetes >
   Proteobacteria > Actinobacteria in all animals ([103]Figure 4A). The
   abundance of Actinobacteria was significantly lower in MSG-treated rats
   (0.30%) than in control rats (1.32%) ([104]Figure 5A).

Figure 4.

   [105]Figure 4
   [106]Open in a new tab

   Faecal microbial composition at the different taxonomic levels in
   MSG-treated (M1–M10) and control (C1–C10) rats. (A = phylum, B = class,
   C = order, D = family, E = genus and F = species). Others indicate
   other bacteria and unidentified bacteria <5%.

Figure 5.

   [107]Figure 5
   [108]Open in a new tab

   Heatmap of relative operational taxonomic unit (OTU) abundance at the
   phylum level (A), and genus level (B) from faeces samples of control (n
   = 10) and the MSG group (n = 10). The colour key corresponds to percent
   relative abundance of the gut microbiota in OTU at each expression
   level.

   At the class level, the most abundant microbiota were Bacilli >
   Clostridia > Bacteroidia > Erysipelotrichia in all animals ([109]Figure
   4B). The MSG-treated group had significantly higher Clostridia (31.59%)
   than the control group (19.25%). However, the MSG-treated rats had
   significantly lower Actinobacteria (0.15%) than control rats (1.09%)
   ([110]Figure 4B). When interpreted at the order level, Lactobacillales,
   Clostridiales, Bacteroidales, Erysipelotrichales were abundant in all
   animals ([111]Figure 4C). Clostridiales were significantly higher in
   MSG-treated rats (31.59%) than in control rats (19.25%), while
   Bifidobacteriales were significantly lower in the MSG-treated rats
   (0.06%) compared to control rats (1.06%).

   The four most abundant families commonly observed in all experimental
   animals were in the following order—Lactobacillaceae >
   Erysipelotrichaceae > Clostridiaceae > Prevotellaceae ([112]Figure 4D).
   At the family level, Bifidobacteriaceae was significantly lower, but
   unidentified Clostridiales was significantly higher in MSG-treated rats
   compared to control rats. The relative abundance of top four genera
   observed in all the animals were in the following order—Lactobacillus >
   Turicibacter > Prevotella < Clostridium ([113]Figure 4E). A heatmap of
   the relative abundances at the genus level showed lower abundance of
   Lactobacillus and higher abundance of Clostridium in MSG-treated rats
   ([114]Figure 5B). Moreover, the lower abundance of Bifidobacterium was
   also observed in MSG-treated rats (0.06%) compared to control rats
   (1.06%).

   At the species level, the top four gut bacterial species common in all
   rats were Lactobacillus intestinalis, Turicibacter sanguinis,
   Clostridium saudiense, and Muribaculum intestinale in order
   ([115]Figure 4F). In the MSG-treated group, Flintibacter butyricus was
   significantly more abundant than in the control group, whereas
   Faecalibaculum rodentium and Bifidobacterium pseudolongum were
   significantly less abundant.

4. Discussion

   Several lines of evidence have consistently linked MSG consumption to
   adverse effects on human health, such as higher incidence of metabolic
   syndrome [[116]3], overweight [[117]4,[118]5] and arterial hypertension
   [[119]6]. Nonetheless, data have been conflicting in some cases and the
   underlying mechanisms remain unclear. To tackle this issue, we explored
   the tissue metabolic alterations and the gut microbial changes induced
   by MSG consumption.

   As illustrated in the schematic representation in [120]Figure 6, MSG
   consumption altered the metabolites associated with hepatic
   gluconeogenesis, branched-chain amino acid (BCAA) metabolism, vitamin
   B6 and trimethylamine metabolism. Moreover, the short-term MSG
   consumption suppressed the relative abundance of Bifidobacterium, which
   is considered a beneficial bacteria, and Clostridium related to TMA and
   TMAO production [[121]30].

Figure 6.

   [122]Figure 6
   [123]Open in a new tab

   Schematic illustration of changes in metabolic pathways as observed
   from tissues and biofluid metabolites analysis of Wistar rats. Red
   indicates higher relative concentration, whereas blue indicates lower
   relative concentration of metabolites in control rats compared with
   MSG-treated rats.

   In the liver, metabolites such as glucose, pyridoxine (B6),
   2-deoxyuridine, inosine, unknown 1 and 2 were significantly higher in
   MSG-treated animals than in controls, while eleven metabolites (i.e.,
   leucine, isoleucine, valine, alanine, N-acetylglycoprotein, acetone,
   1,3-dimethylurate, histamine, xanthine, kynurenate, and nicotinamide)
   were significantly lower. First, higher levels of glucose may indicate
   an increase of hepatic gluconeogenesis associated with MSG, which is in
   agreement with previously reported high plasma glucose in newborn pigs
   receiving MSG supplementation (1 g/kg body weight) [[124]31]. Second,
   the reduced levels of alanine and three BCAAs, including valine,
   leucine and isoleucine, may correlate with amino acid catabolism and
   energy metabolism in which their carbon skeletons can be used as
   precursors for gluconeogenesis (alanine, valine and isoleucine as
   glucogenic amino acids) and energy production via the tricarboxylic
   acid (TCA) cycle. In agreement with the previous report, we found that
   the degradation products of leucine and lysine, beta-hydroxyisovalerate
   and 5-aminovalerate, respectively, were significantly higher in the
   urine of MSG-treated rats [[125]13]. Third, elevated pyridoxine in the
   liver and decreased pyridoxine in the kidney of MSG-treated rats
   suggested alteration of vitamin B6 metabolism. We also found decreased
   levels of histamine, kynurenate, and nicotinamide in the liver of
   MSG-treated rats, which are the products of B6-dependent enzymes in
   histidine and tryptophan metabolism. The connection between MSG intake
   and metabolism of vitamin B6, histidine and tryptophan need to be
   further investigated. Fourth, MSG-treated rats had higher levels of
   trimethylamine (TMA) in the kidney. TMA is synthesized from dietary
   constituents, including choline, L-carnitine and betaine by the action
   of microbial enzymes [[126]16]. TMA is absorbed largely in a passive
   way into the portal circulation, and mainly oxidized to
   trimethylamine-N-oxide (TMAO) by hepatic flavin-containing
   monooxygenases (FMOs). A minor fraction of TMA is oxidized to
   dimethylamine (DMA) and methylamine (MA), and finally excreted into
   urine [[127]32,[128]33]. Higher levels of DMA and MA in the urine of
   MSG-treated rats were reported previously [[129]13]. Several studies
   have demonstrated that the elevated TMA precursors, choline, or its
   metabolites, TMAO, can lead to progressive renal tubulointerstitial
   fibrosis, cardiovascular disease, and chronic kidney disease
   [[130]34,[131]35,[132]36]. Therefore, the elevated level of TMA in the
   kidney may constitute the link between MSG intake and renal damage.

   Moving from the observed changes in TMAs, we investigated the gut
   microbiota of MSG-treated rats and demonstrated the alteration of the
   two major phyla, Firmicutes and Bacteroidetes. The higher abundance of
   Firmicutes, but lower abundance of Bacteroidetes, was observed in
   MSG-treated rats. At the genus level, MSG-treated rats had higher
   abundance of Clostridium, and lower abundance of Lactobacillus and
   Bifidobacterium. The genus Clostridium comprises a group of
   microorganisms belonging to the phylum Firmicutes. Some Clostridium
   spp. are associated with TMA metabolism by producing enzymes to convert
   choline or dietary constituents to TMA [[133]33,[134]37,[135]38]. The
   increased TMA-producing gut bacteria, i.e., Clostridium spp., supports
   the elevation of TMA metabolites in the kidney and urine of MSG-treated
   rats. Moreover, MSG consumption reduced the Bifidobacterium population,
   a probiotic bacterium which plays important roles in gut homeostasis
   and health [[136]39]. Simple and digestible carbohydrates such as
   lactose and sucrose are metabolized in the upper gastrointestinal (GI)
   tract by the host and bacteria such as lactobacilli. However,
   non-digestible carbohydrates, i.e., dietary fibre, are metabolized in
   the lower GI tract by the members of gut microbiota, including
   Bifidobacterium. A diet with high contents of saturated fats and simple
   sugars but depleted of dietary fibre such as a western-style diet may
   contribute to a lower number of beneficial bacteria [[137]40]. The
   decrease of Bifidobacterium, a good bacterium, has been reported in
   chronic inflammatory diseases such as obesity [[138]41], hepatitis B
   [[139]42], and diabetes [[140]43,[141]44]. The effect of MSG
   consumption on the gut microbiota in humans was reported earlier with
   no significant changes in microbial composition compared to the
   baseline [[142]45]. The low impact of MSG on the gut microbiota in this
   study may be due to the low dose of MSG supplementation (2 g/day),
   since the average daily MSG intake in our observation is 4 g/day
   [[143]3]. We found that every 1 g of daily MSG intake increased the
   risk of having metabolic syndrome. Although the mechanism of MSG
   consumption leading to metabolic diseases is not established, the
   decrease of Bifidobacterium in MSG-treated animals may be a clue.
   Coincidently, the decreased Bifidobacterium in MSG-treated rats found
   in present study is similar to that of rats with vitamin B6 deficiency
   [[144]46]. The intensive study of how MSG reduces good bacteria and
   alters vitamin B6 status needs further investigation.

   We showed that MSG consumption induced hepatic and renal metabolic
   changes involved in gluconeogenesis and branched-chain amino acid,
   vitamin B6 and TMA metabolism, accompanied by gut microbiota
   compositional shifts. These observations suggest that the altered
   metabolic pathways may be associated with adverse effects of long term
   of MSG consumption, especially on the liver and kidney.

Acknowledgments