Monosodium glutamate stimulates secretion of glucagon-like peptide-1 and reduces postprandial glucose after a lipid-containing meal


Correspondence to:

Prof. M. Kusano, Department of Endoscopy and Endoscopic Surgery, Gunma University Hospital 3-39-15 Showa-machi, Maebashi 371-8511, Japan.




Monosodium l-glutamate (MSG) is known to influence the endocrine system and gastrointestinal (GI) motility. The mechanism of postprandial glycemic control by food in the GI tract is mostly unknown and of great interest.


To investigate the effect of MSG on glucose homeostasis, incretin secretion and gastric emptying in humans after a lipid-containing meal.


Thirteen healthy male volunteers (mean age, 25.5 years) and with no Helicobcter pylori infection were enrolled. A 400 mL (520 kcal) liquid meal with MSG (2 g, 0.5% wt:vol) or NaCl (control) was ingested in a single-blind placebo-controlled cross-over study. Blood glucose, serum insulin, plasma glucagon, plasma glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide were measured. Gastric emptying was monitored by a 13C acetate breath test. Postprandial symptoms were assessed on a visual analogue scale.


The 30-min postprandial glucose concentration was significantly reduced by adding MSG to the test meal. The area under the glucose concentration vs. time curve (0–60 min) was also significantly reduced by adding MSG (40.6 ± 3.51 mg·1 hr/dL with MSG vs. 49.2 ± 3.86 mg·1 hr/dL with NaCl,= 0.047), whereas, the 30-min postprandial plasma GLP-1 level was significantly increased (58.1 ± 15.8 pmol/L with MSG vs. 13.4 ± 15.8 pmol/L with NaCl, = 0.035). MSG did not affect the half gastric emptying time or postprandial symptoms.


Monosodium l-glutamate improved early postprandial glycaemia after a lipid-containing liquid meal. This effect was not associated with a change in gastric emptying, but was possibly related to stimulation of glucagon-like peptide-1 secretion.


Monosodium l-glutamate (MSG) is known to bind to receptors on taste cells in the oral cavity to elicit the unique taste ‘umami’ and is used as a flavour enhancer, especially in Asia. Glutamate is also known to stimulate insulin secretion both in vitro[1] and in vivo in rats.[2, 3] In humans, oral administration of MSG increases the plasma insulin concentration both while fasting[4] and after intake of a glucose solution,[5] although the mechanism involved has not been clarified.

The mechanism of postprandial glycemic control by food components in the gastrointestinal (GI) tract is largely unknown and has attracted considerable interest. A chemosensory system that responds to glutamate has been reported to exist in the GI tract,[6] and candidate receptors for umami signalling have recently been found in the stomach and the small intestine.[7-9] It was also reported that vagal gastric afferent fibres are activated by glutamate, but not by other amino acids.[7] However, the impact of MSG on insulin secretion after a ‘standard meal’ containing a mixture of fat, carbohydrate and protein is still unclear. Moreover, the effect of MSG on incretins [glucagon-like peptide 1 (GLP-1), and glucose-dependent insulinotropic polypeptide (GIP)] has not been established yet.

Monosodium l-glutamate is also known to have a potential influence on GI motility. We previously reported that adding MSG to a protein-rich meal without lipid accelerated gastric emptying in humans.[10] In addition, the intragastric administration of MSG was reported to stimulate upper GI motility and accelerate gastric emptying in dogs, with this effect being mediated via the vagus nerve.[11]

On the other hand, the blood glucose level has been shown to influence the gastric emptying rate in both healthy volunteers[12] and patients with type 2 diabetes,[13] and gastric emptying is also markedly influenced by various other nutrients, especially lipids. To avoid rapid intestinal passage of high-calorie nutrients, gastric emptying is adjusted depending on the contents of a meal. GLP-1 is one of the most important hormones involved in glycemic homeostasis, and it also affects gastric emptying by regulating the motility of the stomach.[14-16] The present study investigated the effect of adding MSG to a lipid-containing meal on gastric emptying, glucose homeostasis and incretin secretion in healthy humans.

Materials and Methods


Thirteen healthy male volunteers with a mean age of 25.5 years (range: 19–36 years), a body mass index (BMI) of 19.5–24.6 kg/m2, and no Helicobacter pylori (H. pylori) infection were enrolled. Before the study, H. pylori infection was excluded by serologic testing (Mitsubishi Kagaku Bio-clinical Laboratories Inc., Tokyo, Japan). The subjects were required to have (i) no history of abdominal surgery (except appendectomy); (ii) no regular medications; (iii) no intake of medications during the previous week that could alter GI motility; (iv) no history of H. pylori eradication therapy; and (v) no history of cardiovascular disease, renal dysfunction, liver dysfunction or diabetes mellitus.

Study design and ethics

This randomised, single-blind, cross-over study was registered at (NCT01009658). Approval was obtained from the institutional review board of Gunma University Graduate School of Medicine. The study was performed at Gunma University Hospital in accordance with the Declaration of Helsinki, and all subjects gave written informed consent.

Experimental protocol

Examinations were performed after an overnight fast (the subjects were allowed to drink water until 30 min before the examination to avoid dehydration). Each subject ingested a liquid meal with added MSG or NaCl on two separate days in random order at least 1 week apart. The order was randomised by the enveloped method. Subjects were asked to ingest the test meal during the first 5 min of the examination and were then instructed to sit in a chair and relax.

Test meals

The liquid test meal (400 mL) consisted of 78.0 g of dextrin (TK16; Matsutani Chemical, Itami, Japan), 32.0 g of casein-calcium (Casein-calcium S; Nippon Shinyaku Co., Ltd., Kyoto, Japan), and 11.4 g of oleinic acid corresponding to 100 kcal of fat, 300 kcal of carbohydrate, and 120 kcal of protein. Monosodium l-glutamate (MSG: 2 g, 0.5% wt:vol; Happo Syokusan Co. Ltd., Osaka, Japan) or 0.63 g of sodium chloride (NaCl; equivalent to the sodium content in 2 g of MSG) was added to the test meal. The purity of MSG was certified as more than 99.0%. The dose of MSG used in these experiments was 0.5% wt:vol (equivalent to 30 mmol/l-glutamate). The test meal was also labelled with 100 mg of (13C) sodium acetate (Cambridge Isotope Laboratories Inc., Cambridge, MA, USA).


Blood glucose, serum insulin, plasma glucagon, plasma glucose-dependent insulinotropic polypeptide (GIP), and plasma glucagon-like peptide-1 (GLP-1) concentrations were measured before and 15, 30, 45, 60, 90 and 120 min after ingestion of the test meal. Glucose was measured by the glucose oxidase method with a One Touch Ultra (LifeScan, Inc., Milpitas, CA, USA). Insulin, glucagon and GIP were analysed by SRL, Inc. (Tokyo, Japan). Serum insulin concentrations were determined by a chemiluminescence enzyme immunoassay and the hexokinase method. Plasma glucagon and GIP concentrations were determined by radioimmunoassay and enzyme-linked immunosorbent assay (ELISA) respectively. Blood (2 mL) for the GLP-1 assay was collected into ice-cooled EDTA tubes and immediately treated with 20 μL of dipeptidyl peptidase IV inhibitor (Cosmic Corporation, Tokyo, Japan). Blood for analysis of glucagon and GIP was collected into tubes containing EDTA. GLP-1 was determined by an ELISA that measured two of its forms (the 7–36 amide and 9–36 amide) (Yanaihara Institute Inc., Shizuoka, Japan).[17, 18] Blood samples were immediately centrifuged for 10 min at 855.3 g at −4°C to obtain plasma and all samples were frozen and stored at −80°C until analysis.

Assessment of gastric emptying

The gastric emptying rate was determined by the continuous 13C acetate breath test[19, 20] using a Breath ID System (Exalenz Bioscience Ltd, Modiin, Israel). Breath samples were obtained automatically and continually through a nasal cannula for 240 min after ingestion of the test meal. The total number of breath sampled ranged between 90 and 100.

Gastric emptying was evaluated by calculation of the half gastric excretion (emptying) time (t1/2ex), which is the time when half of the 13CO2 dose in the test meal has been excreted during cumulative 13CO2 excretion until infinity. The value of t1/2ex and the scintigraphic half emptying time are not identical, but there is a linear correlation between t1/2ex determined by the 13C breath test and the scintigraphic half emptying time.[21, 22]

Assessment of postprandial sensations and symptoms

The severity of postprandial sensations and symptoms was self-assessed by the subjects using a 100-mm visual analogue scale (VAS) before and 5, 15, 30, 45, 60, 90, 120, 150, 180, 210 and 240 min after ingestion of the test meal. Sensations and symptoms were rated on the VAS by each subject, with 0 corresponding to absence of a sensation or symptom and 100 being an unbearably severe symptom. They assessed eight sensations or symptoms, fullness, satiation, bloating, nausea, belching, epigastric pain, epigastric burning and heartburn. The maximum score and cumulative score were evaluated for each sensation or symptom.

Data and statistical analysis

All results are presented as the mean ± S.E. Increments in the concentrations of insulin, glucose, glucagon, GIP and GLP-1 over the fasting level (baseline) were calculated. To evaluate the early responses of glucose and insulin relative to each other, the ΔInsulin 0–30/ΔGlucose 0–30 ratio was calculated as follows: (the concentration of insulin at 30 min postprandially − the baseline concentration of insulin)/(the concentration of glucose at 30 min postprandially − the baseline concentration of glucose). Ratios were also calculated for glucose vs. other hormones (glucagon, GLP-1, and GIP). The areas under the concentration vs. time curve (AUCs) for blood glucose and each hormone were evaluated over appropriate time periods. Parameters of gastric emptying, the total secretion (i.e. the AUC) of blood glucose and each hormone, and indices of the early response of each hormone were analysed with Student's t-test. The peak time for each parameter was defined as the time when the maximum concentration was detected, and peak times were compared by the Wilcoxon signed rank test. Comparison of the profiles of serum insulin, plasma glucose, plasma GIP and plasma glucagon between the two test meals (with MSG or with NaCl) was done by two-way analysis of variance (anova), with post hoc contrast analysis. All statistical analyses were performed with spss 19 software for Windows (SPSS, Inc., Chicago, IL, USA). Differences were accepted as significant at P < 0.05.


All thirteen subjects completed both parts of the study with no side effects. None of the subjects could differentiate the two meals appropriately by their taste.

Blood glucose

As shown in Figure 1a, after ingestion of either test meal, the blood glucose concentration increased rapidly. The increase of blood glucose was significantly smaller at 30 min after the test meal with MSG (58.8 ± 4.43 mg/dL) than after that with NaCl (68.2 ± 4.30 mg/dL, = 0.045). Although the difference was not significant at 45 min, the blood glucose concentration was lower with MSG (53.6 ± 6.13 mg/dL vs. 66.4 ± 6.76 mg/dL, P = 0.080). The AUC data obtained after the test meals are shown in Figure 1b. The area under the concentration vs. time curve from 0 to 60 min [AUC (0–60)] for blood glucose was significantly reduced by MSG (40.6 ± 3.51 mg·1 h/dL with MSG vs. 49.2 ± 3.86 mg·1 h/dL with NaCl, = 0.047), although the AUC (0–240) of glucose was not altered by adding MSG to the test meal (63.3 ± 11.8 mg·4 h/dL with MSG vs. 76.8 ± 14.1 mg·4 h/dL with NaCl, P = 0.364).

Figure 1.

(a) Glucose concentration profile after ingestion of the test meals. The increment over the fasting level (baseline) is shown in each graph (mean ± S.E.). *< 0.05; MSG, monosodium l-glutamate. The increase in blood glucose was significantly smaller at 30 min after the meal with MSG (58.8 ± 4.43 mg/dL with MSG vs. 68.2 ± 4.30 mg/dL with NaCl,= 0.045 at 30 min postprandially, 53.6 ± 6.13 mg/dL vs. 66.4 ± 6.76 mg/dL, P = 0.080 at 45 min). (b) AUC of glucose for each 60-min period. *P < 0.05; each point represents the AUCs for each individual. The AUC (0–60) was significantly decreased by MSG (40.6 ± 3.51 mg·1 h/dL with MSG vs. 49.2 ± 3.86 mg·1 h/dL with NaCl, P = 0.047), although the AUC (60–120) was not influenced.


Plasma GLP-1 reached a peak at 30 min postprandially and then decreased rapidly after the test meal containing MSG, whereas GLP-1 increased gradually after the meal containing NaCl (Figure 2a). The GLP-1 concentration was significantly higher at 30 min after the test meal containing MSG (58.1 ± 15.8 pmol/L) than after that containing NaCl (13.4 ± 15.8 pmol/L, = 0.035).

Figure 2.

Profile of plasma GLP-1 (a) and serum insulin (b) after ingestion of the test meals. The increment over the fasting level (baseline) is shown in each graph (mean ± S.E.). The increase in GLP-1 was significantly greater at 30 min after the meal with MSG (58.1 ± 15.8 pmol/L with MSG vs. 13.4 ± 15.8 pmol/L with NaCl, = 0.035). There is no significant difference of insulin between MSG and NaCl. *< 0.05; MSG, monosodium l-glutamate.


After ingestion of the test meal, the serum insulin concentration increased rapidly to reach a peak (Figure 2b). Although the serum insulin concentration declined more slowly from the peak value after the test meal with NaCl, there was no significant difference between the two test meals. The AUC (0–120) for insulin, representing total insulin secretion, was not altered by MSG (80.0 ± 7.5 mg·4 h/dL with MSG vs. 85.1 ± 7.2 mg·4 h/dL with NaCl, = 0.51), and neither were the AUCs for each 60 min period.

Factors associated with early postprandial glucose levels

Figure 3 shows the indices of hormonal secretion in the early postprandial period. The ratio of the increment of GLP-1 to the increment of glucose during the first 30 min (ΔGLP-1 0–30/ΔGlucose 0–30) was significantly increased by MSG (0.94 ± 0.3 with MSG vs. 0.21 ± 0.2 with NaCl). In contrast, the values of the ΔInsulin 0–30/ΔGlucose 0–30, ΔGlucagon 0–30/ΔGlucose 0–30 and ΔGIP 0–30/ΔGlucose 0–30 ratios were not affected by MSG.

Figure 3.

Factors associated with glucose in the early postprandial period. Ratios for the increment of each hormone relative to the increment of glucose during the first 30 min after intake of the test meal. For example, ΔInsulin 0–30/ΔGlucose 0–30 = (the concentration of insulin at 30 min postprandially − the baseline concentration of insulin)/(the concentration of glucose at 30 min postprandially − the baseline concentration of glucose). ΔGLP-1 0–30/ΔGlucose 0–30 was significantly increased by MSG (0.94 ± 0.3 with MSG vs. 0.21 ± 0.2 with NaCl, P = 0.047) (a). In contrast, ΔInsulin 0–30/ΔGlucose 0–30 (b), ΔGlucagon 0–30/ΔGlucose 0–30 (c) and ΔGIP 0–30/ΔGlucose 0–30 (d) were not affected by MSG. *< 0.05; MSG, monosodium l-glutamate.

Correlations between GLP-1 and other hormones

The correlations between the concentrations of GLP-1 and other hormones at 30 min were evaluated. It was found that serum insulin, plasma glucagon and plasma GIP levels were not significantly correlated with GLP-1(Figure S1).

Gastric emptying

Figure 4 shows the profile of 13CO2 excretion after ingestion of the test meals, which is considered to be the gastric emptying rate, was almost completely identical to one another. The t1/2ex after the test meal with MSG (135.5 ± 6.8 min) was not significantly different neither from that after the test meal with NaCl (125.6 ± 4.5 min) (P = 0.131).

Figure 4.

Effect of MSG on gastric emptying after meal intake. Curves of 13CO2 excretion (%dose/h) after ingestion were not altered by MSG. MSG, monosodium l-glutamate.

Postprandial sensations and symptoms

Each sensation and symptom rapidly became more severe after intake of either test meal and most symptoms peaked at 5 min (or up to 15 min), followed by a gradual return to baseline (Figure 5). One of the 13 participants assigned very high scores for almost all of his symptoms. We excluded these scores from analysis because the values were more than two standard deviations above the mean in every case. Fullness was the most commonly reported sensation or symptom. The score for each symptom at any time point and the AUC (0–240) of each symptom were not altered by adding MSG to the test meal and were also not related to the gastric emptying (t1/2).

Figure 5.

Effect of MSG on postprandial sensations and symptoms. The increment over the fasting level (baseline) is shown in each graph. Most of the postprandial sensations and symptoms increased immediately after the intake of either test meal, peaked by 5–15 min, and then gradually returned to baseline. The score for each symptom was not altered by MSG at any time. MSG, monosodium l-glutamate.


The present study revealed that the postprandial glucose level was significantly reduced by MSG at 30 min after the test meal and the AUC for 0–60 min was also significantly reduced by MSG. In addition, MSG significantly increased the concentration of GLP-1 at 30 min after the test meal, and the ΔGLP-1 0–30/ΔGlucose 0–30 ratio (an estimate of relative early GLP-1 secretion) was also significantly increased by MSG. These results suggest that addition of MSG amplified the response of GLP-1 to the test meal.

Glutamine is an amino acid formed from glutamate that potently stimulates GLP-1 secretion by GLUTag cells (a GLP-1-secreting cell line),[23] and Reimann et al. noted that glutamate also stimulates the release of GLP-1 from these cells. The same group also recently reported that glutamine reduces postprandial glycemia and augments the response of GLP-1 to meals in type 2 diabetes.[24, 25] In the present study, we showed that GLP-1 secretion was greater in the early postprandial period after the test meal with MSG, whereas GLP-1 secretion increased gradually after the test meal with NaCl. Both test meals had the same calorie content and the same nutrients, except for 2 g of MSG or NaCl, so the change of GLP-1 secretion was considered to be related to the addition of this small amount of MSG. The postprandial insulin and GLP-1 responses after intake of MSG were similar to those after 30 g of glutamine in a previous study.[25]

GLP-1 is secreted by intestinal endocrine L cells, which are mainly located in the distal ileum and colon.[26] In contrast, GIP is released by intestinal K-cells that are located in more proximal regions (the duodenum and jejunum) of the small intestine. GLP-1 secretion shows a biphasic pattern, with initial rapid release after oral nutrient ingestion (early phase) being followed by a longer late phase.[27] Because the majority of L cells are located in the distal small intestine, the early phase of GLP-1 secretion may be mediated by stimuli other than direct contact of nutrients with L cells. Indeed, several studies have suggested some candidates, including the autonomic nervous system and GIP.[26] However, the effect of MSG on GIP secretion has not attracted much attention previously, and the present study revealed that there was no difference of GIP secretion after the two test meals.

The role of the vagus nerve as an important mediator of nutrient-induced GLP-1 secretion has been established by studies in rats, which showed that bilateral subdiaphragmatic vagotomy completely blocks fat-induced GLP-1 secretion, while direct electrical stimulation of the coeliac branches of the vagus increases GLP-1 secretion.[28] Existence of glutamate receptors has been confirmed in the upper GI tract of rats, mice and humans.[7-9, 29, 30] Also, ingestion of glutamate into the stomach increases vagal afferent and efferent activity in rats,[7] and evokes forebrain activity in rats.[31] Thus, glutamate might have an important role in the sensing of nutrients in the GI tract and may be involved in the gut-brain axis.

Although our study did not find any correlation between GLP-1 and insulin in the early postprandial period, a change of GLP-1 secretion was shown to be one of the key mechanisms for reducing postprandial glucose. We also found that the insulin peak time was significantly shortened by addition of MSG to the test meal (39.2 ± 5.2 min with MSG vs. 53.1 ± 7.3 min with NaCl, = 0.048) (data partially shown in Figure 2b). This suggests that intake of MSG might have altered the profile of postprandial insulin secretion by more rapid release of insulin due to promotion of GLP-1 secretion, but the change was not prominent enough to be detected by sampling at 15-min interval.

Other than its main effect of stimulating insulin secretion, GLP-1 has various effects on extrapancreal tissues. GLP-1 is expected to improve insulin sensitivity, however, only repeated administration of GLP-1 receptor agonist has been demonstrated. Some studies have proposed that GLP-1 appears as enhanced glucose disposal in healthy and type 2 diabetic humans independently of islet hormone action.[32-34] Furthermore, GLP-1 may suppress hepatic glucose production.[35] Some of these effects by GLP-1 are plausible ways to reduce early postprandial glucose levels.

Glutamate is known to stimulate insulin secretion.[2-4, 36] Graham et al.[4] reported that an increase in insulin occurs immediately before the increase in plasma glutamate after oral ingestion of MSG, and they proposed that a signalling pathway mediating the response of insulin exists in the GI tract. The early response of serum insulin in the present study supports the proposal of Graham et al.,[4] and GLP-1 could possibly be a key player in the signalling pathway.

The clinical significance of this study is that adding glutamate to the test meal had an influence on sensing nutrients, glycemic homeostasis and the gut-brain axis, suggesting that MSG could potentially improve glucose tolerance after a lipid-containing high-calorie meal.

Because the capability of GLP-1 secretion of patients with type 2 diabetes is controversial, it remains conjectural about the effect of MSG on postprandial glucose levels in these patients. However, something might be needed to change the way of ingestion to maximise its effect to put to practical use in diabetic patients. For example, the timing of MSG ingestion or co-ingestion with nucleotide monophosphates might be useful. Protein preload was reported effective to lower postprandial glycemia in type II diabetes patients, increasing the early postprandial GLP-1 secretion.[37] Alternatively, co-ingestion of glutamate and inositol monophosphate might be effective because inositol monophosphate and glutamate have a synergistic effect on umami-taste sensation,[38] and co-ingestion of the two compounds enhanced GLP-2 secretion into rat portal veins.[39]

In this study, MSG did not significantly accelerate gastric emptying after the lipid-containing test meal. We previously reported that adding MSG to a protein-rich meal accelerated gastric emptying, while there was no effect of adding MSG to a pure carbohydrate meal.[10] In general, lipids strongly inhibit gastric emptying,[40, 41] so the lipid content of the test meal in the present study might have overcome the potential of MSG to accelerate gastric emptying. The effect of MSG on gastric juice secretion has been reported to be dependent on the nutritional composition of the test meal in dogs.[42] Thus, glutamate not only accelerates GI motility but also has the potential to regulate glycemic homeostasis in a manner dependent on the co-existing nutrients in a meal.

Some limitations of this study need to be considered. The concentrations of hormones were only measured at 15-min interval due to technical limitations, so we were unable to identify differences in the peak time of insulin secretion. Also, we did not examine the effect of different doses of MSG, because the dose selected was very similar to that in daily meals in Japan.

In conclusion, MSG reduced the glucose level in the early postprandial period after intake of a lipid-containing liquid meal, without altering total insulin secretion. This effect was not associated with a change of gastric emptying, but was possibly related to promotion of GLP-1 secretion by MSG. In addition to its effect on GI motility reported in previous studies, glutamate could be an important regulator of glycemic homeostasis.


Dextrin and casein-calcium used in this study were kind gifts from Matsutani Chemical (Itami, Japan) and Nippon Shinyaku Co., Ltd (Kyoto, Japan). Declaration of personal and funding interests: None.