Leuconostoc mesenteroides fermentation produces butyric acid and mediates Ffar2 to regulate blood glucose and insulin in type 1 diabetic mice.

Type 1 diabetic patients have lower counts of butyric acid-producing bacteria in the dysbiotic gut microbiome. In this study, we demonstrate that a butyric acid-producing Leuconostoc mesenteroides (L. mesenteroides) EH-1 strain isolated from Mongolian curd cheese can reduce blood glucose and IL-6 in the type 1 diabetic mouse model. L. mesenteroides EH-1 fermentation yielded high concentrations of butyric acid both in vitro and in vivo. Butyric acid or L. mesenteroides EH-1 increased the amounts of insulin in Min6 cell culture and streptozotocin (STZ)-induced diabetic mice. Inhibition or siRNA knockdown of free fatty acid receptor 2 (Ffar2) considerably reduced the anti-diabetic effect of probiotic L. mesenteroides EH-1 or butyric acid by lowering the level of blood glucose. We here demonstrate that Ffar2 mediated the effects of L. mesenteroides EH-1 and butryic acid on regulation of blood glucose and insulin in type 1 diabetic mice.


Methods
Bacterial culture and identification. Mongolian curd cheese was homogenized in 500 µL of sterile PBS with a grinder. Bacteria in the homogenate were cultured by plating on a tryptic soy broth (TSB) (Sigma, St. Louis, MO, USA) agar plate and incubated for 3 days at 37 °C. Sequence analysis of 16S ribosomal RNA (rRNA) genes was utilized for bacterial identification 30 . A single colony of bacteria from a TSB agar plate was isolated with a sterile toothpick and boiled at 100 °C for DNA extraction. Identification of L. mesenteroides EH-1 strain was validated by rRNA sequencing using the 16S rRNA 27F and 534R primers for polymerase chain reaction (PCR) (Supplementary Fig. S1) 31 . The 16S rRNA gene sequences were analyzed using the basic local alignment search tool (BLASTn, National Library of Medicine 8600 Rockville Pike, Bethesda, MD, USA). L. mesenteroides was cultured in TSB (Sigma) overnight at 37 °C. The cultures were diluted 1:100 and cultured to an optical density 600 nm (OD 600 ) = 1.0. Bacteria were harvested by centrifugation at 5000 rpm for 10 min, washed with PBS, and suspended in PBS for further experiments.
Glucose fermentation of L. mesenteroides EH-1. To induce fermentation, L. mesenteroides EH-1 [10 7 colony-forming unit (CFU)/mL] was incubated in rich media [10 g/L yeast extract (Biokar Diagnostics, Beauvais, France), 5 g/L TSB, 2.5 g/L K 2 HPO 4 and 1.5 g/L KH 2 PO 4 ] in the absence or presence of 20 g/L (2%) glucose at 37 °C for 24 h. Rich media or rich media plus 20 g/L glucose without bacteria were included as a control. Phenol red [0.001% (w/v), Sigma] in rich media with 20 g/L glucose served as an indicator of fermentation, converting from red-orange to yellow when fermentation occurred. High performance liquid chromatography (HPLC) was used to quantify the level of butyric acid in cultured media.
Streptozotocin (STZ)-induced type 1 diabetic mice. The Institute Cancer Research (ICR) mice (8)(9)(10)(11)(12) week-old males; National Laboratory Animal Center, Taiwan) were housed at 25 °C with a 12:12 h light-dark cycle, fed a normal chow diet, and given water ad libitum. Mice (n = 4/group) were acclimatized for 5-7 days before the experiment. To induce a rapid ablation of the beta cells and hyperglycemia 32 and avoid the interruption of STZ with L. mesenteroides EH-1 or butyric acid, mice were injected with a single dose, instead of low multiple doses 33 , of STZ. Diabetes was induced following an 8-h fast using a single intraperitoneal (IP) injection of STZ (200 mg/kg body weight) (Sigma) 32 , which was dissolved in acidified citrate buffer (0.1 mol/L, pH 4.5). Two days later, after a 4-h-fast, the level of blood glucose from the tail blood was measured using a glucometer (Advantage Glucometer, Roche, Mannheim, Germany). Injection of STZ for 2 days induced a weight lost (32.8 ± 0.9 vs 29.6 ± 1.2 mg; with vs without STZ). Mice with glucose levels of 200 mg/dL or greater were recruited into the diabetic group 34 . This research was carried out in strict accordance with an approved Institutional Animal Care and Use Committee (IACUC) protocol at National Central University (NCU), Taiwan (NCU-106-015, 19 December 2017). Fasting retro-orbital sinus blood was collected in heparinized tubes and then centrifuged at 3,000 rpm for 15 min. Plasma was stored at −80 °C until use. A mouse IL-6 ELISA kit (R&D systems, Minneapolis, MN, USA) was used to detect the levels of IL-6. The level of insulin was detected using a mouse insulin ELISA kit (Mercodia, Uppsala, Sweden) ( Supplementary Fig. S4).
Diabetic mice treated with butyric acid. Butyric acid at 4 mmol/L in water (5 mL/kg body weight) was administrated to diabetic mice via IP injection. Mice in control groups received water alone. Blood glucose was detected every day. Seven days after butyric acid injection, fasting blood was collected for detection of insulin and IL-6 as described above. For Ffar2 inhibition, a Ffar2 antagonist (GLPG-0974, Tocris Bioscience, Bristol, UK) was dissolved in dimethylsulfoxide (DMSO) to make a stock solution. Diabetic mice were injected with butyric acid (5 mL/kg body weight) taken from a stock solution of 4 mmol/L. GLPG-0974 (1 mg/kg body weight) was diluted in saline then was given at 1 mL/kg body weight 35 by gastric gavage just before butyric acid injection. DMSO Feeding mice with L. mesenteroides EH-1. ICR mice were fed with live or heat (100 °C)-killed L. mesenteroides EH-1 (8 × 10 9 CFU/50 µL) once a day for 2 days. Water (50 µL) was given as control. Two days after the L. mesenteroides EH-1 feeding, cecum was homogenized in water (50 mg/500 µL), followed by a vortex mixing step, and stored at −80 °C until the detection of butyric acid by HPLC. Two weeks after daily feeding diabetic mice with L. mesenteroides EH-1, blood glucose was detected once a week, fasting blood was collected for the detection of insulin and IL-6 levels. Mice were sacrificed in a CO 2 chamber and the pancreases were collected for immunohistochemical analysis. For Ffar2 inhibition, diabetic mice were fed with L. mesenteroides EH-1 (8 × 10 9 CFU/50 µL) once a day for 2 weeks. GLPG-0974 was given at 1 mL/kg body weight by gastric gavage just before L. mesenteroides EH-1 administration and weekly. A vehicle was 0.1% DMSO in saline. Blood glucose was detected once a week. Two weeks after treatment, blood was collected for detection of insulin and IL-6 levels.

Statistical analysis.
To determine significance between groups, comparisons were made using the two-tailed Student's t-test. Data are presented as mean values ± standard deviation (SD). The mean values ± SD for all figures with bar charts were shown in Supplementary Table 1. Statistical analyses were performed using GraphPad Prism 5 software. Unpaired Student's t-test was used to compare two groups. When appropriate, ANOVA was used and post hoc analysis was performed with Tukey's test to compare more than two groups. A p value < 0.05 was regarded as statistically significant.

Results
fermentation properties of L. mesenteroides EH-1. A single colony was isolated from a TSB agar plate spread with Mongolian curd cheese and evaluated by 16S rRNA sequencing. The 16S rRNA gene (Supplementary Fig. S1) of this colony shares 99% identity to that of L. mesenteroides ATCC 8293. This isolated strain was here named as L. mesenteroides EH-1. L. mesenteroides is a lactic acid bacterium that is currently used as a starter for kimchi and kefir 36 . L. mesenteroides EH-1 grew well at the temperatures of 25 °C and 37 °C, but not 4 °C ( Supplementary Fig. S2). The growth of L. mesenteroides EH-1 was unaffected by acidification of the media, as growth curves were similar from pH 3 to pH 7 ( Supplementary Fig. S3). These results indicate that the strain of L. mesenteroides EH-1 isolated from Mongolian curd cheese is stable at room temperature and tolerant of low pH. To examine the fermentative capabilities, L. mesenteroides EH-1 was cultured in rich media in the presence of 2% glucose for 24 h. Rich media with glucose alone or L. mesenteroides EH-1 alone served as controls. The media in the culture of L. mesenteroides EH-1 with glucose turned yellow after incubation for 24 h, while the media in the other three conditions maintained their original colors (Fig. 1a). As shown in Fig. 1b,c, the OD 562 and pH values of media with L. mesenteroides EH-1 plus glucose demonstrated significant decreases compared to controls, indicating that L. mesenteroides EH-1 has a capability of fermenting glucose. HPLC analysis was conducted to quantify the level of butyric acid in fermentation media of L. mesenteroides EH-1. Butyric acid is detectable in media from glucose fermentation of L. mesenteroides EH-1, but not media from controls (Fig. 1d). The different concentrations of butyric acid (0-100 mmol/L) were subjected to HPLC for establishment of a quantitative standard curve. As shown in Fig. 1e, glucose fermentation of L. mesenteroides EH-1 for 24 h yielded approximately 1.6 mmol/L of butyric acid.

Effects of butyric acid on insulin secretion from Min6 cells.
To investigate the effect of butyric acid on insulin secretion, Min6 cells were treated with butyric acid (100 mmol/L) for 24 h. Treatment of glucose or water served as positive and negative controls, respectively. Results in Fig. 2a showed that treatment of cells with butyric acid, like glucose, markedly elevated insulin levels in culture media of Min6 cells. To determine whether Ffar2 mediated the regulation of butyric acid on insulin secretion, cells were pre-treated with Ffar2 or scrambled siRNA before addition of water alone, glucose alone, or butyric acid plus glucose for 24 h. The Ffar2 siRNA induced a 74.46 ± 2.27% knockdown of Ffar2 gene ( Supplementary Fig. S5). As shown in Fig. 2b, the knockdown of Ffar2 with its specific siRNA, but not scrambled siRNA, considerably blocked the effect of butyric acid on induction of insulin secretion. On the other hand, the Ffar2 knockdown had no influence on glucose-induced insulin secretion from Min6 cells. The results suggest that Ffar2 is essential for the action of butyric acid at induction of insulin secretion from Min6 cells.

Involvement of Ffar2 in the effect of butyric acid on the levels of blood glucose, and insulin in type 1 diabetic mice.
To establish a type 1 diabetic mouse model, STZ was administered to ICR mice via IP injection. Compared to control mice, injection of STZ for two days led to higher levels of glucose and IL-6 in the blood. STZ-induced diabetic mice were injected intraperitoneally with butyric acid or water once a day for a week. Injection of butyric acid resulted in a remarkable decrease in fasting blood glucose (Fig. 3a) as well as IL-6 ( Fig. 3c). Furthermore, the amount of insulin in the plasma of butyric acid-injected mice was higher than that of water-injected mice (Fig. 3b). To further confirm the essential role of Ffar2 in mediating the action of butyric acid in vivo, the STZ-induced diabetic mice were administered intragastrically with GLPG-0974, an Ffar2 antagonist, or DMSO control before injection of butyric acid or water. Administration of GLPG-0974, but not DMSO, counteracted the effect of butyric acid on the down-regulation of glucose and the up-regulation of insulin (Fig. 3d,e), although no effect on butyric acid-induced IL-6 reduction was observed (Fig. 3f). These results indicate that butyric acid may regulate the levels of glucose and insulin in blood of STZ-induced diabetic mice via binding to Ffar2. production of butyric acid by L. mesenteroides EH-1 in vivo. To explore whether butyric acid can be produced in vivo by L. mesenteroides EH-1, ICR mice were fed with L. mesenteroides EH-1 or water once a day for 2 days. Butyric acid in cecum homogenates was measured by HPLC. Butyric acid is detectable in the cecum  www.nature.com/scientificreports www.nature.com/scientificreports/ of L. mesenteroides EH-1-fed mice, but not in that of control mice fed with water ( Fig. 4a,b). Approximately 0.7 mmol/L butyric acid was detected in the cecum, suggesting that L. mesenteroides EH-1 can produce butyric acid at a high concentration in a cecum microenvironment. To assess whether butyric acid-producing L. mesenteroides EH-1 can lower the blood glucose in diabetes, STZ-induced diabetic mice were fed with L. mesenteroides EH-1 or water. Feeding mice with L. mesenteroides EH-1 once a day for 2 weeks substantially reduced the levels of glucose (Fig. 4c) and IL-6 in the blood (Fig. 4e) and increased the amounts of insulin (Fig. 4d) in the blood and pancreas (Fig. 4d, inserted panels). These results demonstrate the probiotic activity of L. mesenteroides EH-1 in regulating the levels of glucose and IL-6 in STZ-induced type 1 diabetic mice. To validate the contribution of Ffar2 to the effect of butyric acid-producing L. mesenteroides EH-1 on lowering blood glucose, GLPG-0974 was given to STZ-induced diabetic mice to antagonize the Ffar2 before feeding mice with L. mesenteroides EH-1. Compared to mice treated with DMSO control, mice given GLPG-0974 displayed no difference in glucose (Fig. 5a), IL-6 ( Fig. 5c), as well as insulin (Fig. 5c) in blood. Taken together, the data in Figs. 3 and 5 strongly suggest that Ffar2 mediates the signaling of butyric acid produced by L. mesenteroides EH-1 to diminish the elevated glucose in STZ-induced diabetic mice.

Discussion
Probiotic L. mesenteroides is currently used for food fermentation. An addition of L. mesenteroides to cabbage fermentation ensured that texture and flavor quality were retained, while providing a 50% reduction in sodium chloride 37 . In addition, this probiotic bacterium has demonstrated a number of beneficial effects, including inhibition of three fish pathogens 38 , suppression of low-pathogenic avian influenza (H9N2) virus in chickens 39 , and reduction of Streptococcus thermophilus induced IL-12 and interferon (IFN)-γ production in human peripheral blood mononuclear cells 40 . Previous studies have identified potential antidiabetic effects of other commensal bacteria. Treatment with Lactobacillus casei CCFM419 improved impaired pancreatic function and attenuated type 2 diabetes 41 in a mouse model. Carrot juice fermentation of Lactobacillus plantarum NCU116 reduced pancreatic injuries 42 . Oral administration Lactobacillus rhamnosus CCFM0528 improved glucose intolerance by protecting islet cells 43 . Our current study demonstrated for the first time that L. mesenteroides EH-1 exerts a probiotic activity that lowers blood glucose in the STZ-induced type 1 diabetic mouse model. To validate the essential role of L. mesenteroides EH-1 in regulation of glucose and insulin levels in diabetic mice, we fed STZ-induced type 1 diabetic mice with heat-killed L. mesenteroides EH-1. As shown in Supplementary Fig. S7, the killed L. mesenteroides EH-1 bacteria lost their capacity to lower glucose and increase insulin, indicating that L. mesenteroides EH-1, not endogenous gut microbes indirectly affected by probiotic L. mesenteroides EH-1, exerted the anti-diabetic properties. Future works will use germ-free mice 44 or mice with the gut microbiome depleted by antibiotics 45 to study the effects of probiotic bacteria and endogenous gut microbes on the regulation of blood glucose and insulin.
Ffar2 is highly expressed on pancreatic islets [46][47][48] and up-regulated in islets during pregnancy 49 . The potency rank order of SCFAs for Ffar2 is acetate (C2) ∼ propionate (C3) > butyrate (C4) > valerate (C5) > formate 50 . Moreover, Ffar2 is a novel effector of glucose homeostasis in part due to its direct effect on insulin secretion and beta cell proliferation 51,52 . In agreement with previous studies, our data revealed that treatment of Min6 cells with 100 mmol/L butyric acid for 24 h provoked the insulin secretion (Fig. 2a). By conducting the oral glucose tolerance test (OGTT) in diabetic mice ( Supplementary Fig. 8), we found that injection of butyric acid, but not water, before oral administration of glucose promoted the metabolism of plasma glucose and elevated the level of insulin in the blood. The high level of insulin in the blood could be due to an increase in insulin secretion from pancreas or a decrease in insulin clearance in the liver 53 . It has been reported that sodium butyrate supplementary diet down regulated insulin receptor (IR), and IR substrate 1 (IRS-1) expression 54 involved in insulin signaling in the mouse liver 55 . Application of butyrate also associated with decreased insulin receptor beta subunit (IR-bata) www.nature.com/scientificreports www.nature.com/scientificreports/ expression of in hepatic tissue 56 . Thus, administration of butyric acid into mice may enhance the insulin secretion and prolong the insulin clearance in mice.
Butyric acid is one of Ffar2's known agonists. Previous studies have shown that Ffar2 agonism can trigger an increase in intracellular inositol triphosphate and Ca 2+ levels, and potentiate insulin secretion 52 . In addition, sodium butyrate treatment improved glucose homeostasis and reduced beta cell apoptosis in diabetic rats 25 . Knockdown of Ffar2 significantly reduced butyric acid-induced insulin secretion from Min6 cells, clearly illustrating that Ffar2 signaling is an important effector of insulin secretion induced by butyric acid. Previous studies have demonstrated a strong reduction in plasma glucose by feeding with ketogenic diets. Both propionic acid and butyric acid 57 are ketogenic substrates and can bind to Ffar2 50 . The effects of propionic acid and butyric acid on insulin secretion from Min6 cells were compared side-by-side. Propionic acid induced detectable amounts of insulin secreted from Min6 cells although it is less effective than butyric acid ( Supplementary Fig. S6). Knockdown of Ffar2 considerably lowered propionic acid-induced insulin secretion from Min6 cells. Butyric acid, but not propionic acid, is a potent inhibitor of histone deacetylases (HDAC) 58 . Previous studies demonstrated that butyrate can improve insulin sensitivity via HDAC inhibition 24 . Butyrate as a dietary supplement can prevent high fat diet-induced insulin resistance in mice by promotion of energy expenditure and induction of mitochondria function 24 . Fasting insulin was significantly lower in the butyrate-treated high-fat diet mice. The signal of phosphorylation of IRS-1 in the skeletal muscle was increased in butyrate-treated mice, suggesting a molecular mechanism of insulin sensitization. Our data revealed that butyric acid can mediate Ffar2 to increase blood insulin in STZ-induced type 1 diabetic mice. Thus, it is worth investigating whether butyric acid controls the activity of Ffar2 or HDAC to regulate the insulin secretion in different types of diabetes.
Loss of Ffar2 in mice increases the risk of diabetic status, since Ffar2 knockout (KO) mice exhibit fasting hyperglycemia, reduced insulin levels, and glucose intolerance, despite exhibiting normal insulin sensitivity 52 . Exposing Ffar2 KO mice to a high fat diet resulted in a decrease in both islet number and size, leading to reduced beta cell mass and total pancreatic insulin content. A butyrate-enriched diet could partially protect Ffar2 KO mice in a non-obese diabetic background from type 1 diabetic islet inflammation 59 . As shown in Fig. 3, a single dose of GLPG-0974, a Ffar2 antagonist, suppressed butyric acid-induced increase of fasting blood insulin and decreased fasting blood glucose, confirming the indispensable role of Ffar2 in the action of butyric acid in vivo. Furthermore, Ffar2 activation on intestinal enteroendocrine cells induces the production of glucagon-like peptide www.nature.com/scientificreports www.nature.com/scientificreports/ (GLP)-1 which is anorexigenic and stimulates insulin secretion 60 . Future work is required to determine if GLP-1 mediates the Ffar2-butyric acid induced insulin secretion in the STZ-induced type 1 diabetic mouse model. STZ specifically damages pancreatic islet beta cells. The damaged beta cells passively relate to high mobility group box 1 (HMGB1) release 61 . In parallel, the inflammatory cells infiltrated pancreatic islets, such as macrophages, dendritic cells and T cells. The differentiation of Naïve CD4+ T cells into effector T helper (Th)1 and/or Th17 cells based on current cytokine microenvironment. In addition, the released HMGB1 targets macrophage or dendritic cells via the corresponding surface receptor(s), which induces a cascade signal that activates the nuclear factor κB (NF-κB) pathway 62 . This results in the activation of Th17 cells, therefore leading to the production of proinflammatory cytokines IL-1β, TNF-α, and IL-6 63 . It has been reported that IL-6 induces beta cell apoptosis via signal transducer and activator of transcription (STAT)-3-mediated the production of nitric oxide 64 . The dysregulation of proinflammatory cytokines and Th1/Th2/Th17 further accelerate an inflammation of islets and destruction of beta cells, and leads to type 1 diabetes. Sodium butyrate can heal the balance of Th1/Th2 and block Th17 cells 65 .
In this study, we found that the acute blocking of Ffar2 by GLPG-0974 at a single dose for 24 h showed a minor effect of Ffar2-butyric acid signaling on IL-6 reduction (Fig. 3f), whereas the long term blocking of Ffar2 using multiple doses of GLPG-0974 for two weeks resulted in a significant the suppression effect on the L. mesenteroides EH-1-reduced blood IL-6 level in type 1 diabetic mice (Fig. 5c). Although other Ffar2 antagonists from azetidine derivatives 66 at different doses can be used to completely block Ffar2 in vivo, it has been documented that sodium butyrate, as a direct HMGB1 antagonist, could down-regulate the expression of HMGB1 and mediate the balance of Th1/Th2/Th17 paradigm, thus attenuating type 1 diabetes. Thus, it is possible that butyric acid reduced IL-6 production by directly down-regulating HMGB1 expression and bypassing Ffar2.
STZ does not influence the function of pancreatic beta cells of humans when used in the treatment of islet-cell carcinomas 67 . Insulin can be detected in STZ-injected mice (Fig. 4d), indicating that STZ injection results in incomplete damage to the pancreas in mice. It has been shown that Ffar2 directly mediates both the stimulatory effects of sodium acetate and propionate on insulin secretion and their protection against islet apoptosis 68 . In addition to butyric acid, other SCFAs in the fermentation media of L. mesenteroides EH-1 will be measured in the future. One of limitations of using SCFAs as therapeutics includes their typically short in vivo half-life, with clearance from plasma occurring within a few hours 69 . A high concentration (about 0.7 mmol/L) of butyric acid was detected in the cecum of mice fed with L. mesenteroides EH-1 (Fig. 4a), demonstrating that L. mesenteroides EH-1 is a potent strain for producing butyric acid. Data in literature demonstrated that high-fat-fed mice treated with butyrate showed enhancement of the insulin secretion, which was related to a substantial reduction in lipid accumulation within the pancreas 70 . Furthermore, butyrate can prevent type 1 diabetes in non-obese diabetic (NOD) mice 59 Results from previous studies above supported that butyric acid has a beneficial effect on prevention of type 1 diabetes. Future works will test the anti-diabetic activity of probiotic L. mesenteroides EH-1 using NOD mice or high-fat diet-induced type 1 diabetes.
In summary, our study characterized a new probiotic bacterial strain, L. mesenteroides EH-1, which was originally isolated from Mongolian curd cheese. L. mesenteroides EH-1 produces high concentrations of butyric acid which can activate Ffar2 to raise insulin levels but mitigate glucose amounts in the blood of type 1 diabetic mice.