A novel botanical formula prevents diabetes by improving insulin resistance

A 96-well, cell-free α-amylase assay was established by using 2-chloro-p-nitrophenyl-α-D-maltotrioside (CNPG3) as a substrate in an assay buffer which was prepared by adding 14.5 g sodium chloride, 0.15 g calcium chloride dihydrate and 20 g of non-fat dry milk in 1 L deionized water, pH 5.0. Twenty-five μL of each extract were premixed with 25 μL of α-amylase (1 U/mL) and incubated for 30 min at room temperature. Then 200 μL of 37 °C CNPG3 was added to all wells and the plate was immediately read at 405 nm and every 30 s for 5 min at 37 °C using a Spectramax plate reader (Molecular Devices, Sunnyvale, CA). Acarbose served as positive control (100 μM) and defined 100% inhibition. All reagents, chemicals and enzyme were purchased from Sigma-Aldrich Corporation (St. Louis, MO).

A 96-well, cell-free α-glucosidase assay was established using rat intestinal acetone powder as the enzyme source and sucrose as the substrate. The reaction was carried out by adding 60 μL of sucrose (20 mM in water), 30 μL of extract (in water at 4X of final dose tested) and 30 μL of enzyme solution (3.75 mg/mL in a buffer of 15 mM potassium phosphate and 30 mM sodium phosphate, pH 7.0) to each well, sealed and incubated for 25 min at 37 °C. Following incubation, 120 μL of glucose assay reagent (Wako Chemicals, Tokyo, Japan) was added to each well, sealed and incubated for 5 min at 37 °C. Absorbance was then read in a Spectramax plate reader at 510 nM (Molecular Devices, Sunnyvale, CA). 1-deoxynojirimycin hydrochloride served as positive control (50 μM) and defined 100% inhibition. All other reagents, chemicals and enzyme were purchased from Sigma-Aldrich Corporation (St. Louis, MO).

The assay was performed by Zen-Bio Inc., (Research Triangle Park, NC), using primary human subcutaneous adipocytes that were differentiated in a 96-well format at 37 °C with 5% CO₂. Two weeks after cells were differentiated, they were serum starved in the presence of extracts for 20 h. Following the treatment, 1 nM insulin and a cocktail containing 2-deoxyglucose (2-DOG) and ³H-2-DOG were applied to cells and allowed to incubate for 2 h. Cells were then washed with phosphate buffered saline and lysed, and glucose uptake was determined as counts/well. Cytochalasin B (10 μM) served as a non-specific 2-DOG uptake to account for non-insulin induced glucose uptake. Change in insulin sensitivity was calculated by setting the effect of 1 nM insulin as 0% and 100 nM insulin as 100%.

The botanical test formula was prepared by Nutrilite Health Institute through mixture of commercially available fenugreek seed extract (from India, extracted with 55% to 75% ethanol and standardized to 4.5% 4-OH-IIe), mulberry leaf extract (from China, extracted with water and standardized to 1.25% DNJ) and American ginseng extract (from USA, extracted with water and standardized to 5.5% ginsenoside). The botanical test formula contained 88 mg of fenugreek seed extract, 120 mg of mulberry leaf extract and 300 mg of American ginseng extract.

Male Sprague Dawley rats (150–200 g) were purchased from Charles River (Beijing, China). Animals were housed in a temperature and humidity-controlled room (22–23 °C and 46–63%, respectively) and had free access to food and water. Blood samples were taken from the tail vein and all animals received humane care. The experimental protocols were approved by the Institutional Animal Care and Use Committee of Beijing Institute for Drug Control.

A T2DM model was established by following the protocol from the China Food and Drug Administration for health food registration with the claim of “Assist in Reducing Blood Glucose Level” [46]. After one-week acclimation on a standard control chow diet, the animals were fasted 3 h and levels of serum glucose and postprandial serum glucose at 0.5 h after an oral glucose load (2.5 g/kg) were measured to establish baseline glycemia. Animals were glycemia balanced and divided into 5 groups: 1) naïve control (n = 10); 2) model control (n = 10); 3) 42.33 mg/kg body weight (BW) (1X) (n = 10); 4) 84.66 mg/kg BW (2X) (n = 9); 5) 169.33 mg/kg BW (4X) (n = 7) of the test formula. Model control and treatment groups were given either vehicle or test formula by daily oral gavage for 33 d. After one-week of treatment (day 7), both model control and treatment groups were fed a high fat diet (HFD) (Charles River) for an additional three weeks. HFD was comprised of control chow diet (52.6% w/w) plus lard (10% w/w), sucrose (15% w/w), egg yolk powder (15% w/w), casein (5% w/w), sodium cholate (0.2% w/w), calcium bicarbonate (0.6% w/w), calcium carbonate (0.4% w/w) and cholesterol (1.2% w/w). After three weeks of HFD, animals were fasted for 24 h and a single i.p. injection of alloxan (103–105 mg/kg) was given to induce hyperglycemia. Animals continued the HFD and botanical test formula for another 4 d. Animals were then fasted for 3 h and oral glucose tolerance test (OGTT) was performed. Blood was collected at baseline (0 h) for determination of the serum glucose, insulin, triglyceride and total cholesterol, and at 0.5 h and 2 h after OGTT (2.5 g/kg) for postprandial serum glucose. Serum glucose area under the curve (AUC) was calculated by the following equation, where SG = serum glucose:

Fasting insulin resistance index was assessed using the homeostasis model assessment for insulin resistance (HOMA-IR) originally described by Mathew [47]. HOMA-IR was calculated using the following formula:

HOMA IR = fasting glucose (mmol/L) × fasting insulin (μU/mL)/22.5.

Serum glucose, total cholesterol, triglycerides (BioSino, Beijing, China) and insulin levels (Jiancheng, Nanjing, China) were determined with commercially available kits in a 96-well format per manufacturer’s instructions. In brief, the glucose assay kit uses an enzyme mix that oxidizes glucose to generate a product that reacts with a dye to generate color that is read in a plate reader at 570 nm. The total cholesterol assay kit uses a mix of enzymes to first hydrolyze cholesterol esters followed by oxidation of the free cholesterol to produce hydrogen peroxide that reacts with a probe to produce color that is read in a plate reader at 570 nm. In the triglyceride assay kit, samples are exposed to lipases that generate free fatty acids and glycerol. The glycerol is then oxidized to generate a product which reacts with the probe to generate color that is read in a plate reader at 570 nm.

The insulin kit is a sandwich enzyme linked immunosorbent assay. The principles of the assay are to capture the insulin molecules from samples in the wells of a microtiter plate coated by pre-titered amount of a monoclonal mouse anti-rat insulin antibody and then the binding of biotinylated polyclonal antibodies to the captured insulin. Wells are then washed to remove unbound materials. Next, horseradish peroxidase is added to wells to bind to the immobilized biotinylated antibodies. The wells are then washed to remove any free enzyme conjugates, and quantification of immobilized antibody-enzyme conjugates are determined by horseradish peroxidase activities in the presence of the substrate 3,3′,5,5′-tetramethylbenzidine. The enzyme activity is measured spectrophotometrically by the increased absorbency at 450 nm, corrected from the absorbency at 590 nm, after acidification of formed products.

Epididymal adipose tissue was homogenized in lysis buffer, which was comprised of 50 mM Tris-HCl (pH 7.4), 0.1% sodium dodecyl sulfate (SDS), 2 M phenylmethylsulfonyl fluoride, 10 g/mL leupeptin (Amresco Solon, OH) and boiled in 5× loading buffer (Beyotime Shanghai, China). Twenty μg of protein was then separated by SDS-PAGE and transferred to 0.45 μm polyvinylidene difluoride membrane (Millipore, Billerica, MA). After transfer, membranes were blocked in 5% non-fat powdered milk in phosphate buffered saline with 0.1% Tween 20 and probed with 3-phosphoinositide-dependent kinase-1 (PDK1) (Cell Signaling, Danvers, MA), glucose transporter 4 (GlUT4), (Santa Cruz Biotechnology, Santa Cruz, CA) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Tdybio Beijing, China). Horseradish peroxidase-conjugated secondary antibodies and a chemiluminescence substrate kit were used in detection of specific proteins (Millipore, Billerica, MA).

Quantitative data are presented as mean ± standard deviation (SD). Comparisons between groups were made using two-way repeated measures ANOVA with Holm-Sidak’s multiple comparisons test or one-way ANOVA with Dunnett’s multiple comparison test using GraphPad Prism version 6.00 for Windows (GraphPad Software, La Jolla, CA, www.​graphpad.​com). A significant effect was defined as P < 0.05.

Fenugreek seed and mulberry leaf extracts concentration dependently inhibited α-amylase (IC50 = 73.2 μg/mL) and α-glucosidase (IC50 = 111.8 ng/mL), respectively (Fig. 1a, b). American ginseng extract had no inhibitory activity on either enzyme, fenugreek seed extract did not inhibit α-glucosidase and mulberry leaf extract did not inhibit α-amylase (data not shown).

All extracts at lower doses, 3 and 11 μg/mL, were effective at increasing glucose uptake and insulin sensitivity at submaximal insulin levels (Fig. 2a). Mulberry leaf and American ginseng maintained effectiveness at higher doses (33 and 100 μg/mL). Higher doses of fenugreek seed caused cells to detach from the plate which were lost during the washing step, resulting in a negative effect on glucose uptake and insulin sensitivity (Fig. 2a). Examination of low doses of test formulas at different ratios of fenugreek seed: mulberry leaf: American ginseng resulted in a significant increase in insulin sensitivity at ratios of 1:1:3.6 and 1:3.6:3.6, compared to 1:1:1 (Fig. 2b; P < 0.05). Based on these data, commercial cost viability and regulatory limits on doses, an optimized test formula was designed by combining fenugreek seed, mulberry leaf and American ginseng at a ratio of 1:1.3:3.4.

Fasting serum glucose was significantly higher, 17.1 mmol/L ± 7.5 vs 7.5 mmol/L ± 0.75 (Fig. 3a; P < 0.001), while insulin was lower, 2.9 mIU/L ± 0.56 vs 3.5 mIU/L ± 0.24 (Fig. 3b; P < 0.05) in model control compared to naïve control. This resulted in a significant elevation in HOMA-IR, which was nearly doubled (2.1 ± 0.99 vs. 1.2 ± 0.16; Fig. 3c). After 33 d of treatment with test formula, the induction of insulin resistance and T2DM with an alloxan injection was prevented at all doses. Both fasting serum glucose and HOMA-IR values were similar to naïve control and significantly lower when compared to model control (Fig. 3a, c). These effects were achieved without a change in fasting insulin compared to model (Fig. 3b). In addition, maximal efficacy appeared to be present at the lowest dose tested, as higher doses resulted in equal efficacy.

Postprandial serum glucose levels at 0.5 h and 2 h after glucose challenge and AUC were significantly higher in model control compared with naïve control group (Fig. 4; P < 0.001). All doses of the test formula significantly improved postprandial serum glucose (Fig. 4a) resulting in a significantly lower AUC, compared to model control (Fig. 4b).

Serum triglyceride and total cholesterol levels were significantly elevated in model control compared to naïve control, indicating that lipid metabolism was dysfunctional in this model (Fig. 5). However, while values tended to be lower, none of the botanical test formula doses significantly affected serum triglyceride and total cholesterol levels.

Both PDK1 and GlUT 4 were significantly reduced in model control compared to naïve control (P < 0.05 and P < 0.001, respectively), and the botanical test formula blunted this reduction (Fig. 6).