Voglibose (>98.0%) was purchased from Tokyo Chemical Industries (Tokyo, Japan). Erythromycin, oxytetracycline hydrochloride, cefadroxil, 4-nitrophenyl β-D-glucopyranoside, 4-nitrophenyl β-D-glucuronide, 4-nitrophenyl sulfate, soluble starch, ammonium acetate, streptozotocin, and telmisartan, the internal standard (IS), were purchased from Sigma-Aldrich (St. Louis, MO, USA). High-performance liquid chromatography (HPLC)-grade acetonitrile was purchased from J. T. Bakers (Central Valley, PA, USA). An anaerobe gas generating system was purchased from Becton, Dickinson and Company (Sparks, MD, USA). A blood glucose monitoring system was purchased from Bayer Contour TS (Mishawaka, IN, USA). Potassium phosphate monobasic and potassium phosphate dibasic were purchased from Duksan (Seoul, Korea). All other chemicals were of analytical grades and used as received.

Male Institute of Cancer Research (ICR) mice at 7 weeks of age were purchased from Orient Bio (Seoul, Korea). Animals upon received were randomly placed five in a cage. Animals were acclimatized for at least 1 week under controlled temperature of 22℃±2℃, relative humidity of 50%±10%, and air change of 10 to 20 cycles per hour with a light and dark cycle of 12-hour. The study was performed following procedures of the Institutional Animal Care and Use Committee of Yeungnam University (approved No., 2014-008). Unless specified, all animals were allowed to access food and water ad libitum.

For in vitro studies, mice were grouped as antibiotics-treated or vehicle-treated groups (n=5). In antibiotics-treated group, mice were orally administered with a mixture of three antibiotics in saline, such as erythromycin (300 mg/kg), oxytetracycline HCl (300 mg/kg), and cefadroxil (100 mg/kg), for 3 successive days. Saline was administered in the vehicle-treated group for 3 days. Twenty-four hours after the final dose of either antibiotics or vehicle, intestinal contents were harvested. Collected intestinal contents were mixed with two volumes of potassium phosphate buffer, pH 7.4, vortexed, and homogenized at 11,000 rpm for 2 minutes. Then, the sample was centrifuged at 500 g for 10 minutes at 4℃, and the supernatant was collected for in vitro tests. Aliquots were prepared and stored at −70℃ until their use.

The enzyme assays were carried out according to a previously published report with some modifications [17]. Briefly, 0.4 mL of either 2.5 mM 4-nitrophenyl β-D-glucopyranoside, 4-nitrophenyl β-D-glucuronide, or 4-nitrophenyl sulfate in potassium phosphate buffer, pH 7.4, was mixed with 0.4 mL of potassium phosphate buffer, pH 7.4, and 0.2 mL of intestinal contents obtained from both antibiotics- and vehicle-treated mice, followed by the incubation at 37℃ for 30 minutes with shaking. The reaction was terminated by adding 0.25 mL of 1 N NaOH. Following the centrifugation at 3,000 ×g for 10 minutes, the absorbance of supernatant was measured at 405 nm by using a UV spectrophotometer (Eppendorf, Hamburg, Germany). Standard calibration curves of 4-nitrophenol were prepared for the calculation of individual enzyme activities.

An HPLC connected to a mass spectrometer (API-4000; SCIEX, Framingham, MA, USA) was used. The mobile phase consisted of 10 mM ammonium acetate buffer (mobile phase A) and acetonitrile (mobile phase B). A gradient condition was used for the quantitative analysis of voglibose in which the mobile phase B started at 0.5% was held for 1.0 minute, and increased to 10% during the following 2.0 minutes. Then, mobile phase B was increased to 90% during the following 1.0 minute, held at 90% for 1.5 minutes, and returned to 1% during the following 1.5 minutes. Column equilibration time was 7.0 minutes and the total run time was 14.0 minutes. Analyte peaks were identified with an Agilent Eclipse plus C₈ column (2.1×150 mm, 3.5 µm). The column temperature was maintained at 30℃ with a flow rate of 0.25 mL/min and injection volume of 5.0 µL. Multiple reaction monitoring mode was used for the quantitative detection of voglibose using the ratio of areas of analyte to IS peaks. Voglibose and telmisartan (IS) was detected in positive ion modes. Mass transitions for voglibose and telmisartan were m/z 268→92 and m/z 515→276, respectively. The declustering potential, collision energy and collision cell exit potential values were 70, 30, and 18 for voglibose and 156, 65, and 20 for telmisartan, respectively.

Inter-day and intra-day validations were performed according to the Food and Drug Adminstration guideline [18]. Accuracy and coefficient of variation (CV %) was calculated based on the analysis of spiked quality control (QC) samples of different concentrations. QC samples of 4, 100, and 400 ng/mL voglibose were analyzed. For intra-day validation, five freshly prepared samples of each concentration were analyzed in a single day. For inter-day validation, QC samples were freshly prepared in each day and analyzed for 5 consecutive days. Accuracy and precision of analyzed QC samples were determined.

To investigate the possible metabolism of voglibose by the intestinal microbiota, the remaining level of voglibose was measured following the incubation of voglibose with intestinal contents compared with the initial level of voglibose. Prior to this, the stability of voglibose was tested with a target CV (%) of 15. Voglibose in potassium phosphate buffer, pH 7.4, at 5, 10 and 20 µg/mL was incubated in various conditions for long-term and short-term stability studies. For the short-term stability test, voglibose was incubated in both 25℃ and 37℃ for 24 hours. Similarly, voglibose was tested for long-term stability after storage in a freezer for 30 days at −20℃ and 5 days at 4℃. The stability of voglibose was also tested following three freeze-thaw cycles from −20℃ to room temperature.

For time- and concentration-dependent metabolism of voglibose, 10 µL of either 0.25, 0.5, or 1 mg/mL voglibose was incubated with 490 µL of intestinal contents to prepare the final concentrations of 5, 10, and 20 µg/mL of voglibose. The mixtures were incubated in an anaerobe gas generating system at 37℃ for 3, 6, 12, and 24 hours. Following incubation, 10 µL of each incubation mixture was transferred to another tube containing 490 µL of methanol containing 10 ng/mL of telmisartan (IS). Thus, the final concentrations for liquid chromatography-mass spectrometry (LC-MS/MS) analyses were 100, 200, and 400 ng/mL. The mixture was vortexed and centrifuged at 12,000 ×g for 10 minutes. Then, 200 µL of supernatant was pipetted into vials and analyzed using LC-MS/MS. Calibration curves were prepared accordingly and concentrations of voglibose in incubated samples from antibiotics- and vehicle-treated mice were determined.

To investigate the role of intestinal microbiota in the pharmacodynamics of voglibose in non-diabetic conditions, mice were divided into antibiotics-treated and vehicle-treated groups. In the antibiotic-treated group, a mixture of three antibiotics was administered orally for 3 successive days. Saline was administered to vehicle-treated animals for 3 days. Twenty-four hours after the final treatment, both antibiotics- and vehicle-treated mice were subdivided into two subgroups, i.e., starch+voglibose and starch only, followed by a fasting period of 6 hours. For the main experiment, voglibose at 10 mg/kg was administered orally to the starch+voglibose group and saline to the starch only group. Thirty minutes later, 50 mg/kg of soluble starch was orally administered to all animals and the blood glucose levels were determined at 0, 30, 60, 90, 120, 180, and 240 minutes after starch administration.

To prepare the diabetic model, mice fasted for 6 hours were intraperitoneally injected with 40 mg/kg streptozotocin for 5 successive days. Body weight and blood glucose levels of individual animals were measured daily before each streptozotocin injection. Once blood glucose level reached >250 mg/dL, mice were considered diabetic and were divided into two groups for antibiotics treatment. Mice in the antibiotic-treated group received the mixture of three antibiotics by oral administration for 3 consecutive days. In the vehicle-treated group, mice received saline for 3 days. Twenty-four hours after the last dose of antibiotics, mice were further divided into two groups, starch+voglibose group and starch only group, followed by a fasting period of 6 hours. For the main experiment, voglibose at 10 mg/kg was administered to the starch+voglibose group and saline was administered to the starch only group to 6-hour fasted animals. After 30 minutes, 50 mg/kg of soluble starch was orally administered to all mice and blood glucose levels were measured at 0, 30, 60, 90, 120, 180, and 240 minutes after starch administration.

All data were presented as mean±standard deviation. Student's t-test was used to determine the statistical significance of the data. The results of P<0.05 were represented with alphabets.

For the accurate quantification of voglibose, an LC-MS/MS method was developed. The separation of voglibose was more effective with C₈ column than with C₁₈ or C₁₆ columns (data not shown). In addition, for better resolution, a gradient eluting system was established. As depicted in Fig. 1A, voglibose was ionized in positive ion mode [M+H]⁺ to form a protonated ion [M+H]⁺ at m/z 268.0 with a daughter ion at m/z 92.0. Telmisartan, the IS, was detected at m/z 515.2 and further fragmented to the ion at m/z 276.2 (Fig. 1B). There was no interfering peak at the retention time of voglibose at 1.4 minutes and telmisartan at 7.93 minutes (Fig. 2). The inter- and intra-day accuracy and precision of voglibose are summarized in Supplementary Table 1. A good linearity was observed with the calibration curves constructed with the ratio of areas of analyte to IS peaks. Acceptable linearity (R²=0.9938) was observed between the constructed calibration standards with 4 to 400 ng/mL voglibose. The accuracy (%) and CV (%) of inter-day and intra-day were within 15% for the lowest concentration and 10% for other concentrations tested.

A stability test was conducted for three different concentrations of voglibose. Following exposure of voglibose to various ambient conditions, LC-MS/MS analysis was conducted to measure the remaining voglibose from the initial level. As shown in Table 1, voglibose was stable for 24 hours at both room temperature and 37℃. Similarly, voglibose was stable at freezing and refrigerating conditions. However, voglibose was less stable in samples with more than three freeze-thaw cycles. Nevertheless, it was within the limit of 15%. Taken together, the current analytical method was appropriate to study the in vitro metabolism of voglibose by intestinal microbiota.

Prior to the characterization of voglibose metabolism by intestinal microbiota, three enzyme activities, i.e., β-D-glucosidase, β-D-glucuronidase and sulfatase, in the intestinal contents isolated from both vehicle- and antibiotics-treated mice were determined. As shown in Fig. 3, the enzyme activities in antibiotics-treated mice were significantly lower than in vehicle-treated mice, indicating that the present antibiotic pretreatment model would be effective to study the role of intestinal microbiota in voglibose metabolism and their effects on the pharmacodynamics of voglibose.

Thereafter, several concentrations of voglibose were incubated with intestinal contents from vehicle- and antibiotics-treated mice at several time points. Then, the remaining voglibose was measured. Following incubation, almost 40% of voglibose in the control was metabolized within 3 hours of incubation compared to the level of voglibose in the antibiotics-treated group. Moreover, the level of voglibose in the antibiotics-treated group was always higher than the level of voglibose in the control up to the period of sample testing (Fig. 4). The results indicated that the intestinal microbiota plays a critical role in the metabolism of voglibose. However, other effects of antibiotics, such as changing the gut epithelial enzyme activity and modulating the blood glucose level in animals, are yet to be studied. In addition, a small but gradual increase in metabolism of voglibose over time in the antibiotics-treated group indicated a reduced number of bacteria in intestinal contents following antibiotic treatment. The results indicated that the pharmacodynamic action of voglibose would be significantly modulated by the condition of intestinal microbiota.

To support this hypothesis, the in vivo pharmacodynamic action of voglibose was investigated in conditionally modulated intestinal microbiota by antibiotic treatment.

Because antibiotic pretreatment would cause the reduction of the intestinal microbial population, the enzymes available for voglibose metabolism would be reduced and, thereby, the metabolism of voglibose by the intestinal microbiota would be altered. Due to the low intestinal absorption of voglibose, the pharmacodynamic effects of voglibose in starch-supplemented animals were investigated in the present study. To illustrate the effect of the intestinal microbiota on the metabolism of voglibose in vivo, the blood glucose-lowering effect of voglibose was measured in vehicle- and antibiotic-treated non-diabetic and diabetic mice models (Fig. 5). The blood glucose levels rapidly increased following starch administration to 6-hour fasted animals. The changes in blood glucose levels were also clear when the area under the curve (AUC) for 6 hours was calculated from Fig. 5. In addition, the baseline level of blood glucose was much higher in diabetic mice than in non-diabetic mice (AUC=97,457 vs. AUC=38,745). Following voglibose administration in both antibiotics- and vehicle-treated mice of both non-diabetic and diabetic models, the blood glucose levels, even after starch administration, were precisely controlled and similar to the levels at 0 from 2 hours. Particularly, a significant difference in blood glucose level was observed between the voglibose administered antibiotics- and vehicle-treated groups of non-diabetic model (AUC=27,486 vs. AUC=35,262) (Fig. 5A). The blood glucose levels in the antibiotics-treated group were attenuated by voglibose in diabetic mice (AUC=91,837 vs. AUC=81,513), although the results were not statistically significant (Fig. 5B). The results suggested that voglibose could control the blood glucose level by inhibiting α-glucosidase enzyme present in the intestine, particularly in the antibiotics-treated animals, possibly due to the limited availability of microbial enzymes to metabolize voglibose by antibiotics administration. Therefore, more voglibose available showed its effect as a blood glucose-controlling agent via the inhibition of α-glucosidase enzymes.