Salacinol and Related Analogs: New Leads for Type 2 Diabetes Therapeutic Candidates from the Thai Traditional Natural Medicine Salacia chinensis

The antidiabetic effect of a hot water extract of stems of Salacia chinensis (SCE) was evaluated in vivo in KK-Ay mice, a typical type 2 diabetes mellitus mice model. Administration of CE-2 dietary feed containing 0.25 and/or 0.50% of SCE for three weeks to KK-Ay mice significantly suppressed the elevation of both blood glucose and HbA1c levels without significant changes in body weight or food intake. Glucose tolerance was improved by administration to KK-Ay mice for 27 days of AIN93M purified dietary feed containing 0.12% of SCE. No suppressive effect with respect to HbA1c level was observed when AIN93M/Glc dietary feed in which all digestible glucides were replaced with glucose was administered with SCE. Thus, α-glucosidase inhibitory activity approved as the mechanism of action of the antidiabetic effect of SCE by in vitro investigation was reconfirmed also in in vivo studies. Evaluation of the α-glucosidase inhibitory activity of the active constituents, salacinol (1), kotalanol (3), and neokotalanol (4), by employing human α-glucosidases revealed that these compounds inhibited them as potently (IC50 = 3.9–4.9 μM for maltase) as they inhibited rat small intestinal α-glucosidase. The principal sulfonium constituents (1–4) were highly stable in an artificial gastric juice. In addition, 1–4 were hardly absorbed from the intestine in an experiment using the in situ rat ligated intestinal loop model. The results indicate that these sulfoniums are promising leads for a new type of anti-diabetic agents.

reconfirmed also in in vivo studies. Evaluation of the α-glucosidase inhibitory activity of the active constituents, salacinol (1), kotalanol (3), and neokotalanol (4), by employing human α-glucosidases revealed that these compounds inhibited them as potently (IC50 = 3.9-4.9 μM for maltase) as they inhibited rat small intestinal α-glucosidase. The principal sulfonium constituents (1)(2)(3)(4) were highly stable in an artificial gastric juice. In addition, 1-4 were hardly absorbed from the intestine in an experiment using the in situ rat ligated intestinal loop model. The results indicate that these sulfoniums are promising leads for a new type of anti-diabetic agents.

Introduction
A healthful eating pattern, adequate nutrients, regular physical activity, and often pharmacotherapy are key components of diabetes management [1]. Based on a large number of chemical and pharmacological research work, numerous bioactive compounds have been found in nutritious herbal food ingredients for diabetes [2]. The genus Salacia (Hippocrateaceae) contains woody climbing plants and is widely distributed in countries, such as India, Sri Lanka, China, and Thailand. The stems and roots of the plants have been used for the prevention or cure of diabetes in these countries [3][4][5][6][7]. From a methanol and/or 80% aqueous methanol extract of the genus Salacia plants, we isolated a novel thiosugar sulfonium sulfate inner salt, salacinol (1) [7,8], as a potent α-glucosidase inhibitor. The inhibitory activities of 1 against rat small intestinal maltase and sucrose were as potent as those of acarbose or voglibose, widely used clinical inhibitors. We also isolated its analogs, neosalacinol (2) [9,10], kotalanol (3) [11,12], neokotalanol (4) [13], ponkoranol (5) [14], neoponkoranol (6) [15], salaprinol (7) [14,16], and neosalaprinol (8) [15], and found that compounds 2-6 were as potent as 1 ( Figure 1). Methanol, 80% aqueous methanol, and aqueous extracts from the roots and stems of S. reticulata had been shown to suppress the increase in blood glucose levels in maltose, sucrose, and starch-loaded rats [6,7,[17][18][19][20]. An 80% aqueous methanol extract of the roots of S. oblonga and stems of S. chinensis also exhibited anti-hyperglycemic activity in maltose and sucrose-loaded rats [18]. Several clinical trials of the Salacia extracts have demonstrated the efficacy of S. retuculata [21,22] and S. oblonga [23] to the patient of type 2 diabetes and that of S. chinensis to people with high normal blood glucose and/or volunteers in borderline hyperglycemia [24]. Safety profiles of Salacia extracts to acute toxicity and mutagenicity [25], and those of S. chinensis to reproductive outcome in rats [26] have also been demonstrated. Based on these findings, interest in the genus Salacia plants as a possible nutraceutical product for diabetic patients is increasing, and there has been a strong demand for efficient quality control to ensure the authenticity and the active contents of these products, as well as to verify the claims on product labels. Quantitative analyses of the sulfonium constituents (1)(2)(3)(4)(5)(6)(7)(8) in the extracts have been developed as two separate protocols using LCMS: one for the sulfonates (1, 3, 5, and 7) [27,28] and the other for their de-O-sulfonates (2, 4, 6, and 8) [28,29]. As a result, the distributions of the sulfonium constituents in stems and roots of these plants were found to differ for the different collecting areas. Compound 4 was the major constituent in samples from Thailand, whereas 1 was the major constituent in the samples from Sri Lanka and India. An effort to discriminate the species of genus Salacia by referring to their RNA sequence of the internal transcribed spacer (ITS) region in the nuclear ribosomal RNA gene in an authentic specimen was conducted, and a genotype characteristic of S. chinensis, which is distinguishable from those of S. reticulata and S. oblonga was identified [30]. In the present study, the following in vivo assays of hot water extracts of Thai S. chinensis (SCE) and/or its sulfonium constituents (1)(2)(3)(4) were performed: (i) suppressive effects on blood glucose level elevation in starch-loaded rats; (ii) suppressive effects on elevation of both blood glucose and HbA1c levels after chronic administration to KK-A y mice; (iii) improvement effect of glucose tolerance after chronic administration to KK-A y mice; and (iv) stability of these sulfoniums (1)(2)(3)(4) in the digestive organ as well as their absorption from the digestive tract.

Plant Material
Extracts obtained from the stem of Salacia chinensis, collected in southern region of Thailand, were investigated in this study. The plant material was identified by one of the authors (Yutana Pongpiriyadacha, Rajamangala University of Technology Srivijaya). A voucher specimen of this plant is on file in our laboratory.
Hot water extraction of the de-O-sulfonated constituents (2, 4, 6, and 8) was more efficient than those by methanol and aqueous methanol [28,29]. Dried stems of S. chinensis were crushed and extracted with hot water. The aqueous extract was evaporated under reduced pressure to obtain a hot water extract powder (abbreviated as SCE), which was used with no additives.

Animals
Male Sprague-Dawley (SD) rats were purchased from Japan SLC, Inc., Shizuoka, Japan, and male KK-A y mice were from CLEA Japan, Inc., Tokyo, Japan. The animals were housed at a constant  (7) neosalaprinol (8) temperature of 23 ± 2 °C, at 55% ± 15% humidity, and 12 h of illumination per day. All experiments were performed with conscious animals unless otherwise mentioned. The experimental protocol was approved by the Experimental Animal Research Committee at Kinki University.

Effects of SCE and Sulfonium Constituents (1, 3, and 4) on Blood Glucose Levels in Starch-Loaded Rats
Five-week-old male SD rats were housed for one week in metal cages. After overnight fasting (20 h), the rats were orally administered a 5% (w/v) α-starch solution (1 g/kg) with or without a sample (SCE: 10-300 mg/kg, 1: 0.15-1.48 mg/kg, 3: 0.21-2.06 mg/kg, and 4: 0.07-0.68 mg/kg) using a stomach tube. At 0, 0.5, 1.0, 2.0, and 3.0 h after the administration of α-starch, blood samples were taken from the tail vein and immediately subjected to the measurement of blood glucose using the glucose oxidase method. As a baseline, distilled water was administrated to rats as a "normal group". Median effective dose (ED50) was determined by plotting the inhibition rate of incremental AUC0-2 h (iAUC0-2 h; the AUC above baseline) versus corresponding inhibitor dosage.

Effects on Blood Glucose Levels in SCE-Pretreated Starch-Loaded Rats
Five-week-old male SD rats were housed for one week in metal cages. After overnight fasting (20 h), the rats were orally administered SCE (75 mg/kg) using a stomach tube at various times (0, 0.5, 1.0, and 2.0 h) before loading of 5% (w/v) α-starch solution (1 g/kg). At 0, 0.5, 1.0, 2.0, and 3.0 h after the administration of α-starch, blood glucose levels were measured using the glucose oxidase method.

Effects of Blood Glucose and HbA1c Levels after Three Weeks Administration of SCE in CE-2 Diet-fed KK-A y Mice
Five-week-old male KK-A y mice were housed for one week in individual metal cages. They were divided into four groups based on body weight, blood glucose, and HbA1c levels. The control group was fed a standard diet (CE-2, CLEA Japan, Inc., Tokyo, Japan) and the three SCE-treated groups were fed diets supplemented with 0.10, 0.25, and 0.50% (w/w) SCE, respectively. On day 15 and at the end of the treatment period, blood samples were collected from the tail vein under non-fasting conditions. Blood glucose and HbA1c levels were measured using glucose oxidase method and a DCA Vantage Analyzer TM (Siemens, New York, USA), respectively. The HbA1c values were recorded as Japan Diabetes Society (JDS) values and were then converted to National Glycohemoglobin Standardization Program (NGSP) values as follows: HbA1c (NGSP) = 1.02 × HbA1c (JDS) + 0.25% [31].

Effects of HbA1c Levels after Chronic Administration of SCE in AIN93M Purified and AIN93M/Glc (All Digestible Glucides Replaced with Glucose) Diet-fed KK-A y Mice
Five-week-old male KK-A y mice were housed for one week in individual metal cages. These mice were divided into two groups based on body weight, blood glucose and HbA1c levels; a control group was fed a standard diet (AIN93M purified, CLEA Japan, Inc., Tokyo, Japan) and the SCE-treated group was fed the same diet supplemented with 0.03, 0.06, or 0.12% SCE. Given that the food intake of the SCE-treated group decreased, the control group was pair-fed the amount of food consumed by the SCE-treated group. At regular intervals, blood samples were collected from the tail vein under non-fasting conditions. Blood glucose and HbA1c levels were measured using the glucose oxidase method and a DCA Vantage Analyzer ™ , respectively. In the oral glucose tolerance test, 28 days after the start of the experiment, the 0.12% SCE-treated group and the corresponding control group were orally given a D-glucose solution (2.0 g/kg) after fasting for 20 h. Blood glucose levels were measured at 0, 0.5, 1.0, 2.0, and 3.0 h after administration. To identify the α-glucosidase inhibitory activity of SCE in the chronic experiments, a customized AIN93M diet (AIN93M/Glc, Table 1), in which all the digestible glucides in AIN93M were substituted by D-glucose, was administered to the KK-A y mice under the same conditions.

Effects on Human Intestinal α-Glucosidase
The experiment was performed according to the method as described in our previous report [32].

Statistics
Values are expressed as means ± SEM. For statistical analysis, one-way analysis of variance followed by Student's t-test or Dunnett's test was used.
In order to estimate the duration of the effect of SCE as the α-glucosidase inhibitor, the rats were pretreated by SCE (75 mg/kg, p.o.) at 0, 0.5, 1.0, and 2.0 h before administration of starch. As shown in Table 3, a group of rats, which were administrated by SCE 0.5 h before starch-loading tended to suppress the increase of blood glucose levels. On the other hand, administration of SCE 1.0 h before starch loading, no significant effect was observed. Therefore, the suppressive effect of the SCE against blood glucose elevation was estimated to last for ca. 0.5 h.  Values are means ± SEM (n = 8); Significantly different from control, a p < 0.01 (Dunnett's test).
Next, effect of three weeks of administration of SCE on both blood glucose and HbA1c levels by using Kuo Kondo (KK)-yellow agouti (A y ) mice. The KK mice originating from Japan is a polygenic model of obesity and type 2 diabetes mellitus. KK-A y mice, also named as Yellow KK mice, developed as a result of the dominant mutation of A y gene in KK [33]. As the results, SCE significantly suppressed the increase of both blood glucose and HbA1c levels at doses of 0.25 and/or 0.50% (w/w) in the CE-2 diet without significant changes in body weight and food intake as shown in Table 4. The effects of AIN93M purified diet containing 0.03%-0.12% SCE on blood glucose and HbA1c levels of KK-A y mice were evaluated. As shown in Table 5B, administration of the diet containing 0.06% SCE for 11 and 18 days caused significant suppression of both blood glucose and HbA1c levels without significant changes in body weight. Furthermore, 27 days of administration of the diet of 0.12% SCE significantly suppressed not only blood glucose and HbA1c level elevations but body weight gain as well, as shown in Table 5C. Values are means ± SEM (n = 6-7); Significantly different from control, a p < 0.05, b p < 0.01 (Student's t-test); Control group was pair-fed the amount of food consumed by SCE-treated group: (A) 5.4, (B) 4.6, and (C) 4.6 g/day, respectively.
Next, the effect of SCE on glucose tolerance was evaluated in glucose-loaded KK-A y mice. Continuous administration to the glucose-loaded KK-A y mice for 27 days of 0.12% SCE in AIN93M purified diet significantly suppressed blood glucose level elevation, apparently improving glucose tolerance ( Figure 2). These results suggest that SCE is effective for both inhibiting postprandial glucose elevation and improving glucose tolerance. When the AIN93M/Glc diet, in which all of the digestible glucides were replaced by glucose, containing 0.30% SCE was administered instead of the AIN93M diet, no anti-hyperglycemic activity was observed ( Figure 3). Thus, an α-glucosidase inhibitory effect, identified as the mechanism of action in in vitro studies was also confirmed in the present in vivo examination.  Pint and co-workers reported the stereostructure elucidation and synthesis of 3 [34,35] and also enzymatic inhibitory effects of a series of sulfoniums and their analogs on human recombinant glucosidases, such as catalytic N-and C-terminal subunits of two retaining exo-glucosidases, maltase-glucoamylase (MGAM) and sucrase-isomaltase (SI), ntMSAM, ctMGAM, ntSI, and ctSI [34][35][36][37]. Activities of the active sulfonium constituents (1-6) against human intestinal maltase were also evaluated in the present study. As shown in Table 6, 1 (IC50 = 4.9 μM), 2 (9.0 μM), 3 (3.9 μM), 4 (3.9 μM), 5 (5.0 μM), and 6 (4.0 μM) inhibited the human maltase as potently as they inhibited rat small intestinal maltase, and their activities were almost equivalent to those of voglibose (1.3 μM) and miglitol (3.7 μM), and even more potent than that of acarbose (15.2 μM). Next, the stability of the principal sulfoniums (1)(2)(3)(4) in an artificial gastric juice was examined. More than 96% of each of these sulfoniums survived when they were treated at 37 °C for 1.0 h, as shown in Table 7. More than 90% of them survived even after 3 h of treatment under the conditions, and high stability of these sulfoniums (1)(2)(3)(4) in the artificial gastric juice was observed (Table 7). Their bioavailability through the intestine was also evaluated using an in situ rat ligated intestinal loop model. As a result, these sulfoniums were hardly absorbed from the intestine (residual rate (%) of 1: 97.6 ± 1.8; 2: 94.5 ± 1.8; 3: 99.7 ± 2.7; 4: 96.6 ± 1.7, treated for 2.0 h) as shown in Table 8.  (1) 100.0 ± 4.6 100.0 ± 6.9 92.5 ± 6.1 Neosalacinol (2) 100.0 ± 6.0 96. 5  Values are means ± SEM (n = 3). Values are means ± SEM (n = 5).

Conclusions
An antidiabetic effect of hot water extracts of Salacia chinensis was observed in in vivo studies using KK-A y mice. Elevation of both postprandial blood glucose and HbA1c levels were significantly suppressed by SCE. Glucose tolerance was also improved by the administration of SCE. The active constituents, salacinol (1), kotalanol (3), and neokotalanol (4) inhibited human α-glucosidases as potently as they inhibited rat small intestinal α-glucosidase. The principal sulfonium constituents (1)(2)(3)(4) were highly stable in an artificial gastric juice and were hardly absorbed from the intestine. The results indicate that these sulfoniums are promising leads for a new type of anti-diabetic agents.