Chemical profile and in vivo hypoglycemic effects of Syzygium jambos, Costus speciosus and Tapeinochilos ananassae plant extracts used as diabetes adjuvants in Puerto Rico

Background The increasing numbers of people who use plant-based remedies as alternative or complementary medicine call for the validation of less known herbal formulations used to treat their ailments. Since Puerto Rico has the highest rate of Type 2 diabetes within all the states and territories of the United States, and Puerto Ricans commonly use plants as diabetes adjuvants, it is important to study the plants’ physiological effects, and identify their bioactive compounds to understand their role in modulation of blood glucose levels. We present the phytochemical profiles and hypoglycemic effects of Tapeinochilus ananassae, Costus speciosus and Syzygium jambos. Methods Phytochemicals in methanolic and aqueous extracts were analyzed by thin layer chromatography (TLC). Alkaloids (Bromocresol green, λ = 470 nm), flavonoids (AlCl3, λ = 415 nm), saponins (DNS, λ = 760 nm), tannins (FeCl3/K4Fe(CN)6, λ = 395 nm) and phenolics (Folin-Ciocalteau, λ = 765 nm) were quantified. Male C57BLKS/J (db/db) and C57BL/J (ob/ob) genetically obese mice were orally gavaged with aqueous extracts of lyophilized plant decoctions for 10wks. Results Our results show that T. ananassae had significantly greater amounts of flavonoids and tannins, while S. jambos showed the greatest concentration of phenolics and C. speciosus exhibited higher amounts of alkaloids. C57BLKS/J db/db treated with plant extracts show better glucose modulation when the extracts are administered in complement with an insulin injection. Finally, C57BL/J ob/ob mice on T. ananassae and S. jambos treatments show better blood glucose modulation over time. Conclusion These results document for the first time the chemical profile of T. ananassae and provide evidence for a potential anti-diabetic efficacy of T. ananassae and S. jambos. Electronic supplementary material The online version of this article (doi:10.1186/s12906-015-0772-7) contains supplementary material, which is available to authorized users.


Background
The widespread use of medicinal plants to mitigate or cure medical conditions such as diabetes calls for their scientific investigation. According to the Annual Survey by the Centers for Disease Control and Prevention (CDC) in Atlanta; since 1996, Puerto Rico has the highest Type 2 diabetes rate within the United States in individuals 18 years or older [1,2]. After conducting an ethnopharmacological survey in 11 municipalities of southeast Puerto Rico and carrying out a literature review about herbal remedies used to manage hyperglycemia, we identified that people living in the southeast rural region of Puerto Rico frequently use four medicinal plants as alternative or complementary treatments for diabetes: Costus speciosus (J. Koening) Sm. (Zingiberaceae), Syzygium jambos (L.) Alston (Myrtaceae), Tradescantia spathacea and Tapeinochilos ananassae K. Schum. (Zingiberaceae) [3][4][5].
In vitro studies using these plants to screen their activities as inhibitors of processes related to the physiopathology (oxidative stress, protein glycosylation and aldose reductase activity) of diabetes (data not shown) led us to perform chemical profiling and functional in vivo studies of C. speciosus, S. jambos and T. ananassae.
C. speciosus, a small plant with scarlet flowers, commonly known as "insulin" in Puerto Rico, is a plant widely used in indigenous systems of medicine for the treatment of various ailments. C. speciosus rhizomes are used as an alternative source for diosgenin, and also to control diabetes due to the antidiabetic, antilipidemic and antioxidant effects of its compounds including sesquiterpene lactones [6][7][8]. S. jambos, "rose apple", is an ornamental fruit tree from Southeast Asia that is also found in tropical regions. This medicinal plant is used to treat diabetes, inflammation, and gastrointestinal disorders [9,10]. Diabetic animals treated with aqueous and ethanolic extracts from the seeds of S. jambos show reduced glycemic effect [11]. S. jambos leaf and fruit extracts contain high concentrations of tannins and phenolic acids, respectively [12,13]. Another plant commonly referred to as "insulin" by populations of southeast Puerto Rico is T. ananassae, an ornamental garden plant native to Malasia, Indonesia, New Guinea and Australia. There are no previous studies describing T. ananassae's phytochemical profile or anti-diabetic activity, although we previously reported that this plant is currently being used as a herbal remedy for diabetes in the southeast region of Puerto Rico [5].
The beneficial effects of diverse groups of phytochemicals in herbal remedies have been related to activities consistent with their potential use in treating diabetic disorders and complications. The in vivo anti-diabetic activity of plant extracts has been correlated with their flavonoid and total phenolic content [14,15]. Glycosides, flavonoids, tannins and alkaloids have shown reliable activities that may be useful for the treatment of Type 2 diabetes [16]. Also saponins, such as oleanolic acid, exhibit hypoglycemic activity and resveratrol, a phenolic compound, shows insulin-like effects in streptozotocininduced diabetic rats [17,18]. In the case of tannins, two modes of action have been proposed to explain their anti-diabetic potential. At the protein level, tannins act via insulin receptor activation leading to an increase in glucose uptake rate and lower glucose levels. At the molecular level, tannins have significant superoxide scavenging and antioxidant activity [19]. These facts are relevant since high levels of superoxide ions in pancreatic β-cells, block insulin signaling, affecting glucose regulation [20].
Although the traditional use of decoctions of C. speciosus and S. jambos has been reported as both complementary and alternative treatments for diabetes, the antidiabetic effects of the decoctions of these plants, including T. ananassae have not been validated in animal models. Consequently, this report presents the results of in vivo studies using C57BLKS/J db/db and C57BLKS/J ob/ob genetic diabetes animal models. Moreover, we present the systematic characterization of major phytochemicals and potential markers of anti-diabetic activity in the plant aqueous extracts.

Ethnopharmacological survey
The questionnaire used in the ethnopharmacological survey in Puerto Rico was adapted from the one published by TRAMIL Network (www.tramil.net). The second section of the questionnaire follows a structured interview, asking participants to provide information about the botanical remedies used by the family as the first treatment for the ailments included in the survey. Open-ended questions were used to obtain a detailed description of the health problem, treatment preparation, application and results obtained, including dosage and contraindications or side effects for adults and children. Plants having a usage frequency of 20 % or more for a particular ailment were selected for the in vitro or in vivo studies. The Institutional Review Board at the University of Puerto Rico-Cayey approved the questionnaire used for the survey and the informed consent forms. After completing the survey and reviewing the ethnobotanical literature for herbal remedies used for their hypoglycemic effects, medicinal plants used as diabetes adjuvants were identified [3][4][5].

Plant collection and identification
Leaves of C. speciosus, S. jambos and T. ananassae were collected in Puerto Rico. Vouchers of Costus specious (019660), Tapeinochilus ananassae (019553), Syzygium jambos (019663) were numbered and deposited at the George Proctor Herbarium (SJ) in Puerto Rico. José Sustache, Botanist and Head of the PR Department of Natural and Environmental Resources classified the botanical species.

Preparation of decoctions, methanolic and aqueous extracts
For in vitro and quantitative analysis, fresh leaves (50 g) were weighed, freeze-dried using a Freezone 4.5 lyophilizer and extracted overnight with methanol (MeOH) (350 mL) using a Soxhlet apparatus. The resulting extracts were concentrated to dryness by rotatory evaporation (Yamato RE-200) at room temperature to yield from 1 -11 % w/w of extract (1-5 g extract/ 45 -200 g fresh leaves). Stock solutions of dry plant extracts were re-dissolved in various solvents (depending on the phytochemical to be tested). Alternatively, for the preparation of decoctions used in in vitro, in vivo and quantitative analysis, fresh leaves (30 g) were boiled in 100 mL of distilled, deionized water, concentrated to 15 mL, to lyophilize 3 × 5 mL replicates, filtered through cheesecloth and freeze-dried. The resulting solids were re-dissolved in ddH 2 O and used to prepare aqueous extracts of known concentrations. All extracts were filtered using a 0.4 μm Nanopure® filter syringe before assayed [21].

Dosage calculation
A tea of C. speciosus leaves was prepared according to a consensus of dosage and administration reported during the TRAMIL interviews from Puerto Ricans who selfmedicate with this medicinal plant [5]. The preparation consisted of boiling 18.00 g of fresh leaves (six medium size leaves) in 1.89 L of ddH 2 O for 15 min. The resulting extract was filtered through cheesecloth and aliquots of 5-10 mL were freeze-dried to calculate the total solids. The %w/w yield was 6.6 %, which corresponds to 1.14 g of total solids in 1.89 L. According to the survey the daily dosage of the tea is 250 mL (8 oz) or 158.3 mg of tea solids. Assuming an average body weight of 70Kg, the dose of tea solids per Kg of body weight is 2.3 mg/ kg_BW, which was the dose used as reference for preparing the remaining teas and used in all in vivo studies.

Quantitative phytochemical analysis
Total phenolic content (TPC) in methanolic and aqueous extracts dissolved in DMSO was determined using Folin-Ciocalteu method (765 nm) with some modifications, while flavonoids in methanolic extracts were determined by the aluminum chloride method (415 nm) using quercetin as standard, as described by [25]. The bromocresol green method [26] and the Prussian Blue reaction [27] were used to quantify total alkaloids (470 nm) and tannins (395 nm) in methanolic and aqueous extracts using nicotine and tannic acid as standards, respectively. Saponins were determined by the dinitrosalicylic acid (DNS) method (540 nm) as reported previously [17,28].

Study 1
Male BKS.Cg-Dock7m +/+ Leprdb/J (C57BLKS/J db/db) 5 wks old mice (Jackson Laboratories, ME) were maintained under a constant 12-h light:12-h dark cycle with a temperature of 22-24°C. Four to five animals per cage were divided into 12 treatment groups, for a total of 50 mice. The mice had free access to NIH#31 M Rodent Diet pellets (Agricultural Exports Inc., Doylestown, PA) and water. Consumption of food and water were monitored daily. Mice were weighed weekly for 10 wks.

Study 2
Male B6.Cg-Lepob/J (C57BL/J ob/ob), 5 wk old mice (Jackson Laboratories, ME) were maintained as those described in Study 1. Three to four animals per cage were divided into 4 treatment groups (n = 11 or 12 per treatment). The Institutional Animal Care and Use Committees at Universidad Central del Caribe-School of Medicine approved the protocols for Studies 1 and 2.

Treatment administration Study 1
db/db mice were orally gavaged daily for 10 wks with control (sterile water) or 0.2, 2.2, 22 mg/kg_BW) in a 100 μL volume of T. ananassae, C. speciosus and S. jambos aqueous extracts or 220 mg/kg_BW in a 100 μL volume of T. ananassae, and C. speciosus aqueous extracts, for a total of 12 treatments.

Study 2
ob/ob mice were orally gavaged daily for 10 wks with control (sterile water) or 2.2 mg/kg_BW in a 100 μL volume of T. ananassae, C. speciosus and S. jambos aqueous extracts.

Glucose tolerance test (IP-GTT)
IP-GTT was performed after a 12-h fast as depicted in Fig. 1. A 10 % glucose solution was injected i.p. (1 mg/g BW) and blood samples were obtained from the tail vein at 0, 15, 30, 60, 90, and 120 min after glucose administration as in [29]. Mice were tail bled at 0 wks (before treatment intervention), at 5 wks and at 10 wks (end of the study). Blood glucose was determined using a glucometer (Optium EZ, Abbot Laboratories and One Touch Ultra Mini, Johnson and Johnson) via test strips. Limit of detection of Optium EZ is 500 mg/dL (used in Study 1) and of One Touch Ultra Mini is 600 mg/dL (used in Study 1 and 2).

Insulin tolerance test (IP-ITT)
IP-ITT was performed at the end of the study to determine the effects of insulin on blood glucose modulation between control and treatment groups as depicted in Fig. 1. The db/db mice become severely diabetic by 4 wks of age as described by Buchanan [30] thus we used an insulin (Humulin R U-500, Lilly, Indianapolis, IN) dose of 8units/g BW to achieve adequate glucose lowering. Mice were fasted (12-h), on a different day than that of the IP-GTT; they were tail bled pre-injection and then glucose concentrations were determined at 15, 30, 60, 90, and 120 min post-injection as by Keller et al. [29].

Plasma insulin detection
Plasma insulin levels were analyzed using the Ultrasensitive Mouse Insulin ELISA Kit (Crystal Chem Inc., IL) jambos treated mice at a concentration of 2.2 mg/ kg_BW at the IP-GTT and IP-ITT at 0 or 15 min after glucose or insulin injection, respectively (see Fig. 1). For study 2 we used plasma samples collected at 10 wks of S. jambos and T. ananassae treated mice compared to controls at the IP-GTT and IP-ITT at 0 or 15 min after glucose or insulin injection, respectively (see Fig. 1). The plates were read using the Glomax Multi Detection System (Promega, WI) at an absorbance of 450 nm.

Statistical analysis
All in vitro tests were performed in triplicate. Results are expressed as means ± standard error of the mean. The concentrations of phytochemicals obtained from the quantitative analyses were estimated from the leastsquares regression lines (r 2 98.99 -1.00 %). In vitro statistical analyses were done using GraphPad Prism version 5.0b (San Diego, CA). The data were analyzed using regular one-way analysis of variance procedures. Values P < 0.05 were considered significant. For in vivo studies, to assess the behavior of all the variables in our models, a normality diagnostic was performed.

Results
Profile of surveyed population using medicinal plants Table 2 shows the profile of the surveyed population that uses herbal remedies as the first treatment for the ailments assessed during the ethnopharmacological survey.
A total of 10 out of 118 (8.5 %) families in the study population reported using herbal remedies for diabetes as alternative or complementary treatments. The median Fig. 1 Schematic drawing of in vivo experimental procedures. In study 1 and study 2, the mice (5 wks old) were randomly assigned to the various experimental treatments that were administered daily for 10 wks. An IP-GTT was performed at 0, 5 and 10 wks of the study, while an IP-ITT was performed at 10 wks. In each intervention blood was collected at 0 and 15 min post injection to detect insulin plasma levels Nine different plant remedies were used among the study population the last time that one of the family members was treated for diabetes. The more frequent descriptions of diabetes as a health problem were uncontrolled sugar levels (high or low) and dizziness. Other descriptions included dry mouth, excessive thirst, sleepiness, restlessness, not feeling the legs, blindness or poor vision and frequent urination. About half of the treatments were reported once (55 %) and the use was classified as not significant according to TRAMIL methodology [31]. Most of the plants (60 %) were obtained at the store or market. Decoction of fresh leaves was the most frequent mode of preparation of the herbal remedies. Dosages were variable, with most families reporting the use of one cup of tea (decoction) daily or during one week. The herbal remedies prepared from Costus speciosus (J. Koening) Sm. (Zingiberaceae) and Tapeinochilos ananassae K. Schum. (Zingiberaceae) were selected for the present study as their use was significant according to TRAMIL and they are commonly known as "insulin" in the studied population.

Major groups of phytochemicals were found in plant extracts
To establish product integrity during the scientific investigation of complex botanical products, the National Institutes of Health (NIH), National Center for Complementary and Integrative Health (NCCIH) has published guidelines that required the characterization (chemical profile or fingerprint) and identification of relevant marker compounds used for standardization [32]. Qualitative TLC analysis of methanolic and aqueous extracts of the leaves of S. jambos, T. ananassae and C. speciosus showed the presence of flavonoids, alkaloids, phenolic compounds, saponins, terpenoids, tannins and cardiac glycosides (Table 1). Aqueous extracts showed fewer phytochemicals than methanolic extracts. TLC analysis confirmed the presence of a significant number of phenolics in S. jambos methanolic and aqueous extracts, alkaloids in C. speciosus methanolic extracts, flavonoids in T. ananassae methanolic extracts and C. speciosus aqueous extracts, and tannins in T. ananassae aqueous extracts. The results for the quantitative analysis of the extracts presented in Table 3, confirms the qualitative chemical profile and shows a significant concentration of phenolics in S. jambos (0.45 mg QE/mg extract), flavonoids and tannins in T. ananassae (24.3 mg QE/g DW and 3.361 mg Tannic acid/g FW, respectively) and alkaloids in C. speciosus dry or fresh leaves (806.0 % μg Nicotine/mg FW).
Plant extracts do not affect weight gain, food or water intake in C57BLKS/J (db/db) and C57BL/J (ob/ob) mice In Study 1, three out of the 50 mice died from causes unrelated to treatment toxicity or to an unsuccessful gavage. Results show that there is no significant effect of treatment on weight gain, average food or water intake throughout the 10 wks of the study, an indication that the extracts were not toxic to mice (Additional file 1: Figure S1). In Study 2, none of the ob/ob mice died and there was no significant effect of treatments on weight gain, or on average food or water intake. (Additional file 2: Figure S2). Interestingly, overall db/ db mice weighed less, and consumed more water than ob/ob mice.

Plant decoctions modulate C57BLKS/J (db/db) mice blood glucose levels when administered in complement to insulin
Blood glucose levels (BGL) were monitored in mice treated with plant extracts or control (total of 12 treatments) at 5 and 10 wks post-treatment. In Study 1, there was no significant effect of treatments in lowering blood glucose levels at 5 or 10 wks post-treatment. At 5 wks, S. jambos treated mice showed better blood glucose modulation compared to the mice receiving other plant extract treatments (Fig. 2a-d). Mice on C. speciosus 22 (BGL 273 mg/dL), and 220 mg/kg_BW (BGL 211 mg/dL) had lower baseline BGL compared to vehicle treated mice. After the glucose injection, most of the animals treated with C. speciosus 2.2 mg/kg_BW had glucose levels greater than 500 mg/dL and were not able to reduce their BGL to basal state, indicating that glucose tolerance in these mice is impaired (Fig. 2b). Moreover, mice that received C. speciosus 220 mg/kg_BW, tended to show lower BGL (BGL 392 mg/dL) when compared to the animals on the control treatment at the last reading (120 min, BGL 467 mg/dL). Interestingly, mice treated with S. jambos 2.2 mg/kg showed decreased BGL at 90 min (BGL 267 mg/dL) post injection. These results suggest that mice receiving 2.2 mg/dL of S. jambos display gradual blood glucose modulation (increased and decreased) peaks. At 10 wks, baseline BGL of mice on C. speciosus 2.2, 22, 220 mg/kg_BW, and T. ananassae 2.2 mg/kg_BW were lower compared to control mice ( Fig. 2e-h). Furthermore, at 60 min post-glucose mice on all S. jambos treatments had lower blood glucose levels compared to C. speciosus mice, and by 120 min, all of the glucose readings for T. ananassae and S. jambos mice were below the limit of detection of the glucometer (500 mg/dL). Interestingly, after 15 min of glucose administration plasma insulin levels were significantly higher (P < 0.05) in both control and S. jambos 2.2 mg/kg_BW treated mice at 10 wks post-treatment (Fig. 5a). The IP-ITT data showed that baseline BGL for all mice were in a range from BGL 325 to 491 mg/dL ( Fig. 3a- (Fig. 5b).
T. ananassae (5 wks) and S. jambos (5 and 10 wks) significantly reduce BGL in ob/ob mice There were differences in blood sugar level modulation in ob/ob mice treated with different plant extracts. At 5 wks, the T. ananassae (P < 0.006) and S. jambos (P < 0.049) treated mice were significantly more efficient in modulating the effects of the glucose compared to control or C. speciosus treated mice (Fig. 4a). All of the plant treated mice had lower baseline blood glucose levels (S. jambos BGL 297 mg/dL, T. ananassae BGL 262 mg/dL, C. speciosus BGL 337 mg/dL) compared to vehicle (BGL 402 mg/dL) treated mice. Moreover, at 10 wks, S. jambos (P < 0.056) treated mice were more efficient in modulating blood glucose levels when compared with control or mice treated with the remaining plant extracts (Fig. 4b).
The baseline blood glucose of S. jambos mice was lower (BGL 208 mg/dL) overall. Upon injection of glucose, S. jambos treated mice displayed overall lower peaks in blood glucose. There were no differences in the levels of plasma insulin in response to a glucose injection in control, T. ananassae or S. jambos treated mice (Fig. 5c). In Study 2, the IP-ITT test blood glucose baseline values were lower than in Study 1 and ranged from BGL 221 to 292 mg/dL. By 60 min post-injection glucose concentrations were similar in S. jambos (BGL 99 mg/dL) and control (BGL 91 mg/dL) groups. In addition, at 120 min BGL remained at non-diabetic levels (<200 mg/dL). These values were BGL 78 mg/dL (control), BGL 88 mg/dL (C. speciosus), BGL 107 mg/dL (T. ananassae) and BGL 92 mg/dL (S. jambos). Importantly, plasma insulin levels were significantly higher (P < 0.02) after the insulin injection in mice regardless of treatment (Fig. 5d).

Discussion
Globally, as of 2013, an estimated 382 million people have diabetes. In 2012 and 2013 diabetes resulted in 1.5 to 5.1 million deaths per year, making it the 8th leading cause of death worldwide [33]. Diabetes was the third cause (10 %) of deaths in 2010 in Puerto Rico, highlighting the importance of diabetes as a health problem in the Island [34]. Interestingly, traditional medicine users in Puerto Rico rely on the use of herbs to lower blood glucose levels [5]. Thus, in an effort to validate the traditional use of these herbs associated with diabetes management and treatment, we undertook an ethnopharmacological approach to screen commonly used but less studied herbal remedies, their biological activities and possible bioactive compounds. Plant extracts were qualitatively screened by thin layer chromatography (TLC), while flavonoids, phenolics, saponins, tannins and alkaloids were quantified. We also assessed biological activities in vivo using two genetic diabetes models.
Our results show that S. jambos aqueous extract had significantly higher concentrations of phenolic compounds relative to T. ananassae and C. speciosus. Phenolics can be classified into phenolic acids (hydroxycinammic acids, phenylpropanoids), flavonoid polyphenolics (flavonoids), and non-flavonoid polyphenolics (tannins). Plant polyphenols are powerful antioxidants that exert their effects either by scavenging reactive oxygen and nitrogen species, or by functioning as chain-breaking peroxyl radical scavengers [35,36]. Growing evidence shows that excess generation of reactive free radicals, largely due to hyperglycemia, causes oxidative stress in pancreatic β-cells, which further exacerbates the development and progression of diabetes and its complications [37]. Reynertson and collaborators analysis of tropical edible fruits by HPLC-PDA showed that ellagic acid and derivatives were the most abundant phenolic compounds in the methanol-formic acid lyophilized extracts of S. jambos fruits [13]. This phenolic compound along with t-cinnamic acid, a phenolic acid found in cinnamon bark, accounted for most of the Total Phenolic Content and antiradical activity of the extracts [38]. Also, several ellagitannins were detected in plant extracts of S. jambos from Japan and flavonoids have been isolated from aqueous, methanolic and ethanolic extracts of S. jambos leaves among which myricetin and quercetin 3-O-β-D-xylopyranosyl(1-2)-α-L-rhamnopiranoside showed potent anti-inflammatory activity [12]. Although we did not find other published studies that reference the in vivo hypoglycemic activity of aqueous extracts of S. jambos leaves, an infusion of the leaves of S. jambos/S. cumini was tested for its antidiabetic activity by a glucose tolerance test in a randomized, parallel, double-blind clinical trial in non-diabetic and diabetic subjects, with no significant effect in blood glucose levels [39,40]. S. cumini plus placebo tablets, placebo tea plus glyburide tablets (5 mg twice a day), or placebo tea plus placebo tablets were evaluated in these studies. The authors used two different botanical species indistinctively to prepare the teas, thus the activities reported for each botanical species is uncertain. Moreover, S. cumini has been thoroughly studied at doses similar to the ones reported in this study for its hypoglycemic, antioxidant and antihyperlipidaemic activities, among others [16,41]. Again, the pharmacological activities of S. cumini have been attributed mainly to the presence of flavonoids and phenolics.
Our studies show that T. ananassae had higher concentrations of flavonoids and tannins compared to S. jambos a Plasma insulin levels of db/db mice before (0 min) and after (15 min) a 10 % glucose injection 10 wks post-treatment with water (Control) or 2.2 mg/kg_BW S. jambos, aqueous extracts. b Plasma insulin levels of db/db mice before (0 min) and after (15 min) a 8U/g BW insulin injection 10 wks post-treatment with water (Control) or 2.2 mg/kg_BW S. jambos, aqueous extracts. c Plasma insulin levels of ob/ob mice before (0 min) and after (15 min) a 10 % glucose injection 10 wks post-treatment with water (Control) or 2.2 mg/kg_BW S. jambos or T. ananassae, aqueous extracts. d Plasma insulin levels of ob/ob mice before (0 min) and after (15 min) a 8U/g BW insulin injection 10 wks post-treatment with water (Control) or 2.2 mg/kg_BW S. jambos or T. ananassae, aqueous extracts. Columns represent means ± SEM. Differences are significant when *P < 0.05 and C. speciosus extracts. Among the phenolic compounds, the flavonoids, besides functioning as antioxidants, also act as insulin secretagogues that may improve glucose uptake in peripheral tissues. In addition, flavonoids may regulate the activity or expression of rate-limiting enzymes involved in carbohydrate metabolism pathways [42]. Tannins are also phenolic compounds of higher molecular weight commonly used as healing agents. In Ayurveda formulations tannin-rich plants are used for the treatment of diarrhea, while other studies showed that tannins enhance glucose uptake and inhibit adipogenesis [24]. Finally, C. speciosus extracts showed a greater amount of alkaloids when compared to T. ananassae and S. jambos. This could explain the toxicity of C. speciosus teas reported in ethnobotanical accounts [4]. Alkaloid fractions have shown beneficial effects in treating hypercholesterolemic, hypertriglyceridemic, hyperlipidemic and/or dyslipidemic conditions and their related complications linked to metabolic disorders such as obesity and diabetes [43]. Phytochemicals reported from C. speciosus and other species of the genus Costus that have shown hypoglycemic activity and increased plasma insulin concentrations in alloxaninduced diabetic rats include Quercetine glycosides and the pentacyclic triterpene β-Amyrin [44,45]. An extract of C. speciosus was reported to reduce blood glucose concentrations of streptozotocin-induced hyperglycemic rats. A methanolic extract of dried Costus afer Ker Gawl also reduced blood glucose concentrations in streptozotocin induced hyperglycemic rats, stimulating glucose transport in adipocyte cells, suggesting an ability to improve glucose uptake in vivo [29].
In the current study, we present the differential effects of S. jambos, T. ananassae or C. speciosus aqueous plant decoctions on glucose levels of two different genetic mouse models of type-2 diabetes. Our first mouse study assessed the efficacy of the plant extracts using a mouse model [C57BLKS/J (db/db)] that develops hyperglycemia as young as 6 wks of age [30]. Concentrations of the plant extracts in teas administered to mice were evaluated at a dose smaller (0.2 mg/kg_BW) and commonly consumed by humans (2.2 mg/kg_BW), and at 10 (22 mg/kg_BW) and 100 (220 mg/kg_BW) times greater the concentration used by humans. Our results show that the plant extracts do not affect the weight, food or water intake of db/db mice. The glucose peak in nondiabetic individuals rises within one hour of food consumption, but after two hours, blood glucose returns to baseline levels. However, in people with diabetes the glucose peak is higher and it takes longer to decline back to baseline levels upon food consumption. Importantly, studies describe that the experimental basal BGL limit for diabetic animals is 200 mg/dL [8,46]. Herein we show a severe hyperglycemic animal model where BGL drop to values lower than 200 mg/dL after administration of plant decoctions. Insulin tolerance tests show that mice gavaged with all of the doses of S. jambos (0.2, 2.2 and 22 mg/kg_BW), the higher doses of C. speciosus (22 and 220 mg/kg_BW), and one dose of T. ananassae (2.2 mg/ kg_BW) plant extracts tend to modulate glucose better as compared to controls. This suggests that the effect these plants exert in the db/db mouse model can be complementary to conventional therapy (insulin injection), rather than when administered as alternative therapy (extract administration alone). Moreover, our data shows that plasma insulin levels are significantly higher after the insulin injection. Since the increase of plasma insulin levels occurs regardless of treatment, it is possible that the mechanistic reason for obtaining lower blood glucose levels in response to S. jambos treatment in this mouse model is not due to increases in insulin secretion but to other mechanisms such as lower gluconeogenesis or better glucose uptake by sensitive tissues such as liver, muscle, and adipose tissue. It is known that a balance between insulin secretion and insulin action maintains normal glucose tolerance [47]. The lack of statistical significance of blood glucose modulation in Study 1 could be attributed to the small sample size and to the mouse model chosen. The C57BLKS/J (db/db) model is obese via a mutation in the leptin receptor, develops insulin resistance very quickly and has progressive β-cell depletion [29]. This mouse model is commonly used to detect effects of test products in the preservation of βcell or insulin resistance.
Since in Study 1 we wanted to test a wide range of doses with various plant extracts, the sample size used per treatment was smaller that the one in Study 2. Previous to conducting mouse Study 2, we carried out a statistical power analysis to assess the sample size needed to detect changes in blood glucose taking into account the variability in blood glucose values obtained in Study 1. In Study 2 we used the C57BL/J (ob/ob) Type-2 diabetes mouse model. This model is also genetically obese but the mutation lies in the production of leptin [48,49]. The C57BL/J (ob/ob) mice show tendencies in better modulation of euglycemia due to a robust and persistent compensatory pancreatic β-cell response, matching the insulin resistance with hyperinsulinemia. Studies with ob/ob mice show that at 4 and 8 wks of age these mice have not developed hyperglycemia [50]. In Study 2, we show that plant extracts, especially T. ananassae and S. jambos, lowered glucose peaks and kept these glucose peaks stable over the 10 wks trial. In healthy people, blood glucose levels are stable, with the exception of the increased level after meals. Lowering peaks and maintaining stable glucose levels demonstrate the capacity of the body to break glucose, secrete insulin and uptake glucose efficiently. Studies also show that by 15 wks the ob/ob mice develop hyperglycemia [30,50]. In contrast to Study 1, in Study 2 mice treated with the plant extracts or controls have stable non-diabetic glucose values (200 mg/dL or less) at 60 min following an i.p. insulin injection. Also, ob/ob mice show plasma insulin levels that are 20 times greater than those detected in db/db mice. These results suggest that in the ob/ob mice model we are detecting both endogenous and injected insulin levels at the time the IP-ITT was performed. The ob/ob mice show mild hyperglycemia that is apparent from 8 to 12 wks of age. At this point, pancreatic β-cell compensation occurs and increased insulin levels are secreted for glucose homeostasis [51]. Our results suggest that the insulin injected is responsible for lowering blood glucose levels in the ob/ob mice and not the plant extract. These results differ from that observed in the insulin resistant model. Potential explanations as to why the plant decoctions resulted in mild modulation of BGL include the severity of the insulin resistance associated with these particular animal models. In addition, contrary to other studies that show anti-diabetic efficacy of plant extracts [8,39,40,46,52] our study aimed to mimic the effects of using traditional remedies by the local Puerto Rican population instead of detecting the activities of these compounds extracted by organic solvents of different polarities [53].

Conclusions
Our results are the first to show the qualitative and quantitative chemical profile of three commonly used plants by the Puerto Rican population to lower BGL. This report documents for the first time, the in vivo mouse studies and the identification and quantification of phenolic compounds, flavonoids, tannins, alkaloids and saponins in T. ananassae plant extracts. Moreover, we document a detailed protocol that was used to detect these compounds. Out of the three plants extracts, S. jambos shows better in vivo efficacy in lowering BGL, when evaluated in diabetes genetic mouse models (db/db). At 5 wks, individual mouse BGL readings show that depending on the timepoint either none (90 min), or only one or two animals have BGL > 500 mg/dL. These data substantiate our results that show that S. jambos at levels consumed by humans (2.2 mg/kg) tends to modulate BGL better than the rest of the plants. Future studies will evaluate the efficacy of these extracts as complementary therapy. Studies will investigate blood glucose modulation when the extracts are administered in addition to other glucose modulators. Our results demonstrate that in an insulin resistant model (db/db) when the extracts are used with insulin (complementary therapy), the plant extracts tend to modulate glucose better than controls, while in ob/ob mice this effect is not seen. In some cases, we see that lower doses in complement with insulin have better glucose modulation than the same plant at higher doses (C. speciosus 22 vs. C. speciosus 220). It is possible that the plant extracts display a synergistic or additive effect with insulin that is lost by increasing its concentrations. However, without performing proper combinatorial index analysis [54][55][56], we cannot be certain that such effects indeed occur. Furthermore, the study reveals that herbal remedies used as diabetes adjuvants contain anti-diabetic drug principles that require further characterization and studies to assess mode of action to provide a rich source for new drug discovery.