Evaluation of the Antihyperglycemic Activity of Juglans Mollis and Hamelia Patens Through in Vitro and in Vivo Methods


 Objectives: Diabetes mellitus is one of the most common noncontagious diseases in the world. The International Diabetes Federation (IDF) reported that in 2017, approximately 425 million people worldwide suffered from diabetes. Drugs used for diabetes treatment have unwanted side effects, so new safe drugs are needed. Some natural products have antihyperglycemic activity and are less toxic than currently used drugs. In this work, we evaluated the antihyperglycemic activity of extracts of Juglans mollis and Hamelia patens as well as their cytotoxicity through in vitro and in vivo methods.Materials and methods: Five extracts of each plant were subjected to in vitro amylase and glucosidase inhibition tests and subsequently analyzed by the ex vivo everted sac test. Additionally, their in vivo antihyperglycemic activity was evaluated.Results: Each of the extracts of J. mollis and the polar extracts of H. patens showed antihyperglycemic activity in the in vivo model, but in the in vitro model, the extracts showed different effects; some of the extracts inhibited one or both digestive enzymes, and others reduced the absorption of glucose through the intestine.Conclusions: In this article, we contribute to elucidating the antihyperglycemic mechanism of H. patens, and we report for the first time the antihyperglycemic activity of J. mollis and its possible mechanisms of action.


Introduction
Diabetes mellitus is one of the most common noncommunicable diseases, along with cardiovascular diseases and cancer 1 . Approximately 463 million people between the ages of 20 and 79 worldwide have diabetes 2 . The main characteristic of the disease is the inability to regulate blood glucose levels due to insulin resistance in peripheral tissues and/or due to the quantity or quality of insulin secreted by the beta cells of the pancreatic islets 3 . Long-term uncontrolled diabetes mellitus can cause ketoacidosis, retinopathy, nephropathy, neuropathy, hypertension, cerebrovascular accident, etc 4 . Treatments for lowering or controlling blood glucose levels include insulin (obtained from different sources) and hypoglycemic agents with various mechanisms of action that either inhibit the action of digestive enzymes or inhibit the action of enzymes responsible for absorption, carbohydrate reabsorption, etc.
Almost all of these medications generate side effects and are associated with a high cost of treatment 5,6 .
For this reason, it is necessary to identify new affordable products that can help control hyperglycemia with fewer side effects.
Results obtained in in vivo models are widely accepted in both the laboratory and the clinic since these models involve the entire organism and therefore consider the entire metabolic process. However, the use of in vivo models is expensive, is time consuming and requires a large number of products to be evaluated and a large number of animals 7 . On the other hand, in vitro tests are faster, simpler and more economical. Such tests have been used to predict in vivo activity and help elucidate the mechanism of action of drugs.
Natural products have been shown to be good sources of molecules with various biological activities and fewer side effects 8 . A number of authors have reported the e cacy of plant extracts in lowering blood glucose using different in vivo models, while others have demonstrated the potential hypoglycemic or antihyperglycemic activity of extracts or natural products using in vitro or ex vivo models 6 .
Hamelia patens belongs to the Rubiaceae family and is commonly known as chacloco, coral plant or re bush; it is a species of large evergreen shrub from subtropical and tropical America. It is found from In the present work, we evaluate the antihyperglycemic activity of different extracts of J. mollis and H. patens through in vitro, ex vivo and in vivo methods and suggest possible mechanisms of action of these plants.

Results
The cytotoxic activity of the extracts was evaluated by a cell viability assay using MTT as an indicator of growth. None of the extracts except the methanolic extract of J. mollis, which showed a CC 50 of 285.7 ± 9.1 µg/mL, showed cytotoxicity at the maximum concentration of 500 µg/mL. The doxorubicin positive control showed a CC 50 of 12.8 ± 0.5 µg/mL. The inhibitory effects of the extracts on the activity of enzymes involved in the digestion of carbohydrates is shown in Table 1. All the extracts of J. mollis except the hexane extract inhibited the activity of glucosidase. Furthermore, the butanolic extracts and the aqueous residue inhibited amylase activity. Interestingly, the H. patens extracts were not able to inhibit glucosidase activity at the maximum concentration of 330 µg/mL, although the methanol and butanol extracts strongly inhibited amylase activity.
To evaluate glucose absorption through the inverted intestinal sac, glucose absorption kinetics were rst analyzed in the absence of inhibition, that is, in the negative control condition. In the absence of inhibition, the kinetics showed a correlation coe cient of 0.9943, with an approximate ux of glucose absorption through the intestinal wall of 50 mg/dL every hour (not shown). When 5 mg/dL empagli ozin, an inhibitor of glucose transport through SGLUT, was added, the kinetics showed an excellent correlation coe cient of 0.9932, with an absorption ux of approximately 33 mg/dL every hour (not shown). To determine the reduction in glucose uptake induced by the positive control, the area under the curve was calculated. The area under the curve of the negative control condition was taken as 100%, and it was determined that at the concentration used, empagli ozin, reduced glucose absorption through the intestine by 38.4%.
The extracts were subjected to the inverted sac absorption test at three different concentrations, and their absorption kinetics were determined. The area under the curve was determined for each experiment and statistically compared to that obtained for the negative control. Compared to no treatment, the methanolic, butanolic and aqueous residue extracts of J. mollis signi cantly inhibited absorption. The ethyl acetate and butanolic extracts of H. patens at low concentrations and the aqueous residue of H. patens at high concentrations also caused a signi cant reduction in intestinal absorption (Fig. 1). The concentration of glucose absorbed after half of the kinetics, that is, at 90 min, is shown in Fig. 2.
All in vivo tests were carried out following the Good Practices for the Care and Management of Experimental Animals. The kinetics of oral starch tolerance was rst determined. The hyperglycemic peak was reached 30 min after the administration of starch; subsequently, the blood glucose levels returned to the initial values. When acarbose (positive control) was administered together with starch, the hyperglycemic peak (30 min) signi cantly decreased by approximately 65.7% (not shown).
To evaluate the effect of acarbose on kinetics, the reduction in hyperglycemia induced was determined by calculating the area under the curve, with that of the starch curve being considered 100%. Acarbose was determined to induce an approximately 29.3% reduction in all kinetics.
On the other hand, the control PPHT extract behaved statistically the same as acarbose, with reducing the hyperglycemic peak at 30 min by approximately 65.7% and the total kinetics by approximately 33.4%.
All the extracts were evaluated at three different doses under the same conditions, and the results are shown in Fig. 3. For all extracts except the hexanic extracts of both J. mollis and H. patens, a signi cant decrease in the area under the curve compared with that of starch was found. Interestingly, most of the J. mollis extracts induced a signi cant reduction in the area under the curve similar to that induced by the positive controls.
The normalized glucose level during the hyperglycemic peak (30 min) for each extract at each concentration is shown in Fig. 4. Four extracts had statistically the same effect as starch, but most exhibited signi cantly reduced the hyperglycemic peak.

Limitations
In the present work, only extracts of different polarity obtained from plants were included, not compounds was isolated. Hexane extracts could not be evaluated in the ex-vivo assay.

Discussion
The extracts used in this work were obtained from J. mollis and H. patens, both of which were collected in Nuevo León, Mexico.
J. mollis bark extract has been reported to have powerful antioxidant, hepatoprotective, and antimycobacterial activity 20,21 . To date, its use as an antihyperglycemic agent has not been reported, but recently, the use of J. nigra and J. neotropica as antidiabetic agents was reported 22,23 .
The methanol extract of J. mollis had a moderate cytotoxic effect on the Vero cell line. For this reason, the residue was fractionated with solvents of increasing polarity. The butanolic, methanolic, ethyl acetate extracts and the residue more strongly inhibited glucosidase activity than acarbose as a control, while only the butanolic extract and the residue inhibited amylase activity. The residue and the methanolic and butanolic extracts signi cantly reduced glucose uptake both in full and half kinetics (90 min). In the inverted sack test, the activity of the hexane extract could not be evaluated because the solvent damages the integrity of the intestine. On the other hand, all the extracts were subjected to an in vivo oral starch tolerance test. Interestingly, all the extracts except hexanic at low concentrations decreased the area under the curve in the starch tolerance test. The hexane and butanolic extracts at low concentrations and the residue and the hexane extract at high concentrations failed to reduce the hyperglycemic peak (at 30 min), while the methanolic and ethyl acetate extracts showed a similar inhibitory effect as the control.
To our knowledge, this is the rst time that the antihyperglycemic activity of J. mollis has been reported. The methanolic extract exerted a potent antihyperglycemic effect similar to that of the positive control, possibly by inhibiting glucosidase and reducing absorption through the intestine; however, the extract induced moderate cytotoxicity, although none of the animals treated with this extract showed signs of intoxication in the in vivo experiment. The ethyl acetate extract also had a strong effect in vivo, and the ndings suggest that it exerted this effect through glucosidase inhibition. The butanolic residue and extract also exhibited in vivo antihyperglycemic activity comparable to that of the positive control. According to the ndings, both the residue and extract exert this effect through inhibiting digestive enzymes and delaying intestinal absorption.
On the other hand, various authors have demonstrated that H. patens has antidiabetic and hypoglycemic activity. Andrade-Cetto et al. 24 . postulated that the hypoglycemic activity of this plant is due to the inhibition of glucosidase; however, we found that none of the extracts inhibited glucosidase at concentrations below 330 µg/mL, which was the maximum concentration evaluated. As we found for the rst time that the methanolic and butanolic extracts of H. patens inhibit amylase, these ndings suggest that enzyme inhibition could be the mechanism underlying the antihyperglycemic activity of H. patens extracts. To our knowledge, this is the rst time that glucose absorption was evaluated as a mechanism of the antidiabetic activity of H. patens. Both the residue and the butanolic and ethyl acetate extracts of H. patents signi cantly reduced the absorption of glucose, indicating that reduced glucose absorption could be the mechanism through which these extracts exert their antihyperglycemic activity. Taken together, these ndings could suggest that the H. patens exerts hypoglycemic or antihyperglycemic effect through inhibition of amylase and a reduction in carbohydrate absorption.
In conclusion, in this article, we contribute to elucidating the antihyperglycemic mechanism of H. patens and report for the rst time the antihyperglycemic activity of J. mollis and its possible mechanisms of action. We are currently working on isolating the compound(s) responsible for this activity. Both plants were dried at laboratory temperature for 7 days. Two hundred grams of powdered material was removed with methanol at room temperature on a shaker for 60 min (3 times). The extract obtained was dried under reduced pressure. The methanolic extract was kept at -4°C until use. Twenty grams of the methanolic extract was resuspended in 250 ml of water and subjected to differential extraction with hexane (100 ml x 3), ethyl acetate (100 ml x 3) and butanol (100 ml x 3). The hexane, ethyl acetate and butanol extracts were evaporated under reduced pressure to dryness and stored at -4°C until use.

α-Glucosidase inhibition assay
The optimized and validated method described by Granados-Guzmán et al. 25 . was followed exactly.

α-Amylase inhibition assay
The optimized and validated method described by Granados-Guzmán 26 was followed exactly.

Cytotoxicity
Vero cells were maintained in DMEM supplemented with 10% FBS, 1% L-glutamine and 1% antibiotics and incubated at 37 ºC and 5% CO 2 . When they reached 90% con uence, the cells were detached with trypsin, and after centrifugation at 3500 rpm for 5 min, the cells were counted in a Neubauer chamber. Approximately 50,000 cells were seeded in each well of a 96-well microplate and incubated overnight under the same conditions 27 . Subsequently, the medium was discarded, and 200 µL of medium (negative control), medium supplemented with doxorubicin (positive control) or medium supplemented with extract ( nal concentration between 0.5 and 500 µg/mL) was added to each well. The plates were incubated under the same conditions for 48 hours. Growth was checked under a light microscope to rule out contamination, and the contents of the plate were decanted. Cell growth was evaluated by adding 200 µL of MTT solution (0.5 mg/mL in buffer) to each well and incubating the plates for 3 hours under standard conditions. The supernatant was decanted, and the crystals were dissolved in 200 µL of DMSO 28 . The plates were shaken, and the absorbance of each well at 540 nm was measured. All experiments were repeated ve times on 3 different plates. The absorbance of each well was compared to that obtained of the negative control (100% viability or 0% inhibition), and the percentage of cytotoxicity was calculated. The data for each plate were handled separately. Cytotoxicity percentages were plotted as a function of extract or doxorubicin concentration, and the mean cytotoxic concentration (CC 50 ) was determined by interpolation. The animals were sacri ced by cervical dislocation. An abdominal incision was made, and the small intestine was removed. Approximately 5 cm long pieces of intestine from the duodenum and the rst part of the jejunum were isolated. Each piece was washed with saline solution, and one end was tied with surgical thread. With the help of a glass rod, the ligature was moved to the opposite side of the tissue, and once the piece of intestine had been inverted, the other end was tied to a Pasteur pipette. The intestinal sac was lled with physiological solution (0.15 M NaCl, pH 7.4), and absence of leaks was veri ed.

Intestinal glucose absorption assay
One of the intestinal sacs was incubated in saline at 37°C with constant bubbling for 3 hours to obtain baseline absorption data (negative control). The other sacs were incubated in saline solution supplemented with empagli ozin (5 mg/dL, positive control) or test extract (5, 10 and 20 mg/dL) under the same conditions. During the incubation, aliquots were taken from the inside of the bag at 0, 30, 60, 90, 120, 150 and 180 min. The glucose concentration absorbed by the intestine was measured with an Accuchek Performa portable glucometer (Roche). A graph of the intestinal glucose absorption kinetics was obtained. The area under the curve was calculated using the midpoint Riemann sum to determine the effect of the extracts on the kinetics. In addition, the absorption percentage was determined at each time point, and a signi cant difference with respect to the negative control was observed at 90 min.
Oral starch tolerance test Normoglycemic male Wistar rats with an average weight of 220 g were used. All animals were kept in the vivarium throughout the study period at a temperature of 25°C and a relative humidity of 40 to 50% on a 12-hour light and dark cycle and were also provided food and water ad libitum. All the experiments were The procedure reported by Yusoff et al. 32 . was followed with slight modi cations. Brie y, the rats were divided into groups of 5 rats. Each group received oral treatment: Group 1 was administered water (1 mL/kg) Group 2 was given corn starch (1 g/kg), group 3 was administered corn starch plus acarbose (0.5 mg/kg), and Group 4 was given starch plus protein hydrolysate from the legume Mucuna pruriens (PPHT; 0.5 mg/kg), which was included as a control extract for comparison 29 . Each of the 8 extracts was evaluated at concentrations of 0.5, 2.5 and 5 mg/kg. After a 5-hour fast, basal glucose levels in whole blood (drawn from the tip of the tail) were recorded using an Accu-chek Performa portable glucometer (Roche). Subsequently, carbohydrate (1 g/kg corn starch) with or without inhibitor (acarbose, PPHT or extract) was administered orally, and glucose levels were measured at 15, 30, 45, 60 and 120 min.
Normalized glycemic curves (ratio of the glucose concentration at each time point and at time zero for each rat) were constructed for each group, and the area under the curve was calculated using the midpoint Riemann sum to determine the effect of each agent on the kinetics. In addition, the percentage reduction in the glucose peak was determined after 30 min, with the peak generated by starch taken as 100% elevation.

Statistical analysis
In vivo experiments were performed ve times, those in vitro and ex vivo models were done in triplicate. Results are reported as the mean ± SD. The effect of each extract or positive control was compared that of the negative control by one-way analysis of variance (ANOVA) followed by Dunnett's test. A p value < 0.05 indicated statistical signi cance. Area Under the Curve Oral Tolerance to the controls and starch extracts. * no signi cant difference (p <0.05)