(−)-Tarchonanthuslactone Exerts a Blood Glucose-Increasing Effect in Experimental Type 2 Diabetes Mellitus

A number of studies have proposed an anti-diabetic effect for tarchonanthuslactone based on its structural similarity with caffeic acid, a compound known for its blood glucose-reducing properties. However, the actual effect of tarchonanthuslactone on blood glucose level has never been tested. Here, we report that, in opposition to the common sense, tarchonanthuslactone has a glucose-increasing effect in a mouse model of obesity and type 2 diabetes mellitus. The effect is acute and non-cumulative and is present only in diabetic mice. In lean, glucose-tolerant mice, despite a slight increase in blood glucose levels, the effect was not significant.


Introduction
Type 2 diabetes mellitus (T2DM) is a metabolic disease defined by the presence of chronic hyperglycemia due to a simultaneous development of insulin resistance and a relative defect of the pancreatic islet to secrete insulin [1]. It is currently known that obesity-related metabolic inflammation plays an important role both in the installation of insulin resistance and damaging of the pancreatic islets [1]. Therefore, it is expected that therapeutic methods aimed at reducing inflammation may impact positively on the control of glucose homeostasis [2]. Currently, there is only one report of a clinical trial showing the metabolic benefits of reducing inflammation in diabetes using salsalate, which is derived from the bark of the willow tree [3]. However, additional studies, using different animal models [4] have evaluated other natural compounds for their potential as therapy for this disease [5][6][7].
Tarchonanthuslactone (1), an ester of dihydrocaffeic acid (2), was isolated in 1979 from the leaves of the tree Tarchonanthus trilobus [8]. The genus Tarchonanthus is found in southern Africa and is used in folk medicine. Traditional health practitioners have used T. camphorate for diabetes treatment; moreover, its anti-inflammatory and cytotoxic activities have been reported by van de Venter and coworkers [9]. Tarchonanthuslactone (1) has a privileged structure, presenting the usually bioactive α,β-unsaturated δ-lactone motif [10][11][12][13], which makes it a good target for new asymmetric synthetic approaches [14][15][16][17][18]. Interestingly, Hsu and coworkers [19] have reported that caffeic acid (3) has an antidiabetic effect and, because of the structural similarity between 1 and 3, a number of reports have, thereafter, assigned a putative antidiabetic effect to 1 as well [14][15][16][17][18][20][21][22]. Here, we employed a mouse model of diet-induced diabetes to evaluate the effect of 1 and related compounds ( Figure 1) on blood glucose levels. Such compounds may have effects on other metabolic parameters, however, we have focused on blood glucose levels.

Results and Discussion
Swiss mice belong to an outbread strain related to the diabetes prone Akr mouse [23]. Upon feeding on a high-fat diet (31% fat from lard) Swiss mice rapidly develop obesity accompained by insulin resistance and hyperglycemia [24]. In the present study, six-week old male Swiss mice were fed for eight weeks on a high-fat diet and then employed in the experiments.
Compound 3, in the same dose as previously reported [19] was employed as a positive control. Six-hour fasting diabetic mice (median fasting blood glucose levels = 200 mg/dL) were randomly divided into three groups treated, by an intraperitoneal (ip) injection, with a single dose of either 1 (3.0 mg/kg), 2 (3.0 mg/kg) or 3 (3.0 mg/kg) and blood glucose levels were determined over the following 90 min. As depicted in Figure 2A,B, both 2 and 3 exerted a blood glucose-reducing effect, which was significant 30 min after the ip injection of the compounds but resulted in no significant reduction of the area under the glucose curve during the 90 min evaluation. Conversely, 1 exerted an unexpected blood glucose-increasing effect, which was significant at 90 min but resulted in no significant change in the area under the glucose curve during the 90 min evaluation. The effect of 3 obtained in our experiments matches the results reported previously [19], as maximal effect was obtained as early as 30 min after the injection of the compound. Since 2 presents similar glucose-reducing effect as 3, we hypothesized that the presence of the double bond at C2-C3 (Figure 1), which is the only structural difference between 2 and 3, would be involved in this biological effect. In order to test this hypothesis, we synthetized an analogue of 1 that possesses the C2-C3 double bond (4) and determined its effect on blood glucose levels. As depicted in Figure 3A,B, in diabetic Swiss mice, 4 (3.0 mg/kg, ip) resulted in an even more potent glucose-increasing effect than 1. Blood glucose levels were significantly higher than control at 30 and 60 min, leading to a significantly higher area under the glucose curve during the 90 min evaluation period. Thus, we propose that the presence/abscence of the C2-C3 double bond plays no role in either the glucose-reducing effect of 3 or the glucose-increasing effect of 1.
Next, we hypothesized that the alcohol part of esther 1 could be responsible for the glucose-increasing effect produced by this compound. In order to test this hypothesis, diabetic Swiss mice were treated with the synthetic intermediate 5 (3.0 mg/kg, ip) and glucose levels were assessed. As depicted in Figure 3A,B, 5 had no effect on blood glucose levels in diabetic Swiss mice. Thus, we propose that neither the presence/absence of the C2-C3 double bond nor the alcohol part of 1 and 4 are required, per se, for the glucose regulatory effects of the compounds, rather, the whole molecules are required for the glucose-increasing effect. In addition, we evaluated the long-term and cumulative effect of 1 on blood glucose levels. For that, diabetic Swiss mice were treated three times a week, for four weeks with 1 (3.0 mg/kg per dose, ip). Two days after the last dose of the compound, the mice were submitted to a 6-hour fasting and blood glucose levels were determined. As depicted in Figure 4A, the prolonged treatment with 1 resulted in no significant change in fasting blood glucose level, suggesting that its blood-glucose increasing effect is acute and non-cumulative. Finally, lean, non-diabetic Swiss mice, fed on chow (containing 4% fat), were accutely treated with 1 (3.0 mg/kg, ip) and blood glucose levels were determined over a 90 min time-frame. The median six-hour fasting glucose levels was 120 mg/dL and as depicted in Figure 4B, despite a slight increase in glucose levels at 60 min, 1 produced no statistically significant change in blood levels in lean mice.
In summary, 1 has no anti-diabetic effect as previously suggested [14][15][16][17][18]. In fact, despite its structural similarities with 2 and 3, both of which capable of transiently reducing the blood glucose levels of diabetic animals, the accute treatment of diabetic mice with 1 results in a transient increase in blood glucose. Of note, the glucose-increasing effect of 1 is not due to particular features of the molecule, such as double bond between carbons 2 and 3 or the alcohol part of the molecule, but rather, to the whole molecule. The present study highlights how subtle structural modifications in chemical compounds can affect profoundly and unexpectedly its biological activity. The results described herein has potential impact on the design of more potent anti-diabetic compounds.

General Procedures
Starting materials and reagents were obtained from commercial sources and used as received unless otherwise specified. Dichloromethane was treated with calcium hydride and distilled before use. Tetrahydrofuran was treated with metallic sodium and benzophenone and distilled before use. Anhydrous reactions were carried out with continuous stirring under atmosphere of dry nitrogen. Progress of the reactions was monitored by thin-layer chromatography (TLC) analysis (silica gel 60 F 254 on aluminum plates, Merck, Darmstadt, Hesse, Germany). 1 H-NMR and 13 C-NMR were recorded on Bruker 250, 400, 500 or 600 (Rheinstetten, Baden-Württemberg, Germany), the chemical shifts (δ) were reported in parts per million (ppm) relative to deuterated solvent as the internal standard (CDCl3 7.26 ppm, 77.00 ppm, acetone-d6 2.05 ppm, 29.92 ppm, methanol-d4 3.31 ppm, 49.15 ppm), coupling constants (J) are in hertz (Hz). Peaks multiplicities are reported as follows: singlet (s), doublet (d), triplet (t), quartet (q), sextet (sext), multiplet (m), broad singlet (br s). Mass spectra were recorded on a Waters Xevo Q-Tof apparatus operating in electrospray mode (ES). Infrared spectra with Fourier transform (FT-IR) were recorded on a Therm Scientific Nicolet iS5, the principal absorptions are listed in cm −1 . The values of optical rotation were measured at 25 °C in a polarimeter Perkin-Elmer 341, with sodium lamp, the measure is described as follow {[α]D (c = g/100 mL), solvent}.

Biological Experiments
All experimental procedures were performed in accordance with the guidelines of the Brazilian College for Animal Experimentation and were approved by the Ethics Committee at the University of Campinas. Six-week old male Swiss albinus mice were obtained from the University of Campinas Animal Breeding Center and maintained in individual cages at 21 ± 2 °C, with a 12/12 h dark-light cycle with food and water ad libitum. Mice were fed on a high fat diet containing 31% (w/w) of fat from lard.
After 6 h of fasting (starting 0800 h), mice were divided into three groups treated by an intraperitoneal (ip) injection with a single dose of either tarchonanthuslactone (3.0 mg/kg), dihydrocaffeic acid (3.0 mg/kg) or caffeic acid (3.0 mg/kg). Then, blood glucose levels were measured after 30, 60 and 90 min.
Long-term and cumulative effects of tarchonanthuslactone were evaluated in Swiss mice fed on high fat diet for eight weeks and treated three times a week, for four weeks with 3.0 mg/kg per dose ip of tarchonanthuslactone. Two days after the last dose of the compound, the mice were submitted to a 6-hour fasting and blood glucose levels were determined and expressed as percentage of saline treated group. Furthermore, lean non-diabetic Swiss mice fed on a commercial chow diet with approximately 4% (w/w) of fat were treated by an intraperitoneal injection with a single dose of tarchonanthuslactone (3.0 mg/kg) and the control group received saline. Blood glucose levels were measured after 30, 60 and 90 min and the area under this curve was calculated and expressed as percentage of saline treated group.
The number of animals in each group was at least six. Single blood samples were obtained from the tip of the tail and glucose levels were immediately measured using a glucometer from Abbott (Opptimum, Abbott Diabetes Care Inc., Alameda, CA, USA).