Identification of Glycyrrhiza as the rikkunshito constituent with the highest antagonistic potential on heterologously expressed 5-HT3A receptors due to the action of flavonoids

The traditional Japanese phytomedicine rikkunshito is traditionally used for the treatment of gastrointestinal motility disorders, cachexia and nausea. These effects indicate 5-HT3 receptor antagonism, due to the involvement of these receptors in such pathophysiological processes. E.g., setrons, specific 5-HT3 receptor antagonists are the strongest antiemetics, developed so far. Therefore, the antagonistic effects of the eight rikkunshito constituents at heterologously expressed 5-HT3Areceptors were analyzed using the two-electrode voltage-clamp technique. The results indicate that tinctures from Aurantii, Ginseng, Zingiberis, Atractylodis and Glycyrrhiza inhibited the 5-HT3A receptor response, whereas the tinctures of Poria cocos, Jujubae and Pinellia exhibited no effect. Surprisingly, the strongest antagonism was found for Glycyrrhiza, whereas the Zingiberis tincture, which is considered to be primarily responsible for the effect of rikkunshito, exhibited the weakest antagonism of 5-HT3A receptors. Rikkunshito contains various vanilloids, ginsenosides and flavonoids, a portion of which show an antagonistic effect on 5-HT3 receptors. A screening of the established ingredients of the active rikkunshito constituents and related substances lead to the identification of new antagonists within the class of flavonoids. The flavonoids (-)-liquiritigenin, glabridin and licochalcone A from Glycyrrhiza species were found to be the most effective inhibitors of the 5-HT-induced currents in the screening. The flavonoids (-)-liquiritigenin and hesperetin from Aurantii inhibited the receptor response in a non-competitive manner, whereas glabridin and licochalcone A exhibited a potential competitive antagonism. Furthermore, licochalcone A acts as a partial antagonist of 5-HT3A receptors. Thus, this study reveals new 5-HT3A receptor antagonists with the aid of increasing the comprehension of the complex effects of rikkunshito.


Introduction
The 5-HT 3 receptor channels are the only ionotropic receptors within the 5-HT receptor family and belong to the cys-loop family of ligand-gated ion channels (Derkach et al., 1989;Hannon and Hoyer, 2008). These channels occur in the PNS and are highly expressed in the trigeminus (Manteniotis et al., 2013) and the enteric nervous system (Niesler et al., 2003). In the CNS, they are expressed in the striatum, substantia nigra, amygdala, hippocampus and nucleus accumbens (Boess and Martin, 1994). Other 5-HT 3 receptor expressing structures are the nucleus tractus solitarius and the area postrema (Boess and Martin, 1994), which are parts of the vomiting center that trigger nausea and vomiting. 5-HT 3 receptors are involved in many pathophysiological processes, such as nociception and gastrointestinal motility disorders, and the development of nausea and vomiting, therefore showing broad clinical relevance (Doak and Sawynok, 1997;Gershon, 2004;Jeggo et al., 2005;Costedio et al., 2007). Specific 5-HT 3 receptor antagonists, such as ondansetron, are mainly used for the treatment of nausea in various conditions, such as chemotherapy-induced nausea and vomiting (CINV), and nausea during the postoperative phase (PONV) (Cubeddu et al., 1994;Gyermek, 1995).
Kampo is a traditional Japanese phytomedicine that has its seeds in traditional Chinese medicine. Rikkunshito, a combination of eight constituents, is one of the most famous and prescribed kampo medicines (Tominaga and Arakawa, 2015), and it has a well-known physiological effect on the gastrointestinal system, shows orexigen and antiemetic effects, and takes part in the regulation of peristalsis (Takeda et al., 2008;Tominaga et al., 2011;Yanai et al., 2013;Fujitsuka and Uezono, 2014). These effects indicate an antagonism of 5-HT 3 receptors. Therefore, the antagonistic effects of the eight constituents of rikkunshito (Aurantii pericarpium, Ginseng radix, Zingiberis rhizoma, Jujubae (Zizyphi) frucutus, Pinellia tuber, Atractylodis rhizoma, Glycyrrhiza radix and Poria cocos (Hoelen) were investigated as ethanol tinctures. Furthermore, we investigated the established ingredients of the active rikkunshito constituents to identify new 5-HT 3A receptor antagonists. Although the antagonistic and hence the antiemetic effect of Ginseng and Zingiberis due to the action of ginsenosides, gingerols and shogaols is well-described (Ernst and Pittler, 2000;Kim et al., 2005;Lee et al., 2007;Haniadka et al., 2012;Ding et al., 2013), there is currently little knowledge of the effect of the residual rikkunshito constituents on 5-HT 3 receptors.
The aim of this study was the evaluation of the relative contribution of the single constituents of rikkunshito to 5-HT 3 receptor antagonism and the identification of new antagonists. Therefore, we tested the modulatory effect of tinctures and single substances on heterologously expressed human 5-HT 3A receptors using the two-electrode voltageclamp technique. Surprisingly, Glycyrrhiza was identified as the most effective antagonistic tincture among the rikkunshito constituents. Therefore, we concentrated on the investigation of Glycyrrhiza ingredients and identified several new flavonoids as 5-HT 3A receptor antagonists. The drug Radix Glycyrrhiza is used in Kampo medicine for the treatment of pain, gastric ulcers and inflammations of the gastrointestinal and respiratory systems due to its antiphlogistic effect (Kim et al., 2008). A contribution of Radix Glycyrrhiza to the antiemetic effect of rikkunshito due to the action of flavonoids is conceivable.

Solvent Controls
To exclude any unrequested effects of the solvents ethanol and DMSO, we tested their direct activation on non-injected and 5-HT 3A receptor-expressing oocytes. At the maximal used concentration (1 Vol.-%), a negligible direct activation was observed. Moreover, the modulatory effect on the 5-HT 3A receptor response was tested at concentrations of 1.0 Vol.-% for ethanol and DMSO. Ethanol exhibited an inhibition of 14.1 ± 2.6%, and DMSO exhibited an inhibition of 29.1 ± 4.7% (n = 6 − 11). Equivalent volumes of ethanol and DMSO were added to the reference 5-HT solutions. To resolve glycyrrhizin, the solution had to be acidified (pH 5.5). Therefore, we checked the modulatory effect of the pH values on 5-HT 3A receptors. Low extracellular pH values inhibited the currents but high pH showed potentiating effects (Supplementary Figure 2).

Action of the Tinctures on Non-injected Xenopus Oocytes
In the control experiments, at a concentration of 1 Vol.-%, the tinctures of Ginseng, Zingiberis and Atractylodis evoked currents in some non-injected oocytes with desensitizing responses (data not shown). In our experiments with 5-HT 3A receptor-expressing cells, oocytes were rejected if the amplitude of this direct activation was greater than 10% of the 5-HT-induced current; thus, the direct action of the extracts could not prevent the identification of pronounced blocking effects. Moreover, in our blocking experiments, these currents were desensitized during the 30-s preincubation with the tincture. At a concentration of 0.1 Vol.-%, none of the tinctures evoked any current different from that obtained from the control application of Ringer's solution.

Evaluation of Competitive and Non-competitive Antagonists
To determine the apparent mode of antagonism of the identified tinctures and substances with antagonistic effects, we tested the inhibition of currents induced by low (2.5 µM) and high (30 µM) 5-HT concentrations. In the case of a non-competitive mechanism, the inhibition should be independent of the 5-HT concentration, whereas the efficacy of competitive antagonists decreases with increasing 5-HT concentrations. Alternatively to a competitive mechanism, the dependence of the inhibition on the agonist concentration can also be caused by an allosteric modulation. As a control, we tested the competitive antagonists ondansetron (1 nM) and d-tubocurarine (20 µM) (Hope et al., 1996) and the non-competitive antagonist picrotoxin (50 µM) (Das and Dillon, 2005). As expected, the inhibition of picrotoxin was independent of the 5-HT concentration, and the inhibition obtained with d-tubocurarine and ondansetron was reduced at higher 5-HT concentrations (Table 1). Thus, the method used may indicate the mode of antagonism. Nevertheless, a definitive differentiation between competitive, allosteric, and concentration-dependent antagonists must be performed using ligand binding assays.

Tinctures and Substances
Ethanol tinctures of the rikkunshito constituents were obtained from Dr. Peter Lepke (Kronen Apotheke Wuppertal, Germany). Thus, plant preparations at appropriate quality for Japanese kampo medicine were extracted [200 g crushed plant material in 1 l ethanol (45-90% v/v)] for 10 days at room temperature. The tinctures were obtained by filtrating the supernatant, therefore containing no large solid parts of the plants. The dry weight of the extracted substances was determined by removing the solvent under vacuum (Supplementary Table 2). We used the following tinctures in our study: Aurantii Pericarpium [Citrus reticulata Blanco (Rutaceae)] (chinpi, Chen Pi), White Ginseng Radix [Panax ginseng C.A.Mey. The inhibitions of the 5-HT 3A receptor responses by the controls (picrotoxin, ondansetron and d-tubocurarine) and different tinctures and substances at 5-HT concentrations of 2.5 and 30 µM are shown. The level of significance between the inhibitions obtained with two different 5-HT concentrations is indicated by asterisks in the last column (ns, not significant). Furthermore, the apparent mode of antagonism is listed in the last column (c, apparently competitive; nc, non-competitive) (n = 5-7). Unsweetened licorice (Liquirizia purissima from R. De Rosa, Italy) was inlayed in ethanol (70% v/v) under the same conditions. The chemicals were obtained from Sigma Aldrich (5-HT hydrochloride, niflumic acid (blocker for Ca 2+ activated chloride channels), picrotoxin (non-competitive ion channel blocker), d-tubocurarine (competitive nACh and 5-HT 3 receptor antagonist), ondansetron (specific, competitive 5-HT 3 receptor antagonist), (-)-liquiritigenin, licochalcone A, hesperidin, hesperetin, glabridin and glycyrrhizin), Carl Roth (rutin) and PhytoLab (atractylenolide III). The substances were diluted in water, dimethyl sulfoxide (DMSO) or ethanol.

Data Analysis
The test substances were applied in an alternating manner with 5-HT. Therefore, the currents of the test substances or the modulated currents were normalized to the mean of the 5-HT-induced currents before and after the test substance was applied. The concentration-response data were fitted with the Hill equation with variable slope using SigmaPlot 8.0 (SPSS). Thereby, the calculation of the EC 50 and IC 50 -values was done. The deviations are represented by the standard error of the mean (SEM). The datasets were tested for statistically significant differences through Student's t-test using Excel 2010 (Microsoft) ( * p < 0.05; * * p < 0.005; * * * p < 0.0005). For multiple comparisons, the significance levels were corrected via Bonferroni-correction.
Frontiers in Pharmacology | www.frontiersin.org showed a huge rebound when co-applied with Glycyrrhiza ( Figure 1C). All of the six inhibitory tinctures were also tested at a lower concentration (0.1 Vol.-%). In these experiments, only the tincture of Glycyrrhiza showed a significant inhibition (26.6 ± 4.4%) ( Figure 1A). All of the inhibitions were reversible after a 150-s washout.

Modulatory Effect of the Rikkunshito Ingredients and Tincture of Unsweetened Licorice
The tinctures contain many different chemical substances. To identify the active compounds, we assessed the modulatory effects of some known ingredients of these plants on 5-HT 3A receptors. We focused on the ingredients of Glycyrrhiza uralensis that showed the strongest antagonism within the rikkunshito constituents and other ingredients of further Glycyrrhiza species. Thus, we tested a tincture of unsweetened licorice because it is obtained from Glycyrrhiza glabra L. (Fabaceae). We used a concentration of 1 mM for all of the substances with the exception of glabridin (100 µM) and 1 Vol-% for the licorice tincture.

Competitive and Non-competitive Action of the Identified Antagonists
The tinctures of Atractylodis and unsweetened licorice, as well as glabridin and licochalcone A, exhibited an apparent competitive antagonism. The tincture of Zingiberis and Glycyrrhiza and (-)liquiritigenin showed increased inhibition at the higher agonist concentration (30 µM), indicating a non-competitive blocking mechanism (Table 1).
Frontiers in Pharmacology | www.frontiersin.org Uezono, 2014; Tominaga and Arakawa, 2015). Moreover, it is involved in the regulation of peristalsis and digestion (Tominaga et al., 2011) and therefore ameliorates symptoms of functional dyspepsia and irritable bowel syndrome (Oka et al., 2014). These effects and the antiemetic properties of rikkunshito may be explained by 5-HT 3 receptor antagonism. Therefore, we assessed the modulatory effect of the eight rikkunshito constituents as ethanol tinctures to identify the most effective constituents and to find new specific antagonists for the 5-HT 3A receptor.
We detected an antagonistic effect on 5-HT 3A receptors exerted by the tinctures of rikkunshito, Aurantii, Ginseng, Zingiberis, Atractylodis, and Glycyrrhiza. Surprisingly, Zingiberis, which was initially thought to be mainly responsible for the effects of rikkunshito, was the weakest identified antagonist. However, we identified Radix Glycyrrhiza uralensis as the strongest 5-HT 3A receptor antagonist among the tinctures (Figures 1A, 3A), presumably through the action of (-)liquiritigenin.
The flavonoid glycoside hesperidin from Aurantii is one of the most abundant flavonoids in rikkunshito. It shows a gastroprokinetic effect similar to that of the specific 5-HT 3 receptor antagonist ondansetron, suggesting that hesperidin is a 5-HT 3 receptor antagonist (Tominaga et al., 2011). However, in our experiments, only hesperetin, the aglycone of hesperidin, exhibited 5-HT 3A receptor antagonism. Tominaga et al. used an in vivo animal experimental paradigm in which hesperidin was applied orally. Therefore, it is possible that hesperidin is converted into the active substance by deglycosylation in vivo. The lack of 5-HT 3A receptor antagonism for hesperidin may be explained by steric problems caused by glycosylation with the disaccharide rutinose. However, other glycosylated flavonoids are known, e.g., rutin from Aurantii, which shows an antagonistic effect on 5-HT 3 receptors. However, also in this case, quercetin, the aglycone of rutin, is the more potent substance .
The tincture of Atractylodis, which is used in kampo medicine for the treatment of nausea and cachexia, showed a strong, apparently competitive antagonism. In our study, atractylenolide III, a weak antagonist (IC 50 = 1322 ± 145 µM), was the only ingredient that was investigated. However, investigations of other ingredients, such as atractylol, atractylon or biatractylolid (Shao et al., 2014), could lead to the identification of competitive 5-HT 3A receptor antagonists with higher potency.
Strong 5-HT 3A receptor antagonism was also observed for the tincture of Ginseng, which is used due to its antiemetic effect. Kim et al. showed that Ginseng extracts reduce cisplatin-induced nausea in ferrets (Kim et al., 2005). Steroid glycosides, called ginsenosides, are accountable for the observed 5-HT 3A receptor antagonism (Min et al., 2003;Lee et al., 2007), and their binding site in the pore region of 5-HT 3A receptors has been identified (Lee et al., 2007).
The vanilloids gingerol and shogaol (Abdel-Aziz et al., 2005, 2006Walstab et al., 2013) as well as the diterpene lactone galanolactone (Huang et al., 1991) are responsible for the antagonistic effect of the Zingiberis tincture. Abdel-Aziz et al. showed that gingerols and shogaols inhibit the contractions of isolated guinea pig and rat ilea induced by a specific 5-HT 3 receptor agonist (Abdel-Aziz et al., 2005, 2006, indicating 5-HT 3 receptor antagonism for the spasmolytic effects of Zingiberis and rikkunshito. Our study supports the proposal of these vanilloids as the active principles of Zingiberis due to the non-competitive antagonism of this tincture, which was previously reported for gingerol and shogaol (Walstab et al., 2013). In many clinical trials, Zingiberis was able to reduce nausea under various conditions, such as motion sickness, hyperemesis gravidarum, CINV and PONV (Ernst and Pittler, 2000;Haniadka et al., 2012;Ding et al., 2013).
Radix Glycyrrhiza is used for the treatment of gastric ulcer and inflammations of the gastrointestinal and respiratory system (Kim et al., 2008). Among the rikkunshito constituents, Glycyrrhiza uralensis is the antagonistic tincture with the highest efficacy and potency. Moreover, we tested a tincture of unsweetened licorice, which is sourced from Glycyrrhiza glabra and usually consumed. This tincture was also able to inhibit the 5-HT 3A receptor responses. We tested four Glycyrrhiza ingredients, namely the glycoside glycyrrhizin and the flavonoids glabridin, (-)-liquiritigenin and licochalcone A. Although (-)liquiritigenin and glycyrrhizin are ubiquitous within the plants of the genus Glycyrrhiza, glabridin and licochalcone A are restricted to specific species. Glabridin is a specific ingredient of Glycyrrhiza glabra, and licochalcone A is detectable only in the species G. inflata and G. eurycarpa (Xu et al., 1997;Rauchensteiner et al., 2005;Kondo et al., 2007). Glycyrrhizin, which is used for the treatment of epilepsy, chronic gastritis and obstipation (Hänsel et al., 1993), showed no 5-HT 3A receptor antagonism. The flavonoids glabridin, (-)-liquiritigenin and licochalcone A exhibited the strongest antagonism among our tested substances. (-)-Liquiritigenin appears to be the active component in Glycyrrhiza uralensis. Both showed non-competitive antagonism and lead to a huge rebound of the inhibited 5-HT 3A receptor responses. In addition to (-)-liquiritigenin, hesperetin exhibited similar kinetics at high concentrations (Figures 3E,F). Because both flavonoids are chemically related (Figure 4), possess similar IC 50 values and share a non-competitive antagonism, a common binding site at the receptor is hypothesized.
Kim et al. investigated the antiphlogistic effect of liquiritigenin and attributed this effect to the inhibition of NF-K B, an important transcription factor in the immune response (Kim et al., 2008). Therefore, liquiritigenin may contribute to the antiemetic and antiphlogistic effects of rikkunshito. In contrast, (-)-liquiritigenin cannot be accountable for the concentration-dependent 5-HT 3A receptor antagonism of the licorice tincture. Instead, the apparently competitive blocker glabridin, which is a potentiator of the closely related GABA A receptor (Jin et al., 2013), may be the main antagonist of the unsweetened licorice, which is sourced from G. glabra. The blocker licochalcone A differs from hesperetin and (-)-liquiritigenin by its chemical structure and its action as a partial antagonist (Figures 3D-F, 4). The combination of concentration-dependency and partial blocker properties is unusual. Although these properties allude to the action of licochalcone A as a partial agonist, this consideration can be ruled out due to the absence of the direct activation of 5-HT 3A receptors. The hypothesis of a concentration-dependent, allosteric antagonist whose maximal inhibition decreases with increasing agonist concentrations is more likely.
We showed that the Glycyrrhiza uralensis tincture exhibits the strongest inhibition of 5-HT 3A receptor responses compared with the rest of the rikkunshito constituents, possibly due to the action of the flavonoid (-)-liquiritigenin. Other Glycyrrhiza species share this flavonoid-based 5-HT 3A receptor antagonism, which is attributable to the antagonists glabridin and licochalcone A, a partial 5-HT 3A receptor antagonist. These results contribute to a better understanding of the action of rikkunshito at a pharmacological level and allow the establishment of flavonoids as a new potent class of plant ingredients with regard to 5-HT 3 receptor antagonism. Therefore, flavonoids appear to be at least equally active antagonists as gingerols and shogaols from Zingiberis and ginsenosides from Ginseng, which were thought to be responsible for the antiemetic properties of rikkunshito prior to this study. However, it should be mentioned, that the identified antagonists are inferior to already commonly used drugs like setrons due to their lower potency. Hence, a comparatively higher concentration of flavonoids must be reached to cause physiologically relevant effects. If those concentrations of substances with 5-HT 3 receptor antagonism can be reached by rikkunshito under naturally-occuring conditions is hard to define. However, rikkunshito contains flavonoids, ginsenosides, and vanilloids, three well-investigated classes of plant ingredients that inhibit 5-HT 3A receptors by binding to the receptor, presumably at independent binding sites. Hence, a synergistic drug interaction with additional or maybe mutual potentiating character is conceivable. In an exemplary experiment, we demonstrated that the combined action of distinct plant derived blockers can lead to an increased block of the 5-HT 3A receptor (Supplementary Figure 3). In addition to that, a contribution of structurally related, unemployed substances, e.g., liquiritin and its apiosides and glycosides to the 5-HT 3 receptor antagonism seems to be likely. In this study, we refuted the assumption that hesperidin, the main ingredient of rikkunshito, promotes gastric emptying via 5-HT 3 receptor antagonism due to the lack of antagonism obtained in our screening. Instead, hesperetin, the aglycone of hesperidin, acts as a 5-HT 3A receptor antagonist.
Nevertheless, a 5-HT 3 receptor mediated effect of hesperidin is assumed due to the occurrence of deglycosylation in vivo. This study contributes to a better understanding of the action of rikkunshito at a pharmacological level and emphasizes the importance of Glycyrrhiza and Aurantii for the antagonism of 5-HT 3A receptors.