The Association of Tanacetum parthenium and Salix alba Extracts Reduces Cortex Serotonin Turnover, in an Ex Vivo Experimental Model of Migraine

: The mixture of water extracts from Tanacetum parthenium and Salix alba was studied in an ex vivo assessment of neurotoxicity constituted by isolated mouse cortex specimens challenged with K + 60 mM Krebs–Ringer buffer (neurotoxicity stimulus). The effects of the mixture on lactate dehydrogenase (LDH), nitrite and serotonin levels were investigated. The phytochemical profile of the mixture was also evaluated. A docking approach was conducted to predict, albeit partially, the putative mechanism underlying the observed effects. The extracts displayed a good profile of polyphenolic compounds (22 chromatographic peaks detected), with caftaric acid and epicatechin being the prominent phenols. In isolated cortex, the association of T. parthenium and S. alba extracts was effective in reducing the K + 60 mM-induced levels of LDH and nitrites, whereas the neurotoxicity stimulus-induced serotonin depletion was prevented by the treatment. Regarding the inhibition of serotonin catabolism, epicatechin (44.65 µg/mg) and caftaric acid (10.51 µg/mg) were putatively the main compounds involved in the inhibition of monoamineoxidase-A, which is known to play a master role in serotonin turnover. Collectively, the results of the present study point to the efficacy of the present extract mixture as an innovative pharmacological tool to prevent the onset of migraine.


Introduction
Migraine is one of the most widespread neurovascular disorders, with an incidence which increases during aging, especially in women [1]. Serotonin (5-HT) has long been involved in migraine etiopathogenesis, with clinical data pointing to tight relationships between neurotransmitter activity and migraine attacks [2]. In this context, the activation of 5-HT1B and 5-HT1D receptors could lead to the control of migraine attacks [3]; by contrast, the blockade of the 5-HT2A and 5-HT2C receptors was found effective in the prevention therapy [4]. The involvement of the trigemino-vascular system and neuroinflammation has also been demonstrated in the pathophysiology of migraine [5,6], although the site of origin of migraine is still unsettled. However, cortical spreading depression (CSD), a supraphysiological neurotoxicity stimulus, has been related to the migraine onset. In particular, CSD would represent a possible link between 5-HT depletion and trigeminal nociception [7]. Currently, the pharmacotherapy options for migraine include aborting attack treatments, namely analgesics, nonsteroidal anti-inflammatory drugs, and triptans [8], and preventive medications, including anti-epileptics, anti-depressants and anti-hypertensives [9]. More recently, monoclonal antibody therapy has been demonstrated to be effective and tolerated by migraineurs [10]. Nevertheless, despite there being different and efficacious pharmacological treatments for the management of migraine, the frequent occurrence of side effects is an impulse to search for innovative options with a minor impact, in terms of side effects [11,12]. Specifically, a new perspective is arising from the antimigraine use of nutraceuticals and botanicals [12]. Among botanicals, a promising approach is represented by the use of extracts from T. parthenium and Salix alba, which are known to possess anti-inflammatory effects [13,14]. T. parthenium has been used with success for preventing migraine attacks in both adult and child patients [15,16], whereas S. alba improved 5-HT level in rat hippocampus after in vivo administration [17]. Both were found effective, as single treatments, to prevent 5-HT turnover, in a preclinical experimental model of CSD [18,19]. The scientific literature has suggested their pharmacological association for migraine prophylaxis [20]. The co-administration of magnesium and coenzyme Q10 would further improve the anti-migraine efficacy of these botanicals [15].
In this context, the aim of the present study was to investigate the potential use of the extract mixture of T. parthenium and S. alba water extracts in an ex vivo model of CSD constituted by isolated mouse cortex specimens exposed to an supraphysiological depolarizing stimulus (K + 60 mM). The abovementioned ingredients have been reconstituted in water, and their effects on 5-HT turnover and markers of nitrosative stress and tissue damage, namely nitrites and lactate dehydrogenase, were assayed as well [17,18,21]. The tolerability of the formula was also evaluated through validated toxicological models, namely the Artemia salina and Daphnia magna toxicity tests [22]. Finally, phytochemical and in silico analyses were also conducted in order to shed light the mechanism of action underlying the observed bio-pharmacological effects.

Phytochemical Analysis
The extract mixture was formerly investigated via qualitative MS analysis (scan mode 100-1200 m/z), for confirming the presence of parthenolide and salicin, which are reference compounds in T. parthenium and S. alba, respectively. The m/z 249.3 and 271.1 were monitored in positive scan mode to identify the parthenolide, and the MS spectrum confirmed the presence of parthenolide, in the mixture (Figure 1). The m/z ratio 285.1 was scanned in negative scan mode for the identification of salicin, whose MS spectrum is shown ( Figure 2). The abovementioned m/z ratios were selected according to the MS spectra reported in MassBankEurope databases. Regarding parthenolide, the selection of the m/z 249.3 agrees with the study by Zhao et al. (2016) [23]. In the case of salicin, the m/z ratio was also considered on the basis of the study by Kammerer and colleagues (2005) [24]. These authors observed that ionization-induced salicin derivatives tended partially to decompose to m/z 123, forming its aglycone form, namely saligenin.
The phytochemical composition was also studied through HPLC-DAD-MS, with the aim of elucidating the quali-quantitative profile in phenolic compounds (Table 1), deeply involved in the anti-inflammatory and anti-oxidant effects of herbal extracts, especially from plant materials extracted with polar solvents [25]. Specifically, the quantitative chromatographic analysis, whose experimental conditions are depicted in Table 1, indicated epicatechin (277.89 µg/mL) and caftaric acid (123.93 µg/mL) as the main phenolics ( Figure  3). These compounds were also prominent in other plant materials with intriguing antioxidant properties, namely grape and pomace [26][27][28]. However, if catechins and caftaric acid were identified in plants belonging to the Tanaceutm taxa [29][30][31], Salix species were described to synthesize catechins and other polyphenolic compounds, including apigenin, luteolin, naringenin and quercetin [32]. A qualitative 1 H-NMR analysis was also conducted in the water suppression signal mode [33], and confirmed the presence of phenolics, but also sugars ( Figure 4). The presence of sugars is indicated by the chemical shift values, in the range 3-4 ppm. Considering the elevated level (15% w/w) of salicin in the S. alba extract, we cannot exclude that the signals in this chemical shift range could be related, albeit in part, to this glucoside, as also suggested by the 1 H-NMR spectrum reported in the PubChem database. However, the content of phenolics is witnessed by the signals at chemical shift values in the range 7-8 [34][35][36]. Nevertheless, future studies with independent analytical methods, such as 1 H-NMR over 400 MHz and LC-HR-MS/MS, will improve our knowledge about the phenolic composition of the extract mixture [37]. The phytochemical profile is also consistent with intrinsic antiradical and neuroprotective properties of T. parthenium and S. alba extracts [18,[38][39][40]. Moreover, considering the ability of catechins to cross the blood brain barrier [41], it is rational to evaluate the efficacy of herbal extracts based on these plants in reversing the burden of oxidative stress and neuroinflammation which occurs in neurological diseases, such as migraine. Therefore, after the evaluation of the biocompatibility limits with ecotoxicological models, the extract mixture was administered to isolated mouse cortex specimens exposed to a neurotoxicity stimulus (K + 60 mM), and selected biomarkers, including serotonin, were assayed for evaluating neuroprotective and antimigraine effects.

Toxicological and Pharmacological Studies
The toxicological properties of the mixture T. parthenium/S. alba (1:2, w/w) were initially assayed through the Artemia salina (brine shrimp) lethality assay, to define the biocompatibility of the product in the concentration range 0.1-20 mg/mL [22]. This test is considered, albeit partially, as a reliable alternative model for predicting cytotoxicity, in eukaryotic cells [42]. Specifically, the mixture yielded a LC50 > 2 mg/mL ( Figure 5), which mirrored the toxicity of T. parthenium and S. alba, in the same experimental model [17]. Considering the results of the brine shrimp assay, the mixture was tested at the concentration of 2 mg/mL in the Daphnia magna toxicity test, for evaluating eventual cardiotoxicity effects, in respect to ethanol 10% solution, as representing the reference cardiotoxicity stimulus. At the concentration of 2 mg/mL, the mixture was able to prevent the ethanolinduced reduction of heart rate, in Daphnia magna ( Figure 6); thus scaling back its toxicity impact in eukaryotic organisms. This is also consistent with the abovementioned phytochemical profile; the Toxicity Estimation Software Tool (T.E.S.T.) predicted LC50 values > 3 mg/mL for salicin, catechins and caftaric acid, in this experimental model [22]. Nevertheless, a concentration of the mixture at least ten-fold lower (200 µg/mL) was selected for the subsequent evaluation on cortex tissue. This choice is motivated by the high susceptibility of neurons to oxidative and nitrosative stress [42], and also by the intrinsic capability of phenolics to induce mild oxidative stress [43].
Specifically, the mixture (200 µg/mL) was tested in isolated mouse cortex specimens exposed to different concentrations (3, 15, 60 mM) of K + , in the Krebs-Ringer buffer. If K + 3 and 15 mM are classical basal and depolarizing conditions in neuronal endings, respectively [44], K + 60 mM represents a supraphysiological depolarizing stimulus, also called CSD and able to induce a massive calcium-independent release, with concomitant increase of neurotransmitter degradation [7]. The CSD is common to different neuroinflammatory conditions, including migraine, and it is characterized by increased oxidative stress and serotonin degradation in the cortex [21]. In the present study, the mixture was found effective in inducing protective effects in mouse cortex. Indeed, the treatment determined a significant reduction of lactate dehydrogenase (LDH) level (Figure 7), a wellknown marker of cytotoxicity [45], at all pharmacological conditions; thus suggesting not only the capability to prevent the tissue damage induced by K + 60 mM, but also an intrinsic biocompatibility when cortex was exposed to physiological K + 3 and 15 mM. This biocompatibility is consistent with the abovementioned ecotoxicological predictions, and also mirrored by the inhibition of nitrite level (Figure 8), a reliable index of nitric oxide production and nitrosative stress [46], at all considered depolarizing conditions. This indicated not only the capability to contrast the burden of nitrosative stress induced by K + 60 mM, but also protective effects occurring at physiological conditions. Finally, the mixture was also effective in blunting the K + 60 mM-induced turnover of 5-HT, measured as 5HIIA/5-HT ratio (Figure 9), which is a reliable marker of MAO-A activity [47]. In this case, the mixture was effective at all depolarizing conditions to which the cortex was exposed. These findings further support the use of food supplements based on T. parthenium and S. alba for contrasting and/or preventing migraine attacks. The blunting effects on nitrite level and serotonin degradation could be due to more than one reason. On the one hand, we have to consider the intrinsic antiradical properties of both T. parthenium and S. alba extracts [17], present in this mixture. On the other hand, we have to highlight how the antiradical properties of herbal extracts are strictly related to enzyme inhibition effects, especially in the case of phenolic compounds [48]. Additionally, catechin has been suggested to act as a MAO-A inhibitor [49], whereas there is still a gap in the scientific literature about the effect of caftaric acid on this enzyme. Therefore, a docking approach was conducted with the aim of exploring putative interactions of caftaric acid with MAO-A, which could be predictive of enzyme inhibition effects [50]. The results of the virtual screening indicated the capability of caftaric acid to interact with the pocket/binding site of the enzyme through non-covalent interactions, including hydrogen, electrostatic and hydrophobic bonds ( Figure 10). Specifically, it predicted the ability of caftaric acid to form hydrogen bonds with Ala-111, Phe-112, Thr-204, Pro-114, His-488, and to interact via electrostatic bond with Glu-492. A hydrophobic interaction was predicted with Phe-112, as well. Furthermore, the present and literature data showed micromolar affinities of caftaric acid and epicatechin, respectively, towards the MAO-A enzyme; thus suggesting the occurrence of enzyme inhibition effects by the mixture and underlying the observed reduction of 5-HT turnover. Considering the 5-HT depletion occurring in migraine, the inhibition of 5-HT turnover induced by the mixture could be of relevance for the prevention of attacks, in migraineurs.

Qualitative Mass Spectrometer (MS) and High Performance Liquid Chromatography (HPLC) Analysis of Phenolic Compounds
The extract mixture was analyzed qualitatively using an expression compact mass spectrometer (CMS) (Advion, Ithaca, NY, USA) in negative and positive ion mode (m/z scan mode: 100-1200) [52]. Specifically, MS signal identification was realized through comparison with standard solutions and MS spectra present in the recognized Mass Bank Europe database. The mixture was also analyzed for phenol quantitative determination using a reversed-phase HPLC-DAD-MS in gradient elution mode, in agreement with previous studies [18,53]. The details of the analysis are reported in supplementary material (Tables S1 and S2).

Protonic Magnetic Resonance ( 1 H-NMR) Analysis
The extract mixture was analyzed via protonic magnetic resonance ( 1 H-NMR). The Variant 300 MHz spectrometer in standard proton pulse sequence (s2pul) with pre-saturation was used for the analysis. 60 µL of D2O were added as internal lock to 600 µL of the sample (300 mg/mL). Pre-saturation was carried out with relaxation delay 1 s, pulse 45.0 degrees, acquisition time 1.706 s, width 4803.1 Hz and 128 number of scans; partial water suppression signals occurred around 4.80 ppm.

Artemia Salina and Daphnia Magna Toxicity Tests
The biocompatibility and cardiotoxicity of the extract mixture were predicted through the Artemia salina and Daphnia magna toxicity assays. For the A. salina assay, the mixture was tested in the concentration range 0.1-20 mg/mL. Data were expressed as LC50 values. In Daphnia magna, the cardiotoxicity was induced by ethanol 10% solution. The details about these tests are fully reported in our previous studies [22,54].

Ex Vivo CSD Paradigm
Male adult C57BL6 mice (3-month-old, weight 20-25 g, N = 24) were sacrificed by CO2 inhalation. The experimentation was approved by Local Ethical Committee (University "G. d'Annunzio" of Chieti-Pescara) and Italian Health Ministry (Italian Health Ministry Authorization Number: F4738.N.5QP). After the sacrifice, cortex specimens were immediately collected and maintained in a thermostatic shaking bath at 37 °C for 1 h (incubation period), in Krebs-Ringer buffer at different K + concentrations (3, 15, 60 mM).
Specifically, the different K + concentrations corresponded to basal, physiological depolarizing and neurotoxicity stimulus, respectively [55]. During the experimental procedure, isolated cortex was treated with the extract mixture (200 µg/mL).

High Performance Liquid Chromatography (HPLC) Analysis of Serotonin and 5-Hydroxyindolacetic Acid
Cortex serotonin (5-HT) and 5-hydroxyindolacetic acid (5HIIA) were extracted in HClO4 50 mM and their levels in the tissue homogenate were analyzed through an HPLC-EC apparatus, as previously reported [56]. Details are reported in Supplementary Materials.

In Silico Study
Regarding the docking analysis, the crystal structure of the protein was downloaded from the Protein Data Bank. The PDB code was: 2ZX5 [monoamine oxydase-A: MAO-A]. Putative interactions with caftaric acid were analyzed as previously reported [57]. Details are reported in supplementary materials.

Statistical Analysis
Statistical analysis was performed by GraphPad Prism™ (Version 5.00) software (GraphPad Software, Inc., San Diego, CA, USA). The statistical significance (p < 0.05) was evaluated through analysis of variance (ANOVA) followed by Newman-Keuls comparison multiple test. Regarding the ex vivo studies, the statistical analysis to calculate the number of animals used in each experiment (N = 4 per condition) was performed with the software G*Power (v3.1.9.4). Study potency (1-β) and significance level (α) were 0.8 and 0.05, respectively.

Conclusions
The results of the present study pointed to the efficacy of the mixture T. parthenium/S. alba, in an experimental model of migraine. In this regard, the mixture showed a good tolerability in cortex tissue, as evidenced by the reduction of LDH and nitrites, biomarkers of cytotoxicity and nitrosative stress, respectively. These effects could be related, albeit partially, to the prominent phytocompounds, namely salicin, caftaric acid and epicatechin. The phytochemical profile is also consistent with the inhibition of serotonin degradation, following cortex treatment, and the efficacy was demonstrated both in basal and neurotoxicity conditions, thus suggesting its use as both protection and prevention agent, in migraine. Based on the phytochemical composition, the inhibition of serotonin degradation could be related to the putative MAO-A inhibition by epicatechin and caftaric acid. Since MAO-A is deeply involved in 5-HT catabolism, and considering the 5-HT depletion occurring in migraine attacks, the inhibition of MAO-A activity could be of relevance in the management/prevention of migraine attacks. Moreover, our results add to a previous perspective study suggesting the association of T. parthenium and S. alba as an innovative pharmacological tool to prevent the onset of migraine, with an acceptable toxicological profile [19]. However, further studies conducted with alternative experimental methods are needed to confirm efficacy, especially through in vivo models, and to improve our knowledge about the mechanism of action.

Acknowledgments:
The present study is part of the third mission activities of the Botanic Garden "Giardino dei Semplici" planned for the 20th anniversary of its establishment.

Conflicts of Interest:
The study was funded by Cristalfarma S.r.l.
Sample Availability: Samples of the compounds are not available from the authors.