Antidepressant-like effects of a chlorogenic acid- and cynarine-enriched fraction from Dittrichia viscosa root extract

Dittrichia viscosa is a perennial Mediterranean plant used in traditional medicine for “calming purposes”, pointing at a possible antidepressant activity of the plant. We conducted chromatographic and bioassay-guided fractionation of D. viscosa root extract to isolate a specific fraction (fraction “K”) with antidepressant-like characteristics in vivo and strong antioxidant properties in vitro. A single dose of “K” reduced immobility time in the forced swim test with a mouse model possessing a depressive-like phenotype. Neurochemical profiling for 5-hydroxytryptamine (5-HT) and its primary metabolite, 5-hydroxyindoleacetic acid (5-HIAA), in prefrontal cortex and hippocampus of “K”-treated mice showed reduction in 5-HIAA, indicative of either serotonin uptake transporter or monoamine oxidase-A inhibition, as well as slight increases in 5-HT content. These neurochemical alterations, as well as the behavioral changes observed, were comparable to the effects of paroxetine. “K” also protected PC12 cells in a H2O2 cytotoxicity assay, thus demonstrating antioxidant properties, yet paroxetine augmented oxidative damage and cell death. Identification of the main compounds in “K” by high-performance liquid chromatography-tandem mass spectrometry (HPLC–MS/MS) indicated that chlorogenic acid and cynarine comprised 87% of the total components. D. viscosa root extract appears to produce antidepressant and cytoprotective effects and may serve as an attractive alternative to standard therapies for depression.

www.nature.com/scientificreports/ Numerous plant natural products have successfully entered the clinic as complementary and alternative medicines for the treatment of psychiatric disorders, for example, crude and standardized extracts of Hypericum perforatum L. 8 , Valeriana officinalis L. 9 , Melissa officinalis L., and Verbena officinalis L. 10 . Notably, many plantoriginated drugs were discovered through their use in traditional medicine. Ethnobotanical research, followed by the use of advanced technological tools, molecular and cellular methodology, and valid animal models, serves as a platform for the development of novel active compounds 11,12 .
Dittrichia viscosa (L.) Greuter (Asteraceae), commonly known as False Yellowhead or Woody Fleabane, is a herbaceous perennial Mediterranean plant species 13 formerly belonging to the genus Inula 14 . According to ethnopharmacological surveys and reviews of medicinal herbs in Israel, D. viscosa is commonly used in Arabic traditional medicine [15][16][17][18] . The traditional uses of D. viscosa [Arabic name, tayun (sticky) davik] are vast including treatment of skin diseases, rheumatic pains, prevention of infections, reduction of high blood pressure, reduction of high sugar levels, infertility, and other uses 18,19 . Among different traditional applications of D. viscosa, it is popularly used for "calming purposes", muscle relaxation, analgesia, and fatigue treatment 15,17,18,20 . Fatigue is a residual symptom of depression 21,22 and many commercially-available antidepressants possess relaxing and analgesic properties 23,24 , suggesting a possible antidepressant-like potential in D. viscosa.
In this study, we aimed to determine whether an extract of D. viscosa root does indeed have the properties ascribed to it in the ethnobotanical literature, namely antidepressant effects. Further, we aimed to perform a chromatographic separation and bioassay-guided fractionation of the extract, selecting for fractions with increased pharmacological properties, as an initial effort to identify plant compounds of pharmaceutical potential matching the reported effects. We identified one fraction (named fraction "K") which exhibited cytoprotective qualities as well as antidepressant-like effects in a mouse model with depressive-like phenotype developed in our laboratory 25,26 . Fraction "K" was evaluated by mass spectral analysis for bioactive compounds and two prominent chemicals were identified: chlorogenic acid and cynarine. Here, we present evidence for the pharmacological potential of D. viscosa root extract as an antidepressant natural product which contains at least two compounds of known pharmacology. Ongoing work in our lab is focused on identifying other molecules of therapeutic potential in fraction "K" of D. viscosa root.

Results
Crude extract did not exert a toxic effect in PC12 cells but failed to protect against oxidative stress. PC12 cells incubated with a broad range (250-5000 μg/mL) of crude root extract concentrations did not present with cytotoxic or proliferative effects (Dunnett's test: p > 0.05; Fig. 1 HPLC-guided fractionation resulted in 14 fractions exhibiting diverse properties in a H 2 O 2 -induced cytotoxicity model. As shown (Fig. 1B) root extract showed no protective effect against H 2 O 2 cytotoxicity, however we assumed (taking in mind the behavioral effect in FST, Fig. 4) that extract fractionation may result in identification of active compounds with more potent beneficial pharmacological properties. HPLC-guided separation of crude extract (Suppl. Fig. 1A) resulted in 14 fractions. Each fraction contained between 2 to 8 peaks (Suppl. Fig. 1B, representative fraction), with each peak corresponding to at least one compound. www.nature.com/scientificreports/ To identify fractions with beneficial biological activity, we assumed that cytoprotective effects would accompany the FST behavioral effects observed with the crude root extract. Combinations of separated fractions were tested in a concentration range similar to those of the crude extract (0-1000 μg/mL) in the XTT cell viability oxidative stress assay (400 μM H 2 O 2 ) with PC12 cells (Fig. 2). Our initial assumptions appeared to be correct. Combination of fractions "B + C" and "L + M" were not protective against oxidative stress, with fraction "B + C" conferring no cytoprotective effect and fraction "L + M" concentration-dependently reducing cell viability (Dunnett's test: p = 0.0001, all concentrations). Fraction "J + K" concentration-dependently increased cell viability by 31% over basal H 2 O 2 -induced oxidative stress (Dunnett's test: p = 0.0001, 500-1000 μg/mL). Thus, fractions "J" and "K" were further separated and evaluated separately against H 2 O 2 -evoked cell death.     Fig. 4). These results were similar to the effect of PXT (10 mg/kg), a widely prescribed selective serotonin reuptake inhibitor (SSRI) that was used as a positive control (Fig. 4). Neither the herbal extract nor PXT produced tachyphylaxis, indicated by similar behavioral responses regardless of single or 14-day daily treatments.

Acute and sub-chronic D. viscosa root extract administration decreased mouse immobility time in the forced swim test (FST
Fraction "K" produced antidepressant-like effects similar to paroxetine. Administration of single doses of fraction "K" to Sub mice subjected to the FST indicated similar reductions of depressive-like behavior comparable to the comparison dose of PXT (10 mg/kg) (One-way ANOVA, F [4,45] = 135.5, p < 0.0001; Fig. 5). Doses of 5 mg/kg or 25 mg/kg fraction "K" reduced mouse immobility times by 55.5% and 71.8%, respectively, which were comparable with the 60.7% reduction in immobility times using PXT. Fraction "K" doses as low as 1.0 mg/kg did not produce any measurable antidepressant-like effect in the FST. In experiments with these    Phytochemical analysis of Fraction "K". Two major components in fraction "K", representing approximately 87% of the total content, were putatively identified based on positive and negative mass spectra as 49% chlorogenic acid (CGA) and 38% cynarine (Fig. 7A-C).

Discussion
Among the fractions from D. viscosa crude extract, we highlighted one fraction which we entitled "K", that exhibited a strong cytoprotective effect by dose-dependently inhibiting H 2 O 2 -induced PC12 cell death determined by mitochondrial functionality. Since H 2 O 2 is a potent apoptosis inducer, we may assume that fraction "K"-mediated cytoprotection may have been achieved by a direct anti-apoptotic effect or indirectly through antioxidative mechanisms similar to the effects shown by other plant extracts 29,30 . Although known antidepressants such as amitriptyline and fluoxetine show protective effects in a PC12 cell model 31 , we found that PXT in a non-cytotoxic concentration range for the drug alone produced increased H 2 O 2 -induced PC12 cell death. This is in agreement with other studies showing the cytotoxic activity of PXT against tumor cells of either murine or human origin 32,33 .
Despite the lack of a clear correlation between the cytoprotective effect of herbal/pharmacological agents in vitro and antidepressant activity in behavioral assays in vivo, several studies have reported such phenomena 34,35 . For example, antidepressant properties of Hypericum perforatum have been shown in vivo 36,37 as well as neuroprotective effects observed in a number of in vitro studies 38,39 . We observed an antidepressant-like effect of fraction "K" in Sub mice possessing a strong depressive-like phenotype, which have been previously shown to www.nature.com/scientificreports/ be responsive to antidepressants, as well as plant-derived agents 26,40 In our work, FST was applied as an acute environmental trigger for assessment of stress coping strategy in Sub mice. The decreased immobility in the FST of Sub mice following fraction "K" administration was linked to a reduction in 5-HT turnover in PFC and HPC similar to PXT. PFC and HPC are considered to be hot-spot brain regions of the circuits involved in depression 41 and are critically involved in the mechanism of SSRI action 42 . Both PFC and HPC are densely innervated by serotonergic fibers and the majority of 5-HT receptor subtypes are expressed in these regions 43,44 .
Thus, we suggest that fraction "K"-induced changes in the serotonergic system are associated with serotonin transporter (SERT) inhibition, similar to the established mechanism of PXT. PXT functions as a potent SERT inhibitor and reduces 5-HT reuptake, leading to a redistribution of 5-HT in favor of the synaptic cleft and a consequent decrease in presynaptic cytosolic concentrations. These events may trigger compensatory reduction of 5-HT degradation due to the disinhibition of the activity of tryptophan hydroxylase, reflected by decreased 5-HIAA content and increased 5-HT production 45 as demonstrated here by decreased 5-HIAA/5-HT ratio.
We linked the observed biological activity of fraction "K" to CGA and cynarine, which were identified as major components of the fraction. CGA is a polyphenolic secondary metabolite produced by many plant species, including D. viscosa 46 . As a potent antioxidant, CGA shows high capability in modulating oxidative stress in celland animal-based models 47,48 . CGA is known to penetrate the blood-brain barrier 49 exhibiting antidepressant and anxiolytic properties in rodent models 50,51 . CGA affects plasma β-endorphin 52 , serum 5-HT, and dopamine 51 as well as colonic 5-HT levels 53 . It has also been shown in vitro that CGA stimulates axon and dendrite growth and promotes 5-HT release through augmenting synapsin I expression in rat raphe neurons 49 .
Cynarine has a broad repertoire of biological activities. It is a strong antioxidant compound 54,55 that also produces effects on smooth muscle, which is thought to contribute to reported antihypertensive activity 56,57 . The antioxidant effects of cyanarine are most potently demonstrated in H 2 O 2 cytotoxicity assays and is most likely the reason for cytoprotective effects of fraction "K" observed here. Cynarine also has the potential for neurological effects as the chemical has demonstrated anticholinesterase activity with a K i in the low nM range 55 , however any potential psychotropic properties have not been studied.
In this work, we demonstrated, for the first time, the link between changes in the brain serotonergic system and behavioral phenotype following treatment with a D. viscosa plant root extract fraction (fraction "K"). Although fraction "K" is CGA and cynarine-enriched and despite the predominance of these two compounds in the isolated fraction, fraction "K" is still a crude preparation and we have not yet identified or have assessed the influences of other chemical agents present. Behavioral effects from fraction "K" treatment resemble those reported for CGA, however we cannot exclude the involvement of other, undetermined agents in the preparation. Further work is mandatory to understand the contribution of each compound in the observed biological activity, the contributions of as-yet unidentified compounds in the remaining 13% of the active fraction "K", and their possible interactions. We believe the biggest challenge in future efforts will be the assessment of multiple chemical combinations from this bioactive fraction and identification of synergism effects. It is our intent to continue the study of D. viscosa root to isolate and characterize compounds with antidepressant action and other positive biological properties. Extract was separated into 14 fractions, each combining 5 min of flow, and assigned a letter (A-N). A UV detector visualized the peaks at wavelengths of 280 and 257 nm. Each fraction was lyophilized at − 80 °C for approximately 72 h and reconstituted to an aqueous solution. All reconstituted fractions were standardized to fit the amount of the major parallel peaks in the crude extract by diluting in n times (for example, reconstituted fraction "K" was diluted seven times to achieve its initial concentration in crude extract).

Conclusions
Data were collected and analyzed using ChromNAV 2.0 HPLC Software. Animals. Submissive (Sub) mice used in this study were selectively bred for 36 generations from an outbred Sabra strain (Envigo laboratories, Israel) using social interaction dominant-submissive relationship paradigm 58 . The behavioral profile of Sub mice possesses strong elements of depressive-like behavior and impaired stress coping strategies confirmed by different experimental approaches [25][26][27][28] . Animals were housed in groups of five in a temperature (21 ± 2 °C), humidity (55 ± 5%), and light controlled room (lights on from 7 AM to 7 PM). Standard laboratory chow and water were available ad libitum. All experiments were conducted with male mice (age, 10 weeks), during the light phase of the day-night cycle between 9 AM and 3 PM. Behavioral and neurochemical studies were conducted on the separate cohorts of Sub mice. Housing, care, and experimental procedures involving animal use conformed to NIH/USDA and ARRIVE 2.0 guidelines. The experiments were approved and supervised by the Institutional Animal Care and Use Committee of Ariel University and the Israel Ministry of Health (Protocol IL-74-09-15).

Open field test (OF). Spontaneous locomotor (horizontal) activity was assessed in the Open Field (OF)
test using EthoVision (Noldus, Holland) as previously described 27,59 . Briefly, the OF apparatus consisted of a square black plastic chamber (40 cm × 40 cm). Each mouse was placed individually for 5 min in the center of the chamber and distance travelled was recorded. Between subjects, the apparatus was thoroughly washed with 70% ethanol and dried with tissue paper.

Forced swim test (FST).
FST was performed to access stress coping strategies following psychotropic interventions 60 . Mice (6 per group) were individually tested in a glass cylinder (30 cm in height, 10 cm in diameter) filled with water up to 25 cm height (25 ± 2 °C). Mice were tested for 6 min, the last 4 min were analyzed, where the total time spent immobile was recorded manually. Mice were considered "immobile" if they displayed no activity except for what is required to keep the head above water.
Neurochemical profiling. For sample preparation, mice were anesthetized in a CO 2 chamber and decapitated immediately afterwards. Prefrontal cortex (PFC) and hippocampus (HPC) were dissected, frozen in liquid nitrogen and kept at − 80 °C until use. Serotonin (5-hydroxytryptamine, 5-HT) and its metabolite 5-hydroxyindoleacetic acid (5-HIAA) were assayed in brain tissue samples using HPLC with electrochemical detection (EICOM Co., Kyoto, Japan) as previously described 45 .

HPLC-MS/MS analysis. Chemical analysis of fractions was performed using a Waters 2695 Separation
Module with a Photodiode Array Detector together with a Quattro Micro Mass Spectrometer. Chromatographic separation was achieved using a Luna (Phenomenex) reverse phase C18 (5 µm; 4.6 × 250 mm; 100 Å) column and the binary gradient described above (see Extract fractionation). MS acquisition was conducted in both ESI positive and negative ionization mode under the following conditions: capillary voltage-3.5 kV, cone voltage-45 V, extractor voltage − 3 V, RF lens − 0.2 V, source temperature − 120 °C, desolvation temperature − 350 °C, nitrogen flow rate of 700 L/h for desolvation and 50 L/h cone gas. Tentative identification of compounds was achieved based on comparisons with purchased standards (Cayman Chemical Company) as wells as spectra from an MS library (NIST 2017).
Statistical analysis. Data are expressed as means (± SD). All multiple comparisons were Bonferroni-corrected against α = 0.05 and all analyses were two-tailed. Multiple comparison analyses were performed by oneway ANOVA followed by a Tukey's HSD or Dunnett's test as appropriate or were compared by two-way ANOVA followed by a Sidak or Tukey's HSD test based on the structure of the dataset. All analyses were performed in GraphPad Prism 7.0. Results of all analyses are presented in detail in Supplementary Table S1 www.nature.com/scientificreports/