Antidepressant-Like Effects of 3-(3,4-dihydroxyphenyl)lactic Acid Isolated from Lavender (Lavandula angustifolia) Flowers in Mice

Abstract

In the present study, we clarified the antidepressant-like constituents in an aqueous extract of lavender (Lavandula angustifolia) flowers. The antidepressant-like effects of the lavender extract (LE) and four fractions from the extract (Fr. 1 – 4) were evaluated using the forced swimming test (FST) in mice. Oral administration of the LE or Frs. 1, 3, and 4 significantly reduced immobility in the FST, whereas Fr. 2 had no significant effects. High-performance liquid chromatography using an authentic standard clearly showed that Fr. 3 contained a large amount of the known antidepressant-like substance rosmarinic acid (RA). Moreover, 3-(3,4-dihydroxyphenyl)lactic acid (DLA) was identified in Fr.1 as a new constituent for L. angustifoli. Oral administration of DLA (150 and 300 mg/kg) and of RA (37.5 mg/kg) significantly reduced immobility time in the FST. These results suggest that RA and DLA may participate in the antidepressant-like effects of the LE in mice.

Introduction

Lavender (Lavandula angustifolia) is an aromatic evergreen, perennial, woody shrub belonging to the Lamiaceae family, native to the Mediterranean region (Charles, 2013). The essential oil obtained from the fresh flowering tops of this plant has long been used in the production of perfume and is also used in aromatherapy as a relaxant (Lis-Balchin, 2006). In addition, lavender is used as a tea infusion to treat restlessness, insomnia, and nervous disorders of the stomach and intestines (Blumenthal, 1998). However, despite its long history in folk medicine, the use of lavender as a functional food ingredient is limited by its strong floral odor. In our previous studies, a non-volatile fraction of an aqueous lavender extract (LE), which was free from the main aromatic constituent linalool, showed antidepressant-like (Kageyama et al., 2012; Ueno et al., 2012) and anti-ulcer (Ueno et al., 2012) effects in rodents after oral administration. In addition, rosmarinic acid (RA), a known antidepressant-like substance, appeared to be an active constituent of the LE (Kageyama et al., 2012). The present study was designed to further clarify the active constituents responsible for the antidepressant-like effects of the LE.

Material and Methods

Animals    Six-week-old male ICR mice (29 – 31 g) were obtained from Japan SLC (Hamamatsu, Japan). The mice were group-housed (4 mice/cage; cage size: 210 × 315 × 130 mm height) in a room maintained at a constant temperature (22 – 24°C) and humidity (38 – 51%) with a 12-h light-dark cycle (light: 7:00 – 19:00). They were fed a standard diet, CRF-1 (Oriental Yeast, Tokyo, Japan), and tap water ad libitum, and allowed to acclimatize for 7 days. On the last day of the acclimatization period, the mice were divided into groups comprising eight mice each. Behavioral tests were performed between 13:00 and 17:00 h on the next day. All animal experiments were approved by the Institutional Animal Care and Use Committee of the Bioscience Center of KAC Co., Ltd (Ritto, Japan), and performed in accordance with “Standards Relating to the Care and Management of Laboratory Animals and Relief of Pain” (Notice No. 88, the Ministry of Environment of Japan, April 28, 2006).

Chemicals    Rosmarinic acid (RA) and caffeic acid (CAA) were purchased from Sigma-Aldrich Japan (Tokyo, Japan). Imipramine hydrochloride was purchased from Wako Pure Chemical Industries (Osaka, Japan). (+)-Catechin hydrate was purchased from Nacalai Tesque (Kyoto, Japan), and its water content was determined by the Karl-Fischer method. All other reagents were of the highest grade commercially available.

Preparation of LE    An aqueous extract from lavender (Lavandula angustifolia) flowers (LE) was prepared as described previously (Kageyama et al., 2012). Briefly, 125 g of dried lavender flowers were extracted with 2.5 L of distilled water at 90°C for 1 h. The extract was separated from the residue, evaporated in vacuo, and lyophilized to give 36.4 g of LE.

Preparative high-performance liquid chromatography (prep-HPLC)    The LE was fractionated using a prep-HPLC system (Shimadzu, Kyoto, Japan) that consisted of an SIL-10AP auto-sampler, two LC-6AD solvent delivery units, a DGU-12A degasser unit, an SCL-10A VP system controller, an SPD-M20A diode array detector and an FRC-10A fraction collector. The LE (5 g) was dissolved in 100 mL of water and filtered through a 0.45-µm polytetrafluoroethylene (PTFE) membrane filter (DISMIC-25HP; Toyo Roshi, Tokyo, Japan). The sample solution was repeatedly injected into a Chromatorex ODS SMB 100 – 10 column (250 mm × 20 mm i.d.; particle size, 10 µm; pore size, 10 nm; Fuji Silysia Chemical, Kasugai, Japan). The operating conditions were as follows: mobile phase, a linear gradient from 13% to 30% acetonitrile in 0.1% formic acid in 35 min, then eluted with 100% acetonitrile for 10 min; flow rate: 12 mL/min; and injection volume, 5 mL. Fractions were collected at retention times for 0.3 – 10.8 min (Fr. 1), 10.8 – 33.3 min (Fr. 2), 33.3 – 36.5 min (Fr. 3) and 36.5 – 45.0 min (Fr. 4). After 25 injections, the fractions were evaporated in vacuo and lyophilized to give dried fractions in the yields shown in Table 1.

Table 1. Total phenolic and rosmarinic acid (RA) contents in the fractions obtained from the lavender extract.

Fr. no.	Yield (mg)	Total phenolsa (mg)	RA (mg)
1	3479	133	n.d.b
2	1446	328	n.d.b
3	137	82	66
4	339	71	tr.c

a  Values determined by the Folin-Denis method and expressed as (+)-catechin equivalents.

b  Not detected.

c  Trace.

Isolation and identification of 3-(3,4-dihydroxyphenyl)lactic acid (DLA)    The LE prepared from 125 g of dried lavender flowers as described above was suspended in 1.75 L of 0.1% formic acid, chromatographed on a porous resin column (Sepabeads SP-70; Mitsubishi Chemical, Tokyo, Japan), and eluted with 1.75 L of 0.1% formic acid twice and then 1.75 L of 10% ethanol in 0.1 % formic acid three times. The second 10% ethanol fraction was evaporated in vacuo and lyophilized to give 961 mg of the crude product. The crude product was purified by prep-HPLC under similar conditions to that described above to yield 297 mg of DLA: ¹H MMR (CD₃OD, 400 MHz): δ 6.71 (1H, d, J = 2 Hz, C_2′-H), 6.67 (1H, d, J = 8 Hz, C_5′-H), 6.57 (1H, dd, J = 2, 8 Hz, C_6′-H), 4.26 (1H, dd, J = 4, 8 Hz, C₂-H), 2.94 (1H, dd, J = 4, 14 Hz, C₃-H), 2.75 (1H, dd, J = 8, 14 Hz, C₃-H) [consistent with the literature (Yahara et al., 1985)]; ¹³C MMR (CD₃OD, 100 MHz): δ 177.3 (C₁), 145.9 (C_3′), 144.9 (C_4′), 130.3 (C_1′), 121.9 (C_6′), 117.7 (C_2′), 116.1 (C_5′), 73.1 (C₂), 41.0 (C₃); HRMS (ESI-TOF, negative): m/z 197.0470 [M-H]_- (calcd. for C9H9O5, 197.0450).

Forced swimming test (FST)    The FST was performed according to the method of Porsolt et al. (1977a). Briefly, mice were individually placed into a glass cylinder (40 cm × 7 cm i.d.) filled with water at 25°C to a depth of 20 cm. The total immobility time of the mice was assessed during the last 4 min of a 6-min observation period. Immobility was defined as the absence of movement except for movement to keep the mouse's head above the water. Sample solutions (10 mL/kg body weight) were administrated orally by gavage 1 h before the test. Frs. 1 – 3 obtained from the LE were tested at two different doses, a low and a high dose that corresponded to 2500 and 5000 mg/kg of the LE, respectively. Fr. 4 was tested only at the high dose.

Measurement of locomotor activity    Locomotor activity was measured using a multi-channel activity monitoring system (SUPERMEX, Muromachi Kikai Co. Ltd., Tokyo, Japan). Briefly, mice were individually placed into an empty cage (210 × 315 × 130 mm height) and total motor activity accounts in each 10-min segment were recorded for 60 min. Sample solutions (10 mL/kg body weight) were administrated orally by gavage 1 h before the test.

HPLC    An Agilent 1200 Series HPLC system equipped with a diode array detector (Agilent Technologies, Palo Alto, CA) and a Capcell Pak C18 MG column (250 mm × 4.6 mm i.d.; particle size, 5 µm; Shiseido, Tokyo, Japan) were used. The operating conditions were as follows: column oven temperature, 40°C; mobile phase, a linear gradient from 100% solvent A [water (pH 2.5 adjusted with phosphoric acid):acetonitrile = 90:10] to 100% solvent B (acetonitrile) in 30 min, held at 100% B for 5 min; flow rate, 1 mL/min; and injection volume, 1 µL. The samples were filtered through a 0.45-µm PTFE membrane filter (DISMIC-13HP, Toyo Roshi) prior to injection. Wavelengths of 280 and 310 nm were used for the quantification of DLA and RA, respectively, and the ultraviolet spectra (wavelengths from 200 to 400 nm) plus retention time were used for their identification.

RA enrichment studies    Five g of dried lavender flowers were extracted with 100 mL of distilled water containing RA (0, 22, 44, or 88 mg) at 90°C for 1 h. After the extraction, the concentration of DLA and RA was determined by HPLC analysis as described above.

Measurement of total phenolic content    The total phenolic content was determined using the Folin-Denis method (Folin and Denis, 1915) with minor modifications. Folin-Denis reagent was prepared as follows: the solution containing 12.5 g of sodium tungstate dihydrate, 2.5 g of phosphomolybdic acid hydrate, 6.25 mL of 85% phosphoric acid and 90 mL of distilled water was refluxed for 2 h, cooled, diluted to 500 mL with water and stored in an amber bottle. Sample solution (100 µL) containing the LE or its fractions in water and Folin-Denis reagent (100 µL) was put into a 96-well microplate. (+)-Catechin (3 – 25 mg/L in water) was used as a standard. After 3 min, 10% sodium carbonate (100 µL) was added to each well, and the plate was incubated at room temperature for 60 min. Absorbance was measured at 655 nm using a microplate reader (MTP-310Lab; Corona Electric, Hitachinaka, Japan). The total phenolic content was calculated by linear regression analysis of the dose-response curves and expressed as (+)-catechin equivalent.

Nuclear Magnetic Resonance (NMR) Spectroscopy    NMR spectra (¹H, 400 MHz; ¹³C, 100 MHz) were recorded on a Bruker Avance 400 spectrometer (Bruker, Tsukuba, Japan) in CD₃OD and fully assigned using COSY, HMQC, and HMBC. All chemical shifts are quoted on the δ scale in ppm using residual solvent as the internal standard (CD₃OD: ¹H NMR, δ = 3.30; ¹³C NMR, δ = 49.0). Coupling constants (J) are reported in Hz with splitting abbreviations (d = doublet, dd = double doublet)

High-resolution mass spectrometry (HRMS)    High-resolution mass spectra were obtained on a Waters LCT Premier time-of-flight (TOF) mass spectrometer (Nihon Waters, Tokyo, Japan) using electrospray ionization (ESI) at a cone voltage of 40 V.

Statistical analysis    All data are expressed as the mean ± standard error of the mean (SEM) and were analyzed using Microsoft Excel 2010 (Microsoft Japan, Tokyo, Japan) with the add-in software Statcel 3 (OMS publishing, Saitama, Japan). Comparisons between two groups were performed using the Student's t-test. The Williams test was used for multiple comparisons in dose-response experiments. P < 0.05 was considered to be statistically significant.

Results and Discussion

Total phenolic and rosmarinic acid (RA) contents of fractions obtained from the LE    To investigate whether there are antidepressant-like constituents other than RA in lavender, the LE was fractionated by prep-HPLC. As shown in Fig. 1, the LE was successfully separated into one RA fraction (Fr. 3) and three other fractions (Fr. 1, 2, and 4) based on the retention time on the ODS column. Table 1 shows the total phenolic and RA content in these fractions. RA was mostly confined to Fr. 3, in which 48 % of the yield was composed of RA. Conversely, the total phenolic content was distributed across all of the fractions, among which the RA fraction (Fr. 3) accounts for only 13% of the sum of the total phenolic content in the whole fraction. These results indicate that considerable amounts of phenolic constituents other than RA are present in Frs. 1, 2, and 4.

Fig. 1.

HPLC chromatograms of the lavender extract (LE) and its fractions obtained by prep-HPLC. A wavelength of 210 nm was used for recording the chromatograms.

Antidepressant-like effects of the LE and its fractions    Antidepressant-like effects of the LE and its fractions were evaluated by the forced swimming test (FST). The FST is the most widely used animal model for screening potential antidepressant activity. Immobility displayed by rodents when subjected to an avoidable stress such as forced swimming is assumed to be related to a state of despair, and immobility time in the FST has been shown to be reduced by the administration of clinically effective antidepressants (Porsolt et al., 1978; Porsolt et al., 1977b). In the present study, oral administration of the LE to mice (500, 1000, and 2500 mg/kg) produced a significant reduction in immobility time in the FST (Fig. 2a). In addition, Fr. 1 (1610 and 3221 mg/kg), Fr. 3 (127 mg/kg), and Fr. 4 (314 mg/kg) produced a significant reduction of immobility time in mice after oral administration (Fig. 2a, b). Fr. 3 is composed of a large amount of RA; therefore, these results suggest that important antidepressant-like constituents other than RA might be present in Frs. 1 and 4.

Fig. 2.

Effects of the lavender extract (LE) and its fractions on immobility time in mice. Distilled water (a) or 1% Tween 80 (b) was used as the vehicle. Data are expressed as the mean ± SEM (n = 8). DW, distilled water; IMP, imipramine (positive control); ^**P < 0.01, ^*P < 0.05 vs. control (DW); ^†† P < 0.01 vs. control (Tween 80).

Identification of 3-(3,4-dihydroxyphenyl)lactic acid (DLA) in Fr. 1    As shown in Fig. 1, a single high-abundance peak (peak A) is present in the HPLC chromatogram of Fr. 1, whereas no such clear peak appears in the chromatogram of Fr. 4. Therefore, we decided to first identify peak A in Fr. 1. Chromatographic separation of the LE on the SP-70 porous resin and subsequent purification by prep-HPLC yielded a pure compound identical to peak A in Fr. 1. Based on the ¹H and ¹³C NMR spectral and MS spectral analyses, this compound was identified as DLA (Fig. 3). DLA is a structural element of RA, in which caffeic acid (CAA) is linked to DLA by an ester linkage (Fig. 3). In previous studies, DLA has been found in a number of Lamiaceae plants, such as Salvia miltiorrhiza (Zhao et al., 2008), Origanum vulgare (Fujie et al., 2003), and Mentha haplocalyx (She et al., 2010), whereas, to the best of our knowledge, its presence in Lavandula species has not been previously reported.

Fig. 3.

Chemical structures of 3-(3,4-dihydroxyphenyl)lactic acid (DLA), caffeic acid (CAA), and rosmarinic acid (RA).

Antidepressant-like effects of DLA, RA, and CAA    Antidepressant-like effects of RA and DLA were evaluated using the FST in mice (Fig. 4a). The oral administration of RA (37.5 mg/kg) resulted in a significant reduction in immobility time in the FST. Moreover, DLA produced a dose-dependent reduction in immobility time in mice after oral administration, and significant effects were observed at doses of 150 and 300 mg/kg. Previous studies have demonstrated that the oral and intraperitoneal administration of RA produced a significant reduction in immobility time in the FST in mice (Takeda et al., 2002a; Takeda et al., 2002b). However, to the best of our knowledge, the anti-immobility effects of DLA have not been previously reported. Recently, Kwon et al. (2014) reported that the oral administration of DLA at doses in the range of 1 – 10 mg/kg did not produce any antidepressant-like effects in mice subjected to the FST. This is consistent with our results in which the oral administration of DLA at doses of 37.5 and 75 mg/kg did not produce a significant reduction in immobility time in the FST.

Fig. 4.

Effects of 3-(3,4-dihydroxyphenyl)lactic acid (DLA), rosmarinic acid (RA), and caffeic acid (CAA) on immobility time in mice. Distilled water (a) or 1% carboxymethyl cellulose (b) was used as the vehicle. The data are expressed as the mean ± SEM (n = 8). DW, distilled water; IMP, imipramine (positive control); CMC, carboxymethyl cellulose; ^**P < 0.01, ^*P < 0.05 vs. control (DW); ^† P < 0.05, ^†† P < 0.01 vs. control (1% CMC).

To investigate the structural elements necessary for RA to produce antidepressant-like activity, CAA was evaluated for its potential antidepressant-like effects after oral administration. Previous studies have demonstrated that the intraperitoneal administration of CAA produced a significant reduction in immobility time in the FST in mice (Takeda et al., 2002a); whereas, to the best of our knowledge, the antidepressant-like effects of CAA after oral administration have not been previously reported. As shown in Fig. 4b, the oral administration of CAA (75 and 150 mg/kg) produced a significant reduction in immobility time in the FST. Since both DLA and CAA produced a significant reduction in immobility time in the FST, we suggest that both structural elements may be important for the antidepressant-like effects of RA. Additional experiments using another animal model of depression such as the tail suspension test (Cryan et al., 2005) might provide further evidence on the antidepressant-like activities of these compounds.

Effects of DLA, RA, and CAA on the locomotor activity in mice    Psychostimulants such as d-amphetamine and caffeine have been reported to decrease the immobility time in the FST, apparently due to motor stimulation rather than to persistent attempts to escape (Porsolt et al., 1978; Porsolt et al., 1977b). Such false-positive results can be excluded by measuring the locomotor activity (Porsolt et al., 1978). In the present study, neither DLA (300 mg/kg) nor RA (37.5 mg/kg) affected the locomotor activity in mice after oral administration (Fig. 5a). Conversely, CAA (150 mg/kg) increased the locomotor activity during the 10 – 20 min segment after oral administration (Fig. 5b). These results suggest that the antidepressant-like effect of CAA in the FST might be attributed in part to motor stimulation. A mild psychostimulant effect of CAA after oral administration in mice was also reported by Ohnishi et al. (2006).

Fig. 5.

Effects of DLA, RA (a), and CAA (b) on the locomotor activity in mice. Distilled water (a) or 1% carboxymethyl cellulose (b) was used as the vehicle. The data are expressed as the mean ± SEM (n = 8). DW, distilled water; CMC, carboxymethyl cellulose; ^*P < 0.05 vs. control (1% CMC).

Concentration of DLA and RA in the LE    The concentration of DLA and RA in the LE was determined by HPLC analysis (Table 2), whereas the concentration of CAA was unable to be determined due to peak overlapping. From these data, it can be calculated that 34.3 mg/kg of RA is present in the highest dose of LE (2500 mg/kg) in the FST (Fig. 2a). This value is close to an effective dose of RA (37.5 mg/kg) in the FST when administrated alone (Fig. 4a). Conversely, the same dose of the LE contains 47.5 mg/kg of DLA, which is about one-third of the lower effective dose of DLA (150 mg/kg) in the FST (Fig. 4a). These comparisons suggest that RA might be a more important constituent than DLA, even though RA by itself cannot explain the antidepressant-like effects of the LE at lower doses (500 mg/kg and 1000 mg/kg) in the FST. Combined effects of RA, DA, and other possible contributors such as apigenin and its glycosides (Kageyama et al., 2012) should be investigated for further understanding of the antidepressant-like effects of the LE.

Table 2. Concentration of DLA, RA, and CAA in the LE.

Compound	Concentration (% by dry weight)
3-(3,4-Dihydroxyphenyl)lactic acid (DLA)	1.90
Rosmarinic acid (RA)	1.37
Caffeic acid (CAA)	NDa

a  Not determined due to peak overlapping.

Effects of RA enrichment on the concentration of DLA and RA in the LE after hot-water extraction    The question remains as to how DLA can be produced in the LE. DLA might be accumulated within the plant as an intermediate for the biosynthesis of RA (Di et al., 2013) or inversely produced by the degradation of RA due to hot-water extraction of the dried flowers. In order to clarify whether DLA can be produced from RA during hot-water extraction, the effects of RA enrichment on the concentration of DLA was investigated (Fig. 6). The results clearly showed that the enrichment of RA did not affect the concentration of DLA after extraction and simply resulted in a linear increase in the concentration of RA. These findings suggest that DLA formation from RA during hot-water extraction is negligible. This is consistent with the results of Fujie et al. (2003), who demonstrated that the addition of RA did not affect the concentration of DLA after boiling of oregano leaves in 2% sodium chloride solution for 15 min. Further investigation including analysis of fresh flowers of lavender is needed to clarify whether DLA can be accumulated within the plant before harvesting.

Fig. 6.

Effects of RA enrichment on the concentration of DLA and RA in the LE after hot-water extraction at 90°C for 1 h.

In conclusion, the present study revealed that DLA isolated from lavender flowers is a novel antidepressant-like substance; the oral administration of DLA produced a significant reduction in immobility time in the FST in mice. Further studies are needed to clarify the mechanisms mediating the antidepressant-like action of DLA.

References

•  Blumenthal,  M.,  Busse,  W.R.,  Goldberg,  A.,  Gruenwald,  J.,  Hall,  T.,  Riggins,  C.W., and  Rister,  R.S. (eds.) (1998). Lavender flower. In “The complete German Commission E monographs: therapeutic guide to herbal medicines,” American Botanical Council, Austin, pp. 159-160.
•  Charles,  D. (2013). Lavender. In “Antioxidant Properties of Spices, Herbs and Other Sources,” Springer, New York, pp. 363-369.
•  Cryan,  J.F.,  Mombereau,  C., and  Vassout,  A. (2005). The tail suspension test as a model for assessing antidepressant activity: review of pharmacological and genetic studies in mice. Neurosci. Biobehav. Rev., 29, 571-625.
•  Di,  P.,  Zhang,  L.,  Chen,  J.,  Tan,  H.,  Xiao,  Y.,  Dong,  X.,  Zhou,  X., and  Chen,  W. (2013). ¹³C tracer reveals phenolic acids biosynthesis in hairy root cultures of Salvia miltiorrhiza. ACS Chem. Biol., 8, 1537-48.
•  Folin,  O. and  Denis,  W. (1915). A colorimetric method for the determination of phenols (and phenol derivatives) in urine. J. Biol. Chem., 22, 305-308.
•  Fujie,  A.,  Yoshida,  K., and  Oba,  K. (2003). Antioxidative phenolic acids from oregano (Origanum vulgare L.) leaves. Nippon Shokuhin Kagaku Kogaku Kaishi, 50, 404-410 (in Japanese).
•  Kageyama,  A.,  Ueno,  T.,  Oshio,  M.,  Masuda,  H.,  Horiuchi,  H., and  Yokogoshi,  H. (2012). Antidepressant-like effects of an aqueous extract of lavender (Lavandula angustifolia Mill.) in rats. Food Sci. Technol. Res., 18, 473-479.
•  Kwon,  G.,  Kim,  H.J.,  Park,  S.J.,  Lee,  H.E.,  Woo,  H.,  Ahn,  Y.J.,  Gao,  Q.,  Cheong,  J.H.,  Jang,  D.S., and  Ryu,  J.H. (2014). Anxiolytic-like effect of danshensu [(3-(3,4-dihydroxyphenyl)-lactic acid)] in mice. Life Sci., 101, 73-8.
•  Lis-Balchin,  M. (2006). Lavender oil. In “Aromatherapy science: a guide for healthcare professionals,” Pharmaceutical Press, London, pp. 219-226.
•  Ohnishi,  R.,  Ito,  H.,  Iguchi,  A.,  Shinomiya,  K.,  Kamei,  C.,  Hatano,  T., and  Yoshida,  T. (2006). Effects of chlorogenic acid and its metabolites on spontaneous locomotor activity in mice. Biosci. Biotechnol. Biochem., 70, 2560-2563.
•  Porsolt,  R.D.,  Anton,  G.,  Blavet,  N., and  Jalfre,  M. (1978). Behavioural despair in rats: a new model sensitive to antidepressant treatments. Eur. J. Pharmacol., 47, 379-91.
•  Porsolt,  R.D.,  Bertin,  A., and  Jalfre,  M. (1977a). Behavioral despair in mice: a primary screening test for antidepressants. Arch. Int. Pharmacodyn. Ther., 229, 327-36.
•  Porsolt,  R.D.,  Le Pichon,  M., and  Jalfre,  M. (1977b). Depression: a new animal model sensitive to antidepressant treatments. Nature, 266, 730-2.
•  She,  G.M.,  Xu,  C.,  Liu,  B., and  Shi,  R.B. (2010). Polyphenolic acids from mint (the aerial of Mentha haplocalyx Briq.) with DPPH radical scavenging activity. J. Food Sci., 75, C359-C362.
•  Takeda,  H.,  Tsuji,  M.,  Inazu,  M.,  Egashira,  T., and  Matsumiya,  T. (2002a). Rosmarinic acid and caffeic acid produce antidepressive-like effect in the forced swimming test in mice. Eur. J. Pharmacol., 449, 261-7.
•  Takeda,  H.,  Tsuji,  M.,  Matsumiya,  T., and  Kubo,  M. (2002b). Identification of rosmarinic acid as a novel antidepressive substance in the leaves of Perilla frutescens Britton var. acuta Kudo (Perillae Herba). Nihon Shinkei Seishin Yakurigaku Zasshi, 22, 15-22.
•  Ueno,  T.,  Masuda,  H.,  Mutoh,  A., and  Yokogoshi,  H. (2012). Anti-depressant-like and anti-stress-ulcer effects of an aqueous extract of lavender (Lavandula angustifolia Mill.) on mice. Nippon Shokuhin Kagaku Kogaku Kaishi, 59, 435-441 (in Japanese).
•  Yahara,  S.,  Satoshiro,  M.,  Nishioka,  I.,  Nagasawa,  T., and  Oura,  H. (1985). Isolation and characterization of phenolic compounds from coptidis rhizome. Chem. Pharm. Bull., 33, 527-31.
•  Zhao,  G.R.,  Zhang,  H.M.,  Ye,  T.X.,  Xiang,  Z.J.,  Yuan,  Y.J.,  Guo,  Z.X., and  Zhao,  L.B. (2008). Characterization of the radical scavenging and antioxidant activities of danshensu and salvianolic acid B. Food Chem. Toxicol., 46, 73-81.