Variability, toxicity, and antioxidant activity of Eupatorium cannabinum (hemp agrimony) essential oils.

Content: Eupatorium cannabinum L. (Asteraceae) is as a potential source of biologically active compounds. The plant is used in traditional medicine for the treatment of diarrhea and livers diseases. Objective: The present study provides investigation on pharmacological properties (antioxidant and toxic activities) of essential oils of E. cannabinum, collected from 11 wild populations in Lithuania. Materials and methods: Twenty-two hemp agrimony essential oil samples were prepared by hydrodistillation according to the European Pharmacopoeia, and their chemical composition was determined by GC–FID and GC–MS. Compositional data were subjected to principal components analysis (PCA). Instead of conventional spectrophotometric methods, cyclic voltammetry (CV) and square wave voltammetry (SWV) techniques were applied to determine antioxidant activity of hemp agrimony essential oils. Meanwhile, toxicity of the oils was determined using brine shrimp (Artemia sp.) assay. Results: Chemical profiles of E. cannabinum oils were described according to the first predominant components: germacrene D (≤22.0%), neryl acetate (≤20.0%), spathulenol (≤27.2%), and α-terpinene (11.5%). For the first time, α-zingiberene (≤7.8%) was found to be among three major constituents (as the second one) for hemp agrimony oils. SWV measurements revealed that oxidation potentials of compounds present in the oils are lower (below 0.1 V) compared with that of well-known antioxidant quercetin (0.15 V). Toxicity tests evaluated that hemp agrimony oils containing predominant amounts of germacrene D and neryl acetate were notably toxic (LC₅₀ value 16.3–22.0 μg/mL). Conclusion: The study provided some new data concerning chemical composition and pharmaceutical properties of E. cannabinum essential oils.

Keywords: α-Zingiberene; cyclic voltammetry; Germacrene D; neryl acetate; square wave voltammetry; statistical data analysis

Eupatorium cannabinum L. (Asteraceae), commonly known as hemp agrimony, hemlock parsley, holy rope or Indian ague root, is a perennial plant of 50–170 cm in height, with reddish stems, leaves divided into narrow leaflets, and tiny light pink or purple flowers forming domed clusters. Hemp agrimony populates wide regions in N. America, Europe, and Eastern Asia (Herz, [10]); its preferred growing locations are fertile limy soil, moist loam or clay, mainly near water sources (Lekavicius, [15]).

Eupatorium cannabinum is a medicinal and aromatic herb with antibacterial, immunological, cytostatic, anti-inflammatory, fungicidal, etc. properties. In Lithuanian folk medicine, the plant has been used as a remedy for treatment of liver diseases and against diarrhea (Lekavicius, [15]). Currently hemp agrimony is used as a component for some homeopathic preparations.

Various extracts of E. cannabinum aerial parts contain many biological active constituents, such as sesquiterpene lactones, pyrrolizidine alkaloids, benzofurans and dihydrobenzofurans, polysaccharides, flavonoids, terpenoids (including miscellaneous mono- and sesquiterpenoids, diterpenes, and triterpenes) (Chen et al., [4]; Edgar et al., [7]; Herz, [10] and references therein; Stevens et al., [31]).

Natural herbal products can be considered a valuable alternative for some synthetic drugs. Essential oils of hemp agrimony are known to possess antibacterial, cytotoxic, fungicidal, etc., properties (Dubey et al., [5], [6]; Mehdiyeva et al., [17]; Pirineau et al., [27]; Senatore et al., [29]; Yadav & Dubey, [34]). The main composition of the oils from various countries is summarized in Table 1.

Table 1. Principal composition (at least over 5.0%) of E. cannabinum essential oils investigated during last decade and their activities.

Antioxidant activities of essential oils are usually investigated spectrophotometrically by various assays such as 1,1-diphenyl-2-picrylhydrazyl (DPPH) free radical scavenging and 2,2′-azinobis (3-ethylbenzothiazoline-6-suphonic acid) diammonium salt (ABTS) cation radical decolorization assay or linoleic acid peroxidation, ferric/cupric reducing power, and conjugated autoxidizable triene assay (Mukazayire et al., [20]; Öztürk, [22]; Riahi at al., 2013; Tchobo et al., [32]).

In this investigation of essential oils, electrochemical assay instead of spectrophotometry was proposed. Many antioxidants exhibit inherent electroactivity and can be easily oxidized at an inert electrode (Sochor et al., [30]). Carbon paste electrodes consisting of graphite and liquid binder (Kalcher et al., [13]) have a number of advantages such as chemical inertness, low cost, simple fabrication, low background currents, rapid renewal of the surface, and easy modification. Modified carbon pastes typically contain admixed water-insoluble compounds. Electrodes prepared from graphite and vegetable oils instead of usual inert binder (Apetrei et al., [2]) were suitable to examine the antioxidant properties of these oils. Voltammetric signals obtained under different conditions at these oil-modified carbon paste electrodes allowed discrimination of the oils from different plants. Similar methodology was used in this work to test the antioxidant properties of essential oils obtained from E. cannabinum. The oils were admixed to carbon paste prepared from graphite and paraffin oil.

Previous studies of the oils from hemp agrimony grown in Lithuania were rather fragmental (Judzentiene, [11], [12]). Increasing awareness in the usage of natural herbal products or phytoconstituents for healing remedies encouraged us to evaluate pharmacological properties of hemp agrimony. This research aimed to investigate in detail the variability of essential oils of E. cannabinum growing wild in various localities in Lithuania and for the first time to evaluate their toxicity and antioxidant properties.

Eupatorium cannabinum plants (up to 1.0 kg) were collected at the flowering stage (July–August, during several years) from 11 populations (in Radviliskis, Varena, Svencionys, Moletai, and Vilnius districts) near natural water sources in Lithuania (Figure 1). Plant material was dried at room temperature (20–25 °C); leaves and inflorescences were separated before drying. Plant material has been identified and voucher specimens (Nos. 65233, 65233′, 65234, 65238′, 65240, 65241, 68920–68923, and 68923′) were deposited in the Herbarium (BILAS) of the Institute of Botany, Nature Research Centre (Vilnius, Lithuania).

Graph: Figure 1. Oils (samples 1–22) obtained from Eupatorium cannabinum plants collected at corresponding sampling sites (in Lithuania) characterized by locality (district (d.) and name of natural water source): 1, 2 – Varena d., Griova river; 3, 4 – Vilnius d., Asveja lake; 5, 6 – Moletai d., Asveja lake; 7, 8 – Radviliskis d., Susve river; 9, 10 – Svencionys d., Zeimena lake; 11, 12 – Moletai d., Susedas lake; 13, 14 – Moletai d., Baltieji Lakajai lake; 15, 16 – Moletai d., Luokesai lake; 17, 18 – Svencionys d., Luknele river; 19-22 – Svencionys d., Gilutis lake.

The essential oils were prepared by hydrodistillation (2.5 h) of air-dried material in a Clevenger-type apparatus according to the European Pharmacopoeia. The yields of yellow oils with a strong characteristic odor ranged 0.1–0.38% (v/w, on a dry weight basis).

Quantitative analyses of the essential oils were carried out by GC on a DB-Wax capillary column (polyethylene glycol 30 m × 0.25 mm i.d., film thickness 0.25 μm) using a Perkin–Elmer Clarus 500 chromatograph (PerkinElmer Health Sciences, Inc., Shelton, CT) equipped with a FID. The oven temperature was programmed from 60 °C to 280 °C at a rate of 3 °C/min. Injector and detector temperatures were 250 °C; carrier gas helium (1 mL/min); ratio split 1:60. At least two repetitions (n≥2) per analysis were performed.

Qualitative analyses were performed by GC–MS using a chromatograph HP 5890 interfaced to an HP 5971 mass spectrometer (ionization voltage 70 eV, scan time 0.6 s, scan range 35–400 Da) and equipped with a capillary column DB-5 (50 m × 0.32 mm i.d., film thickness 0.25 μm).

The GC oven temperature was programmed from 60 °C (isothermal for 2 min) to 160 °C (isothermal for 1 min) at a rate of 5 °C/min, then increased to 250 °C at a rate of 10 °C/min, and the final temperature was kept for 3 min. The temperature of the injector and detector was 250 °C. The flow rate of carrier gas (helium) was 1 mL/min, ratio split 1:40. Mass spectra in electron mode were generated at 70 eV, 0.97 scans/s, mass range 35–650 m/z.

The percentage composition of the essential oils was computed from GC peak areas without correction factors. Qualitative analysis was based on comparison of retention indexes on both columns, co-injection of some terpene references and C8–C28n-alkane series; and mass spectra with corresponding data in the literature (Adams, [1]) and computer mass spectra libraries (Wiley and NBS 54K). Identification has been approved when computer matching with the mass spectral libraries was with probability above 90%. The relative proportions of the oil constituents were expressed as percentages obtained by peak area normalization, all relative-response factors being taken as one.

Principal component analysis (PCA) and agglomerative hierarchical clustering (AHC) were performed by XLSTAT (Version 2009, Addinsoft, New York, NY). Both methods were applied using a percentage of 22 oils and 13 variables (individual constituents with quantity ≥5%, at least in one sample). AHC was done using various methods and aggregation criterions, and final definition of the oil groups was based on Pearson dissimilarity and complete linkage as an agglomeration method.

Phosphate buffer prepared from 0.025M KH2PO4 (Fluka, Newport News, VA) contained 0.1M KCl (Fluka, Newport News, VA). The pH value was adjusted with KOH. Quercetin dihydrate (minimum 98% HPLC) was obtained from Sigma (St. Louis, MO).

Electrochemical experiments were carried out with a BAS-Epsilon System (West Lafayette, IN) and a three-electrode cell. Platinum wire and Ag/AgCl, 3M NaCl served as counter- and reference electrodes, respectively. Essential oil- or quercetin- modified carbon paste electrode served as a working electrode. Oil-modified carbon paste was prepared by mixing 100 mg of graphite powder (Merck) with 40 μL of essential oil and 50 μL of paraffin oil (Fluka, Newport News, VA). Quercetin-modified paste was prepared from 97 mg of graphite, 3 mg of quercetin dihydrate, and 50 μL of paraffin oil. The paste was immediately placed into the cavity of a homemade electrode consisting of a plastic tube (diameter 2.9 mm) and a copper wire as an electrode contact. The surface was then smoothened on a weighing paper. The electrochemical measurements were performed immediately after the electrode preparation. Cyclic voltammetry was performed at a potential scan rate 50 mV/s.

The square wave voltammograms were recorded under following conditions: step potential 4 mV, amplitude 50 mV, and frequency 25 Hz.

Toxicity of six samples of hemp agrimony inflorescence and leaf essential oils were tested in vivo, using brine shrimp Artemia sp. (larvae) (McLaughlin et al., [16]). The eggs of shrimps hatch within 48 h to provide larvae (nauplii) in sea water (31 g/L sea salt) at 22–25 °C, after, different concentrations of hemp agrimony essential oils dissolved in dimethyl sulfoxide were added, and survivors were counted after 24 h. Lethality (LC50) of nauplii was calculated (n≥4, with 95% confidence interval).

Quantitative and qualitative compositional characteristics of E. cannabinum oils were evaluated chromatographically. In total, up to 94 compounds were identified in the oil samples, but only 13 components were found in amounts over 5.0%. Principal chemical composition (over 5%, at least in one sample) of hemp agrimony inflorescence (indicated by even numbers) and leaf (odd numbers) essential oils are presented in Table 2 (total 22 samples). The compositional data were subjected to statistical analysis and AHC distinguished four subclusters (Figure 2). Two major subclusters (II and IV) contained the most samples (18 out of 22). The other two (I and III) were comprised of two non-typical pairs of oils, in both cases, inflorescences and leaves of the same plants.

Graph: Figure 2. Dendrogram of 22 essential oils of Eupatorium cannabinum from Lithuania clustered by Pearson dissimilarity and complete linkage.

Table 2. Principal composition (major constituents, ≥5.0%) of essential oils of E. cannabinum from Lithuania.

1 tr, traces (≤0.05%), numbers n (odd) indicate E. cannabinum leaf essential oils, while n+1 (even) – inflorescence oils of the same plants.

• 2 II subcluster comprised the oils indicated by numbers such as 2, 9–12, 16, 18, 21, and 22; while IV subcluster contained the oils: 1, 3-6, 15, 17, 19, and 20.
• 3 Compounds with quantity up to 5.0%.
• 4 0–1.0%: α-pinene, limonene, β-ocimene, α-terpinolene, linalool, n-non-anal, trans-thujone, camphor, borneol, terpinen-4-ol, lavandulol, neral, geraniol, geranial, silphiperfol-5-ene, hexyl tiglate, α-cubebene, silphiperfol-4,7(14)-diene, carvacrol acetate, β-bourbonene, β-cubebene, (E)-caryophyllene, lavandulyl isobutanoate, trans-α-bergamotene, (E)-β-farnesene, allo-aromadendrene, γ-muurolene, trans-muurola-4(14),5-diene, α-muurolene, and germacrene-D-4-ol.
• 5 ≤3.0%: 1,3-dimethyl benzene, camphene, β-pinene, myrcene, o-cymene, α-terpineol, isobornyl acetate, lavandulyl acetate, trans-sabinyl acetate, thymol, carvacrol, δ-elemene, epi-7-silphiperfol-5-ene, β-elemene, 2,5-dimethoxy-p-cymene, β-copaene, γ-elemene, aromadendrene, neryl propionate, α-humulene, (E)-β-ionone, bicyclogermacrene, (E, E)-α-farnesene, α-silphiperfolan-6-ol, geranyl isobutanoate, δ-cadinene, citronellyl butanoate, neryl isovalerate, salvial-4(14)-en-1-one, humulene epoxide, epi-α-muurolol, α-cadinol, (Z)-citronellyl tiglate, α-eudesmol; n-hexadecanol, n-non-adecane, phytol and n-docosane.
• 6 <5%: p-cymene, nerol, geranyl acetate, bornyl isobutanoate, β-ylangene, decyl propionate, isobornyl isovalerate, geranyl butanoate, caryophyllene oxide, caryophylla-4(14),8(15)-dien-5,α-ol, (Z)-γ-atlantone, (E)-γ-atlantone, n-eicosane, and n-heneicosane.

The samples 13 and 14 (from plants grown near lake Baltieji Lakajai, Moletai district, Eastern Lithuania), which were attributed to the Subcluster I in AHC, stood out as the most different. These oils were characterized by significantly higher quantity of spathulenol (19.0% and 27.2%, respectively) and α-zingiberene (6.5% and 7.8%, as the second major component).

In the subcluster II (composed by nine samples, Figure 1), neryl acetate (quantities of 9.7% to 20.0% in seven samples: 10–12, 16, 18, 21, and 22) and germacrene D (9.8% and 14.2% in oils 2 and 9, respectively) were the first predominant constituents. As the second or third principal compounds in subcluster II oils were found to be neryl acetate (8.5–12.0%), germacrene D (5.4–13.3%), methyl thymol (5.5–7.4%), β-bisabolene (4.4%), neryl isobutanoate (6.4%), nerolidol (8.7%), and spathulenol (4.5%). Samples 7 and 8 stood out from the others due to higher amounts of α-terpinene (≤11.5%) and spathulenol (≤8.9%).

The last Subcluster IV of oils 1, 3–6, 15, 17, 19, and 20 with the major compound germacrene D (6.8–22%) contained neryl acetate as the second (6.1–10.5%) or the third constituent (4.1–5.8%) (Figure 2). In the rest of this segment, δ-2-carene (4.5–6.5%), α-terpinene (4.1–5.8%), methyl thymol (5.9%), (E)-nerolidol (5.3%), geranyl tiglate (6.8%), and isoborbyl-2-methyl butanoate (5.0%) were identified among the second and the third main compounds.

In fact, most of the main constituents (germacrene D, neryl acetate, δ-2-carene, α-terpinene, methyl thymol, neryl isobutanoate, spathulenol, etc.) determined in the study had been found in essential oils of E. cannabinum investigated in other countries [such as Italy, Sardinia, and Corsica islands, Iran, India (Table 1)] with the exception of α-zingiberene. To the best of our knowledge, the latter sesquiterpene has been mentioned among principal constituents of hemp agrimony essential oils for the first time.

Electrochemical potential sweep techniques (cyclic voltammetry, differential pulse voltammetry, and square wave voltammetry) are widely used for acquiring information about the properties of redox active substances (Wang, [33]). The voltammetric signals obtained during a positive-going forward potential scan (anodic currents) and a negative-going backward potential scan (cathodic currents) are related, respectively, to the electrochemical oxidation and the reduction of redox active substances both present in the solution or immobilized onto the electrode surface. The resulting plot of current versus potential gives various electrochemical parameters. The peak current values (ipa and ipc, anodic and cathodic, respectively), the areas under the waves (Q) are related to the amount of charge transfer, i.e., concentrations of the electrochemically active substances. The peak potential values (Epa and Epc, anodic and cathodic, respectively) reflect the redox properties of the substances. Compounds with lower Epa values are oxidized more easily compared with those with high Epa values.

Cyclic voltammograms recorded at carbon paste electrodes modified with E. cannabinum essential oils (Figure 3) show similar voltammetric profiles: one anodic and one cathodic peak in the potential region 0.0–0.15V and only one anodic peak in the potential region above 0.6 V. Higher ipa values obtained at electrodes containing three oils (Figure 3, solid line) reflected possibly higher concentration of redox active substances. The substances that oxidize at low potentials (up to 0.4 V) are supposed to possess antioxidant properties. Therefore, by means of square wave voltammetry, the voltammetric signals recorded during a positive-going potential scan in the potential region −0.2 to 0.5 V were examined more closely.

Graph: Figure 3. Cyclic voltammograms recorded at a plain carbon paste electrode (dotted line) and at essential oil-modified (as indicated) electrodes in phosphate buffer at pH 7.2. Potential scan rate 50 mV/s.

The oils indicated by numbers 4, 18, and 22 contained substance or several substances that were oxidized at very similar potentials (Figure 4). Somewhat more positive value of the oxidation potential and considerably higher the Q value of the oil 3 was possibly indicative of other oxidizable substance/substances in greater quantities compared with those in the three other oils. The Epa and Q values defined from voltammetric curves (the mean values of three measurements with three freshly prepared electrodes) are summarized in Table 3.

Graph: Figure 4. Square wave voltammograms recorded at essential oil-modified (as indicated) carbon paste electrodes in phosphate buffer at pH 7.2. Step potential 4 m/V, amplitude 50 mV, and frequency 25 Hz.

Table 3. Epa and Q obtained from square wave voltammograms at carbon paste electrodes modified with E. cannabinum essential oils in phosphate buffer at pH 7.2. Reference electrode Ag/AgCl, 3M NaCl.

It is not known yet exactly what constituents in the essential oils contribute to the oxidation current. To evaluate the antioxidant activities of the oils, data were compared with that of some well-known antioxidant compound. Quercetin is one of the most abundant plant flavonoids and is considered to possess effective antioxidant properties. The mechanism of its electrochemical oxidation has been studied (Brett & Ghica, [3]). Therefore, electrochemical data were compared with that obtained at quercetin-modified carbon paste electrode (Figure 4, dotted line). The Epa value of quercetin was 0.15 V. Lower Epa values obtained at essential oil-modified electrodes suggesting possibly higher antioxidant activity compared with that of quercetin.

Toxicity test of the six oil samples 3, 4, 17, 18, 21, and 22 showed that lethality (LC50) of brine shrimp (Artemia sera) larvae ranged from 16.3 to 22.0 μg/mL (Table 4). These data revealed that essential oils of E. cannabinum containing appreciable amounts of germacrene D (≤22.0%) and neryl acetate (≤19.7%) were notably toxic. There are some reports concerning toxicity of volatile oils rich in germacrene D (Dubey et al., [6]; Haber et al., [9]; Kiran & Devi, [14]; Ogutcu et al., [21]). The leaf essential oils of Talauma gloriensis with the most abundant components germacrene D (43.5%) also showed notable brine shrimp toxicity (LC50 value = 14.1 μg/mL) (Haber et al., [9]). Essential oils from Eupatorium cannabinum, Chloroxylon swietenia, and Trachyspermum ammi containing among main constituents germacrene D displayed remarkable mosquitocidal, larvicidal, and broad spectrum of fungitoxic activities (Dubey et al., [6]; Kiran & Devi, [14]). Volatile oils of Salvia limbata with major constituent spathulenol (29.3%) and germacrene D (24.7%) exhibited antimicrobial, antioxidant, and antiviral activities (Ogutcu et al., [21]).

Table 4. Lethality (LC50) to brine shrimp (Artemia sera) larvae.

This study represents new data on chemical polymorphism of essential oil within E. cannabinum species from Lithuania. The compositions of hemp agrimony oils, analyzed in this study, showed a broader variation of major components compared with those investigated earlier in our laboratory or reported in other countries. Germacrene D, neryl acetate, spathulenol, α-terpinene, δ-2-carene, methyl thymol, α-zingiberene, β-bisabolene, neryl isobutanoate, (E)-nerolidol, geranyl tiglate, and isoborbyl-2-methyl butanoate were the major components indicating individual chemotypes of studied oils. It could be pointed out that α-zingiberene was determined among major constituents for hemp agrimony essential oils for the first time.

Results of toxicity (tests with brine shrimp larvae) and antioxidant activity (performed by electrochemical methods) of the oils provided an evident support for the potential uses of hemp agrimony in pharmacology. Data obtained with essential oil-modified electrodes suggest higher antioxidant activity compared with that of quercetin. To the best of our knowledge, the antioxidant and toxic activity has been evaluated for E. cannabinum oils for the first time.

The authors report that they have no conflicts of interest.