Chemical Variability and In Vitro Anti-Inflammatory Activity of Leaf Essential Oil from Ivorian Isolona dewevrei (De Wild. & T. Durand) Engl. & Diels

The chemical variability and the in vitro anti-inflammatory activity of the leaf essential oil from Ivorian Isolona dewevrei were investigated for the first time. Forty-seven oil samples were analyzed using a combination of CC, GC(RI), GC-MS and 13C-NMR, thus leading to the identification of 113 constituents (90.8–98.9%). As the main components varied drastically from sample to sample, the 47 oil compositions were submitted to hierarchical cluster and principal components analyses. Three distinct groups, each divided into two subgroups, were evidenced. Subgroup I−A was dominated by (Z)-β-ocimene, β-eudesmol, germacrene D and (E)-β-ocimene, while (10βH)-1β,8β-oxido-cadina-4-ene, santalenone, trans-α-bergamotene and trans-β-bergamotene were the main compounds of Subgroup I−B. The prevalent constituents of Subgroup II−A were germacrene B, (E)-β-caryophyllene, (5αH,10βMe)-6,12-oxido-elema-1,3,6,11(12)-tetraene and γ-elemene. Subgroup II−B displayed germacrene B, germacrene D and (Z)-β-ocimene as the majority compounds. Germacrene D was the most abundant constituent of Group III, followed in Subgroup III−A by (E)-β-caryophyllene, (10βH)-1β,8β-oxido-cadina-4-ene, germacrene D-8-one, and then in Subgroup III−B by (Z)-β-ocimene and (E)-β-ocimene. The observed qualitative and quantitative chemical variability was probably due to combined factors, mostly phenology and season, then harvest site to a lesser extent. The lipoxygenase inhibition by a leaf oil sample was also evaluated. The oil IC50 (0.020 ± 0.005 mg/mL) was slightly higher than the non-competitive lipoxygenase inhibitor NDGA IC50 (0.013 ± 0.003 mg/mL), suggesting a significant in vitro anti-inflammatory potential.


Introduction
Isolona Engl. is a genus belonging to the Annonaceae family and comprising about 20 species, widely distributed in tropical rain forests of West and Central Africa, and Madagascar. Five species of this genus grow wild in Côte d'Ivoire: I. cooperi, I. campanulata, I. soubreana, I. deightonii and I. dewevrei [1,2].
Isolona dewevrei (De Wild. & T. Durand) Engl. & Diels (synonym: Monodora dewevrei De Wild. & T. Durand) is an evergreen shrub or a tree that can reach 15 m in height. It has narrowly obovate to obovate or elliptic to narrowly elliptic leaves that are 10-17 cm long and 4-7 cm wide, with an acuminated apex. The inflorescences appear on leafy branches and sometimes on older ones, whereas the fruits are ovoid (6-7 cm long, 4-5 cm in diameter), smooth but very finely ribbed, glabrous, green and yellow at maturity [1]. Although no ethno-pharmacological use of I. dewevrei was reported in the literature, other Isolona species (I. cooperi and I. campanulata) are traditionally used as herbal medicine in Côte d'Ivoire to treat bronchial ailments, skin diseases, hematuria and infertility, and to facilitate childbirth [1,2].

Results and Discussion
The chemical composition of 47 leaf essential oil samples from I. dewevrei growing wild in Côte d'Ivoire was investigated in order to highlight possible variability. Oil samples were isolated by the hydrodistillation of fresh leaves collected at six sites of the Bossématié forest (Eastern Côte d'Ivoire) and two sites of the Haut-Sassandra forest (Western Côte d'Ivoire). The extraction yields calculated on a weight basis (w/w) were in the range of 0.094-0.506%. Analyses were carried out using a combination of GC(RI), GC-MS and 13 C-NMR, following a computerized method developed at the University of Corsica [15,22]. Most of the constituents were identified by the three techniques, including the identification by 13 C-NMR of components present at a content as low as 0.4-0.5% and compiled in our laboratory-made 13 C-NMR spectral data library. However, several minor compounds remained unidentified despite the use of these complementary techniques. Hence, samples S2 (4.115 g), S14 (4.820 g), S24 (3.702 g), S44 (4.226 g) and S47 (3.910 g), which displayed qualitative and quantitative variations of the main constituents and the minor unidentified compounds, were separately submitted to a detailed analysis by column chromatography (CC). The CC fractions were then analyzed by the three above techniques. Finally, the 47 oil compositions were subjected to a statistical analysis, and the in vitro anti-inflammatory activity of the single monoterpene-rich oil sample (S44) was evaluated.
The dendrogram from the HCA revealed three distinct groups within the 47 investigated oil samples: Group I (9 samples), Group II (19 samples) and Group III (19 samples), each consisting of two subgroups (Figure 1). The first principal factor of the PCA (F1: 44.31%), the second (F2: 22.52%) and the third (F3: 13.01%) accounted for 79.12% of the total variance of the chemical composition. The PCA map of the samples' distribution relative to the principal axes F1 and F3 (57.32%) confirmed the three chemical composition groups (Figure 2). The Groups II and III were more homogenous than Group I. The mean contents (M) and the standard deviation (SD) of the majority compounds of the different subgroups are reported in Table 2.   Order of elution and percentages on apolar column (BP-1), except components with an asterisk (*), percentages on polar column (BP-20); (#) percentages calculated by combination of GC(FID) and 13 C NMR; [b] RIa, RIp: Retention indices measured on apolar and polar capillary column, respectively; M% ± SD: mean percentage and standard deviation; (-): not detected; tr: traces level (<0.05%).
The Group II is characterized by its high content of germacrene B, (5αH,10βMe)-6,12-oxido-elema-1,3,6,11 (12) This study demonstrated both qualitative and quantitative variations of constituents within the detected groups and subgroups. The observed chemical variability of the leaf essential oil from I. dewevrei was probably due to several external factors, although genetic factors could not be completely excluded. As all samples were collected from shrubs at different phenological stages, at six sites of the Bossématié forest (Sites 1 to 6; Eastern Côte d'Ivoire) and two sites of the Haut-Sassandra forest (Sites 7 and 8; Western Côte d'Ivoire), during the rainy and dry seasons, the effect of phenology, season and harvest site on samples' distribution into groups was examined (Figure 3). Regarding the phenology, samples collected from shrubs bearing flowers are all included in Group I, where they constituted the Subgroup I−A. Moreover, Group II is exclusively constituted of samples collected from shrubs bearing neither flowers nor fruits, while all the samples collected from shrubs bearing fruits are included in Group III. Therefore, the phenology appeared to be an important factor explaining the chemical variability of the leaf essential oil of the plant. From a seasonal point of view, samples collected during the dry season are more differenciated than those harvested during the rainy season. Indeed, the three barycenters of the samples collected during the dry season, corresponding to the three chemical groups, are furthest from the origin. In addition, Subgroups I−B, II−B and III−B exclusively consisted of samples collected during the rainy season, whereas samples from Subgroup I−A were harvested during the dry season. The effect of seasons on the chemical variability of I. dewevrei leaf oil could not therefore be overlooked. Although all the harvest sites are located in the same climatic zone (mesophilic sector; Celtis spp., Triplochiton scleroxylon and its variant of Nesogordonia papaverifera and Khaya ivorensis forests) with the same soil type (ferrallitic), the sites' effect on the chemical variability was noticeable. In fact, the Subgroup I−A consisted of samples from Site 8, while those collected at Sites 3 and 4 were all included in Subgroup II−A. Samples from Site 7 are all included in Subgroup III−B, whereas those from Site 2 belonged to Group III. The harvest sites being sufficiently distant, their effect on the chemical variability could be related to possible microclimates or genetic differences. It could finally be argued that phenology, season and harvest site really impacted the chemical variability of the leaf essential oil from I. dewevrei.

Evaluation of In Vitro Anti-Inflammatory Activity
The in vitro anti-inflammatory potential of I. dewevrei leaf essential oil (S44) was evaluated by determining its ability to inhibit lipoxygenases (LOX). Indeed, LOXs are nonheme iron-containing dioxygenases that convert linoleic, arachidonic and other polyunsaturated fatty acid into biologically active metabolites involved in the inflammatory and immune responses. Several inflammatory processes such as arthritis, bronchial asthma and cancer are associated with an important production of leukotrienes catalyzed by the LOX path-way from arachidonic acid [31][32][33][34]. The inhibition of the LOX pathway with inhibitors of LOX would prevent the production of leukotrienes and therefore could constitute a therapeutic target for the treating of human inflammation-related diseases. Thus, the search for new LOX inhibitors appears as critical because many exhibit significant in vitro anti-inflammatory activity.
The ability of I. dewevrei leaf essential oil to inhibit soybean lipoxygenase was determined as an indication of potential in vitro anti-inflammatory activity. I. dewevrei leaf essential oil exhibited an inhibition of LOX activity ( Table 3). The percentage of inhibition increases with the concentration of the oil i.e., 10.3% at 0.005 mg/mL to 51.5% at 0.020 mg/mL of the essential oil. The IC 50 values (concentration at which 50% of the lipoxygenase was inhibited) were determined for the I. dewevrei leaf essential oil and for the non-competitive inhibitor of lipoxygenase, the nordihydroguaiaretic acid (NDGA) ( Table 3), usually used as a reference in in vitro anti-inflammatory assays [32][33][34]. Data showed that the IC 50 value of I. dewevrei leaf essential oil (0.020 ± 0.005 mg/mL) is slightly higher than the IC 50 value of NDGA (0.013 ± 0.003 mg/mL). The low ratio between the two values of IC 50 (I. dewevrei leaf essential oil vs. NDGA) makes it possible to consider this essential oil as a high inhibitor of the LOX activity, suggesting an in vitro anti-inflammatory potential [35]. Values are means of triplicates ± standard deviation; * NDGA: NorDihydroGuaiaretic Acid; # mg/mL.

Plant Material
The fresh leaves' samples were collected on individual I. dewevrei shrubs at different phenological stages, at six sufficiently distant sites (Sites 1 to 6) of the Bossématié forest, Region of Abengourou, Eastern Côte d'Ivoire, and at two sites (  Table S2). Plant material was authenticated by botanists from the Centre Suisse de Recherches Scientifiques (CSRS) and Centre National de Floristique (CNF) Abidjan, Côte d'Ivoire. A voucher specimen was deposited at the herbarium of CNF, Abidjan, with the reference LAA 12874.

Essential Oil Isolation and Fractionation
The essential oil samples were obtained by the hydrodistillation of fresh leaves for 3 h each, using a Clevenger-type apparatus. Yields were calculated from fresh material (w/w). Plant material and essential oil extraction data are reported in Table S2 (Supplementary material). The leaf essential oil samples S2 (4.115 g), S14 (4.820 g), S24 (3.702 g), S44 (4.226 g) and S47 (3.910 g) were separately chromatographed on a column with silica gel (Acros Organics, Waltham, MA, USA, 60-200 µm, 120 g each, except S14, 150 g), using a gradient of solvents, distilled n-pentane (VWR Chemicals, Radnor, PA, USA, 99%)/diethyl ether (VWR Chemicals, 100.0%) of increasing polarity (P/DE, 100/0 to 0/100). Seven fractions were eluted for each oil sample: F1 and F2 (eluted with n-pentane) contained hydrocarbons; F3-F6 (eluted with P/DE mixtures) contained medium polar compounds; F7 (eluted with diethyl ether) contained polar compounds. The respective weights of the fractions from the different samples are reported in the Table 4.

Gas Chromatography
Analyses were performed on a Clarus 500 PerkinElmer Chromatograph (PerkinElmer, Courtaboeuf, France), equipped with a flame ionization detector (FID) and two fused-silica capillary columns (50 m × 0.22 mm, film thickness 0.25 µm), BP-1 (polydimethylsiloxane) and BP-20 (polyethylene glycol). The oven temperature was programmed from 60 • C to 220 • C at 2 • C/min and then held isothermal at 220 • C for 20 min; injector temperature: 250 • C; detector temperature: 250 • C; carrier gas: hydrogen (0.8 mL/min); split: 1/60; injected volume: 0.5 µL. Retention indices (RI) were determined relative to the retention times of a series of n-alkanes (C8-C29) with a linear interpolation (« Target Compounds » software from PerkinElmer). The relative response factor (RFF) of each compound was calculated according to the International Organization of the Flavor Industry (IOFI)recommended practice for the use of predicted relative response factors for the rapid quantification of volatile flavoring compounds by GC(FID) [36]. Methyl octanoate was used as an internal reference, and the relative proportion of each constituent (expressed in g/100 g) was calculated using the weight of the essential oil and reference, peak area and relative response factors (RRF).

Nuclear Magnetic Resonance
All 13 C-NMR spectra were recorded on a Bruker AVANCE 400 Fourier transform spectrometer (Bruker, Wissembourg, France) operating at 100.623 MHz for 13 C, equipped with a 5 mm probe, in CDCl 3 , with all shifts referred to via an internal TMS. The following parameters were used: pulse width = 4 µs (flip angle 45 • ); relaxation delay D1 = 0.1 s, acquisition time = 2.7 s for a 128 K data table with a spectral width of 25,000 Hz (250 ppm); CPD mode decoupling; digital resolution = 0.183 Hz/pt. The number of accumulated scans was 3000 for each sample or fraction (40 mg, when available, in 0.5 mL of CDCl 3 ).

Identification of Individual Components
Identification of the individual components was carried out: (i) by a comparison of their GC retention indices on apolar and polar columns, with those of reference compounds [25,37]; (ii) by computer matching against commercial mass spectral libraries [37][38][39]; (iii) by a comparison of the signals in the 13 C-NMR spectra of the samples and fractions with those of reference spectra compiled in the laboratory spectral library, with the help of a laboratory-made software [15,22]. This method allowed for the identification of individual components of the essential oil at contents as low as 0.4-0.5%.

Statistical Analysis
The chemical compositions of 47 leaf essential oil samples from I. dewevrei were submitted to a hierarchical cluster analysis (HCA) and principal component analysis (PCA) using XLSTAT software (Addinsoft, Paris, France) [40]. Only constituents in a concentration higher than 1.0% were used as variables for the PCA analysis. The aptitude of the complete correlation matrix was checked by the Kaiser-Meyer-Olkin criterion. The HCA and dendrogram were made with dissimilarity matrices calculated in Euclidean distance, and the average link was the aggregation method systematically chosen.

In Vitro Anti-Inflammatory Capacity of Isolona dewevrei Leaf Essential Oil
The in vitro anti-inflammatory capacity of I. dewevrei leaf essential oil was evaluated by an in vitro lipoxygenase inhibition assay [41][42][43]. The in vitro analysis for lipoxygenase inhibitory activity was performed using Lipoxidase type I-B (Lipoxygenase, LOX, EC 1.13.11.12) from Glycine max (soybean) purchased from Sigma-Aldrich Chimie (Saint-Quentin-Fallavier, France). It was determined by the kinetic mode of the spectrophotometric determination method, which was performed by recording the rate of change in absorbance at 234 nm. Indeed, the increase in absorbance at 234 nm was due to the formation of 13-hydroperoxides of linoleic acid (substrate used for LOX inhibition activity assay) [41][42][43].
A stock solution of LOX was prepared by dissolving around 5.7 units/mL of LOX in PBS (Phosphate Buffer Solution; 1 unit corresponding to 1 µmol of hydroperoxide formed per min). Four concentrations of I. dewevrei leaf essential oil sample (S44) in dimethylsulfoxide (DMSO) were tested as the inhibitor solution for the LOX inhibition activity assay: 0.005, 0.010, 0.015 and 0.020 mg/mL.
The LOX inhibition assays were performed by mixing 10 µL of LOX solution with 10 µL of inhibitor solution in 970 µL of boric acid buffer (50 mM; pH 9.0) and incubating them briefly at room temperature. The reaction started by the addition of 10 µL of substrate solution (Linoleic acid, 25 mM), and the velocity was recorded for 30 s at 234 nm. One assay was measured in the absence of the inhibitor solution, and one assay was measured with DMSO mixed with distilled water (99.85% of DMSO in distilled water) which made it possible to eliminate the inhibition effect of DMSO. No inhibitor effect of DMSO on the LOX activity was detected, and the LOX activity measured without inhibitor solution was considered as a control (100% enzymatic reaction). All assays were performed in triplicate. The percentage of lipoxygenase inhibition was calculated according to the equation: Inhibition % = (V 0control − V 0sample ) × 100/V 0control V 0control is the activity of LOX in the absence of the inhibitor solution, and V 0sample is the activity of LOX in the presence of the inhibitor solution [43]. The IC 50 was calculated by the concentration of I. dewevrei leaf essential oil in DMSO inhibiting 50% of LOX activity.

Conclusions
The chemical variability of the leaf essential oil from I. dewevrei growing wild in Côte d'Ivoire was investigated through 47 oil samples. A combination of chromatographic (CC, GC(RI)) and spectroscopic (GC/MS, 13 C-NMR) techniques was used to determine the samples' chemical compositions. One hundred and thirteen constituents accounting for 90.8-98.9% of the whole sample compositions were identified, and the main components varied drastically from sample to sample. Therefore, the 47 oil compositions were submitted to hierarchical cluster and principal components analyses, which evidenced three distinct chemical groups, each dividing into two subgroups. The Subgroup I−A was dominated by (Z)-β-ocimene, β-eudesmol, germacrene D and (E)-β-ocimene, while (10βH)-1β,8βoxido-cadina-4-ene, santalenone, trans-α-bergamotene and trans-β-bergamotene were the main compounds of Subgroup I−B. The prevalent constituents of Subgroup II−A were germacrene B, (E)-β-caryophyllene, (5αH,10βMe)-6,12-oxido-elema-1,3,6,11(12)-tetraene and γ-elemene. The Subgroup II−B displayed germacrene B, germacrene D and (Z)-βocimene as the majority compounds. Germacrene D was the most abundant constituent of Group III, followed in Subgroup III−A by (E)-β-caryophyllene, (10βH)-1β,8β-oxidocadina-4-ene, germacrene D-8-one, and then in Subgroup III−B by (Z)-β-ocimene and (E)β-ocimene. Compounds bearing the eudesmane skeleton characterized Subgroup I−A and also Subgroup III−B to a lesser extent. Likewise, santalane, bergamotane and bisabolane skeletons were markers of Subgroup I−B, while the elemane skeleton was specific to Group II. Group III markers were oxygenated germacrane and cadinane compounds. Although genetic factors could not be completely excluded, the observed qualitative and quantitative chemical variability of the leaf essential oil from I. dewevrei could be related to mostly phenology and season, then harvest site to a lesser extent. A leaf oil sample (S44) was tested for its lipoxygenase inhibition ability. The oil IC 50 value (0.020 ± 0.005 mg/mL) was slightly higher than the non-competitive lipoxygenase inhibitor NDGA IC 50 value (0.013 ± 0.003 mg/mL). Therefore, this leaf essential oil exhibited significant in vitro antiinflammatory potential.
Supplementary Materials: The following are available online, Table S1 (Chemical composition of the 47 leaf essential oil samples from Isolona dewevrei) and Table S2 (Plant material, essential oil extraction and climate data).