Evaluation of the Anti-Trypanosomal Activity of Vietnamese Essential Oils, with Emphasis on Curcuma longa L. and Its Components

Human African trypanosomiasis (HAT), known as sleeping sickness and caused by Trypanosoma brucei, is threatening low-income populations in sub-Saharan African countries with 61 million people at risk of infection. In order to discover new natural products against HAT, thirty-seven Vietnamese essential oils (EOs) were screened for their activity in vitro on Trypanosoma brucei brucei (Tbb) and cytotoxicity on mammalian cells (WI38, J774). Based on the selectivity indices (SIs), the more active and selective EOs were analyzed by gas chromatography. The anti-trypanosomal activity and cytotoxicity of some major compounds (isolated or commercial) were also determined. Our results showed for the first time the selective anti-trypanosomal effect of four EOs, extracted from three Zingiberaceae species (Curcuma longa, Curcuma zedoaria, and Zingiber officinale) and one Lauraceae species (Litsea cubeba) with IC50 values of 3.17 ± 0.72, 2.51 ± 1.08, 3.10 ± 0.08, and 2.67 ± 1.12 nL/mL respectively and SI > 10. Identified compounds accounted for more than 85% for each of them. Among the five major components of Curcuma longa EO, curlone is the most promising anti-trypanosomal candidate with an IC50 of 1.38 ± 0.45 µg/mL and SIs of 31.7 and 18.2 compared to WI38 and J774 respectively.


Introduction
Human African trypanosomiasis (HAT) or sleeping sickness is caused by two subspecies of the parasite Trypanosoma brucei, T. brucei gambiense and rhodesiense, while another subspecies, T. brucei brucei affects non-human vertebrates [1]. T. brucei gambiense causes the chronic form in West and Central Africa while T. brucei rhodesiense causes the acute form in Eastern and Southern Africa. Although many efforts were made this past decade to decrease HAT incidence, this fatal disease is still endemic in 36 African countries [2]. Remote rural areas are the most affected partly because of high poverty, higher risk of infection from the livestock reservoir (by the tsetse fly, responsible for parasite transmission, mainly living in rural areas), and lack of health care accessibility and infrastructures for current drugs These nineteen EOs were then analyzed for dose-response activity on Tbb bloodstream form and also on mammalian WI38 and J774 cells to calculate IC50 values and selectivity index (SI). Three samples extracted from three Zingiberaceae species, Curcuma longa, Curcuma zedoaria, and Zingiber officinale, and one sample extracted from a Lauraceae species, Litsea cubeba, showed the most active and selective effects with IC50 values of 3.17, 2.51, 3.10, and 2.67 nL/mL respectively and SI > 10 compared to cytotoxicity (Table 1 in bold).   The chemical composition of these four interesting EOs was analyzed using gas chromatography (GC) with mass spectrometry (MS) and flame ionization detector (FID) in order to control their quality but also to identify some active compounds. As shown in Table 2, more than 85% of each EO composition was characterized. Monoterpenes such as citronellal (43.10%), isopulegol (11.10%), limonene (8.72%), pulegol (6.52%), linalool (5.60%), and citronellol (5.17%) were major components of L. cubeba EO while EOs of the three Zingiberaceae species contained mainly sesquiterpenes. Interestingly, α-zingiberene, β-bisabolene, β-sesquiphellandrene, and ar-curcumene were identified in both EOs extracted from C. longa and Z. officinale with relative percentages respectively of 25.38, 3.38, 18.27, and 5.22% (C. longa) and 27.71, 7.27, 8.08, and 2.71% (Z. officinale). The difference in the composition of these two EOs is that oxygenated sesquiterpenes (i.e., α-turmerone (10.28%), germacrone (3.34%), curlone (5.15%), and ar-turmerone (9.93%)) were found in C. longa EO while monoterpenes including β-phellandrene (14.78%) and camphene (6.94%) were found in Z. officinale EO. 8,9-Dehydro-9-formyl cycloisolongifolene (29.31%), curdione (13.52%), and germacrone (8.95%) were shown as major components of the EO extracted from another curcuma species, C. zedoaria. ( Cyclosativene  The C. longa EO was chosen for further investigations because it was easy to obtain in a high amount as being present in the marc after turmeric starch extraction. The first fractionation using column chromatography with silica gel and gradients of eluents (n-hexane-ethyl acetate) allowed to obtain two important groups, CF1 with sesquiterpenes and CF5 with oxygenated sesquiterpenes. After the second column chromatography using silver nitrate impregnated silica gel of both fractions, three compounds, β-sesquiphellandrene, ar-curcumene, and curlone with a purity respectively of 96.9, 97.4, and 91.7%, were purified.
These isolated compounds along with two commercially available ones, α-zingiberene and ar-turmerone (chemical structures in Figure 2), were analyzed for anti-trypanosomal activity and cytotoxicity. The results are summarized in Table 3. Curlone with an IC 50 of 1.38 µg/mL (6.32 µM) against Tbb bloodstream form and SI > 10 compared to cytotoxicity on mammalian cells could explain a part of the observed activity of the EO (IC 50 = 3.17 nL/mL). This compound may be considered as a promising model for the development of a new treatment of HAT. After the second column chromatography using silver nitrate impregnated silica gel of both fractions, three compounds, β-sesquiphellandrene, ar-curcumene, and curlone with a purity respectively of 96.9, 97.4, and 91.7%, were purified. These isolated compounds along with two commercially available ones, α-zingiberene and ar-turmerone (chemical structures in Figure 2), were analyzed for anti-trypanosomal activity and cytotoxicity. The results are summarized in Table 3. Curlone with an IC50 of 1.38 µ g/mL (6.32 µ M) against Tbb bloodstream form and SI > 10 compared to cytotoxicity on mammalian cells could explain a part of the observed activity of the EO (IC50 = 3.17 nL/mL). This compound may be considered as a promising model for the development of a new treatment of HAT.

Discussion
This is the first time that these thirty-seven EOs extracted from Vietnamese plants were described for anti-trypanosomal activity in vitro. Some of these EOs were already reported to be tested on the same model, such as D. ambrosioides, M. alternifolia, and O. gratissimum [17][18][19], but they were extracted from plants collected in other countries with possibly different compositions. Based on a preliminary screening, half of them showed a potential activity on Tbb with less than 3% of viable parasites at 25 nL/mL (or 10 nL/mL for four of them). Within these nineteen EOs, the one extracted from C. cassia revealed the strongest effect (IC 50 = 1.77 ± 0.15 nL/mL), 17 EOs showed moderate activity with IC 50 values between 2-20 nL/mL, and one EO extracted from P. indica showed less interesting activity with an IC 50 value of 21.29 ± 1.38 nL/mL. In order to identify the most selective EOs, the cytotoxicity on two different mammalian cell lines, WI38 and J774, was evaluated in parallel. Four samples extracted from C. longa, C. zedoaria, L. cubeba, and Z. officinale displayed SI from 14 to > 19 (WI38) and 11 to > 19 (J774). GC analyses led to the identification of more than 85% of their components. We observed the predominance of sesquiterpenes in EOs extracted from rhizomes of the three Zingiberaceae species (C. longa, C. zedoaria, and Z. officinale) while monoterpenes were the major compounds in the EO extracted from fruits of a Lauraceae species (L. cubeba).
The C. zedoaria EO composition including 8,9-dehydro-9-formyl cycloisolongifolene (29.31%), curdione (13.52%), and germacrone (8.95%) differed from both other EOs but also from previous articles. These publications reported the presence in higher amounts of either monoterpenes such as eucalyptol, p-cymene, α-phellandrene, and camphor or other sesquiterpenes such as curzerene, epicurzerene, curdione, curzerenone, and germacrone, depending on the analyzed samples [31]. This variability can be related to the different geographic origins but also to different chemotypes. Germacrone, one of the major compounds identified in our EO, did not show any effect on Tbb bloodstream form in the study of Petrelli et al. (IC 50 > 100 µg/mL) [32], meaning that other components should be responsible for the observed activity of the Vietnamese C. zedoaria EO (IC 50 = 2.51 ± 1.08 nL/mL). Knowing that this EO activity is moderately selective compared to the cytotoxicity on two mammalian cell lines, and that the plant is not difficult to cultivate, C. zedoaria EO and its components should be further studied.
Concerning the mechanism of action of curlone, which is not commercially available, there are no data in the literature. Three other related sesquiterpenoids, α-zingiberene, β-sesquiphellandrene, and ar-turmerone, showed apoptotic effects on different human cancer cells which was associated with the release of mitochondrial cytochrome c and the activation of capase-3 at concentrations of 120 and 160 µg/mL (α-zingiberene), 5.10 µg/mL (β-sesquiphellandrene), and 40, 80, and 120 µg/mL (ar-turmerone) [44][45][46]. The structure of curlone shows the presence of both an exocyclic ethylene group as in β-sesquiphellandrene and a ketone group as in ar-turmerone. This suggests that these two groups may play an important role in anti-trypanosomal activity of the compound.
As mentioned before, it is very important to emphasize the composition complexity of EOs, making it difficult to identify active compounds [11]. Indeed, in these mixtures, most components were found in low percentages, while two or three accounted for 20%-70% of the whole oil [47]. Major compounds can be studied easier and are often considered as responsible for EO biological activities. However other constituents could of course contribute to the activity through a higher efficacy, additional and/or synergistic activity [48]. For example, eucalyptol, which was characterized as the major component (accounts from 27% to 55%) in four of our tested EOs: A. aromaticum, E. blanda, H. coronarium, and L. cubeba (leaves) (data not shown), was not effective on Tbb as shown by its high IC 50 value (83.15 µg/mL) [49]. However these EOs showed a moderate effect in this study with IC 50 values in the range of 8-16 nL/mL. On the contrary, antagonistic effects may also occur between constituents of these mixtures [48]. An interesting example is terpinolene, which accounts for 55% of our C. indica EO (data not shown). This compound was shown to have a very strong activity against Tbb bloodstream form (IC 50 = 0.035 ± 0.05 µg/mL or 0.041 nL/mL) [50], however the IC 50 value of our C. indica EO was 330 times higher (13.22 ± 4.54 nL/mL). The difference in the experiment design could also be another explanation.
Column chromatography (CC) was performed using silica gel 60 (70-230 mesh) (Merck KGaA, Darmstadt, Germany). Thin-layer chromatography (TLC) analysis was done on a sheet pre-coated with silica gel (Merck KGaA, Darmstadt, Germany) and impregnated with AgNO 3 as described by Sliwowski [51]. The GC-MS analyses were carried out on a TRACE GC 2000 series (Thermo-Quest, Rodano, Italy) and the GC-FID analyses were done on a FOCUS GC (Thermo Finnigan, Milan, Italy).

PLANTS Collection and Essential Oils Extraction
The thirty-seven EOs used in this study were extracted from Vietnamese plants as described previously [12]. All EOs were dissolved in dimethyl sulfoxide (DMSO) to obtain stock solutions at 20 µL/mL and then further diluted in fresh medium for anti-trypanosomal and cytotoxicity assays.

Parasites, Cells, and Media
Trypanosoma brucei brucei, although not infecting humans because of its susceptibility to the innate immune system, has been used as a good predictive model in a first screening for the identification of anti-trypanosomal compounds [52,53]. Bloodstream forms of this parasite (Tbb, strain 427) were grown in HMI-9 medium (IMDM-Gibco-Thermo Fisher Scientific, Merelbeke, Belgium-supplemented with 10% heat-inactivated fetal bovine serum (Sigma-Aldrich, Bornem, Belgium) and bloodstream form supporting factors) [54].
Parasites and cells cultures were maintained at 37 • C in 5% CO 2 incubator.

Anti-Trypanosomal Assay
The assay was performed in 96-well plates as previously described [28]. The primary screening was repeated two times in triplicate at the concentrations of 50 nL/mL and 25 nL/mL (for 33 EOs) or 10 nL/mL and 5 nL/mL (for 4 EOs). Fifty µL of parasite culture (5 x 10 4 parasites/mL) was added with 50 µL of diluted EOs in each well. Ten µL of alamar blue (diluted with PBS at the ratio 1:1) was added to each well after 72 h of incubation and the plates were further incubated for 4 h. The fluorescence of the reduced reagent was measured on a spectrophotometer at 530 nm excitation and 590 nm emission wavelengths. Suramin was used as positive control. The EOs that inhibited more than 50% of the parasite growth at 25 or 10 nL/mL were analyzed for IC 50 determination. Samples were tested in eight serial three-fold dilutions ranging from 50-0.02 nL/mL, except Cinnamomum cassia EO, Curcuma zedoaria EO, and Zingiber zerumbet EO (ranging from 10-0.005 nL/mL) and Dysphania ambrosioides EO (ranging from 5-0.002 nL/mL) in duplicate. IC 50 values were calculated from dose response growth inhibition curves using Microsoft Excel files and mean IC 50 values were obtained from at least three repetitions.

Cytotoxicity Assay
The cytotoxicity assays were performed as described previously [12] with concentrations ranging from 50 to 1.40 nL/mL (dilution of 1.67). The selectivity index (SI) values were calculated using the formula: SI = IC 50 on mammalian cells/IC 50 on protozoan parasites

Essential Oils Analysis
Four EOs were analyzed as explained in our previous publication [12].

Components Isolation
C. longa EO obtained by hydro-distillation was subjected to column chromatography on silica gel 60 (70-230 mesh) using n-hexane/ethyl acetate (EtOAc) gradients as the eluent to yield six fractions (CF1-CF6). CF1 and CF5 were further separated to obtain three compounds by column chromatography using AgNO 3 -impregnated silica gel as stationary phase because argentation chromatography is known for the purification of cis-trans-isomers or positional isomers mixtures [55][56][57][58]. This separation relies on the weak interactions between silver ions and the π-orbital of olefins in which cis-olefinic structures complex more tightly with silver ions than the trans-isomers [55]. We modified the procedure of Denyer et al. [58] with 10% (w/w) of silver nitrate instead of 25%, and a gradient of n-hexane and toluene was preferred to hexane and benzene ( Figure 3). These isolated compounds were confirmed as β-sesquiphellandrene, ar-curcumene, and curlone using NMR in comparison with data from previous reports [59][60][61] and their purity was checked by GC-FID. olefins in which cis-olefinic structures complex more tightly with silver ions than the trans-isomers [55]. We modified the procedure of Denyer et al. [58] with 10% (w/w) of silver nitrate instead of 25%, and a gradient of n-hexane and toluene was preferred to hexane and benzene ( Figure 3). These isolated compounds were confirmed as β-sesquiphellandrene, ar-curcumene, and curlone using NMR in comparison with data from previous reports [59][60][61] and their purity was checked by GC-FID.

Conclusions
Our results highlighted for the first time the interesting anti-trypanosomal activity of four EOs extracted from Vietnamese plants, C. longa, C. zedoaria, L. cubeba, and Z. officinale. Monoterpenes were major components of L. cubeba EOs, while the three other EOs contained mostly sesquiterpenes. Among the five major compounds of the C. longa EO, curlone was the most active and selective compound. This compound can explain a part of the observed activity of the EO. Its activity should be further investigated in the research for new anti-trypanosomal compounds.

Conclusions
Our results highlighted for the first time the interesting anti-trypanosomal activity of four EOs extracted from Vietnamese plants, C. longa, C. zedoaria, L. cubeba, and Z. officinale. Monoterpenes were major components of L. cubeba EOs, while the three other EOs contained mostly sesquiterpenes. Among the five major compounds of the C. longa EO, curlone was the most active and selective compound. This compound can explain a part of the observed activity of the EO. Its activity should be further investigated in the research for new anti-trypanosomal compounds.