Aspidosperma (Apocynaceae) plant cytotoxicity and activity towards malaria parasites. Part II: experimental studies with Aspidosperma ramiflorum in vivo and in vitro

Several species of Aspidosperma plants are used to treat diseases in the tropics, including Aspidosperma ramiflorum, which acts against leishmaniasis, an activity that is experimentally confirmed. The species, known as guatambu-yellow, yellow peroba, coffee-peroba andmatiambu, grows in the Atlantic Forest of Brazil in the South to the Southeast regions. Through a guided biofractionation of A. ramiflorum extracts, the plant activity against Plasmodium falciparum was evaluated in vitro for toxicity towards human hepatoma G2 cells, normal monkey kidney cells and nonimmortalised human monocytes isolated from peripheral blood. Six of the seven extracts tested were active at low doses (half-maximal drug inhibitory concentration < 3.8 µg/mL); the aqueous extract was inactive. Overall, the plant extracts and the purified compounds displayed low toxicity in vitro. A nonsoluble extract fraction and one purified alkaloid isositsirikine (compound 5) displayed high selectivity indexes (SI) (= 56 and 113, respectively), whereas compounds 2 and 3 were toxic (SI < 10). The structure, activity and low toxicity of isositsirikine in vitro are described here for the first time in A. ramiflorum, but only the neutral and precipitate plant fractions were tested for activity, which caused up to 53% parasitaemia inhibition of Plasmodium berghei in mice with blood-induced malaria. This plant species is likely to be useful in the further development of an antimalarial drug, but its pharmacological evaluation is still required.

Malaria, one of the most prevalent parasitic diseases in the world, still causes a high morbidity and is responsible for approximately 600,000 deaths yearly worldwide mainly due to Plasmodium falciparum (WHO 2014). This increasing global importance is a result of the spread of drug-resistant parasites, the current limitations of vector control and the absence of an effective vaccine. Treatment remains the main strategy for malaria control and new drugs are required (Ridley 2002, de Ridder et al. 2008, WHO 2014. Popular medicine remains an important source for malaria treatment using plant remedies in the endemic areas of African sub-Saharan countries (Willcox & Bodeker 2004, Bourdy et al. 2007, Adebayo & Krettli 2011) and in Latin America (Oliveira et al. 2009(Oliveira et al. , 2015. The use of plants in the search for new antimalarials is also influenced by the fact that two important antimalarials, quinine and artemis-inin, originated from the Cinchona species native to the Peruvian Amazon (White 2008) and the Artemisia annua native to China (Cui & Su 2009), respectively.
The activities of several plant species used against fever and malaria in Brazil have been tested against malaria parasites after fractionation and chemical characterisation (Krettli et al. 2001, Frausin et al. 2015. Plants of the Aspidosperma genus tested in vitro and in vivo were active, like the compounds from the barks of Artemisia vargasii, Artemisia ulei and Artemisia desmanthum (de Andrade-Neto et al. 2007, Rocha e Silva et al. 2012, Torres et al. 2013, as well as Artemisia nitidum, which is used in the Amazon against fever and malaria (Coutinho et al. 2013), and Artemisia olivaceum, a plant from South Brazil (Aguiar et al. 2012).
The extracts and compounds purified from Aspidosperma ramiflorum (Apocynaceae), known as guatambu-yellow and yellow peroba, matiambu, matambu, guatambu-grande, peroba-café and pequiá, are now being studied. The species is native in the Atlantic river forests of Brazil from the South Region [states of Santa Catarina and Paraná (PR)] to the Southeast Region [states of Rio de Janeiro and Minas Gerais (MG)]. Several plant fractions are active against Leishmania amazonensis parasites. The compounds responsible for the activity are the monoterpenoid indole alkaloids (Cunha et al. 2012), which are present in large amounts in the plant and were again isolated ( Fig. 1) and studied for activity against P. falciparum in vitro. Some extracts were tested against malaria in mice with Plasmodium berghei. Extraction of plant materials -The collected A. ramiflorum materials (stem bark 2 kg and leaves 0.340 g) were dried in a circulating air oven, ground and subjected to extraction by maceration with methanol for seven days. The organic solvent was evaporated at 40°C under rotation and reduced pressure. The crude extract was lyophilised and subjected to acid base fractionation, as previously described (Marques et al. 1996, Cunha et al. 2012. After a complete partition of the plant stem bark methanolic crude extract (138 g), four alkaloid rich fractions were obtained: the acid, the neutral precipitate (NP), the neutral and the basic fractions, which were all concentrated under reduced pressure at 40ºC, lyophilised and provided for biological testing at the René Rachou Research Centre, Fiocruz (Belo Horizonte, MG).

Ethics-
A crude extract of the plant leaves (20.4 g) was subjected to a simplified acid-base partition employing the same solvents (Tanaka et al. 2007). Two alkaloid rich fractions were obtained (the acid fraction and the basic fraction) and were concentrated under reduced pressure at 40ºC for solvent evaporation and then lyophilised (Cunha et al. 2012) (Fig. 2).
The alkaloid extracts from the stem barks and the leaves were analysed by thin-layer chromatography (TLC) on a silica gel 60 GF 254 and developed with CHCl 3 / MeOH 92:8 in an NH 3 atmosphere with the p-anisaldehyde reagent followed by heating to 105ºC for 2-4 min.
Electrospray ionisation mass spectrometric analyses (ESI-MS) -Samples from the extracts of the A. ramiflorum barks and leaves were submitted to off-line ESI-MS analysis of alkaloids. The lyophilised extracts were dissolved in methanol:water (1:1) and 100 mL of trifluoroacetic acid (1.0 mg/mL), filtered through a 0.2 mm size pore membrane and introduced into the mass spectrometer using a syringe pump. The spectra were obtained in the positive ionisation mode after using a triple quadruple Micromass ® Liquid Chromatograph Mass Spectrometer model Quattro Micro TM API setting the capillary voltage at 2,300 V, the cone voltage at 60 V and the source at 100ºC. Each spectrum was produced by accumulation data over 1 min.

Isolation of alkaloids from the stem bark extracts -
The neutral (0.69 g) and basic (0.36 g) fractions from the methanolic crude extract were subjected to a silica gel column chromatography, eluted with CHCl 3 and followed by CHCI 3 with increasing amounts of methanol. Further purification by preparative TLC was performed using a silica gel 60 GF 254 as the stationary phase and CHCl 3 :CH 2 Cl 2 :ethyl acetate:MeOH 4.5:4:1:0.5 (v/v) as 98.8, was previously described in the stem barks of A. ramiflorum and its structure was established by uni and bidimensional nuclear magnetic resonance (NMR). MS and comparison with the literature values confirmed compound 3 as 10-methoxygeissoschizol (Marques et al. 1996, Tanaka et al. 2007).
Isolation of alkaloids from the leave extracts -The basic alkaloid fraction (0.35 g) was placed on a silica gel column, eluted with CHCl 3 followed by CHCI 3 with increasing amounts of MeOH and was further purified using a preparative TLC silica gel 60 GF 254 as the stationary phase and then eluted with CHCl 3 :CH 2 Cl 2 :ethyl acetate:MeOH 4.5:4:1:0.5 (v/v) in an NH 4 OH atmosphere (40 mg of extract/per plate). The steps permitted the isolation of compound 2, an indole alkaloid.
Compound 2 [α] 25 D + 58.5º (c = 1.0 mg/mL, EtOH) was identified by ESI-MS at m/z 467.3 [M + H] with daughter ions at m/z 281.3 and m/z 174.0 as described before in stem barks (Marques et al. 1996). Its structure was established by uni and bidimensional NMR and MS data that confirmed its identity (2) as ramiflorine B by comparison with previous data (Tanaka et al. 2007).
Tests with plant materials against P. falciparum blood parasites in vitro -The activity of the A. ramiflorum extracts and fractions was evaluated against P. falciparum blood parasites [clone W2, chloroquine (CQ)-resistant], which were cultured as previously de-   (Lambros & Vanderberg 1979) were immediately incubated with the test compounds at various concentrations that were previously solubilised in 0.05% dimethyl sulfoxide (DMSO) (v/v). Each test was performed in triplicate and the results were compared with the control cultures in complete medium with no drugs. CQ was used in each experiment as an antimalarial control. The compounds' effects were measured through the [ 3 H]-hypoxanthine incorporation assay (Desjardins et al. 1979) and by the immunoenzymatic test, using specific monoclonal antibodies to a parasite protein histidine and alanine-rich protein (HRPII) that were commercially acquired (MPFM ICLLAB-55A ® and MPFG55P ICLLAB ® , USA) (Noedl et al. 2002).
For the [ 3 H]-hypoxanthine assay, the parasites were maintained at least four days in medium without hypoxanthine and adjusted to 1% parasitaemia and 1% haematocrit. The levels of [ 3 H]-hypoxanthine incorporated into the parasites were measured using a beta counter (PerkinElmer, EUA). For the anti-HRPII test, parasitaemia was adjusted to 0.05% and the haematocrit to 1.5%; binding of the HRPII antibodies was quantified at 450 nm using a spectrophotometer (SpectraMax340PC 384 , Molecular Devices).
The half-maximal drug inhibitory concentration (IC 50 ) was estimated by curve fitting using software from the OriginLab Corporation (USA) and comparing to the parasite growth in the drug-free medium.
Cytotoxicity tests using immortalised or primary cells -The cytotoxicity of the plant extracts and fractions was evaluated in a human hepatoma cell line (HepG2) and a monkey kidney cell line (BGM) using cells cultured in 75cm 2 sterile flasks containing RPMI-1640 medium (supplemented with 10% heat-inactivated foetal calf serum and 40 mg/L gentamicin) under a 5% CO 2 atmosphere at 37ºC. When confluent, the cell monolayer was washed with culture medium, trypsinised, distributed in a flat-bottomed 96-well plate (5 × 10 3 cells/well) and incubated for 18 h at 37ºC for cell adherence (Denizot & Lang 1986). The compounds (20 µL), at various concentrations (1,000-1 µg/mL), were placed in the 96-well plates, incubated with the cultured cells for 24 h under a 5% CO 2 atmosphere at 37ºC and then the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution (5 mg/mL; 20 µL/ well for 3 h) was used to evaluate the mitochondrial viability. The supernatants were carefully removed and 100 µL DMSO was added to each well and mixed to solubilise the formazan crystals. The optical density was determined at 570 nm and 630 nm (background) (Spec-traMax340PC 384 ). The cell viability was expressed as the percentage of the control absorbance in the untreated cells after subtracting the appropriate background.
Cytotoxicity was also tested using normal peripheral blood mononuclear cells (PBMC) isolated from healthy individuals by Ficoll-Histopaque (Sigma-Aldrich) gradient centrifugation (Panda & Ravindran 2013). The cells were washed twice at 100 g in RPMI-1640 (Sigma-Aldrich) and resuspended in 2 mL of RPMI supplemented with 10% heat-inactivated foetal bovine serum.
The viable PBMCs cells were counted using the Trypan blue exclusion test, plated in 96-well cell culture plates at a final concentration of 3 × 10 6 cells/mL and incubated with the compounds (20 µL) at various concentrations (≤ 1,000 µg/mL) for 24 h at 37ºC. The MTT solution (5 mg/mL; 20 µL/ well) was added to evaluate the cell mitochondrial viability as described above. The minimum lethal dose for 50% (MLD 50 ) of the cells was determined as previously described (Madureira et al. 2002). The selectivity index (SI) was calculated as the ratio between the activity and the cytotoxicity.
Antimalarial tests against P. berghei in mice -The drug suppressive test was performed as previously described (Peters 1965) with some modifications (Carvalho et al. 1991). Adult Swiss outbred mice (20 ± 2 g weigh) were inoculated by the intraperitoneal route with 1 × 10 5 red blood cells infected with P. berghei, the NK65 strain. Groups of 20-30 infected mice were kept together for 2-24 h, then were divided randomly in groups of up to six animals per cage and were treated by gavage for three consecutive days with 25 mg/Kg or 50 mg/Kg of the test compounds diluted in DMSO 3% (v/v). Two control groups were used and either received CQ (20 mg/kg, diluted in water) or the drug vehicle. Smears prepared from the tail blood of the mice on days 5-10 post-infection were methanol-fixed, stained with Giemsa and examined microscopically. Parasitaemia was evaluated and the percent inhibition of the parasite growth was calculated in relation to the untreated control group, which was considered 100% growth.

RESULTS
The results of the in vitro activity of the extracts from the stem barks and the leaves of A. ramiflorum against the blood cultures of P. falciparum (W2 clone, CQ-resistant) are shown in Table II. The compounds with an IC 50 below 10 µg/mL were considered active. All but the aqueous extracts were active, particularly the acid, the NP and the basic bark crude extracts, which showed IC 50 values between 0.5-3.8 µg/mL. Both the hypoxanthine and the anti-HRPII assays provided similar results. The methanolic and acetone extracts from the leaves showed IC 50 values of 1.4 and 4.0 µg/mL, respectively. The pure substances (5, 3 and 2) isolated from the bark fractions (neutral and basic) were active. Compound 5 had the lowest IC 50 (0.3 µg/mL). All of these extracts were rich in indole alkaloids.
The plant showed no toxicity towards the HepG2 and BGM cell lines and showed high MLD 50 values for most of the extracts and fractions. The acid and nonsoluble extracts were toxic to the normal cells (BGM) ( Table II). When evaluated against freshly isolated human PBMC, the MLD 50 values were lower for most of the extracts and fractions tested, except for the acid, the aqueous and the acetone extracts, which had similar MLD 50 values in relation to the immortalised cell line. The SI of compound 5 was the highest (SI = 100), but substances 3 and 2 had SI values below 10, which were indicative of toxicity; the other purified compounds had SI values above 22 (Table II).
Two plant fractions were tested in vivo: the neutral and the precipitate fractions from the stem barks. The NP fraction reduced the P. berghei parasitaemia to 53% on day 10 in relation to the nontreated control mice. However, both of the extracts increased the survival of the mice (Table I). The other extracts and purified compounds were not tested in vivo due to the insufficient amounts available.
The mass peaks observed in the chemical characterisation of the crude extract from the barks of A. ramiflorum used in the ESI-MS analysis are shown in Fig.  3. A new molecule was identified, isositsirikine (5) at m/z 355 [M + H]+ (Table III). The other isolated compounds were ramiflorine A (1) and B (2) at m/z 467 [M + H]+, 10-methoxy-geissoschizol (3) at 327 [M + H]+ and β-yohimbine (4) and these compounds correspond to compounds previously described in the species (Marques et al. 1996, Tanaka et al. 2007). All of the fractions from the stem bark had similar mass spectra profiles.

DISCUSSION
The chemical analysis of the A. ramiflorum crude extract showed the mass spectra profiles of two compounds that were similar to what had been described before in the plant leaves (Marques et al. 1996). Their mass peaks corresponded to the lead alkaloid compounds ramiflorine A (1) and ramiflorine B (2) at m/z 467 [M + H]+. Using preparative TLC, the bioactive fractions from the stem bark (neutral and basic) and the leaves (basic) were chromatographed on a silica gel column to generate purified fractions and two previously described compounds in A. ramiflorum were identified, the indole alkaloids ramiflorine B (2) and 10-methoxy-geissoschizol (3), whereas the compound isositsirikine (5) was reported here for the first time in the species.
The structures of 2, 3 and 5 were established in this study using uni and bidimensional NMR and the values of the MS analysis in comparison with previous literature data ( Kan et al. 1981, Marques et al. 1996, Tanaka et al. 2007. The mass spectrum of 5 gave a molecular ion at m/z 354 (M+, EI/MS 70 eV) as well as peaks at m/z 323, 251, 249 and 169, which are typical of the corynanthine skeleton. The IR and UV spectra showed absorption bands characteristic of a 10-methoxyindole chromophore (Marques et al. 1996). The 1 H and 13 C NMR spectra showed downfield signals at δH 7.52 (1H, t, J= 7.5 Hz), 7.12 (1H, d, J = 7.5 Hz), 7.12 (1H, t, J = 7.5 Hz) and 7.48 (1H, d, J = 7.4 Hz), which correlate with the respective carbons at δ C 117.9 (C-9), 121.6 (C-10), 119.5 (C-11) and 111.3 (C-12) (heteronuclear single quantum coherence). Additional signals were detected from four quaternary carbons, one with downfield signals at δ C 175.4 (COOCH 3 ) and the other four at δ C 133.2 (C-2), 107.7 (C-7), 127.6 (C-8) and 137.7 C-20). 1 H NMR spectrum also showed the presence of an ethylidene side chain methyl group (C-19) at δ H 1.67 (3H, d, J = 6.0 Hz) and an olefinic proton (C-18) at δ H 5.64 (1H, q, J = 6.0 Hz) and at δ H 4.30 (1H, sl) that correlated with δ C 49.7. The H-3 α-configuration was confirmed by comparison with the chemical shift of the protons of similar compounds described in the literature (Kan et al. 1981). The 1 H and 13 C NMR data of alkaloid 5 were similar to those reported for a compound  known as isositsirikine (Kan et al. 1981). We suggest that the presence of the indole alkaloid skeleton (3 and 5) explains the decrease in the cytotoxicity of the compounds and the consequent increase in the SI values; indeed, the dimeric ramiflorine B (2) was more cytotoxic than 3 and 5.
The results of the values of the MLD 50 in the PBMC nonimmortalised cells showed a slightly higher cytotoxicity in comparison with the immortalised cells for some samples, which can be attributed to the high multiplication rates of the other cell lines.
It was only possible to test the in vivo antimalarial activity of the neutral and precipitate extracts from the stem barks in mice and both extracts increased survival of the mice, but whether this was due to the presence of the alkaloids in A. ramiflorum having anti-inflammatory and antioxidant activities (Barbosa-Filho et al. 2006) needs to be clarified. It is also necessary to study the in vivo activity of isositsirikine (compound 5), the most active compound against P. falciparum in vitro, which was demonstrated for the first time in the present study.
The pharmacokinetic profile of the active extracts and the alternative routes of drug administration were not determined and should be further evaluated. Other biological processes, such as drug metabolism and/or drug excretion may act on the compounds' activities in the body and thus, decline their amounts in an exponential manner (Alavijeh et al. 2005) thereby altering the in vivo activity of the plant products.