In Vivo Efficacy and Metabolism of the Antimalarial Cycleanine and Improved In Vitro Antiplasmodial Activity of Semisynthetic Analogues

Bisbenzylisoquinoline (BBIQ) alkaloids are a diverse group of natural products that demonstrate a range of biological activities. In this study, the in vitro antiplasmodial activity of three BBIQ alkaloids (cycleanine [compound 1], isochondodendrine [compound 2], and 2′-norcocsuline [compound 3]) isolated from the Triclisia subcordata Oliv. medicinal plant traditionally used for the treatment of malaria in Nigeria are studied alongside two semisynthetic analogues (compounds 4 and 5) of cycleanine.

research (4,5). For example, artemisinin isolated from Artemisia annua, sweet wormwood, a traditional Chinese medicine, together with a series of its semisynthetic derivatives, has become the first-line therapy for P. falciparum malaria (6,7). However, due to the development of artemisinin drug resistance (8), novel therapies are still urgently needed.
Bisbenzylisoquinoline (BBIQ) alkaloids are a diverse group of natural products consisting of two benzylisoquinoline groups (9). BBIQ alkaloids are primarily found in the Berberidaceae, Lauraceae, Menispermaceae, and Ranunculaceae plant families. These alkaloids possess a variety of biological activities, which include antimalarial activities (9,10). For example, BBIQ alkaloids isolated and identified from Triclisia species of the Menispermaceae family have antiproliferative activities (10). In Nigeria, the root of Triclisia subcordata Oliv. is traditionally used for the treatment of a range of diseases, including malaria (11,12). The bioactive components of T. subcordata are the BBIQ alkaloids cycleanine (compound 1), isochondodendrine (compound 2) and 29-norcocsuline (compound 3) (Fig. 1) and have previously been isolated and characterized by our group (13,14). We have also produced synthetic analogues of cycleanine (compounds 4 and 5) ( Fig. 1) (15). The three naturally occurring BBIQ alkaloids, cycleanine (16)(17)(18), isochondodendrine (18,19), and 29-norcocsuline (16,20) have been reported to possess antiplasmodial effects against chloroquine-sensitive and chloroquine-resistant P. falciparum strains. Despite the promising in vitro biological activity of these natural BBIQ alkaloids, the in vivo antimalarial activity of BBIQ alkaloids has not been evaluated, nor has their potential in vivo metabolism. Here, we assess the in vivo antimalarial activity and metabolism of cycleanine (compound 1). The effect of cycleanine analogues (compounds 4 and 5) on antiplasmodial potency and selectivity is also investigated.
This study sets out an evaluation of the in vitro antimalarial activities of BBIQ alkaloid compounds 1 to 3 compared to those of two semisynthetic BBIQ alkaloids (compounds 4 and 5) derived by a modification of cycleanine at the C-5 position by introducing additional secondary or tertiary amine moieties in an attempt to increase potential solubility and potency (15). The most abundant BBIQ alkaloid in T. subcordata extract is cycleanine; this was therefore used to establish in vivo antimalarial activity in a murine malaria model. In addition, the metabolites and metabolic pathways of cycleanine were analyzed after intragastric administration in rats to help understand how cycleanine is eliminated in vivo to guide future optimization of cycleanine for antimalarial development.

RESULTS
The semisynthetic derivatives of cycleanine have improved in vitro antiplasmodial activity and selectivity. The in vitro antiplasmodial activities of the five BBIQ alkaloids (compounds 1 to 5), as well as that of a chloroquine control, were determined against intraerythrocytic stages of the P. falciparum Dd2 chloroquine-resistant strain using a malaria SYBR green I fluorescence assay. These data are provided in Table 1 (see Fig. S1 in the supplemental material) as 50% inhibitory concentration (IC 50 ) values (mean 6 standard deviation [SD] for n = 3 independent biological repeats). While the data for chloroquine in Dd2 are comparable to those of the chloroquine-resistant strain W2, the activities of cycleanine, isochondodendrine, and 29-norcocsuline are significantly lower in Dd2 than that reported in W2, and certainly lower than that in the chloroquine-sensitive strain D6. The semisynthetic products 4 and 5 are relatively more potent than compound 1 to 3 in Dd2, with the most potent, compound 4, being some 25.2-fold more potent than its natural precursor, cycleanine (compound 1).
Data from cytotoxicity studies of BBIQ alkaloids 1 to 3 in human oral epidermoid carcinoma (KB) or HCT-116 human colon carcinoma cells suggest low to moderate selectivity, with a selectivity index (SI) of 14 to .133. Fifty percent cytotoxic concentration (CC 50 ) data for all five compounds are available from human ovarian epithelial (HOE) cells (Table 1). These data reinforce the findings of low selectivity, albeit improved in semisynthetic products 4 and 5.
In vivo antimalarial activity of cycleanine (compound 1). The isolation of the abundant cycleanine (compound 1) in T. subcordata root enabled us to investigate its toxicity in healthy mice and efficacy in murine malaria models after infection with Plasmodium berghei. The acute median lethal dose (LD 50 ) of cycleanine after 24-h oral administration was determined to be 4.5 g/kg in mice, indicating a good safety profile. The malaria-suppressive activity of cycleanine using two oral doses (25 and 50 mg/kg of body weight/day) following P. berghei infection was demonstrated through a significant suppression of parasitemia and an increased mean survival time (MST) compared to those of untreated controls (Table 2). In particular, the higher dose (50 mg/kg/day) showed efficacy, both in terms of suppression of parasitemia and in MST, comparable to that for chloroquine at a dose of 5 mg/kg/day. The prophylactic activity of cycleanine, with the same 25 and 50 mg/kg dosing regimen during P. berghei infection in mice, was also demonstrated (Table 3). At the higher dose (50 mg/kg), cycleanine showed a suppression of parasitemia by 59.0%, only slightly less than that of 76.2% using the prophylactic pyrimethamine control at a dose of 1.2 mg/kg /day.
The curative activity and MST of mice after initial P. berghei infection and subsequent treatment with cycleanine (compound 1) were determined. After infection of mice for 3 days, cycleanine was administered at both doses of 25 and 50 mg/kg and mice showed decreasing parasitemia in a dose-dependent and time-dependent manner from day 3 to day 7 (Fig. 2). The speed of killing of P. berghei parasites by chloroquine was much faster than that with cycleanine. Chloroquine reached 0% parasitemia after 5 days, while at that time cycleanine at doses of 25 and 50 mg/kg had remaining levels of 13.3 and 10.5%, respectively (Fig. 2). In this curative model, the MST of mice at doses of 25 and 50 mg/kg were 21 and 25 days, respectively,  (16). c IC 50 data against chloroquine-resistant P. falciparum strain W2 and CC 50 values for HCT-116 human colon carcinoma cells were sourced from a previous report (18). d ND, not determined. e CC 50 data for human ovarian epithelial (HOE) cells. Data in this column for compounds 1 to 5 were sourced from our previous reports (13,15). f This selectivity index (SI) was calculated as CC 50 /IC 50 against P. falciparum.
which were significantly longer than that of the control (12 days). However, they were both shorter than that of mice treated with chloroquine (30 days) (see Table  S1 in the supplemental material).
In vivo metabolism of cycleanine. In order to explore the in vivo metabolism of cycleanine, the plasma and urine of Wistar rats following an oral dose of 120 mg/kg body weight/day over a 24-h period were analyzed for cycleanine metabolites. Samples from urine and plasma were prepared and submitted to high-performance liquid chromatography-electrospray ionization-tandem mass spectrometry (HPLC-ESI-MS/MS) analysis. The peak at the retention time of 9.7 min was cycleanine (M0) with the protonated molecular ion m/z of 623.3116 [M1H] 1 (elemental composition C 18 H 43 N 2 O 6 ) in the positive ion mode spectrum (Table 4  Twelve peaks on LC-MS/MS chromatograms relevant to cycleanine were detected in either urine or plasma samples (Table 4 and Fig. S3 in the supplemental material). The original form of cycleanine and 11 metabolites were found from the urine of rats, which were presumed to be hydroxylation (M1 and M2), demethylation and hydroxylation (M3), monodemethylation (M4), didemethylation (M5), dehydrogenation and hydroxylation (M6 and M12), and dehydrogenation and dihydroxylation (M7) metabolites, and the isomeric metabolites of M7 (M8, M9, and M11). From the cycleaninecontaining plasma of rats, the original form cycleanine (M0) and five metabolites were found, which were presumed to be hydroxylation (M2 and M10), dehydrogenation and hydroxylation (M6 and M12), and dehydrogenation and dihydroxylation (M7) metabolites. Among them, the prototype (M0), hydroxylation (M1), and dehydrogenation and hydroxylation (M6, M12) metabolites were detected in both rat urine and plasma (Table 4 and supplemental material). Therefore, the metabolic pathway of cycleanine in rats involves hydroxylation, dehydrogenation, and demethylation or their combination, which are the main means of biotransformation of

DISCUSSION
Natural products (e.g., artemisinin and quinine) have demonstrated their potential as a source of antimalarial drugs. Previously, a number of BBIQ alkaloids were demonstrated to have in vitro antiplasmodial activities (16). Cycleanine had antiplasmodial effects with an IC 50 of 70 nM (16) (or 80 nM [17]) against P. falciparum chloroquine-sensitive clone D6 (or 3D7) and an IC 50 of 4.5 mM against a chloroquine-resistant strain (18). Isochondodendrine showed a low IC 50 of 0.2 mM against a chloroquine-resistant strain (18,19). 29-Norcocsuline also showed potent in vitro antiplasmodial activity, with IC 50 values of 48 and 248 nM against chloroquine-sensitive clone D6 (3D7) and chloroquine-resistant clone W2 (16,20), respectively (Table 1). Our results against P. falciparum chloroquine-resistant strain Dd2 also confirmed the in vitro antimalarial activity of these compounds but with slightly higher IC 50 values (Table 1) compared to the corresponding values reported in literature. Isochododendrine is a structurally demethylated analogue of cycleanine, and showed a greater potency than cycleanine in the chloroquine-resistant W2 strain and the Dd2 strain in this study (Table 1). This indicated that the increase of the hydrophilicity of cycleanine could improve its antiplasmodial activity. The SI values of all three BBIQ alkaloids ranged from 14 to 133 based on the KB or HTC-116 cells and the W2 strain, which were much greater than those based on HOE cells and the Dd2 strain. The discrepancy might be due to the different methodologies (16) used to determine IC 50 or the different mammalian cancer cells or P. falciparum clones used. The semisynthetic analogues of cycleanine (compounds 4 and 5) produced by chemical modification of cycleanine through introduction of dimethylamino and (mono)alkynylamino groups, respectively, at the C-5 position exhibited increase in antiplasmodial potency and SI relative to cycleanine. The presence of a dimethylamino group in compound 4 could also increase the water solubility of the parent compound as is often found in the modification of other natural products, such as camptothecin (21) and thymoquinone (22). Compound 5, with a unique aminoalkynyl group, was used as a chemical probe for exploring the mechanism of action (e.g., cellular uptake) of cycleanine in cancer cells using click chemistry (15), and can also be utilized for identification of the molecular target of cycleanine in parasite-infected blood cells using a chemoproteomic approach (23). By changing the amino substitution groups, additional analogues of cycleanine with a variety of diverse structures can be synthesized for in vitro antiplasmodial evaluation.
To further confirm and validate the efficacy of cycleanine (compound 1) in vivo, its safety in healthy mice and efficacy in a murine malaria model was investigated. The LD 50 (4.5 g/kg) of cycleanine indicated that cycleanine has a good safety profile, in agreement with the LD 50 of 1.1 g/kg found previously in mice (24). Using suppression, prophylactic, and curative murine malaria models after infection with P. berghei (25), cycleanine showed a similar or closer effect at an oral dose of 50 mg/ kg to their positive controls (chloroquine [5 mg/kg] and pyrimethamine [1.2 mg/ kg]). At least, a much higher dose of cycleanine was needed to achieve the effects of these positive controls, indicating a mild efficacy in vivo. However, its low toxicity profile could allow increase of the oral dose (e.g., 100 mg/kg/day), which is expected to improve its efficacy. In the curative model, the slower effect of cycleanine comparing to chloroquine might be due to the metabolism of cycleanine to various metabolites. The in vivo antimalarial activity of cycleanine was consistent with its in vitro antiplasmodial activity. To our knowledge, this is the first demonstration of the in vivo antimalarial efficacy of a BBIQ alkaloid, cycleanine. Overall, three alkaloids (compounds 1 to 3) of T. subcordata could contribute to the antimalarial effects of this medicinal plant used in Nigeria for the treatment of malaria. BBIQ alkaloids of Triclisia gilletii (De Wild) Staner were also reported to be attributed to the in vitro and in vivo antimalarial activity of its plant extract (26). Study on the metabolism of drugs can help to further understand their pharmacokinetics, efficacy, and safety (27). For example, metabolites of piperaquine were shown to have stronger antiplasmodial activity (28). However, there have been only a few in vivo metabolism studies of BBIQ alkaloids. Previously, in vitro metabolites of a BBIQ alkaloid, isoliensinine, from dog hepatic microsomes were identified as 29-Ndesmethylisoliensinine, 2-N-desmethyl-isoliensinine, and 29-N-6-O-didesmethylisoliensinine (29). The study of the pharmacokinetics and metabolism of another BBIQ alkaloid, neferine, indicated that it was partially converted to liensinine, desmethylliensinine, isoliensinine, and desmethyl-isoliensinine by CYP2D6 (30). Tetrandrine was found to be initially biotransformed to a quinone methide-derived metabolite mediated by CYP3A enzymes, which was then trapped by a glutathione molecule to form a glutathione conjugate in mice (31). Metabolism of isotetrandrine in vitro by the rat hepatic system produced a major metabolite, N-desmethyl isotetrandrine (16%), and three minor oxidized metabolites, oxo-isotetrandrine (7%), hydroxy-isotetrandrine (6%), and oxohydroxy-isotetrandrine (7%), via N-demethylation and isoquinoline ring oxidation (32).
Our identification of 12 new metabolites of cycleanine in both plasma and urine in rats using LC-MS/MS has indicated that there are various metabolic pathways of cycleanine. These metabolites of cycleanine found in rats are also likely generated in mice after the same route of oral administration, and therefore they could contribute to its in vivo antimalarial efficacy found in the murine malarial model and its toxicity finding in healthy mice. Hydroxylation and demethylation of cycleanine were the common pathways consistent with those found in isoliensinine, neferine, and isotetrandrine described above. Preparation of these metabolites through chemical synthesis (33) or in vitro biotransformation using hepatic microsomes and/or P450 enzymes (34,35) is possible and necessary to evaluate their potency and toxicity. Such information can be used to further guide the chemical design and modification of cycleanine to improve its potency and pharmacokinetics and to increase metabolic stability (36). Further work is necessary and ongoing in our laboratory to determine the in vivo antimalarial effects of BBIQ alkaloids (compounds 2 and 3) and semisynthetic derivatives (compounds 4 and 5) and the in vitro and/or in vivo antimalarial activity of the metabolites of cycleanine. Novel active drugs, particularly those with a wide safety margin, are required to help alleviate malaria morbidity and mortality and to contribute to the global control of malaria and infectious diseases.
In vitro antiplasmodial activity. The evaluation of in vitro antiplasmodial activity of the alkaloids (compounds 1 to 3) and semisynthetic analogues (compounds 4 and 5) were performed on the intraerythrocytic P. falciparum Dd2 strain (chloroquine-resistant strain) using a SYBR green1 fluorescence dye assay as described (22,37,38). Compounds 1 to 5 were prepared in dimethyl sulfoxide (DMSO) with no greater than 1% of the total solvent concentration in any assay. Normalized fluorescence signals were measured against controls with 1% DMSO (100% growth) and after exposure to a supralethal concentration (10 mM) of chloroquine (0% growth). Determination of the 50% inhibitory concentration (IC 50 ) was performed from a log concentration versus the mean normalized fluorescence signal curve using GraphPad Prism software (v5.0). Each biological replicate consisted of three technical repeats, with three independent biological replicates performed.
Evaluation of the in vivo antimalarial activity of cycleanine. (i) Malaria parasite. Chloroquinesensitive strains of P. berghei were sourced from the National Institute of Medical Research (NIMER), Yaba Lagos, Nigeria, and maintained by subpassage in mice.
(iii) Experimental animals. Female and male Swiss albino mice (18 to 25 g) were obtained from the University of Uyo's animal house. Before use, mice were kept in cages and acclimatized for 10 days. All mice were kept in cross-ventilated rooms at room temperature. The care and use of mice were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (1996). This investigation was approved by the University of Uyo's Animal Ethics Committee.
(iv) Determination of median lethal dose (LD 50 ) of cycleanine. The median lethal dose (LD 50 ) of cycleanine was determined using albino mice by the intraperitoneal (i.p.) route (40). Different doses of cycleanine (10 to 5,000 mg/kg) were intraperitoneally administered to groups of three mice each. The mice were monitored for manifestation of physical signs of toxicity, including decrease of motor activity, writhing, decrease of body/limb tone, weakness, and death. The number of deaths in each group within 24 h was recorded. The LD 50 value was calculated as the geometrical means of the minimum dose producing 100% mortality and the maximum dose producing 0%.
(v) Drug administration. Cycleanine, chloroquine, and pyrimethamine were prepared in water and administered orally with the aid of a stainless metallic feeding cannula.
(vi) Suppressive activity of cycleanine. The schizontocidal activity of cycleanine and chloroquine against early P. berghei infection in mice was measured according to an established protocol (25,41,42). On the first day, 20 to 24 mice were infected with the parasite and randomly separated into four groups. The mice in groups 1 and 2 were given 25 and 50 mg/kg of cycleanine, respectively, those in group 3 were given 5 mg/kg of chloroquine (positive control), and those in group 4 were given distilled water (10 ml/kg, negative control) for four consecutive days. Thin films were made from the tail blood on the fifth day. Parasitized erythrocytes were counted in stained films (Giemsa stain) under a microscope. The average suppression of parasitemia (%) was calculated as follows: average % parasitemia positive control2average % parasitemia negative control average % parasitemia negative control Â100 The MST (in days) of the mice in each group was determined over a period of 30 days.
(vii) Prophylactic activity of cycleanine. The prophylactic activity of cycleanine was evaluated using a previously described method (42,43). The mice were randomly divided into four groups of six mice per group. Groups 1 and 2 were given 25 and 50 mg/kg of cycleanine, respectively, group 3 was given 1.2 mg/kg of pyrimethamine (positive control), and group 4 was given 10 ml/kg of distilled water (negative control). Administration of cycleanine or the drug continued for three consecutive days. On the fourth day, the mice were inoculated with P. berghei. The parasitemia level was evaluated by blood smears after 3 days. The survival times (days) of the mice were recorded over a period of 30 days, and MST were calculated.
(viii) Curative activity of cycleanine. The curative activity of cycleanine was assessed according to a method described previously (42,44). P. berghei was injected intraperitoneally into another 24 mice on the first day. Three days later, the mice were also separated into four groups of six mice per group. Groups 1 and 2 were administered different doses of cycleanine, 25 and 50 mg/kg, respectively, group 3 was given 5 mg/kg chloroquine (positive control), and group 4 was given 10 ml/kg distilled water (negative control). Cycleanine and chloroquine were given once a day for 5 days. Mouse tail blood samples were collected on each day, and Giemsa-stained thin smears were prepared to determine the parasitemia level. The MST of the mice in each group was determined over a period of 30 days.
Metabolism of cycleanine in rats. (i) High-performance liquid chromatography-quadrupole time-of-flight tandem mass spectrometry. Analysis of cycleanine metabolites was performed through high-performance liquid chromatography-quadrupole time-of-flight tandem mass spectrometry (HPLC-QTOF-MS/MS) system that consists of an Agilent 1260 HPLC coupled with an 6530 QTOF mass spectrometer with dual Agilent Jet Stream electrospray ionization source (Agilent Technologies, CA). The mass spectra were recorded in positive auto MS/MS mode, and the parameters were set as follows: temperature of drying and sheath gas, 300°C and 350°C; skimmer, 75 V; capillary voltage, 4,000 V; fragmentor, 110 V; nozzle voltage, 1,000 V; collision energy, 50 eV; pressure of nebulizer, 35 lb/in 2 ; and flow rate of the drying and sheath gas, 5 and 11 liter/min, respectively. The QTOF mass spectra were recorded in high-resolution mode. The range of mass-to-charge ratio (m/z) scanning was set between 100 and 1,200. Samples (5 ml) were loaded onto an Agilent Poroshell 120 EC-C 18 column (100 Â 2.1 mm, 2.7 mm) at 35°C. The mobile phase consisted of water containing 0.1% formic acid (solvent A) and acetonitrile containing 0.1% formic acid (solvent B) at a flow rate of 0.35 ml/min. Gradient separation was achieved by changing the proportion of the solvent B mobile phase as follows: 0 to 2 min, 10% B; 2.1 to 5 min, 18% to 20% B; 30 to 45 min, 70% to 90% B; and 45 to 50 min, 10% B. MassHunter WorkStation software (Agilent Technologies) was utilized for the system operation and data analysis.
(ii) In vivo experiments. In vivo animal experiments were approved by the Animal Ethics Committee of Shanghai Institute of Materia Medica, and performed according to procedures approved by the Institutional Animal Care and Use Committee of Shanghai Institute of Materia Medica, Chinese Academy of Science. Male Wistar rats were obtained from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). The rats were given free access to water and a standard diet under controlled humidity (45% to 55%) and temperature (20°C to 24°C), except in the overnight fasting period before administration of cycleanine. The rats were adapted to the environment for a week.
Cycleanine (compound 1) was suspended in 0.4% carboxymethyl cellulose sodium (CMC-Na) and was formulated at 12 mg/ml for intragastric administration to Wistar rats (male, 220 6 10 g, fasted for 12 h prior to administration) at a dose of 120 mg/kg body weight. Three rats were used for blood collection through the orbital vein using cannulation at 0, 0.5, 1, 2, 4, 6, 8, 12, and 24 h postdose after anaesthetization with isoflurane. The plasma samples were separated from blood by centrifugation at 12,000 rpm and 4°C for 10 min. Another three rats were placed in the metabolism cages, and urine samples were collected into tubes from 0 to 24 h after oral administration of cycleanine. All samples were stored in a 280°C freezer before analysis. Aliquots of 1.2 ml of plasma or urine samples were mixed with 3 times the volume of acetonitrile to precipitate proteins. After centrifugation at 14,000 rpm for 10 min, the supernatant was collected and evaporated under vacuum. The residue was reconstituted in 200 ml methanol, and 5 ml of each sample was injected for HPLC-QTOF-MS/MS analysis.
Statistical analysis. Data were expressed as mean 6 standard error of the mean (SEM) for in vivo antimalarial experiments. Data were subjected to GraphPad Prism software analysis. Results were analyzed using one-way analysis of variance (ANOVA) followed by a post hoc Tukey multiple-comparison test. The difference between the mean of the experimental and control groups was considered significant at a P value of ,0.05 (ANOVA).
Conclusions. Three BBIQ alkaloids of T. subcordata, cycleanine (compound 1), isochondodendrine (compound 2), and 29-norcocsuline (compound 3) and two semisynthetic analogues (compounds 4 and 5) of cycleanine were demonstrated to exert significant in vitro antiplasmodial activities against P. falciparum. Cycleanine (compound 1) was further demonstrated to have safety and efficacy in the treatment of mice infected with P. berghei. Cycleanine was transformed to various metabolites in rats after oral delivery. The findings from this study support the use of T. subcordata as an antimalarial agent in traditional medicine. BBIQ alkaloids could be exploited in novel drug development in search of antimalarial agents/drugs that are urgently needed to challenge resistant plasmodium species that currently present a significant threat to human life.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.8 MB.