Structure-Activity Relationship Study of Sesquiterpene Lactones and Their Semi-Synthetic Amino Derivatives as Potential Antitrypanosomal Products

Sesquiterpene lactones (STLs) are natural products that have potent antitrypanosomal activity in vitro and, in the case of cynaropicrin, also reduce parasitemia in the murine model of trypanosomiasis. To explore their structure-antitrypanosomal activity relationships, a set of 34 natural and semi-synthetic STLs and amino-STLs was tested in vitro against T. b. rhodesiense (which causes East African sleeping sickness) and mammalian cancer cells (rat bone myoblast L6 cells). It was found that the α-methylene-γ-lactone moiety is necessary for both antitrypanosomal effects and cytotoxicity. Antitrypanosomal selectivity is facilitated by 2-(hydroxymethyl)acrylate or 3,4-dihydroxy-2-methylenebutylate side chains, and by the presence of cyclopentenone rings. Semi-synthetic STL amines with morpholino and dimethylamino groups showed improved in vitro activity over the native STLs. The dimethylamino derivative of cynaropicrin was prepared and tested orally in the T. b. rhodesiense acute mouse model, where it showed reduced toxicity over cynaropicrin, but also lost antitrypanosomal activity.


Introduction
Sleeping sickness, or human African trypanosomiasis (HAT), is a deadly protozoal disease caused by Trypanosoma brucei species spread by tsetse flies (Glossina spp.). The two human pathogenic subspecies, T. b. rhodesiense (95% of cases) and T. b. gambiense (5% of cases), differ in terms of their geographic distribution, clinical pictures, and the drugs used to treat parasitemia [1]. Currently, there are about 30,000 new HAT cases annually, and as many as 30 million people live in HAT endemic areas [2]. Despite some recent successes like nifortimox-eflornithine combination therapy (NECT) [3], HAT drugs are still inefficient by modern standards and need to be replaced by compounds that are safer and easier to administer [4].
Natural products from plants have been instrumental in developing drugs to treat protozoal diseases, such as quinine and artemisinin against malaria [5,6]. Currently, though, no natural product-based antitrypanosomal drugs are in use or even in late stage clinical development. Recently, we reported the in vivo activity [7] and mode of action [8] of cynaropicrin (1), a natural sesquiterpene lactone (STL). Although more than 883 plant-derived small molecules have shown antiprotozoal (antitrypanosomal, antiplasmodial, and antileishmanial) effects in vitro, of which 87 were STLs [5,6], this was the first reported plant compound with in vivo anti-T. brucei action specifically. STLs are a chemotaxonomic feature of the largest plant family, the Asteraceae [9], and to date more than 5,000 STLs are known [5,10]. STLs are a promising compound class for antitrypanosomal drug discovery [11][12][13], thus a better understanding of the structural features that contribute to activity is expedient [14,15]. This study explores the antitrypanosomal structure-activity relationships within this compound class using a group of 18 molecules comprising natural STLs and 16 semi-synthetic STL-amines and two other derivatives against T. b. rhodesiense in vitro. Mammalian cancer cells (L6 cell line) were used to evaluate cytotoxicity. The antitrypanosomal effects of the eight compounds 1, 2, 5-9,and 13 have been reported before [7,11,12], but were included for comparison. Additionally, based on the in vivo antitrypanosomal effects of 1 after intraperitoneal application [7,8], 1 and the dimethylamino derivative 19 were tested in vivo in the acute mouse model with oral application. The rationale behind derivative 19 was that masking the α,β-unsaturated enoate in the lactone ring would possibly create a prodrug with increased water solubility, improved pharmacokinetic properties, and reduced unspecific binding to biological thiols via Michael addition to the α-methylene-γ-lactone. Through subsequent bioactivation it would be converted to the parent compounds and, hence, display its biological activity on the target. A similar approach had been previously successfully applied to several STLs with anticancer activity like helenalin, costunolide, and parthenolide [16].
We recently reported in vivo antitrypanosomal effects of 1 [7,8]. Upon intraperitoneal administration, the parasitemia was decreased over several days, but the compound was not able to cure the mice when they were treated with 10 mg/kg/b.i.d. for four consecutive days. In an attempt to improve the bioavailability, the exocyclic double bond in the lactone ring was masked to obtain water soluble dimethylamino derivative 19. Compounds 1 and 19 were orally administered in the acute sleeping sickness mouse model. Four mice, each treated with 50 mg/kg body weight/day of 1, showed reduced parasitemia on day 7 after infection. However, the animals were euthanized on day 10 post infection due to obvious signs of cytotoxicity of the compound. Mice treated with compound 19 exhibited less signs of toxicity. However, the compound showed no in vivo efficacy, since the mean survival time was the same as for the control ( Table 2).

Discussion
This structure-activity relationship (SAR) study showed the STL 2-(hydroxymethyl)acrylates 1, 2, and 4 along with the STL 3,4-dihydroxy-2-methylenebutylate 3 are the most active and selective STLs against T. b. rhodesiense. These compounds have two active α,β-unsaturated enoate moieties in common, one on the lactone ring and the other on the side chain. Compound 10 has no side chain with an enoate moiety, but does possess a cyclopentenone group, which can serve as the additional reactive enone. In fact, vernodalin (4), which has a third reactive α,β-unsaturated enoate group, had slightly higher potency (along with greater toxicity) than 1. STLs lacking a reactive α,β-enoate function in their side chains (12, 13, and 14) or with no side chain (5, 6) showed weaker antitrypanosomal and cytotoxicity effects compared to their analogues (1, 2, and 3). The results are in accordance with findings by others [16][17][18]. Schmidt et al. showed in a SAR study using 40 STLs that bifunctionality as Michael acceptors is very important for a high level of antitrypanosomal activity but that very similar structural features determine both antiprotozoal and cytotoxic activity [14]. This was recently confirmed by the same authors with a larger data set of over 70 STLs [17]. Many bioactivities in STLs have been attributed to a Michael addition of the methylene-γ-lactone moiety to biological thiols [19]. Recent findings on the adduct formation of the two α,β-unsaturated nucleophilic enoate groups at C13 and C3' in 1 with trypanothione and glutathione in trypanosomes corroborated the long presumed mode of action of STLs, in addition to an inhibition of ornithine decarboxylase [8]. The same mode of action can be expected for 2, 3, and 4. Regarding the 14 semisynthetic STL amines tested here, it was observed that the addition of morpholino and dimethylamino groups maintained or even enhanced the in vitro activity compared to their parent structures, whereas tyramino-, 2-(4-chlorophenyl)ethylaminoand 1-(2-chlorophenyl)piperazyl derivatives were generally two to four fold more active than their parent compounds having no enoate moiety, albeit with the disadvantage of higher toxicity (SI in the range 3-7, generally).
Compared to 1, the toxicity and antitrypanosomal activity of 19 in the T. brucei rhodesiense acute mouse model after oral application was poorer. Since the amino derivative 19 -in spite of a reasonably high level of in vitro activity-did not show any in vivo efficacy, it will have to be investigated in further studies whether other types of derivatives could be useful prodrugs of STLs for the p.o. treatment of HAT.

General Experimental Information
NMR spectra were run on a 400 MHz Varian INOVA instrument. Samples were referenced against chloroform at 77.00 ppm for 13 C and against tetramethylsilane at 0.00 ppm for 1 H. High resolution mass spectra were recorded on a Waters SYNAPT G1 HDMS mass spectrometer operated in electrospray mode. Leucine enkephalin (50 pg/mL) was used as reference calibrant to obtain typical mass accuracies between 1 and 3 mDa. Melting points were determined using a Mettler FP62 capillary melting point apparatus and are uncorrected. All reagents were of reagent grade purchased from Sigma-Aldrich (Schnelldorf, Germany) and were used without any further purification. Solvents used for chromatography or extractions were distilled prior to use. Thin-layer chromatography was carried out using pre-coated aluminum-backed plates (Merck Silica Gel 60 F 254 ). Column chromatography was performed on Fluka silica gel 60 (70-230 mesh). Dry solvents were purified as described by Perrin and Armarego [23]. All starting materials were obtained commercially and used without further purification.

Syntheses and Spectral Data of Analogues 15, 21-34
3-Hydroxyisophotosantonin (15). A solution of O-acetylisophotosantonin (1.475 g, 4.813 mmol) in MeOH (49 mL) at 0 °C was treated with sodium borohydride (0.293 g, 7.750 mmol) carefully. The reaction was left at 0 °C for 3 h, then left to warm to room temperature overnight. The mixture was extracted from saturated aqueous NH 4 Cl (50 mL) with ethyl acetate (3 × 50 mL), the extracts pooled and dried (MgSO 4 ). The dried filtrate was concentrated to a tacky white foam, then dissolved in absolute ethanol (16 mL). 5% Aqueous KOH (150 mL) was added and the mixture stirred for 18 h at room temperature. The mixture was acidified to pH < 2 with 18% aqueous HCl, stirred for 30 min, extracted with ethyl acetate (3 × 50 mL) and washed with saturated aqueous K 2 CO 3 . Concentration yielded a yellow solid, which was recrystallized (EtOAc/hexane) to a white amorphous powder (0.313 g, 36%). NMR showed an approximately 2.2:1 mixture of secondary alcohols had been isolated. General Procedure for Preparing Lactone Methylamines: A solution of the appropriate enoate (1 eq.) in ethanol (0.1M) containing the required volatile amine (2.5 eq.) or non-volatile amine (0.6 eq.) and triethylamine (1.1-2.5 eq., for the appropriate hydrochloride salt) was heated at 85 °C under microwave irradiation set at 30 W for 30 min to 1 h, depending on the amine. All were prepared on a sufficiently small scale that the solutions could simply be concentrated and purified by column chromatography. The following compounds were produced this way:

Rat Myoblast Cell L6-Cytotoxicity Assay
The cytotoxicity assay was performed with rat skeletal myoblasts (L6-cells) seeded in 100 μL RPMI 1640 in 96-well micro titre plates, using the Alamar Blue assay described above. After 24 h the medium was removed and replaced by 100 μL of fresh RPMI 1640 with serial threefold drug dilution. Podophyllotoxin (purity > 95%, Sigma-Aldrich) was used as a reference drug. After 70 h of incubation under a humidified 5% CO 2 atmosphere, 10 μL of the Alamar blue marker (see above) was added to all wells. The plates were incubated for an additional 2 h. A Spectramax Gemini XS micro plate fluorescence reader (Molecular Devices) was used to read the plates using an excitation wavelength of 536 nm and an emission wavelength of 588 nm. The IC 50 s were calculated from the sigmoidal growth inhibition curves using Softmax Pro software (Molecular Devices).

Acute Mouse Sleeping Sickness Model
This model mimics the first stage of the human African trypanosomiasis. Adult female NMRI mice were purchased from Janvier (St. Berthevin, France). They weighed between 20 and 25 g at the beginning of the study and were kept under standard conditions at 22 °C and 60%-70% humidity in macrolon type III cages with food pellets and water ad libitum. All protocols and procedures used in this study were reviewed and approved by the local veterinary authorities of the Canton Basel-Stadt, Switzerland (authorization N 739; 11.12.2009). The samples were first dissolved in 100% DMSO, followed by addition of distilled H 2 O to a final DMSO concentration of 10%. For determination of the in vivo antitrypanosomal activity, mice were infected intraperitoneally with 1 × 10 4 STIB900 bloodstream forms. Experimental groups of four mice were treated orally once a day on four consecutive days from day 3 to day 6 post infection. A control group of four mice was infected, but remained untreated. The determination of the parasitaemia was done on day 7 post infection. 6 μL of tail blood were diluted in 24 μL sodium citrate (3.2%), whereby the first μL was discarded to obtain circulating blood. Five μL of this mixture were transferred to a glass slide and covered with an 18 × 18 mm cover slide. The sample was examined under a light microscope (200-fold magnification) and parasites were counted in 3 of the 16 squares of the grid.

Conclusions
The conclusions are that the α-methylene-γ-lactone is necessary for both antitrypanosomal effects and cytotoxicity of these sesquiterpenes. Antitrypanosomal selectivity is facilitated by 2-(hydroxyl-methyl)acrylate or 3,4-dihydroxy-2-methylenebutylate side chains, and by the presence of cyclopentenone rings. Semi-synthetic STL amines with improved activity over the native STLs were those with morpholino and dimethylamino groups. The dimethylamino analogue of cynaropicrin was prepared and tested orally in the T. b. rhodesiense acute mouse model, where it showed reduced toxicity over cynaropicrin, but also reduced antitrypanosomal effects.