Hybrid Gold(I) NHC-Artemether Complexes to Target Falciparum Malaria Parasites

The emergence of Plasmodium falciparum parasites, responsible for malaria disease, resistant to antiplasmodial drugs including the artemisinins, represents a major threat to public health. Therefore, the development of new antimalarial drugs or combinations is urgently required. In this context, several hybrid molecules combining a dihydroartemisinin derivative and gold(I) N-heterocyclic carbene (NHC) complexes have been synthesized based on the different modes of action of the two compounds. The antiplasmodial activity of these molecules was assessed in vitro as well as their cytotoxicity against mammalian cells. All the hybrid molecules tested showed efficacy against P. falciparum, in a nanomolar range for the most active, associated with a low cytotoxicity. However, cross-resistance between artemisinin and these hybrid molecules was evidenced. These results underline a fear about the risk of cross-resistance between artemisinins and new antimalarial drugs based on an endoperoxide part. This study thus raises concerns about the use of such molecules in future therapeutic malaria policies.


Introduction
Plasmodium falciparum, the protozoan parasite causing malaria, was responsible for 228 million cases with 405,000 deaths in 2018, mainly in Sub-Saharan Africa where about 90% of the deaths affect children under five [1]. The current malaria treatments recommended by the WHO are artemisinin-based combination therapies (ACTs) combining an artemisinin derivative with one or two other antimalarial drugs with the particularity of having different modes of action and different pharmacokinetic properties. The use of these drug combinations has contributed to a significant decrease in malaria mortality in all endemic regions these last 20 years. However, since 2008, resistance of P. falciparum to artemisinins and partner drugs has widely spread in South-East Asia, resulting in a loss of efficacy of many antimalarial drugs to treat patients [2][3][4]. This situation is a major threat to global public health [5] and imposes a need to accelerate the discovery of new compounds able to eliminate resistant parasites. Among the strategies followed to improve the efficacy of artemisinin both directly against the parasite or for pharmacokinetic and pharmacodynamic enhancements, the synthesis of artemisinin dimers is proposed [6][7][8]. One of the aims was to amplify oxidative stress through increased production of reactive oxygen species (ROS) to kill the parasites. However, although some of these compounds show antiplasmodial activities in the nanomolar range, they have not been evaluated in an artemisinin resistance framework needing particular assays. Indeed, because artemisinin resistance is based on a quiescence mechanism corresponding to a cell cycle arrest of a small number of parasites during artemisinins treatment and resumption of their growth when the drug is eliminated [9][10][11], specific compounds have to be designed. Artemisinin resistance is also associated with mutations in the propeller domain of the gene pfk13 linked to a complex combination of different biochemical pathways [10,12,13]. It is to note that, at the quiescent state, the parasite's metabolism is greatly downregulated. However, apicoplast and mitochondrion seem still active in quiescent parasites [14]. Targeting these two organelles appears thus as a very promising avenue to kill quiescent artemisinin-resistant parasites. In this context, we have recently synthetized and evaluated hybrid gold(I) N-heterocyclic carbene (NHC) complexes based on triclosan targeting mitochondrion and apicoplast, respectively [12]. These novel hybrids showed a strong antiplasmodial activity, however a cross-resistance trend with artemisinins was noted [12]. Interestingly, we showed that there is no cross-resistance between a gold(I) complex and artemisinins [12]. This is in accordance with former studies showing that atovaquone, a mitochondrial electron transfer inhibitor, was efficient not only on proliferating parasites but also on artemisinin-pretreated and dormant parasites [15,16].
Here, we focused our investigations on the second organelle involved in the quiescence mechanism: the parasite mitochondrion. In addition to its role as an energy supplier, mitochondrion contains a wide variety of additional processes notably the two major antioxidant defence systems: thioredoxin (Trx) and glutathione systems, Trx being essential for the erythrocytic P. falciparum cycle [17].
Our chemical research is mainly focused on the design of bioactive gold(I) NHC complexes for parasitic diseases, P. falciparum [18][19][20] and Leishmania infantum [21,22], and anticancer applications [23][24][25]. Moreover, in the field of anticancer metal-based agents, the ubiquitous selenoenzyme thioredoxin reductase (TrxR), responsible for cell homeostasis regulation, is considered as one of the most relevant targets for gold(I) complexes and inhibition of TrxR could lead to apoptosis though a mitochondrial pathway [24][25][26]. Considering that artemisinins-and dihydroartemisinin, the active metabolite of all artemisinin derivatives-are active against all erythrocytic stages of P. falciparum and are still the best antiplasmodial drugs on the field, we designed hybrid complexes combining an ether derivative of dihydroartemisinin (called here DHA) with a gold(I) cation, covalently integrated in our NHC ligand systems. This approach aims to firstly eliminate most of the parasite stages thanks to the highly active DHA part, thereafter, the remaining resistant quiescent parasites will be treated by the gold(I) moiety. The goal of this work is to determine the activity of these hybrid molecules against P. falciparum, then to evaluate the efficacy of the most active hybrids in a context of artemisinin resistance.

Antiplasmodial Activity and Selectivity
The proligands (Ln-R) and the complexes (Aubis(n-R) and Au(n-R)Cl) as well as reference molecules, artemisinin, artemether, auranofin and the proligand precursor DHA-C3, were screened in vitro against the P. falciparum strain F32-Tanzania and the cytotoxicity on Vero cells was evaluated to determine the selectivity of the most active compounds (Table 1). Globally, the results are extremely interesting, with IC 50 values against Plasmodium ranging from 9 to 935 nM for the proligands and the gold(I) complexes. The DHA-ether derivative used for the synthesis of the proligands containing a C3 lateral chain, DHA-C3, has an IC 50 value of 8.5 nM comparable to that of the antimalarial reference drug artemether (IC 50 = 6.1 nM). Surprisingly, for the imidazolium salts obtained by addition of a substituted imidazole moiety to DHAC3-C5 precursors, a significant loss of the antimalarial activity was observed (IC 50 = 98-935 nM, entries 5-9, 18-21 and 26-29). In contrast, the presence of the gold(I) cation in the corresponding complexes greatly improved the antiplasmodial efficacy (IC 50 = 9-104 nM, entries 10-17, 22-25 and 30-33) compared to the corresponding proligands. The antiplasmodial activity was notable, with IC 50 values below 100 nM for two proligands in the C5 series (L5-Bn (11) and L5-Mes (12) and for the majority of the cationic and neutral gold(I) complexes. By comparison, auranofin, a gold-based reference molecule used for the treatment of rheumatoid arthritis, was not active against P. falciparum parasites with a higher IC 50 value of 1.5 µM.

In Vitro Cross-Resistance between Hybrid Molecules and Artemisinin
Three hybrid molecules with the best selectivity indexes in each series, i.e., Aubis(3-Me) (14), Aubis(4-Me) (19) and Aubis(5-Me) (23), were evaluated in vitro for their efficacy in a context of resistance to artemisinins. For that, the comparison of the recovery capacity between the strain F32-ART, artemisinin-resistant, and its twin artemisinin-sensitive F32-TEM was evaluated after 48 h-treatment with the molecule to be tested via the recrudescence assay [9,15]. When ring-stage resistant parasites are exposed to artemisinin, most of them die, but a small sub-set of parasites is able to escape the treatment by a cell cycle arrest, called quiescence mechanism. This phenomenon is characterized by a halt of DNA synthesis, which explains the very low IC 50 values obtained with all artemisinins even for the artemisinin-resistant strain F32-ART [9,15,30]. Therefore, the standard in vitro chemosensitivity assay, based on the measurement of DNA levels to estimate the inhibition of the parasite proliferation, is irrelevant to study the resistance to artemisinins and evaluate possible cross-resistances.
The validation of the test was here done thanks to the results obtained after 18 µM artemisinin treatment and demonstrating that F32-ART is able to recrudesce faster than F32-TEM with a difference to reach the initial parasitemia between the two strains of 9.7 days ( Table 2). For the three hybrid molecules tested, namely Aubis(3-Me) (14), Aubis(4-Me) (19) and Aubis(5-Me) (23), Table 2 showed a difference of recrudescence between F32-ART and F32-TEM. Whatever the hybrid tested, the differences of recrudescence between both strains rise at the higher doses confirming that increase the doses of the molecules to be tested allows a better discrimination of the recrudescence capacity between the artemisinin-resistant and the -sensitive strains [9]. According to the obtained results, a cross-resistance between artemisinin and the hybrid molecules Aubis(3-Me) (14), Aubis(4-Me) (19) and Aubis(5-Me) (23) was noted. This cross-resistance could be explained by the DHA part of the hybrid, responsible for the quiescence entrance of the parasites and the lack of activity of the NHC part at the mitochondrial level due to limited access or pharmacodynamic properties. These data are in accordance with previously obtained results which highlighted the risks of parasites cross-resistance between artemisinins and endoperoxide-based compounds [31,32]. Synchronized ring-stage parasites have undergone 48 h of drug treatment. After that, cultures were washed and parasitemia was monitored during 30 days or until reaching the initial parasitemia, defined as the recrudescence day. If no parasites were observed at the end of the experiment, the culture was classified as showing no recrudescence, and the recrudescence day was noted as >30.

General Information
All complexation reactions were performed under an inert atmosphere of dry nitrogen by using standard vacuum line and Schlenk tube techniques.

Evaluation of Antiplasmodial Activity
In vitro antiplasmodial activities of the synthesized hybrid molecules were evaluated against the P. falciparum strain F32-TEM (corresponding to the parental strain F32-Tanzania). Each molecule was tested at least in three independent experiments (except Aubis(5-Bn), tested twice) by a SYBR Green Fluorescence assay [33] but also at least once with the standard chemosensitivity assay recommended by the WHO and based on the incorporation of [ 3 H] hypoxanthine [33].
For the SYBR Green Fluorescence assay, ring stage parasites were treated with the drugs for 48 h. Then, the pellets were washed twice with PBS and frozen at −20 • C. Thawed plates were incubated for 2 h at room temperature with the SYBR Green (Thermo-Fisher, Illkirch, France) lysis buffer (20 mm Tris base pH 7.5, 5 mm EDTA, 0.008% w/v saponin, 0.08% w/v Triton X-100). The plates were then read using a fluorescence plate reader (FLx800, BioTek, Illkirch, France) at an excitation wavelength of 485 nm and 535 nm for the emission wavelength. The IC 50 values (concentration inhibiting 50% of the parasite's growth) were determined using GraphPad Prism software 7 (GraphPad Software, San Diego, CA, USA).
For [ 3 H] hypoxanthine assay, ring stage parasites were incubated with the drug dilutions for 24 h, then, [ 3 H] hypoxanthine (50 µL/0.25 µCi; Perkin-Elmer, Courtaboeuf, France) was added for another 24 h period. After that, plates were frozen at −20 • C. The next step is to thaw the plates and to collect the nucleic acids. Tritium incorporation was then determined thanks to a β-counter (Perkin-Elmer, Courtaboeuf, France).
The antiplasmodial activities reported in Table 1 correspond thus to the mean of IC 50 values from at least 4 independent experiments acquired by the two methods, SYBR Green and radioactivity.

Evaluation of the Cytotoxicity
The most active compounds were tested for their cytotoxicity using the Vero cell line (monkey epithelial cell line), according to a previously described method [35] with some modifications. The cells were cultured in MEM medium (Dutscher, Brumath, France) supplemented with 10% fetal bovine serum (Dutscher), 0.7 mM glutamine and 100 µg/mL gentamicin.
The cells were seeded at 10 4 cells per well in a 96-well plates and incubated during 24 h. Then, cells were incubated for additional 48 h with the drugs. Cell proliferation was measured with MTT (1-(4,5-dimethylthiazol-2-yl)-3,5-diphenylformazan, Sigma, Saint-Quentin Fallavier, France) which is added to each well for 1 h at 37 • C and 5% CO 2 . Subsequently, DMSO was added to the wells containing the cells and MTT to dissolve the formed crystals. The plates were then read to determine the absorbance at wavelength of 540 nm (µQuant, BioTek, Illkirch, France). Cell proliferation was calculated from at least three independent experiments using GraphPad Prism software 7 (GraphPad Software, San Diego, CA, USA). The cytotoxic/antiplasmodial activity ratio corresponds to the selectivity index.

Recrudescence Assay
P. falciparum strains F32-ART and F32-TEM, synchronized at ring stage parasites at 3% parasitemia and 2% hematocrit have undergone a 48 h-treatment with the drugs to be tested. The parasites were then washed with RPMI-1640 medium and re-cultivated in drug-free culture conditions with 10% human serum. The parasitemia was then monitored daily to determine the time required for each parasite culture to recover the initial parasitemia of 3%. If no parasites were seen up to 30 days, the parasites were considered as "not recrudescent" [9,15]. The drug doses were chosen in order to discriminate the phenotype response between the artemisinin-resistant and the -sensitive strains. A range from 1-fold to >100-fold the antiplasmodial IC 50 value previously found has been tested to determine the most appropriate drug dose to use in this recrudescence assay.

Conclusions
Three original families of gold(I) NHCs complexes incorporating a covalently attached DHA derivative were synthetized and fully characterized. All the proligands and complexes were tested for their antiplasmodial potency on the P. falciparum strains, F32-TEM artemisinin-sensitive and F32-ART artemisinin-resistant. Among the 29 compounds tested, ten gold(I) complexes have shown high antiplasmodial activities, with IC 50 values less than 50 nM and very low cytotoxicity, with selectivity indexes up to 294. However, even though a simple gold(I) bis(NHC) had shown in a previous work [20] no cross-resistance with artemisinins, the presence of this metal could not prevent cross-resistance in the hybrid gold(I)-DHA complexes Aubis(3-Me), Aubis(4-Me) and Aubis(5-Me) tested. These data confirm that the presence of an endoperoxide moiety whatever the structure and the other parts of the hybrid molecules tested leads to artemisinins cross-resistance. Therefore, these findings raise concerns about the potential development of new artemisinin derivatives and more broadly endoperoxide-based antiplasmodial drugs even if they include a moiety metabolically active on biochemical pathways involved in the quiescence state. Funding: This study was supported in part by the French "Agence Nationale de la Recherche" (ANR grant INMAR ANR16 CE35 0003) and the "Centre National de la Recherche Scientifique".