We first assessed the in vitro activity of amiodarone and dronedarone and their respective metabolites desethylamiodarone (DEA) and N-desbutyldronedarone (NDBD) against T. cruzi Tulahuen WT (DTU VI) amastigotes in mouse 3T3 fibroblasts (Table 1A). Mean IC₅₀ values of 49 µM and 9 µM were found for amiodarone and dronedarone respectively. Interestingly, both metabolites are nearly equipotent compared to the parent compounds when run in parallel in the same assay (mean IC₅₀ of 28 µM for DEA and 10 µM for NDBD) although both metabolites showed poor selectivity to host cells. However, both drugs, and in particular amiodarone, had moderate to no selectivity index (SI) over the 3T3 cells used in the assay (SI values varying between 0.75 and 1.5 for amiodarone and 3.2 and 7.8 for dronedarone (Table 1A), thus diminishing the therapeutic potential of these drugs.

To get a broader idea of the potential of these two drugs, in vitro data were also generated in amastigote intracellular inhibition assays using T. cruzi strains representative of various Discrete Typing Units (DTU): JR and Sylvio X10 strains both belonging to DTU I, Y strain (DTU II) and CL Brener strain (DTU VI). In all cases benznidazole was assessed in parallel as a positive control (IC₅₀ ranging from 2-10 µM). Overall, the averaged IC₅₀ values (µM; mean +/- SD) for amiodarone and dronedarone were 11.2 +/- 15.6 and 4.1 +/- 2.8 respectively, while the IC₉₀s (µM; mean +/- SD) were 17.2 +/- 15.2 and 6.4 +/- 4.4, respectively (Tables 1A, B). Importantly, in each assay, all available calculated SI values highlighted the relatively low selectivity over the mammalian cell lines used (SI ranging from 0.75 to 7.2 for amiodarone and 0.6 to 7.8 for dronedarone). Selectivity should always be considered in the interpretation of in vitro activity of compounds to identify the negative effect of compounds on mammalian host cells. The consistency of this finding of minimal selectivity, irrespective of the parasite strain, specific host cell used or methodology, highlights the fundamental issues with these drugs in treating Chagas disease.

The in vitro activity of amiodarone and dronedarone was also determined against epimastigotes, the insect form of the T. cruzi parasite, and were found to be within the same µM range demonstrated for amastigotes (see Supplementary Table S1). Interestingly, both drugs were more potent than benznidazole against the epimastigote stage of the parasite.

To assess in vivo efficacy, we first determined the impact of amiodarone and dronedarone on acute stage murine infections with T. cruzi CL Brener (DTU VI) and JR (DTU I) in BALB/c mice using bioluminescence imaging. BALB/c mice at the peak of the acute stage (day 14 post-infection) were treated with 5 daily oral doses of 75 mg kg^-1 amiodarone or with 100 mg kg^-1 dronedarone (same batches as those used in the in vitro studies to eliminate potential effects of batch variation). Neither of the treatment schedules used had any significant effect on the bioluminescence-inferred parasite burden in acute stage disease irrespective of the T. cruzi strain used (Figure 1).

In contrast, parallel treatment of infected mice with the front-line drug benznidazole (100 mg kg^-1) reduced the parasite burden by >99.9%, close to the limit of detection. To further explore the potential anti-T. cruzi activity of amiodarone and dronedarone, we then treated mice during the chronic stage of infection with the CL Brener strain (104 days post-infection) when immune pressure had reduced the parasite burden by more than 2 orders of magnitude (Figure 2).

We did not observe any significant reduction in parasitemia at the end of a 5-day treatment regimen in the chronic BALB/c infection model with the T. cruzi CL strain. Therefore, highly sensitive bioluminescence imaging demonstrates that neither amiodarone or dronedarone has significant activity against acute or chronic stage T. cruzi infections at the doses used. In contrast, treatment with benznidazole as a positive control reduced the parasite burden close to the limit of detection as expected. This benznidazole treatment schedule is curative when applied to chronic stage infections (34). Given the modest in vitro activity, the limited selectivity over mammalian cells (Table 1), the poor drug exposure (see Figure 3, Table 2 below), and the lack of in vivo activity against the CL Brener and JR strains (Figures 1, 2), further animal experimentation was not undertaken.

In parallel to the in vivo efficacy studies, the exposure of amiodarone and dronedarone was assessed in non-infected female BALB/c mice following oral administration at the same doses (75 and 100 mg kg^-1, respectively) and using the same formulation as for the in vivo efficacy studies. Plasma protein binding in mouse and human plasma and binding to the in vitro testing medium were also measured to determine the respective unbound concentrations. Plasma concentration versus time profiles is shown in Figure 3 and pharmacokinetic and binding properties are tabulated in Table 2.

Amiodarone was detected in plasma for the duration of the 48 h sampling period. The apparent half-life was long (15 h), however this is an approximate value only as the terminal phase was not clearly defined by the available data. Maximum plasma concentrations were observed at 0.5 h indicative of rapid absorption. At 24 h (at which time the next dose was administered in the efficacy study), the concentration of amiodarone was ~0.48 µM. The primary metabolite, DEA, was also detected in plasma for the duration of the sampling period, however plasma exposure was low (C_max < 0.1 µM, Figure 3, Table 2).

Dronedarone plasma concentrations were detected up to 24 h post dose, but concentrations were at the LLOQ (~0.001 µM) at the 24 h time point. The half-life was approximately 2 h and maximum plasma concentrations were observed at 2 h post dose. Plasma concentrations of the primary metabolite, NDBD, were comparable to concentrations of the parent compound and were below the LLOQ at the 24 h time point (Figure 3, Table 2).

Plasma protein binding studies indicated very high binding in both species, with >99.97% bound for amiodarone and >99.7% bound for dronedarone. These results are consistent with data previously reported for amiodarone binding in human plasma (35). To assess the likely concentration present over the full 5-day dosing period, amiodarone and dronedarone single dose concentration versus time profiles (Figure 3) were fitted to a 2-compartment model with first order absorption and elimination (Figure S1). Repeat dose profiles were then simulated with the assumption that there were no changes to the PK properties following repeat administration. Figure 4 shows the simulated profiles for the total (solid lines) and unbound (dashed lines) plasma concentrations for both compounds along with the unbound in vitro IC₅₀ value for reference (red dotted lines).

Amiodarone HCL, its metabolite DEA and dronedarone HCL, used for in vitro and in vivo studies, were purchased from AK Scientific, Union City, USA. Dronedarone metabolite, aka N-desbutyldronedarone, was purchased from BDG Synthesis, Wellington, New Zealand. Compounds were dissolved in dimethylsulfoxide (DMSO) to 10 mM stock solutions. Amiodarone HCL and dronedarone HCL were formulated as a 2.5 to 10 mg ml^-1 suspension (depending on dose) in a vehicle containing 0.5% (w/v) hydroxypropyl methylcellulose (HPMC) and 0.4% (v/v) Tween 80 in Milli-Q H₂O. Benznidazole, was synthesized by Epichem Pty Ltd., Bentley, Australia.

Direct comparison of the in vitro potency of amiodarone, dronedarone and their respective metabolites DEA and NDBD (Table 1A) against T. cruzi Tulahuen WT (DTU VI) amastigotes in mouse 3T3 fibroblasts, was performed as described earlier (46).

In vitro potency of amiodarone and dronedarone against different T. cruzi intracellular amastigotes and selectivity towards various host cells (Table 1B) was determined using several different assays (different host cells, multiplicity of infection, parasite strain, drug incubation time and endpoint used e.g. HCS), and performed in different laboratories. Detailed information of the methods used has been published previously (47–50). When presented as averaged IC₅₀ and IC₉₀ values, data from amastigote intracellular inhibition assays was calculated as the mean value obtained from 3-5 different assays. Determination of the in vitro potency of amiodarone and dronedarone against epimastigotes and amastigotes forms of the parasites used in the in vivo studies was performed as described previously (50, 51).

Animal infections were performed under UK Home Office licence PPL P9AEE04E4 and approved by the London School of Hygiene and Tropical Medicine Animal Welfare and Ethical Review Board (AWERB). All protocols and procedures were conducted in accordance with the UK Animals (Scientific Procedures) Act 1986. Female BALB/c mice were obtained from Charles River (UK) and CB17 SCID mice were bred in-house. BALB/c mice aged 7–8 weeks were maintained in groups of 3, 4 or 5, where they experienced a 12-hour light/dark cycle and had access to food and water ad libitum. At time points indicated, mice were infected with bioluminescent T. cruzi CL Brener or JR strains engineered to express a red-shifted luciferase that facilitates highly sensitive non-invasive monitoring of the parasite burden in real-time (52, 53). Typically, 1×10⁴ bloodstream trypomastigotes (BTs) in 0.2 ml PBS were used to infect SCID mice via intraperitoneal (i.p.) inoculation. Parasitaemic blood from these mice was obtained 17 (CL Brener) and 20 (JR) days later and adjusted to 5×10³ BTs ml⁻¹ with Dulbecco’s Ca²⁺/Mg²⁺ free PBS (D-PBS). BALB/c mice were then infected i.p. with 1×10³ BTs. Amiodarone and dronedarone were prepared for administration in an aqueous suspension vehicle containing 0.5% (w/v) hydroxypropyl methylcellulose and 0.4% (v/v) Tween 80. The dosing regimens were selected based on previously published in vivo work (25, 38). Drugs were administered by oral gavage (adjusted by weight), and vehicle only was administered to control mice.

Mice were injected i.p. with 150 mg kg⁻¹ d-luciferin in D-PBS, then anaesthetized using 2.5% (v/v) isofluorane in oxygen for 2-3 minutes. Images were obtained using an IVIS Spectrum system (Caliper Life Sciences) 5-10 minutes after d-luciferin administration. Exposure times varied from 10 seconds to 5 minutes, depending on signal intensity. To estimate parasite burden, whole body regions of interest were drawn using Living Image^® 4.5.5 to quantify bioluminescence expressed as total flux (photons/second; p/s). The detection threshold was established from uninfected mice.

Pharmacokinetic studies in non-infected mice were conducted at Monash University using established procedures in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, and the study protocols were reviewed and approved by the Monash Institute of Pharmaceutical Sciences Animal Ethics Committee. Female BALB/c mice (17-21 g) were obtained from the Monash Animal Research Platform and mice had access to food and water ad libitum throughout the pre- and post-dosing periods.

Amiodarone and dronedarone were formulated for oral administration as suspensions in 0.5% (w/v) hydroxypropyl methylcellulose, 0.5% benzyl alcohol, and 0.4% (v/v) polysorbate 80 in water. Compounds were dosed orally by gavage (0.2 mL per mouse). Blood samples were collected up to 24-48 h (n = 3 mice per time point) with a maximum of three samples from each mouse. Samples were collected via submandibular bleed (approximately 120 μL). Blood was collected into polypropylene Eppendorf tubes containing heparin as anticoagulant and centrifuged immediately for the separation of plasma, which was stored at -80°C until analysis by LC-MS.

Plasma sample and standard analysis was conducted using LC/MS/MS following protein precipitation with acetonitrile (4- or 5-fold volume ratio). Blank mouse plasma was used to prepare calibration standards over the concentration range of 2.5-10,000 ng mL^-1 (amiodarone) or 0.5-5000 ng mL^-1 (dronedarone) and diazepam was added as an internal standard. Following protein precipitation, samples were centrifuged and the supernatant collected for analysis. Samples and standards (1 µL) were injected onto a Supelco AscentisExpress RP C8 (amiodarone, desethylamiodarone) or RP amide (dronedarone, N-desbutyldronedarone) column (50 x 2.1 mm, 2.7 µm) and eluted using a mobile phase of 0.05% formic acid in water and 0.5% formic acid in acetonitrile mixed using a gradient from 5 to 95% (amiodarone, desethylamiodarone), or 10 to 80% (dronedarone, N-desbutyldronedarone) acetonitrile, over a 4 min cycle time with a flow rate of 0.4 mL min^-1. The LCMS system included a Waters Xevo TQD MS coupled to a Waters Acquity UPLC. Sample detection was conducted using positive electrospray ionization in multiple reaction monitoring mode. The MS/MS transitions were 646.03 > 100.07 (cone voltage of 60 V, CID 30V) for amiodarone, 618.07 > 72.05 (cone voltage of 55 V, CID 25V) for desethylamiodarone, 557.36 > 100.04 (cone voltage of 75 V, CID 40 V) for dronedarone, 501.27 > 114.09 (cone voltage of 60 V, CID 35 V) for N-desbutyldronedarone and 285.12 > 154.09 (cone voltage of 50 V, CID of 25 V) for diazepam. Assays were validated for linearity, matrix effects, accuracy, precision and recovery.

Plasma concentration versus time data were analyzed using standard non-compartmental methods and were also fitted with a two compartment (amiodarone) or one compartment (dronedarone) model with first order absorption and elimination using PK Solver (ver 2.0). A weighting factor of 1/concentration² provided the best fit to the experimental data. Using the derived PK parameters, repeat dose profiles were simulated assuming no change to the PK properties upon repeat dosing (i.e. minimal enzyme inhibition or induction).

Protein binding was determined in diluted plasma via rapid equilibrium dialysis with incorporation of a pre-saturation step based on methods reported previously (35). Prior to the dialysis experiment, RED inserts (Thermo Fisher Scientific Inc.) were presaturated by adding solutions containing each compound (at 500 nM in isotonic phosphate buffered saline (PBS), pH 7.4) to both the donor and the dialysate chambers and incubating at 37°C on an orbital plate shaker (450 rpm; ThermoMixer C, Eppendorf) for 2 x 30 min cycles and an overnight cycle. At the end of each presaturation cycle, solutions from both donor and dialysate chambers were removed and discarded. Following the presaturation, diluted human and mouse plasma (10% v/v in pH 7.4 PBS) were spiked with each compound (at 5000 nM) and dialyzed against pH 7.4 PBS containing a low concentration of compound (100 nM) for 24 h. Concentrations of each compound in dialysate and donor samples collected at the end of the dialysis period were determined via LC-MS as described above. Compound binding was assessed on the basis of the measured concentrations in dialysate and donor samples at the end of the dialysis period assuming that the system was at steady state by 24 h. As the binding assay was performed using diluted plasma, data were corrected for the dilution factor to give a binding value for neat matrix via an established approach which accounts for the shift in equilibria that occurs with protein dilution (54).