Ethics statement
The maintenance and care of mice used in this work were conformed to institutional guidelines (CETS n° 123). Protocols to perform studies on in vivo toxicity in mice and quantification of compounds in mice plasma were approved by the Instituto de Parasitología y Biomedicina “López Neyra” (CSIC) Ethical Committee. In vivo efficacy was monitored at New York University School of Medicine (US). This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (US). The protocol was approved by the Institutional Animal Care and Use Committee of New York University School of Medicine, which is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International [5].
Reagents and chemicals
Reduced nicotinamide adenine dinucleotide phosphate tetrasodium salt (NADPHNa4), testosterone, diclofenac and dextromethorphan (standard Cytochrome P450 probe substrates), cortisone and levallorphan (LC–MS–MS internal standards), ketoconazole, quinidine and sulfaphenazole (Cytochrome P450 control inhibitors) were obtained from Sigma Aldrich (St. Louis, MO, USA). 4’-hydroxydiclofenac 13C6 (LC–MS–MS internal standard) was purchased from Toronto Research Chemical (Toronto). Human liver microsomes (HLM) mixed gender pool of 22 were purchased from BD Gentest Corp (Bedford, MA). Materials related to the chemistry part of the experiments include acetonitrile (gradient grade, Merck KGaA), formic acid (reagent plus, Sigma Aldrich) and milli-Q water 18.2 Mega Ohm (Milli-Q gradient system, Merck Millipore).
In vitro cytotoxicity assay
Cytotoxicity of strasseriolides A–D was determined against immortalized THLE-2 (CRL-2706) liver cells by the MTT assay. The cells were first seeded in 96-well plates at 1 × 104 cells/well in 200 µL MEM culture medium and incubated at 37 °C in 5% CO2 for 24 h. The spent medium was replaced with fresh 200 µL medium containing 1 µL of compounds and controls. The controls were, 8 mM methyl methanesulfonate (MMS, positive control), DMSO 0.5% (negative control) and the standard drug doxorubicin. Compounds were tested in triplicate using twelve-point ½ dose-response curves with maximum concentrations at 50 µM. The cells were further incubated with the compounds for 72 h. After incubation, the MTT solution was prepared at 5 mg/mL in PBS and then diluted to 0.5 mg/mL in medium without phenol red. The spent medium was then replaced with 100 µL of MTT solution and the plates gently shaken and incubated at 37 °C in 5% CO2 for 3 h. The supernatant was then carefully removed without disturbing the cells and 100 µL DMSO added. The plates were gently shaken to solubilize the formed formazan and absorbance measured at 570 nm in a Victor2TM multireader (Perkin Elmer, USA). The percentage inhibition of cell growth was computed by the formula:
$$Percentage \, inhibition=\left[1-\left(\frac{{Abs}_{well}-{Abs}_{neg}}{{Abs}_{pos}-{Abs}_{neg}}\right)\right] \times 100$$
where Abswell is the absorbance measured per specific well, and Abspos and Absneg are the average absorbance measured for the positive and negative controls respectively.
In vitro cardiotoxicity screening
Cardiotoxicity screening of strasseriolides A–D was performed in hERG expressed HEK cells using the FluxOR™ potassium assay. The FluxOR™ potassium assay was performed on a FLIPR TETRA (Molecular Devices) as outlined in the product information sheet from Invitrogen. As directed by the kit, the Powerload™ concentrate and water-soluble probenecid were respectively added in the first step to enhance the dye solubility and retention, after which the FluxOR™ dye was added and mixed. The FluxOR™ loading buffer (165 mM NaCl, 4.5 mM KCl, 2 mM CaCl2, 1 mM MgCl, 10 mM HEPES, 10 mM Glucose) was adjusted to a pH of 7.4. Spent medium from cells previously seeded as described above for the MTT assay, was removed and replaced with 80 µL loading buffer containing the FluxOR™ dye mix. The dye was removed after 60 min incubation at room temperature and the plates subsequently washed once with assay buffer. The compounds were dissolved in DMSO after which 2 µL of compound solution was added to 398 µL assay buffer using a Biomek liquid handling unit (Beckman Coulter). Hundred microlitres (100 µL) of the diluted compounds were added to the cells and the plates incubated for 30 min at room temperature (23–25 °C) to allow equilibration of the compounds. The thallium stimulation buffer (Tl2SO4 + K2SO4) was prepared according to the manufacturer´s instruction and injected into the plates on the FLIPR TETRA, to allow kinetic analysis from time zero (t0) to time 120 s (t120). In each well, the average of values for 2-5 s was used as background and that for 90-92 s was used as maximum channel activity. The ratio of maximum channel activity to background values recorded in each well was used for calculating the percentage of inhibition of each compound with respect to the control wells in which no compounds were added. The compounds were tested in triplicate using twelve-point ½ dose-response curves with maximum concentrations at 50 µM.
Cytochrome P450 inhibition assay (drug–drug interaction)
The Cytochrome P450 inhibition assay was performed using substrates for the CYP3A4, CYP2D6 and CYP2C9 isoforms of the enzyme [6, 7]. The 96-well ABgen-0765 plates used in the assay were first pretreated by being washed with acetonitrile, sonicated for 5 min and rinsed with deionized water, after which they were centrifuged upside down to dryness and kept warmed at 37°C in a water bath before used. As the test compounds were initially dissolved in DMSO, the final DMSO content of each reaction mix was established at 0.35% to minimize its interference with the reactions. Two microlitres (2 µL) of compounds were pre-incubated with 98 µL of cofactor/buffer solution at 37°C for 10 min (resulting final conc. of NADPH/phosphate buffer was 1 mM/100 mM at pH 7.4 when HLM/substrate was added). The reaction was then initiated by the addition of 100 µL of HLM/substrate solution (at final protein concentration of 0.25 mg/mL). For CYP3A4, CYP2D6 and CYP2C9 enzyme isoforms, 50 µM testosterone, 22 µM dextromethorphan and 10 µM diclofenac were respectively used as substrates in the HLM/substrate solution. No inhibitor and inhibitor controls (i.e. ketoconazole for CYP3A4, quinidine for CYP2D6 and sulfaphenazole for CYP2C9) were included in all plates at the compound addition stage. The reactions were incubated for 15, 30 and 45 min with shaking at 180 Hz, for CYP3A4, CYP2D6 and CYP2C9 respectively, after which they were quenched by adding 90 µL acetonitrile containing internal LC–MS–MS standards (0.166 µM cortisone, 0.314 µM 4’-hydroxydiclofenac 13C6 and 0.212 µM levallorphan for CYP3A4, CYP2D6 and CYP2C9, respectively). The plates were gently mixed on a microtiter shaker for 1 min and centrifuged at 4000 rpm for 20 min at 4 °C. Fifty microlitre (50 µL) supernatants were then transferred from the reaction wells to new 96-well round bottom plates and 25 µL of 0.3% formic acid/water added. The new plates were centrifuged at 4000 rpm for 15 min at 4 °C, sealed and placed in a mass spectrometer with an auto-sampler. LC–MS–MS analysis was performed to quantify the relative amounts of 6β-hydroxy-testosterone (CYP3A4), dextrorphan (CYP2D6) or 4’-hydroxy-diclofenac (CYP2C9) products formed from the respective substrates. A triple quadrupole mass spectrometer (AB SCIEX API4000) coupled to an Agilent 1290 HPLC system (Agilent Technologies, USA) was used for the quantitative analysis. Analysis was performed on a Discovery HS C18 (50 mm, 2 mm, 3 μm) column (Supelco, Torrance, CA) at 25 °C with mobile phase solvents A (water/acetonitrile 90/10 (v/v) + 0.1% formic acid), B (90/10 acetonitrile/water + 0.1% formic acid) and flow rate of 1 mL/min. Mobile phase B was initially held for 0.5 min at 10%, then increased linearly to 45% during a period of 1.50 min, after which it was maintained at this percentage for 0.30 min, and increased linearly again to 95% during a period of 0.30 min. It was then held at 95% for 0.50 min, before bringing it down to 10% B in 0.10 min and maintained there for 0.61 min to re-equilibrate the column. In all, the total run time was 3.51 min. The mass spectrometer was operated with an atmospheric pressure chemical ionization (APCI) probe in positive mode using multiple reaction monitoring (MRM) scanning mode. Integration of reaction product and internal standard peak areas was performed using the Analyst® software version 1.5.2. Each compound was tested in triplicate using ten-point ½ dose-response curves with maximum concentrations at 83 µM.
Drug metabolic stability assay in human liver microsomes
Pretreatment of the 96-well ABgen-0765 reaction plates was the same as previously described above. Two microlitres (2 µL) of compounds (previously dissolved in DMSO) were pre-incubated with 98 µL of cofactor/buffer solution (0.1 M phosphate buffer of pH 7.4, 100 mM K3PO4, 3.3 mM MgCl2, 3.3 mM G6P, 1.3 mM NADPH, and 0.4 u/mL G6PDH) at 37 °C for 10 min. The reactions were then initiated by the addition of 100 µL of HLM solution (at final protein concentration of 1 mg/mL), after which they were stopped at pre-determined incubation times of 0, 10, 15, 30, 45 and 60 min by adding 90 µL acetonitrile containing internal LC–MS–MS standards and shaking at 180 Hz. The samples were centrifuged to sediment the protein precipitates and the supernatants were analysed by LC–MS–MS as previously described above. The integration of each analyte (the ionized metabolite of each compound) and internal standard peak area was performed by the Analyst® software version 1.5.2 and used to calculate the percentage of remaining compound at each incubation time. The natural logarithm of the percentage of the remaining parent compound was plotted against the incubation time and used to calculate the half-life by the following equation:
Half-life (T ½ min) = 0.693/k, where k is the slope of a plot of the natural log of the percentage parent compound remaining vs. time.
The intrinsic clearance (CLint) was determined by the equation:
CLint
\(=\frac{Ln 2}{{\text{t}}_{\frac{1}{2}\left(\text{m}\text{i}\text{n}\right)}}* \frac{volume \, incubation \left(mL\right)}{microsomal \, protein \left(mg\right)}*45\left(MPPGL\right)\frac{1500 \, g \, human \, liver}{70 \, kg \, human \, body \, weight}j\)
Units for CLint were expressed as mL/min/mg microsomal protein and MPPGL = mg microsomal protein per gram liver.
In vivo toxicity and quantification of compounds in mice plasma by LC–MS–MS
Three Balb/C mice (6–8 weeks old) were intravenously injected with 25 mg/kg of strasseriolide D dissolved in 200 µL of vehicle (0.5%HPMC, 0.5% Tween 80 and 0.5% Benzyl alcohol in water). Blood samples were collected by pricking the jugular vein on the leg of each mouse at time points 1.5 (T 1.5 h), 5 (T 5 h) and 24 h (T 24 h). From the harvested blood, 50 µL plasma were prepared for all the samples and stored at -80 °C until needed. A previously developed LC–MS–MS method (see supporting information) was used to determine the plasma concentration of strasseriolide D as described below.
Ten millimolar (10 mM) DMSO stocks of strasseriolide D (the test compound) and strasseriolide A (used as internal standard,) were prepared. From the strasseriolide D stock, a 59.7 µM (27,000 ppb) working concentration was made in DMSO and an aliquot of this was serially diluted (½) in reconstitution solution (1:1 of acetonitrile: H2O with 0.1% formic acid) to create a six-point calibration curve in triplicate. Three different quality control (QC) concentrations of strasseriolide D at 59.7, 0.74, and 0.25 µM (i.e. 27,000 ppb, 333 ppb and 111 ppb) were also prepared in duplicate. From the strasseriolide A (internal standard) stock, a 6.2 µM (2700 ppb) working concentration was prepared in DMSO.
Blank mouse plasma (prepared from non-injected mouse) was retrieved from the freezer (− 80 °C) and kept on ice to thaw. The sample was vortexed adequately and 15 µL aliquots transferred into 0.7 mL Eppendorf tubes after which 35 µL of Milli-Q water were added and vortexed. Triplicate sets of the diluted plasma were then respectively spiked with 2 uL of each concentration of the calibration curves of strasseriolide D. Another triplicate set of the diluted plasma was spiked with 2 µL of each concentration of the standard QC solutions. All the above samples were also spiked with 2 µL of strasseriolide A (the internal standard)). To serve as blank samples, a fresh triplicate set of diluted plasma were spiked with 4 µL of DMSO. Non-matrix samples were also prepared by spiking 100 µL aliquots of reconstitution solution with 2 µL of the standard QC solutions and 2 µL of internal standard.
Fifteen microlitres (15 µL) plasma from the various time points of the strasseriolide D-injected mice, were respectively transferred into 0.7 mL Eppendorf tubes. To each sample, 37 µL of Milli-Q water were added, vortexed and spiked with 2 µL of strasseriolide A (internal standard). Two hundred microlitres (200 µL) of ice-cold acetonitrile were added to all of the matrix (plasma-containing) samples and centrifuged at 13,300 rpm for 15 min at 4 °C. Subsequently, 180 µL aliquots were carefully transferred (without disturbing the protein precipitate) from each tube into clean HPLC vials and evaporated to dryness in an EZ2 Genevac at room temperature. The dried samples were then reconstituted in 40 µL of reconstitution solution by vortexing. The samples were placed in an auto-sampler drawer for LC–MS–MS analysis and quantification of strasseriolide D (the test compound) in mice plasma.
Similar sample preparation and LC–MS–MS analysis were respectively performed for mice intravenously injected with strasseriolides A and C as test compounds and mice blood samples collected at 30 min (T 30 M), 2 h (T 2 h), 6 h (T 6 h) and 24 h (T 24 h). In the case of strasseriolide B, the three mice injected with this compound died about eight minutes after their injection, hence, blood samples were immediately collected, and plasma prepared for analysis as previously described. For strasseriolides A, B and C as test compounds; strasseriolides C, D and B were respectively used as internal standards. After LC–MS–MS analysis, the plasma concentration of compound vs. time data was determined for each strasseriolide. This data was fed into the PK solver 2.0 add-in program in an Excel file with “non-compartmental analysis of plasma data after intravenous bolus input” [8], to compute the PK parameters of each tested compound and to plot plasma concentration vs. time curves from which the PK (pharmacokinetic) parameters were computed for the strasseriolides (Table 2; Fig. 2).
Animals were observed individually at 15, 30 and 60 min after dosing and periodically during the first 24 h for any clinical signs of toxicity or mortality. The toxicity evaluation was carried out by observing behaviour parameters, including the Grimace scale parameters. The mice were sacrificed, and the main organs were examined by visual inspection during autopsy.
In vivo efficacy evaluation of strasseriolides C and D in mice
In vivo efficacy experiments were performed at the Anti-infectives Screening Core Services facility of the New York University School of Medicine [5]. The transgenic P. berghei line 676m1cl1 parasite (PbGFP-Luccon) expressing the fusion GFP (mutant 3) and firefly luciferase (LucIAV) proteins generated in the reference clone of ANKA strain cl15cy1 was used [9]. Swiss Webster female mice were infected with transgenic P. berghei line 676m1cl1 (PbGFP-Luccon) ANKA-infected 10³ red blood cells. Treatment was then started 48 h post-parasite infection by i.p. dosage of 22 mg/kg of strasseriolide D and 50 mg/kg of strasseriolide C. Vehicle only (PBS 2% methylcellulose, 0.5% Tween 80) and 20 mg/kg chloroquine were used as negative and positive controls. The infected mice were treated for four days, after which luciferase activity of the parasites was quantified 5 to 10 min after i.p. injection of 150 mg/kg D-Luciferin Potassium-salt (Goldbio) dissolved in PBS using an IVIS® Lumina II imager on the fifth day.
Data analysis and statistics
The Genedata Screener online application (Genedata AG, Basel, Switzerland) was used to compute IC50 values. The Smart Fit strategy with Hill model was selected as the choice for curve fitting. Where applicable, data are presented as mean ± standard deviation (SD). In Fig. 3, the significant differences between the vehicle controls and test compounds were calculated by the Student’s t test.