Impact of chrysosplenetin on the pharmacokinetics and anti-malarial efficacy of artemisinin against Plasmodium berghei as well as in vitro CYP450 enzymatic activities in rat liver microsome

Artemisinin and internal standard artemether (ARM, Fig. 1b) were purchased from Chongqing Huali Konggu Co., Ltd. (purity >99.0 %, Chongqing, China). CHR (purity >98.0 %) was previously isolated and purified from the industrial wastes of ART (acetone layer) with about 1 % yield in the authors’ laboratory. Industrial waste was kindly gifted by Chongqing Huali Konggu Co. Ltd, and the voucher specimen of the waste has been deposited in the College of Pharmacy, Ningxia Medical University, for further references.

Six metabolic probes including phenacetin (PN), coumarin (CA), omeprazole (OMP), dextromethophan (DM), and chlorzoxazone (CLZ) were purchased from National Institutes for Food and Drug Control, China. Midazolam (MDZ) injection was provided by Jiangsu Enhua Pharmaceutical Co., Ltd (10 mg/2 mL). NADPH, EDTA, BCA kit, DTT, and Tris were of analytical grade and purchased locally. Acetonitrile and methanol (HPLC grade) were purchased from Fisher Chemicals (Fairlawn, NJ, USA).

Rat liver microsomes (RLM, Sprague–Dawley) were purchased from Wuhan Pulaite Medical and Technical Co. Ltd (M10011). Protein concentration was determined as 20 mg/mL by bicinchoninic acid (BCA) method using bovine serum albumin as standard.

The LC system for LC–MS/MS was Shimadzu Nexera UPLC LC-30A containing a binary LC-30AD pump, DGU-20A5 vacuum degasser, SIL-30AC autosampler and CTO-30A thermostat column compartment. An API 4000 triple quadrupole mass spectrometer (AB SCIEX, Foster, USA) with an electrospray ionization (ESI) operated in the positive ion mode was used for analysis. Quantification was performed using multiple reaction monitoring (MRM) for the transitions m/z 300. 1–209.0 for ARN and 316.2–163.0 for ARM (Fig. 2). The system was controlled by Analyst software version 1.5.1. Separation was performed on a Shimadzu XR-ODS C₁₈ column (2.0 mm × 100 mm, 2.2 µm) with a Shimadzu ODS C₁₈ security guard column (5 mm × 2.0 mm, 2.2 µm) maintained at 30 °C using a mobile phase containing acetonitrile and 0.1 % formic acid in 10 mM ammonium acetate (85:15, v/v) at a flow rate of 300 µL/min. The source temperature was maintained at 600 °C and the ESI source voltage was set at 5500 V. Collision gas pressure was 3 units and collision energy was 17 V.

The six enzymatic probe substrates were standardized by using Agilent 1200 (Agilent, USA) RP-HPLC system consisted of an on-line G1322A vacuum degasser, a G1311A quaternary pump, a G1329A injection valve (USA) with a sample loop of 20 μL, a G1314B UV–visible diode-array detector (DAD). A phenomenex C₁₈ column (Synergi Hydro-RP 80A, 150 mm × 4.6 mm, 4 μm) was used as stationary phase with a flow rate of 1.2 mL/min at 30 °C. The isocratic mobile phase consisted of acetonitrile and purified water containing 1 % triethylamine and 0.02 M sodium dihydrogen phosphate (40:60 for PN, CA, DM, CLZ, and MDZ; 35:65 for OMP, v/v, PH = 3.5) was respectively used for assay of PN (wavelength: 250 nm), CLZ (282 nm), MDZ (230 nm), OMP (302 nm), CA (278 nm), and DM (202 nm).

ART and CHR were separately suspended in 0.5 % carboxy methyl cellulose (CMC-Na) by sufficient emulsification to get the stock solution of 2 mg/mL strength and diluted to get the desired concentrations for each drug before it was administrated by the intramuscular injection or gavage perfusion. Master stock solutions for assay of blood concentration were individually prepared by dissolving ARN and ARM standards in acetonitrile at equivalent concentrations of 1000 μg/mL and were gradually diluted to 2 μg/mL by mobile phase for the preparation of calibration curve (0.2–200 ng/mL) and quality control (QC) samples (0.5, 10 and 160 ng/mL for ART), respectively.

For in vitro hepatic metabolic study, CHR and six enzymatic probes were separately prepared in methanol to strength of 1 mg/mL stock solutions and were diluted to desired concentrations by phosphate buffers (PBS, 0.1 M, PH = 7) before use. QC samples were solved in inactive RLM with three concentration levels (3.0, 28.0 and 89.0 for PN; 3.4, 34.2 and 100.0 for CA; 6.0, 29.5 and 118.0 for CLZ; 3.0, 15.0 and 49.0 for MDZ; 2.7, 13.5 and 43.2 for DM). A serial of RLMs in strength of 1.25, 1, 0.75, 0.5, 0.25 mg/mL were gained for optimization of protein concentration experiment by diluting the stock protein solution in strength of 20 mg/mL (stored at −80 °C). All other stock solutions were stored at 4 °C when not in use.

Calibration curves for artemisinin were constructed by weighted least-squares linear regression (1/x ²) analysis of an eight-point calibration curve by plotting peak area of the analyte versus the peak area of the internal standard. The assay was validated through linearity, inter- and intra-day accuracy and precision, recovery, lower limits of quantification (LLOQ), stability, and matrix effect as usual.

Calibrations curves for six probe substrates in inactive RLM were obtained and the assay was evaluated through linearity, inter- and intra-day precision and accuracy, and recovery.

Plasma samples were separated by centrifugation 3500g for 10 min and 200 µL plasmas were stored at −20 °C until analysis. To 100 μL blank plasma, 50 μL ARM working solution (IS, 2 μg/mL), 100 μL methanol aqueous solution (50:50, v/v), and a standard solution or quality control (QC) solution were added. After vortex mixing for 1 min, the sample was extracted with 3 × 3 mL of methyl tert-butyl ether with gentle shaking for 10 min followed by centrifugation at 3500g for 10 min. The methyl tert-butyl ether extracts were combined and evaporated to dryness by a Pressure Gas Blowing Concentrator (multivap nitrogen evaporation system 118, Organomation Co., Ltd, USA) at 40 °C. The residue was reconstituted in 100 μL methanol and was filtered through a 0.22 µm nylon filter. Aliquots (10 μL) of the solution were injected onto the LC–MS analysis.

Twenty-four hours after the last dosing, the blood was collected from caudal vein of mice. Thin blood smear slides were air-dried, methanol-fixed, and stained in Giemsa for 40 min. The Giemsa-stained slides were examined for counting the number of parasites in random three microscopic fields, equivalent to over 200 erythrocytes each field at 1000× magnification. Percent parasitaemia (%) was calculated through dividing the number of total red blood cells by that of infected red blood cells. Percentage inhibition (%) for each drug was calculated relative to placebo. Reproducibility of counts was checked by two other readers to maintain the quality control.

The incubation system included 40 μL of RLM (protein concentration was optimized as 1 mg/mL for PN; 0.75 mg/mL for OPZ; 0.5 mg/mL for CA, DM, CLZ, and MDZ), 5 μL of substrate solution (substrate concentration was optimized as 25 μM for PN; 5 μM for CA; 20 μM for OPZ; 10 μM for DM; 15 μM for CLZ; 5 μM for MDZ), and NADPH-regenerating system (1.3 mM NADP⁺, 3.3 mM G-6-P, 3.3 mM MgCl₂, 4 U/mL G-6-P-D) and a test compound (CHR, 0–50 μM). Total reaction volume was complemented to 200 μL by PBS. After pre-incubating for 3 min at 37 °C, the reaction was initiated by the addition of NADPH solution (incubation time was optimized as 45 min for PN; 60 min for CA, 15 min for OPZ; 20 min for DM; 90 min for CLZ; 15 min for MDZ). 200 μL ice acetonitrile was added as stop reagent. After a vortex for 3 min and centrifugation for 10 min at 14,000g, an aliquot of 20 μL supernatant was injected into the HPLC system at an interval of 10 min. The amount of methanol in diluted concentrations (<1 %) had no effect on liver microsomes. All the testing groups were in three replicates (n = 3). Under the optimized condition, V_max, Michaelis constant (K _m ), enzyme activity (E%), and IC₅₀ values were obtained.

Incubation systems including 1.0 mg/mL RLM, PN (12.5, 25.0, and 50.0 μM), NPDPH, and CHR (0, 2, 4, 8 μM) for 1A2 or 0.5 mg/mL RLM, MDZ (2.5, 5.0, 10.0 μM), NPDPH, and CHR (0, 2, 4, 8 μM) for CYP3A were employed as above (n = 3). A serial of Lineweaver–Burk plots by 1/[S] versus 1/V were prepared individually under different concentrations of CHR. The location of intersection point for the linear lines decided the type of inhibition. Namely, if the point lied in Y axis it was competitive inhibition and if it lied in X axis it was non-competitive inhibition. Finally, if the linear lines were paralleled, it was uncompetitive inhibition. Concomitantly, inhibition constant (K _i ) was calculated when a second plotting was given by intercepts of the four lines versus [I] (inhibitor concentration).

The electrospray ionization mass spectra for ART and ARM were shown in Fig. 2a (MS¹ for ART), b (MS² for ART), c (MS¹ for ARM), and d (MS² for ARM). Under optimized UPLC conditions, ART and ARM were eluted within 2.5 min (Fig. 3b, c). Blank plasma showed no significant interfering peaks at the retention times of each analyte (Fig. 3a). The calibration curve of ART was linear over the concentration range of 0.2–200.0 ng/mL (y = 0.0336x + 0.0259, r = 0.9965, w = 1/x ²) with LLOQ 0.2 ng/mL (RSD% = 3.91, RE% = −5 %, n = 5). Inter- and intra-day precision and accuracy, stability, recovery, and matrix effect data were showed in Table 1. The precision and accuracy of this method indicated that all coefficients of variation at each concentration level are below 10 %. The normalized matrix effects of ART at low, middle, and high concentrations were close to 1 with a low variation in accordance with international guidelines [20]. ART was found stable in plasma for at least 6 and 24 h at ambient temperature before and after treatment and was also stable during three freeze/thaw cycles.

Six probe drugs were well analysed within 10 min by HPLC–UV method without interference. Standard curves equations of six substrates in inactive RLM were displayed in Table 2. Details of method validation data were elaborately arranged in Table 3a, b in different strength of QC samples.

Plasma concentration–time profiles of ART, in untreated rats and in pre-treated rats with different combinatory doses of CHR, were shown in Fig. 4a, b. Corresponding pharmacokinetic parameters were reported in Table 4. This study has demonstrated that the AUC_0-t of ART in ART-CHR-L (1:1), ART-CHR-M (1:2), and ART-CHR-H (1:4) groups, respectively, increased 1.44-, 1.40-, and 1.29-fold compared with ART alone (P < 0.01) without dose–response manner (P > 0.05 among the combination groups). The C_max in ART-CHR-L (1:1) and ART-CHR-M (1:2) group increased 1.64- and 1.65-fold versus control (P < 0.05) while no increase was found in ART-CHR-H (1:4) group. The t _1/2 in ART-CHR-M (1:2) group enhanced 1.68 folds compared with ART alone (P < 0.05). It showed that oral co-administration of CHR in combinatory ratio of 1:2 prior to intramuscular injection of ART achieved a longer t _1/2, a higher AUC and C_max, and lower CLz than ART alone.

The results showed in Table 5 suggested that parasitaemia (%) in ART alone, ART-CHR-L, ART-CHR-M, and ART-CHR-H groups significantly decreased 1.59, 2.23, 2.53, and 2.01 folds (P < 0.05) compared with placebo, whereas no significance was observed when CHR was solely used (P > 0.05). CHR, therefore, has no anti-malarial effect itself. Only in ART-CHR-M group, parasitaemia (%) and inhibition (%), respectively, achieved 1.59-fold reduction (P < 0.05) and 1.63-fold augment compared with ART alone. The result completely conformed to the pharmacokinetics study.

K _m values for PN, CA, OPZ, DM, CLZ, and MDZ were calculated to be 30.88, 8.90, 24.07, 28.94, 19.97, and 7.88 μM with V_max 0.14, 0.12, 0.58, 0.20, 0.18, 0.81 nmol/min/mg proteins, respectively. The working substrate concentrations for PN, CA, OPZ, DM, CLZ, and MDZ, therefore, were confirmed to be 25, 5, 20, 10, 15, and 5 μmol L⁻¹ (lower than K _m ).

Graphs showed in Fig. 5 and data in Table 6 demonstrated that CHR has a significant inhibition (P < 0.01) on CYP1A2 and CYP2C19 in a concentration-dependence manner and enzyme activity (%) decreased from (21.64 ± 1.05) % to (4.05 ± 1.41) % for CYP1A2, and from (41.98 ± 0.35) % to (0.60 ± 0.11) % for CYP2C19 when CHR concentration increased from 0 to 50 μM. As for CYP2E1, CHR had a significant inhibition in a dose–response manner only after CHR concentration exceeded 4 μM (P < 0.05). Most interestingly, 0.5 μM of CHR significantly increased the enzyme activity (%) of CYP3A from 48.27 ± 0.93 (0.0 μM) to 59.44 ± 2.55 (P < 0.01), and it slowly decreased to 57.15 ± 0.93 (P < 0.01), 52.02 ± 3.07 (P > 0.05), and 51.37 ± 0.84 (P > 0.05) after CHR concentration added from 1.0 to 2.0 μM. An inhibition of CHR on CYP3A subsequently (P < 0.01) emerged after the concentration exceeded 4 μM as well as CYP2E1. No impact of CHR on CYP2A6 and CYP2D6 were found in this paper (P > 0.05).

IC₅₀ values were computed to 4.61 μM for 1A2, 6.23 μM for 2C19, 3.38 μM for 3A, and 28.17 μM for 2E1 by GraphPad Prism 5.0 software. CHR is a moderate inhibitor on 1A2, 3A, and 2C19 (1 < IC₅₀ < 10 μM) and a weak inhibitor on 2E1 (IC₅₀ >10 μM) according to the research [21].

Figure 6a, b showed that the type of CHR as an inhibitor on CYP1A2 belonged to non-competitive (intersection point lied in X axis) and that on CYP3A was uncompetitive inhibition (parallel lines). K _i was determined to 4.07 μM for CYP1A2 (y = 33.221x + 135.2, r = 0.9904) and 3.71 μM for CYP3A (y = 0.5416x + 2.012, r = 0.9600).

Inhibition of inhibitors on enzymes has two types, namely irreversible and reversible inhibition. The latter includes competitive, noncompetitive, uncompetitive and mixed competitive inhibition (showed in Fig. 7). The dynamic characteristic of uncompetitive inhibition (substrate is indispensable for the inhibition) is that both V_max and K _m decreased when inhibitor concentration increased. This type of inhibition is infrequency in single substrate catalysis of enzyme. This study demonstrated that CHR acted as an on uncompetitive inhibitor of CYP3A activity and as a non-competitive inhibitor of CYP1A activity. The results are in accordance with the literatures [22, 23]. Especially, the fact that CHR inhibited CYP3A in rat liver microsome in an uncompetitive manner might be meaningful for us to partly explain why CHR has a partial dose-dependent manner in pharmacokinetics and anti-malarial efficacy studies. More interestingly, initially increased enzyme activity of CYP3A under a low concentration of CHR treatment was observed in the inhibition experiment using RLM. The authors speculate that the phenomenon might be closely related to the uncompetitive inhibition type of CHR on CYP3A. CHR exclusively bonded to enzyme-substrate complex (ES) instead of enzyme itself; therefore, when inhibitor is added into the reaction, the equilibrium reaction will shift to the product generation direction, which enhances the production of ES complex instead. When CHR concentration is very low, ES complex may not be fully bonded. The surplus will continue to be decomposed to enzyme and product by enzymolysis. So the enzyme activity initially increased a little compared with the control. As CHR concentrations increased, however, the activity slowly decreased and then inhibition effect was finally observed due to the binding between CHR and ES complex and therefore the hindering of the enzymolysis. The increased activity might, therefore, not be activated by induction of CHR on CYP3A but possibly depends on the shifting of equilibrium reaction. Further research should be carried out to demonstrate the hypothesis.