Large Volume Direct Injection Ultra-High Performance Liquid Chromatography–Tandem Mass Spectrometry-Based Comparative Pharmacokinetic Study between Single and Combinatory Uses of Carthamus tinctorius Extract and Notoginseng Total Saponins

The combination of Carthamus tinctorius extract (CTE) and notoginseng total saponins (NGTS), namely, CNP, presents a synergistic effect on myocardial ischemia protection. Herein, comparative pharmacokinetic studies between CNP and CTE/NGTS were conducted to clarify their synergistic mechanisms. A large volume direct injection ultra-high performance liquid chromatography–tandem mass spectrometry (LVDI-UHPLC-MS/MS) platform was developed for sensitively assaying the multi-component pharmacokinetic and in vitro cocktail assay of cytochrome p450 (CYP450) before and after compatibility of CTE and NGTS. The pharmacokinetic profiles of six predominantly efficacious components of CNP, including hydroxysafflor yellow A (HSYA); ginsenosides Rg1 (GRg1), Re (GRe), Rb1 (GRb1), and Rd (GRd); and notoginsenoside R1 (NGR1), were obtained, and the results disclosed that CNP could increase the exposure levels of HSYA, GRg1, GRe, GRb1, and NGR1 at varying degrees. The in vitro cocktail assay demonstrated that CNP exhibited more potent inhibition on CYP1A2 than CTE and NGTS, and GRg1, GRb1, GRd, quercetin, kaempferol, and 6-hydroxykaempferol were found to be the major inhibitory compounds. The developed pharmacokinetic interaction-based strategy provides a viable orientation for the compatibility investigation of herb medicines.


Supplementary tables and figures
. The total ion current chromatogram (TIC) of CTE, the corresponding chemical composition information were reported on the previous researches (Chen, et al., 2014; Analyst 139, 6474-6485). Figure S2. The optimization of sample solvents for the pharmacokinetic analysis (A) and the cocktail assay (B) (n = 3). Figure S3. Kinetic profiles for the enzymatic turnover of CYP450-mediated probe reactions. Figure S4. Inhibition curves of the seven positive inhibitors obtained from the substrate cocktail incubation. Table S1 Multiple reaction monitoring transitions and fragmentation parameters of six standards and IS1 for PK analysis. Table S2 Multiple reaction monitoring transitions and fragmentation parameters of seven metabolites and two internal standards (IS2 and IS3) for cocktail assay. Table S3 The instrument stability of the LVDI-UHPLC-MS/MS setup. Table S4 Regression equations, linear ranges, and low limits of quantification (LLOQ) of the six standards in rat plasma for the PK study. Table S5 Intra-and inter-day precisions and determination accuracies of six standards for the pharmacokinetic study. Table S6 Extract recoveries and matrix effects of six target constituents in rat plasma samples for the PK study. Table S7 Stability of the six CNP constituents in rat plasma samples for the PK study Table S8 Plasma concentration-time of the six target constituents after oral administration of CTE, NGTS, and CNP, respectively. Table S9 Regression equations, linear ranges, LLOQs of the seven metabolites for the cocktail analysis. Table S10 Intra-and inter-day precisions and determination accuracies of the seven metabolites for cocktail analysis. Table S11 Extract recoveries and matrix effects of seven target constituents and two IS for the cocktail analysis. Table S12 Km values determined for the enzymatic reaction of the probe substrates and the inhibition IC50 values measured for the positive inhibitors to seven CYP450s.

Mass spectrometric parameter optimization for pharmacokinetic study and in vitro cocktail assay
Stock solution of each reference standard was diluted to appropriate concentration (100 -200 ng·mL -1 ) with 50% aqueous methanol and directly infused (flow rate, 10 µL·min -1 ) into the ion source of a QTRAP-MS via a syringe pump for mass parameter optimization. For the pharmacokinetic analysis, seven analysts, including HSYA, GRb1, GRg1, GRd, NGR1, GRe and linarin (IS1) were involved as the targeted components. Negative polarity could afford better mass responses for those components in comparison with the positive ionization mode. Regarding the seven metabolites (dEtPHE, OHMID, OHTOL, dMeDEX, OHCHL, OHBUP, and OHOME) of the cocktail analysis, the optimum mass parameters were also obtained by manual tuning via directly infusing pure compounds into mass spectrometer; both positive and negative polarities were applied according to the results. Quantitative analyses were monitored in MRM mode. Mass axis was calibrated using standard polypropylene glycol (PPG) dilution solvents. The ion-spray voltages were maintained at -4500 V and 5500 V for the negative and positive polarities, respectively. Nitrogen was used as the nebulizer (GS1), curtain (CUR), heater (GS2), and collision gases. While the GS1, GS2, and CUR for the PK study were set as 45, 45, and 35 psi, respectively. GS1, GS2, and CUR for the cocktail assay were set as 50, 50, and 35 psi, respectively. The ion sources of PK and cocktail studies were separately heated to 450 °C and 500 °C. The precursor-to-product transition, optimized declustering potential (DP) values, and collision energy (CE) values of the PK and cocktail studies are separately shown in Table  S1 and Table S2, whereas the dwell time, entrance potential (EP), and collision cell exit potential (CXP) values of all ion transitions were fixed at 30 ms, 10 V, and 12 V, respectively.
The injection volume of PK study was set as 100 µL (50 µL sample for two times by LVDI). The preparation and measurement of the drug-free samples were performed in parallel with those of the treated samples. For the cocktail study, the injection volume was set as 50 µL. The preparation and measurement of the drug-free samples were performed in parallel with those of the treated samples.

Method validations
Mixed standard stock solutions were individually obtained by pooling all stock solutions (HSYA/GRg1/GRb1/GRd/GRe/NGR1 for the PK study, dEtPHE/OHOME /OHTOL/dMeDEX/OHMID/OHBUP/OHCHL for the cocktail assay), and the obtained solutions were then sequentially diluted using 50% aqueous methanol to afford serial mixed standard solutions with desired concentration levels. Four concentration levels of calibration samples, including high, medium (two concentration levels), and low levels, were selected as quality control (QC) samples. The method validation, in terms of selectivity, linearity and sensitivity, precision and accuracy, recovery and matrix, and stability, was conducted following the US Food and Drug Administration (FDA) Guidance on Bioanalytical Method Validation and Drug Interaction studies [2,3].

The instrument precision of the LVDI-UHPLC-MS/MS
It is well known the instrument stability is very important for establishing a quantitative method. Contrary to the common UHPLC-MS methods, the setup of LVDI-UHPLC-MS/MS was installed using additional pipelines to connect the UHPLC, the 6-port/2-channel switching valve, and the QTRAP-MS. Therefore, the stability of LVDI-UHPLC-MS/MS setup was firstly tested and verified by injecting 5 µL HQC and LQC samples before optimizing the chromatographic programs of the loading phases. The results (Table S3) indicated that the instrument stability of the LVDI-UHPLC-MS/MS setup could meet the demands for developing a quantitative method.

Optimization the elution phase program of LVDI-UHPLC-MS/MS
The elution programs of the LVDI-UHPLC-MS/MS for the PK and cocktail studies were separately optimized. Because of the pivotal role for the chromatographic performances, the analytical columns were carefully screened. For the PK study, the HSS T3 column (50 × 2.1 mm, I.D, 1.8 µm) was advantageous at resolution, peak shape, and chromatographic retention of HSYA in comparison with BEH C18 (50 × 2.1 mm, I.D, 1.7 µm) and RP shield C18 (50 × 2.1 mm, I.D, 1.7 µm) columns. In the cocktail assay, the T3 column showed more strong retention of dEtPHE than the BEH C18 column, and obtained better peak shapes of OHBUP and OHMID than the RP shield C18 column. Thus, both PK and cocktail studies employed HSS T3 column based on the resolution, peak shape, and chromatographic retention. Regarding the HSS T3 column, the gradient water and acetonitrile were employed as elution solvents after careful assessments between water and acetonitrile and water and methanol. Both PK and cocktail studies introduced formic acid (0.01%, v/v) as the solvent additive since it could afford better peak shapes along with overall MRM responses than ammonium formate (1, 5, 10 mM). In total, the ammonium formate additives can induce peak shape distortions of HSYA, dEtPHE, OHBUP, and OHMID. Afterwards, the gradient programs of the elution phases were individually customized to afford satisfactory chromatographic separations for the pharmacokinetic and cocktail studies. Furthermore, a relative lower temperature (25 °C) was applied for the analytical column, which could significantly modify the peak shapes of these analytes, especially for the HSYA and dEtPHE, in comparison with those higher temperatures, e.g. 40 °C and 50 °C. Consequently, the gradient programs of the loading phases were optimized based on the above chromatography programs of the elution phases.

Results of method validation 2.3.1 Specificity
For the PK study, representative MRM chromatograms obtained from blank rat plasma, a blank plasma sample spiked with six analytes and an internal standard (IS1), and the plasma sample after oral administration of CNP were respectively shown in Figure 2. For the cocktail assay, representative MRM chromatograms obtained from the incubation matrix, an incubation matrix spiked with seven metabolites, and two internal standards (IS2 and IS3), were respectively shown in Figure 3. No significant interferences of endogenous ingredients were observed for the LVDI-UHPLC-MS-based methods for the PK and cocktail studies.

Linearity and Sensitivity
Linear regression equations for calibration curves of the six standards for the PK study and the seven metabolites for the cocktail assay were respectively summarized in Table S4 and Table  S9. The calibration curves covered a wide dynamic range and the correlation coefficients of all constituents were more than 0.9911 in the linear range.

Precision and Accuracy
As shown in Table S5 and Table S10, RSDs of intra-and inter-day precisions were found to be lower than 15.15% for the PK and cocktail studies. The accuracy of the PK and cocktail studies were respectively in the ranges of 88.21-104.79% and 86.52-107.60% at four-level QC samples. All the assay values satisfied the acceptable criteria, indicating the favorable data for precision and accuracy of this developed LVDI-UHPLC-MS/MS method.

Extraction Recovery and Matrix Effect
Matrix effects and extraction efficiency were examined in duplicate by three groups of standard addition experiments. Each group included four concentration levels. For the PK study, the extraction efficiencies of HSYA, GRg1, NR1, and GRe ranged from 90.23% to 110.26% at all the four concentrations (Table S6). Their matrix effect led to weak ion suppression, ranging from −2.4% to 15.4% for all the four concentrations (Table S6). Considering the higher plasma concentrations of Rb1 and Rd, their extraction efficiency and matrix effects were compromised (around 70%) to improve the sensitivity of the other four analytes. The extraction recoveries of the seven metabolites for the cocktail assay at four concentration levels ranged from 82.06% to 114.70% (Table S11), indicating the recovery of protein precipitation with methanol was precious and proper for various levels. And the matrix effects were in the range of 83.03% to 114.17% at four QC levels (Table S11). Therefore, there were no obvious matrix effects for the analysis of target compounds and two internal standards (IS2, IS3) in the cocktail investigation, showing that the endogenous ingredients did not interfere with the ionization of the target analytes.

Stability
The stabilities of the six target constituents in the rat plasma samples were listed in Table  S7. The results showed that these constituents in plasma were all stable in autosampler at 4 °C for 24 h, at −80 °C for 60 days, and three freeze/thaw cycles, with RSD values in the range of 0.22% to 16.34%.
Above all, the newly developed methods based on LVDI-UHPLC-MS/MS were sensitive, precise, and accurate for the pharmacokinetic and cocktail assays.

Optimization and verification of the incubation system for the cocktail assay
The probe compounds, viz. PHE, OME, TOL, DEX, MID, CHL, and BUP were finally chosen after incubating all the recruited substrates. In order to assure the linear relationship between enzyme activity and metabolic transformation, the protein concentration should be in the range of 0.05−0.20 mg/mL, and the incubation time should be among 0−20 min. Therefore, the incubation was conducted using 0.20 mg/mL protein for 15 min.    For abbreviations of analytes please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the "Chemicals and reagents" section. For abbreviations of six standards please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the "Chemicals and reagents" section.
For abbreviations of analytes please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the "Chemicals and reagents" section.   For abbreviations of substrates, metabolites, and inhibitors please refer to the "Chemicals and reagents" section. For abbreviations of analytes please refer to the ″Chemicals and reagents″ section of the Supporting information.