Cell culture medium and reagents, propidium iodide (PI), DMSO (molecular biology grade), sodium carboxymethyl cellulose (CMC-Na) and formic acid were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Fetal bovine serum (FBS) was obtained from HyClone (Logan, UT, USA). RNase was bought from Roche (Basel, Switzerland). Meisoindigo was provided by the Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College (Beijing, China). Indirubin (internal standard) was purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). HPLC grade ethyl acetate, methanol and acetonitrile were purchased from Fisher Scientific Co. (Fair Lawn, NY, USA). Milli-Q water was obtained from a Millipore water purification system (Billerica, MA, USA) and used to prepare buffer solutions and other aqueous solutions.

NB4 (APL cells, M3 subtype according to FAB) (7), NB4.007/6 (retinoic acid resistant cells developed through continuously culturing NB4 in the medium containing retinoic acid (9), HL60 (AML cells, M2 subtype according to FAB) (14), and U937 (myelomonocytic leukemia cells, M5 subtype according to FAB) (10,15), were maintained in RPMI-1640 medium supplemented with 10% heat-inactive fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified 5% CO₂ atmosphere at 37°C. The cells were subjected to sub-culture by 1:2 dilution with fresh medium every 3 days. Cultured cells with a passage number of 10–20 were used in the experiments to reduce variability due to cell culture conditions.

The inhibitory effects of meisoindigo on the four human leukemic cells were tested by treating the cells with meisoindigo at various concentrations (0, 1, 2, 4, 8, 12, 16, 20 μM). The cells were cultured in 24-well plates. The number of the cells was counted by trypan blue exclusion method. The initial cell numbers (NB4, NB4.007/6, HL60 and U937) were 3×10⁶ cell/ml. After incubation for 24, 48 and 72 h, the plates were taken out and the number of viable cells was counted. All experiments were carried out in eight replicates. The cell activity was defined as the number of viable cells in the test culture over the mean number of viable cells in the control. The 50% inhibitory concentration (IC₅₀) on cell activity was estimated based on the cell number obtained after incubation for 72 h. The data were fitted to the Inhibitory Effect Sigmoid E_max Model with the software, WinNonlin version 1.0 (Lexington, KY, USA). All regressions were carried out with equal weighting.

Apoptosis and cell cycle distribution of the tested leukemic cells were examined by flow cytometry analysis. The leukemic cells were treated with meisoindigo of various concentrations (0, 4, 8, 12, 16 and 20 μM) for 24, 48 and 72 h in the respective 6-well plates. At the end of incubation, the cells were harvested, washed twice with 3 ml PBS (pH 7.4 contain 1% FBS), and then fixed with 70% ice-cold ethanol and kept at −20°C for at least 24 h. On the day before flow cytometry assay, the fixed leukemic cells were spun down, washed with 3 ml PBS with 1% FBS and spun down again. The cells were then dyed in 500 μl PBS containing 100 μg/ml propidium iodide (PI) and 100 μg/ml RNase and stored at 4°C overnight. On the following day, the samples were first filtered through a 30 μm pore size nylon mesh prior to the flow cytometry analysis. The samples were analyzed on a Coulter Epics Elite ESP flow cytometer (Beckman Coulter, FL, USA) equipped with a 15 mW argon-ion laser source of 488 nm. Red fluorescence of propidium iodide (PI) was collected with a 610 nm bandpass filter. Typical flow rates were about 200 to 300 particles per sec. A total of 10,000 cells were analyzed per sample. The experiment was carried in triplicate. The data obtained were analyzed using WinMDI 2.8 software (La Jolla, CA, USA). From the DNA histogram, the percentages of cells in different cell cycle phases were determined. Cells with DNA concentration less than the G1 phase were regarded as apoptotic cells.

Statistical analysis was performed by using SPSS 10.0 (Chicago, IL, USA) or GraphPad Prism 2.00 (La Jolla, CA, USA). All experimental data were analyzed with One-Sample Kolmogorov-Smirnov Test for their distribution. As the parameters were found to be normally distributed, data were expressed as mean ± standard derivation (SD). EC₅₀ on growth inhibition was compared at their 95% confidential level. The percentages of apoptotic cells at different meisoindigo concentrations were compared with the one-way ANOVA and the posthoc Tukey’s test. A value of p<0.05 was adopted to indicate statistical significance.

Synthetic routes of reductive metabolite standards (M1+M2 and M3+M4) are shown in Fig. 1 as previously described (11). The starting material of meisoindigo (50 mg) was dissolved in methanol (100 ml). After adding 10% palladium on carbon (10 mg), the mixture was placed inside the hydrogenator apparatus and kept shaking at 40 psi for 2 h at room temperature. The reaction mixture was filtered through filter paper and the filter pad was washed with distilled water. The filtrate was evaporated under reduced pressure and the precipitate was purified by Agilent 1100 preparative HPLC system (Palo Alto, CA, USA) on a Peeke Scientific Combi-A C18 column (50 × 20 mm, 5 μm) (Redwood City, CA, USA) with a guard cartridge at ambient temperature. The isocratic mobile phase consisted of 50% water (solvent A) and 50% methanol (solvent B) and was delivered at a flow rate of 4 ml/min within 20 min. The UV absorbance was monitored at 254 nm. The combined metabolites fractions (total M1+M2 and M3+M4) were evaporated under reduced pressure and stored at −20°C.

The LC-MS/MS system consisted of an Agilent 1200 HPLC (Palo Alto, CA, USA) with a Q TRAP™ 3200 hybrid triple quadrupole linear ion trap mass spectrometer from Applied Biosystems/MDS Sciex (Concord, Ontario, Canada). Chromatographic separations for rat plasma samples were performed on a Phenomenex Luna C18 column (150 mm × 2.00 mm i.d., 5 μm) (Torrance, CA, USA) with a guard cartridge. The injection volume and column temperature were 5 μl and 25°C, respectively. The mobile phases consisted of 0.1% formic acid in water (solvent A) and acetonitrile (solvent B). The gradient elution was set at a flow rate of 0.3 ml/min. The optimized linear gradient elution conditions were: 30 to 95% B over 13 min, followed by an isocratic hold at 95% B for another 2 min. At 15 min, B was returned to 30% in 1 min and the column was equilibrated for 14 min before the next injection. The total run time was 30 min. During LC-MS/MS analysis, up to 4 min of the initial flow was diverted away from the mass spectrometer before the data acquisition. The mass spectrometer was operated in the positive ion mode with a TurboIonSpray source. Enhanced product ion (EPI) scans were used to investigate the fragmentation patterns of authentic standard compounds. A list of three Multiple Reaction Monitoring (MRM) transitions in Table I was selected for the quantification of meisoindigo, its reductive metabolites and indirubin (internal standard). The Q1, Q3, declustering potential (DP), and collision energy (CE) values in Table I were based on the MS/MS results of those respective standards. The other ionization parameters were as follows: curtain gas (CUR), 20 (arbitrary units); ion source gas 1 (GS1), 40 (arbitrary units); ion source gas 2 (GS2), 50 (arbitrary units); source temperature (TEM), 550°C; entrance potential (EP), 10 V. The dwell time of each MRM transition was 150 msec. The HPLC system and the mass spectrometer were controlled by Analyst™ 1.4.2 software from Applied Biosystems/MDS Sciex.

Stock solutions of meisoindigo and its reductive metabolites (M1+M2 and M3+M4) were prepared by dissolving the accurately weighed compounds in methanol to give a final concentration of 1 mg/ml for both. Solution of indirubin (internal standard) was prepared in methanol at the concentration 1 mg/ml and diluted to 1 μg/ml with methanol. Blank rat plasma (drug free) was obtained by centrifugation of predose rat blood. Calibration curves were prepared by spiking appropriate standard solutions of the parent drug meisoindigo and its reductive metabolites, respectively, to the blank plasma. Concentrations in plasma samples were 1, 2.5, 5, 10, 50, 250 and 500 ng/ml for meisoindigo and 2.5, 5, 50, 100, 250, 500 and 1,000 ng/ml for its reductive metabolites. Quality control (QC) samples were separately prepared in blank plasma samples at the concentration of 5, 50 and 500 ng/ml for meisoindigo and its reductive metabolites, respectively.

Each rat plasma sample (45 μl) was added with 5 μl of internal standard indirubin (100 ng/ml). The samples were briefly mixed before extraction with 2 × 2-fold volume of cold ethyl acetate saturated with H₂O. The organic phases from the extracted plasma were combined and dried at 35°C under a gentle stream of nitrogen. The residue was reconstituted with 50 μl acetonitrile and H₂O (1:1) for LC-MS/MS analysis.

The pharmacokinetic profiles of meisoindigo and its reductive metabolites after oral administration were studied in rats. The animal experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the National University of Singapore (NUS). For the study, four healthy male Sprague-Dawley (SD) rats (7–8 weeks of age, average 250 g) were purchased from Animal Holding Unit, National University of Singapore. They were provided with a standard diet and water ad libitum. The room was kept on a 12/12-h light/dark cycle at a temperature of 23±1°C and relative humidity of 50±10%. At least one week of acclimatization period was allowed for the rats prior to drug administration. The rats were fasted 12 h prior to administration of the dose and were fed 12 h after the dose. The dose was formulated in a mixture of meisoindigo suspended in 1% sodium carboxymethyl cellulose (CMC-Na) solution at a target concentration of 3.75 mg/ml. The rats received meisoindigo solution by gavage administration with a single dosage of 10 mg/kg body weight. Blood samples were collected in heparinized tubes via caudal vein predose and at 1, 2, 3, 4, 5, 6, 8, 10 and 24 h postdose. Blood samples were immediately centrifuged at 6,000 rpm for 10 min to obtain the plasma. Plasma samples were transferred to clean tubes and underwent the sample preparation procedures as described above before LC-MS/MS analysis. Pharmacokinetic parameters were calculated by non-compartmental analysis using the software of WinNonlin standard version 1.0 from Scientific Consulting Inc. (Apex, NC, USA).

As shown in Fig. 2, the leukemic cells grew well in the absence of meisoindigo with a relatively rapid proliferation rate, suggesting appropriate cell culture conditions were applied. The NB4, NB4.007/6 and HL60 cells proliferated 4–5 times during a 3-day period, while U937 proliferated about 7 times during the same period. There was always a delay in growth on the first day, before the growth was accelerated. In our preliminary study, the growth and/or proliferation of the test cells were observed to decline on the 4th day after culture, if no fresh medium was added. In order to avoid this fluctuation in growth due to change of media, the inhibitory effects of meisoindigo was studied over three days.

Meisoindigo effectively inhibited the growth and proliferation of the retinoic acid sensitive cells (NB4 and HL60), as well as the retinoic acid resistant cells (NB4.007/6 and U937). The inhibitory effects of meisoindigo on the growth and proliferation of various leukemic cells exhibited a dose-dependent characteristic. The 50% inhibitory concentrations (IC₅₀) were estimated based on the values of growth obtained after 72 h incubation and the data fitted to an Inhibitory Effect Sigmoid E_max Model. The correlation of regression of the model fittings was 0.986, indicating an appropriate pharmacodynamic model was applied. The IC₅₀ values were 7.9, 7.1, 7.1 and 7.5 μM in NB4, NB4.007/6, HL60 and U937 cells, respectively. No statistically significant difference among these IC₅₀ values was observed. After exposure to meisoindigo at 8 μM or above for 72 h, all leukemic cells were clearly inhibited and there was a greater than 50% decrease in the number of viable cells in the meisoindigo treated group over the control (Fig. 2). When treated with 20 μM meisoindigo for 48 to 72 h, most of the cells did not survive. The antileukemic effects of meisoindigo seemed to be time-dependent. Prolonged incubation with high concentrations of meisoindgo (12, 16 or 20 μM) led to enhanced inhibitory effects. The calculation of IC₅₀ with counts on day 1, 2 or 3 resulted in very similar finding in the four leukemic cells. Interestingly, the finding that meisoindigo was effective in inhibiting the growth and proliferation of retinoic acid resistant NB4.007/6 and U937 cells, suggests a less possibility of cross-resistance with retinoic acid. Thus meisoindigo could be a potential treatment for leukemia, either sensitive or resistant to retinoic acid.

The flow cytometry analysis showed that NB4, NB4.007/6, HL60 and U937 cells grew well in the absence of meisoindigo with about 20–30% of the cells distributed in the respective M/G2 phase and S phase (Table II), suggesting that these cells have active DNA synthesis and cell division. This phenomenon was in good accord with the rapid growth and proliferation of these cells. After treatment with meisoindigo, the percentage of sub-G1 cells increased substantially over time in all the four cell types (Table II). The percentage of sub-G1 cells, which is regarded as the percentage of apoptosis shown in Fig. 3, indicated that the induction of apoptosis was both time- and dose-dependent. The percentages of sub-G1 cells in all the four cell types were quite similar at high concentrations of meisoindigo (12, 16, 20 μM), suggesting that the apoptosis-inductive effect of meisoindigo was saturable at high concentrations. Also, the percentage of G1/G0, S and M/G2 dropped substantially as the dosages increased (Table II), which suggested that no phase arrest or cell phase accumulation occurred in these meisoindigo-treated cells, and thus the antileukemic effects of meisoindigo in the cells might be independent of cell cycle arrest. However, the apoptotic effect is not associated with DNA fragmentation (data not shown).

The LC conditions were optimized by varying the mobile phase, gradient elution, flow rate and columns. The MS conditions were optimized by tuning with authentic standards of meisoindigo, its reductive metabolites and indirubin. The proposed fragmentation scheme and MS/MS spectrum of protonated meisoindigo (m/z 277) are shown in Fig. 4A. Loss of CONH (43 Da) following gain of one H generated a major product ion at m/z 234. Loss of CONCH₃ (57 Da) following loss of one H generated the most abundant product ion at m/z 218. The product ion at m/z 234 was selected to form the MRM transition for quantification due to its better selectivity without compromising the sensitivity. As is shown in Fig. 1, the reductive metabolites of meisoindigo M1+M2 were a pair of (3-R, 3′-R) and (3-S, 3′-S) reduced-meisoindigo enantiomers with two hydrogens located at the same side of 3,3′-single bond, whereas M3+M4 were another pair of (3-R, 3′-S) and (3-S, 3′-R) enantiomers with two hydrogens located at the opposite sides. The MS/MS spectra of M1+M2 and M3+M4 were identical and are shown in Fig. 4B. The proposed fragmentation scheme of protonated reductive metabolites (m/z 279) showed that cleavage of the 3,3′ bond generated the dominant product ion at m/z 147, which was thus selected to form the MRM transition for quantification. The proposed fragmentation scheme and MS/MS spectrum of protonated indirubin (m/z 263) are shown in Fig. 4C. Loss of CONH (43 Da) and CO (28 Da) from indirubin following loss of one H generated the most abundant product ion at m/z 190. Thus the product ion at m/z 190 was selected to form the MRM transition for quantification.

Full validation was conducted for the pharmacokinetic study of meisoindigo and its reductive metabolites in rat plasma because this bioanalytical method was developed for the first time. The assay selectivity was confirmed as no significant interference was observed at the retention time of meisoindigo in the blank rat plasma. The matrix effect on the analyte determination was not taken into account due to the relative slow chromatographic separation. The linear range of calibration curve was 1–500 ng/ml with correlation coefficients greater than 0.9998. LLOQ was 1 ng/ml for meisoindigo (Fig. 5A). The linear range of calibration curve was 2.5–1,000 ng/ml with correlation coefficients greater than 0.9928. LLOQ was 0.5 ng/ml for the reductive metabolites of meisoindigo (Fig. 5B).

Table III shows both intra- and inter-day accuracy and precision data determined by the analysis of meisoindigo and its reductive metabolites. These results fulfilled the criteria of validation with accuracy and precision not more than ±20%, indicating that this assay is consistent and reliable with good accuracy and precision. The extraction recoveries of low, medium, high concentrations were within ±20% and reproducible. The results of short-term, long-term, freeze and thaw, stock solution and post-preparative stability evaluated at a low concentration of 5 ng/ml and a high concentration of 500 ng/ml of meisoindigo and its reductive metabolites in rat plasma were summarized in Table IV.

The pharmacokinetic profiles of meisoindigo and its reductive metabolites in rat plasma are shown in Fig. 6. Data points were the average of four measurements with standard deviation (SD) as the error bars. The major pharmacokinetic parameters are listed in Table V. It is shown in Fig. 6 that meisoindigo was converted to its reductive metabolites rapidly, indicating the 3,3′ double bond within the molecule of meisoindigo is metabolically unstable. Plasma level of the parent drug was low, whereas plasma level of its reductive metabolites was much higher suggesting the occurrence of extensive first pass effect. The wide variation in plasma levels of the drug and its reductive metabolites also indicated possible incomplete absorption due to the poor biopharmaceutical properties of the drug. In Table V, the low AUC_{0→24 h} and long t_1/2 (a composite parameter determined by volume of distribution and clearance) could be partly caused by a large volume of distribution of meisoindigo due to its lipophilicity.