1 Shakaran

Zhang Yun Jing Gender Reassignment

The incidence of STZ-HFD induced HCC is significantly higher in male mice than in female mice

STZ-primed neonatal mice fed with HFD resulted in HCC at week 20. In 100% of the male mice (n = 8), HCC liver tumors were observed (Fig. 1, arrowhead). However, we observed that 1 out of 8 female mice developed liver tumors and the number of tumors in the single female mouse was significantly lower than those found in male mice (Fig. 1A and B). Regardless of sex, liver to body weight ratio, fasting serum glucose, serum triglyceride (TG), serum lipopolysaccharide (LPS), ALT, alpha-fetoprotein (AFP), and mRNA expression of Collagen type I (Col I) and Glypican-3 (Gpc-3) were significantly higher in STZ-HFD-exposed mice than the controls (Fig. 1C). When grouped by sex, no significant differences were observed in controls whereas male mice that underwent STZ-HFD intervention had statistically significant higher liver to body weight ratio, fasting serum glucose, serum TG, serum LPS, ALT, AFP, mRNA levels of Gpc-3 and Col I relative to females. Based on our results, the incidence of HCC in male STZ-HFD mice was 100% vs. a 12.5% HCC incidence observed in female STZ-HFD mice, thus revealing a clear sex disparity for development of HCC.

STZ-HFD intervention induced significant alteration in gut microbiota

To monitor shifts in the composition of fecal microbiota in the development of HCC, Illumina MiSeq sequencing was performed. In total, 969585 valid sequences were generated and a total of 639057 reads (average of 31953 ± 3692 S.D. reads per sample) were obtained for 20 samples (n = 5 in each group) after quality control. A total of 1159 operational taxonomic units (OTUs) were then identified by grouping reads at the 97% similarity level. The Shannon and Chao1 indices all reached stable values as indicated by the observed plateaus seen in for each group (Supplementary Fig. S1A,B). This indicated that most of the bacterial richness, ie., the number of taxa (species) present in a sample at a particular phylogenetic level (Chao1 index) and diversity, ie., a metric that combines both richness and the evenness of abundance of different taxa (Shannon index) in these communities were covered (Supplementary Fig. S1A,B). The Rarefaction curves revealed that although new rare phylotypes would be expected with additional sequencing, most of the diversity had already been captured as each curve has started to plateau (Supplementary Fig. S1C). Compared with the controls, the STZ-HFD group exhibited lower alpha-diversity as indicated by Chao1 (t test, P = 0.005), ACE (t test, P = 0.006) and Shannon (t test, P = 0.14) for both males and females (Supplementary Table S1). The Simpson (t test, P = 0.04) index is also a measure of diversity and was also significantly different between STZ-HFD mice and controls but the interpretation of this index with respect to our data is that control mice had a slightly higher value indicating more dominance from one taxa relative to the STZ-HFD groups. This was confirmed at the level of phylum in Fig. 2A. ACE or Chao1 were significantly different between control and STZ-HFD group in female mice but with no significant difference in male mice, highlighting sex differences in community richness with STZ-HFD female mice showing lower community richness relative to males. This result is also evident from the rarefaction curve (Supplementary Fig. S1C). In contrast, a significant difference in the Simpson index was only observed between control and STZ-HFD male mice (Supplementary Table S1). All of the indices describing microbiota α-diversity were found be significant when comparing STZ-HFD male vs. female mice. Female STZ-HFD mice scored lower in both diversity and richness relative to the male STZ-HFD mice. These results highlight the sex specific shifts in gut microbiota that occurred upon STZ-HFD treatment.

At the phylum level, the majority of the bacterial phyla identified in the fecal samples were encompassed by Bacteriodetes (73.1% in control male mice and 67.4% in control female mice, 59.9% in STZ-HFD male mice and 57.3% in STZ-HFD female mice, on average) and Firmicutes (18.9% in control male mice and 26.5% in control female mice, 24.3% in STZ-HFD male mice and 13.5% in STZ-HFD female mice, on average) as depicted in Fig. 2A. This is also reflected by the relatively high Simpson index (Supplementary Table S1).

The relative amounts measured for other bacteria were; (1) Proteobacteria (6.7% in control male mice and 5.7% in control female mice, 13.9% in STZ-HFD male mice and 28.5% in STZ-HFD female mice, on average), (2) Deferribacteres (1.0% in control male mice and 0.2% in control female mice, 1.0% in STZ-HFD male mice and 0.3% in STZ-HFD female mice, on average), and (3) Actinobacteria (0.1% in control male mice and 0.1% in control female mice, 0.7% in STZ-HFD male mice and 0.1% in STZ-HFD female mice, on average). Twenty weeks of HFD feeding induced widespread changes in gut microbial community structure at the phylum level, with abundances of Proteobacteria increased and abundances of Bacteroidetes decreased in all mice. Interestingly, Firmicutes were decreased significantly after 20 weeks of HFD feeding in female mice, in contrast to a significantly increased Firmicutes population in male mice. The ratio of Firmicutes to Bacteroidetes was markedly increased upon HFD in male mice (0.26 to 0.41) and decreased significantly in female mice (0.39 to 0.24). Verrucomicrobia was significantly decreased in male mice but was increased in female mice. As shown in Fig. 2B, differences in gut microbiota at the phylum level were observed between males and females in the controls and the difference remained after STZ-HFD intervention.

Identification of bacterial taxa abundances associated with STZ-HFD intervention and sex

Microbial compositions of STZ-HFD in male and female mice were compared by applying the linear discriminant analysis (LDA) effect size (LEfSe) algorithm on relative taxonomic abundances at different phylogenetic levels (from phylum until genus level). When compared to controls (Supplementary Fig. S2A,B), STZ-HFD mice showed decreased abundance of Coriobacteriaceae, Bacteroidaceae, Paraprevotellaceae, Prevotella, Lactobacillus, Lactobacillaceae, Anaerostipes, Coprobacillus, and Erysipelotrichaceae. On the other hand, Corynebacterium, Corynebacteriaceae, Rhodococcus, Nocardiaceae, Streptophyta, Bacillus, Bacillaceae, Staphyiococcus, Aerococcus, Enterococcus, Allobaculum, Erysipelotrichales, Klebsiella, Acinetobacter, Pseudomonadales, Enterobacteriales and Turicibacteraies were significantly increased in STZ-HFD-exposed mice, compared to control mice, based on the alpha-values for the factorial Kruskal-Wallis test between groups (p < 0.05) and the logarithmic LDA score (>2.0). Next, sex-dependent differences in taxa were identified by directly comparing STZ-HFD exposed males with STZ-HFD-exposed females (Fig. 2C and Supplementary Fig. S2C). This revealed a higher abundance of Corynebacterium, Corynebacteriaceae, Rhodococcus, Nocardiaceae, Adlercreutzia which belong to the phylum Actinobacteria, Bacillus, Bacillaceae, Staphylococcus, and Staphylococcaceae within the class of Bacilli, Desulfovibrio and Desulvibrionales within the phylum of Proteobacteria, and Clostrodium within the phylum of Firmicutes in male mice when compared to female mice. In particular, we observed that the bacteria involved in BA metabolism were different between males and females and became significantly different after STZ-HFD intervention (Fig. 2D).

As revealed by the OPLS-DA scores plot established using gut microbiota involved in BA metabolism (R2X = 766, R2Y = 0.957, Q2(cum) = 0.721), the control male, control female and STZ-HFD female mice were located in the first and second quadrant while STZ-HFD male mice were located at the fourth quadrant away from the controls (Fig. 2E).

We also performed the MANOVA on the first three weighted microbial PCoA axes and found that the influence of STZ-HFD intervention (p < 0.0001), sex (p = 0.001) and the interaction of STZ-HFD intervention and sex were significant (p < 0.0001) per the Wilks’ test. Thus far we have established; (1) that male mice are more susceptible to HCC, (2) that there are significant sex disparities in gut microbiota in STZ-HFD treated mice, (3) significant differences at the phylum level exist between male and females both in control and STZ-HFD mice, (4) significant differences in BA metabolizing microbiota were present in male vs. female mice for both control and STZ-HFD groups.

STZ-HFD resulted in significantly higher levels of hepatic BAs in male mice than in female mice

Given the significant sex-associated differences in BA metabolizing microbiota, we next investigated the hepatic BA profiles in the mice. STZ-HFD treatment led to significantly altered liver BA concentrations in both sexes (Fig. 3A) as revealed by the OPLS-DA scores plot established using hepatic BA data (R2X = 0.801, R2Y = 0.739, Q2(cum) = 0.607). The hepatic BAs, 3-ketodeoxycholic acid (3-ketoDCA), taurocholic acid (TCA), taurolithocholic acid (TLCA), taurochenodeoxycholic acid (TCDCA), and 7-ketodeoxycholic acid (7-ketoDCA), were significantly increased in STZ-HFD-exposed mice compared to controls (Fig. 3 and Supplementary Fig. S3). More substantial increases in hepatic BA levels were observed in male STZ-HFD mice. Moreover, the increase in TLCA was only observed in males exposed to STZ-HFD and decreased levels of TLCA was observed, but with no statistical significance, in females (Fig. 3B). Among the significantly altered liver BAs, TCA, TCDCA, TLCA, 7-ketoDCA and 3-ketoDCA were significantly higher in males than in females after STZ-HFD intervention (Fig. 3C).

STZ-HFD also led to significant increases in fecal and serum BA levels. Fecal BAs, TDCA, GLCA, GDCA, and GCA were increased in male STZ-HFD mice relative to control. The results were more variable for female mice with TDCA, GDCA and GCA showing increases with STZ-HFD and GLCA slightly but significantly decreased in the model vs. control (Supplementary Fig. S3). GDCA, TDCA, and GLCA, secondary, microbiota metabolized BAs were significantly higher in male relative to female model mice (Supplementary Fig. S3).

Serum concentrations of TCDCA, TCA, ACA, TLCA, 3-ketoDCA, and 7-ketoDCA were lower in control male vs. female mice (Supplementary Fig. S4). Notably, serum levels of TCDCA, ACA, 3-ketoDCA and 7-ketoDCA were found to be significantly higher in STZ-HFD treated male relative to female mice. This flip from low to high concentration of specific BAs in male relative to female model mice reminds us of the flip in abundance discussed earlier for the Firmicutes/Bacteriodetes ratio which showed the ratio to go from low in control to high in STZ-HFD males and vice versa in female mice. Both of these results indicate sex specific changes upon STZ-HFD treatment. In order to determine whether BA transport into and out of the liver was affected by STZ-HFD and was responsible for the greater increase in hepatic BAs for STZ-HFD male mice, we next examined mRNA expression for the BA transporter genes.

Sex disparity was found in the expression of hepatic BA transporter mRNA

A qRT-PCR analysis revealed that genes involved in hepatic BA transport and synthesis were significantly different between sexes. In STZ-HFD treated male mice, hepatic FXR expression was significantly decreased. In STZ-HFD female mice FXR showed a decreasing trend that was not statistically significant. FXR is known to regulate the SHP and thus, accompanying the decrease in FXR mRNA expression was a decrease in mRNA expression for SHP for both male and female STZ-HFD mice. A depressed expression of FXR mRNA could also explain decreased expression of mRNA for BA transporters. The expression of mRNA for the major BA uptake transporter, the sodium-taurocholate cotransporting polypeptide (NTCP), was suppressed by STZ-HFD treatment (Fig. 3D). In addition, the bile salt export pump (BSEP) mRNA was found to be significantly decreased in male model mice relative to control. The female model mice exhibited BSEP mRNA levels that were significantly decreased relative to control but significantly increased with respect to male model results. Thus, these alterations in BA transport may lead to increased BA accumulation in hepatocytes and BA-induced liver injury. The expression of mRNA for BA synthesis, CYP7A1 and CYP7B1, was sinificantly down-regulated after STZ-HFD intervention in male STZ-HFD mice relative to control but the smaller decrease observed for female STZ-HFD mice was not statistically significant. Notably, in female mice, no significant difference in the mRNA expression of hepatic SHP, CYP7A1, and CYP7B1 was found between model and normal mice (Fig. 3D).

The mRNA expression of FXR, CYP7B1, BSEP and SHP, was lower and expression of NTCP and CYP7A1 were higher in normal female mice when compared to normal male mice. The expression of the above-mentioned genes was less altered in female mice than in male mice when exposed to STZ-HFD (Fig. 3D).

Hepatic expression of miRNAs was significantly different between STZ-HFD treated male and female mice

Since the expression of miRNAs are different between men and women with HCC and can be regulated by BAs20,21,22, we further analyzed miRNAs in liver tissues of male and female mice from the STZ-HFD model group and control group. As shown in Fig. 4, the tumor suppressive miRNAs, miR-26a, miR-26a-1, miR-192, miR-122, miR-22, and miR-125b were lower, whereas the tumor-promoting miRNAs, miR-10b and miR-99b were higher in males than in females in both the STZ-HFD group and the control group. As expected, the expression of tumor-suppressive miRNAs were decreased whereas the tumor-promoting miRNAs were increased much more in male mice than in female mice after STZ-HFD treatment, which presumably facilitated the development of liver tumors in male model mice.

BA-binding resin treatment can prevent HCC in male mice with recovered levels of differentially expressed BAs, gut microbiota and miRNAs

The levels of BAs including TCA, TCDCA, TLCA, 3-keto DCA, and 7-keto DCA, and the gut microbiota including Corynebacterium, Corynebacteriaceae, Rhodococcus, Nocardiaceae, Adlercreutzia, Bacillus, Bacillaceae, Staphylococcus, Staphylococcaceae, Lactobacillales, Desulfovibrio, Desulvibrionales, Clostrodium, and Clostridiales, were much higher in male STZ-HFD mice than in female STZ-HFD mice. The miRNAs were also significantly different between males and females. In a separate study using the STZ-HFD mice model we used a BA sequestrant, cholestyramine, to remove the intestinal BAs in male mice. We observed that depletion of secondary BAs in the intestine by cholestyramine prevented the STZ-HFD male mice from developing tumors, none in the cholestyramine treatment group (n = 8) developed tumor while all of the mice in the model group (n = 8) developed liver tumors (Fig. 5A and B). After cholestyramine administration, the levels of BAs, TCA, TCDCA, TLCA, 3-keto DCA and 7-keto DCA, were significantly decreased in the liver (Fig. 5C). The abnormal gut microbial profile and miRNAs were also normalized with cholestyramine intervention (Fig. 5D and E).

The BA metabolic profiles were significantly different between men and women

Results from our recently published data18 showed that the serum BA levels including TCA, TCDCA, TLCA, 7-keto DCA, 3-keto DCA, DCA and GCA were significantly different between healthy men and women, similar to the mice data (Supplementary Fig. S4). To verify the findings from the animal studies that differentially expressed BAs impact liver carcinogenesis in a sex dependent manner, we profiled the serum BAs in age and BMI matched liver disease patients and healthy participants of men and women. Serum BA measurement in liver fibrosis (n = 30, 15 males and 15 females aged 50–75 years), cirrhosis (n = 40, 20 males and 20 females aged 50–75 years), and HCC (n = 40, 30 males and 10 females aged 50–75 years) patients and healthy participants (n = 40, 20 males and 20 females aged 50–75 years) showed that the levels of BAs differentially expressed between healthy men and women were significantly increased in patients (both sexes) but with higher fold changes in men than in women in the development of liver disease (Fig. 6 and Supplementary Fig. S4).

Abstract

A specific and sensitive liquid chromatography–electrospray ionization–tandem mass spectrometric method was developed for the quantification of imatinib and its primary metabolite N-desmethyl imatinib in human plasma. Protein precipitation with methanol was used for sample preparation. High-performance liquid chromatographic separation was performed on a Thermo BDS Hypersil C18 column (4.6 × 100 mm, 2.4 µm) with methanol–water (55:45, v/v) containing 0.1% formic acid and 0.2% ammonium acetate as the mobile phase, using isocratic elution at a flow rate of 0.7 mL/min. Detection was conducted with positive electrospray ionization multiple reaction monitoring of the ion transitions at m/z 494 → 394 for imatinib, 480 → 394 for N-desmethyl imatinib and 297 → 110 for the internal standard (palonosetron). The assay was validated in the concentration ranges of 8–5,000 ng/mL for imatinib and 3–700 ng/mL for N-desmethyl imatinib. The quantification limits for imatinib and N-desmethyl imatinib were 8 and 3 ng/mL, respectively. The intra-day and inter-day precision values of the assay (expressed as percentage relative standard deviation) were less than 15% at all concentration levels within the tested range, and the accuracy values were between 85 and 115%. The established method was successfully applied to the pharmacokinetic study of imatinib mesylate capsules in 24 healthy Chinese volunteers.

Introduction

Chronic myelogenous leukemia is a myeloproliferative disorder associated with an abnormal BCR-ABL tyrosine kinase. Imatinib, a synthetic phenylaminopyrimidine derivative, which inhibits the tyrosine kinase with high selectivity, has been established as a highly effective therapy for chronic myelogenous leukemia (1) and gastrointestinal stromal tumors (2).

Imatinib is predominantly metabolized by CYP3A4 to N-desmethyl imatinib, which shows comparable biological activity to the parent drug (3). The activity of CYP3A4 displays large inter-individual variability. Therefore, a given dose of imatinib can yield very different circulating concentrations of the parent drug and its metabolites (4). Studies show that an adequate plasma concentration of imatinib is important for a good clinical response (5–6), which emphasizes the significance of therapeutic drug monitoring and pharmacokinetics investigation of imatinib.

The quantitation of plasma levels of imatinib and its primary metabolite is the key aspect of the pharmacokinetic study. Until now, the simultaneous quantification of imatinib and N-desmethyl imatinib in human plasma have been accomplished by liquid chromatography–ultraviolet detection (LC–UV) (7–14), LC–mass spectrometry (MS) (15–16) or tandem MS (MS-MS) (17–18) methods. The most frequently used method is LC–UV. However, UV detection suffers from a lack of sensitivity and a limited linear range. LC–MS-MS is considered to be the preferred method for the analysis of imatinib and N-desmethyl imatinib in complex biological samples because of its sensitivity and selectivity. However, some published works have involved complex gradient LC methods (15–18) to obtain good peak shapes and sensitivity, but a longer run time was required due to column re-equilibration. The most recent work (19) reported a gradient elution for the simultaneous determination of imatinib and N-desmethyl imatinib with a re-equilibration time of 3 min and a total run time of 6.0 min; the calibration ranges were from 10 to 2,000 ng/mL and analyte concentrations above the upper limit of quantification (ULOQ) had to be diluted and re-assayed.

Isocratic elution often provides stable baselines and constant LC–MS ionization efficiency for the analytes and requires no time-consuming preconditioning between individual runs. These are important aspects for high throughput bioassays. In this paper, a selective and robust method was developed for the simultaneous determination of imatinib and N-desmethyl imatinib by LC–MS-MS with higher sensitivity and wider linear range through the use of more straightforward isocratic elution using only a 3.8 min runtime The method showed good retention time reproducibility; therefore, it will be easy to transfer between laboratories and highly advantageous for pharmacokinetic studies. The feasibility of the proposed method has been demonstrated by the successful application to a pharmacokinetic study of imatinib mesylate capsules in healthy Chinese volunteers.

Experimental

Chemicals and reagents

Reference standards of imatinib mesylate (purity > 99.8%), N-desmethyl imatinib (purity > 99.8%), and the internal standard (IS) palonosetron hydrochloride (purity > 99.8%) were provided by Jiangsu Chia-Tai Tianqing Pharmacy Co. (Nanjing, China). Methanol [high-performance liquid chromatography (HPLC) grade] was supplied by Tedia Company (Fairfield, OH). All other chemicals and reagents were of analytical grade. All aqueous solutions were prepared with purified water (18.3 MΩ cm; Millipore, Billerica, MA).

Methods

Instrumentation and conditions

The LC–MS-MS system consisted of a Waters 2695 HPLC system (Waters, Milford, MA) with a quaternary gradient pump, an online vacuum degasser, a column oven and an autosampler, coupled to a Micromass Quattro micro triple-quadrupole mass spectrometer (Micromass, Manchester, UK) equipped with an electrospray ionization (ESI) interface. Data acquisition was performed with Masslynx 4.0 software (Micromass).

HPLC separation was performed on a Thermo BDS Hypersil C18 column (4.6 × 100 mm, 2.4 µm) maintained at 40°C with a mobile phase of methanol–water containing 0.1% formic acid and 0.2% ammonium acetate (55:45, v/v), which was delivered at 0.7 mL/min; 30% of the eluent was split into the inlet of the mass spectrometer for detection. A divert valve was used to divert the eluent to waste from 0 to 2.4 min. The autosampler was set at 4°C.

The mass spectrometer was operated in the positive ESI mode with the spray voltage set at 3 kV, nitrogen gas desolvation flow of 500 L/h at a temperature of 350°C and a sweep gas flow of 20 L/h. Quantification was performed with multiple reaction monitoring (MRM) by using argon gas collision induced dissociation and the following ion transitions: m/z 494 → 394, 480 → 394 and 297 → 110 for imatinib, N-desmethyl imatinib and palonosetron (IS), respectively, with the cone voltages all set at 30 V and the collision energy at 28 eV. Figure 1 shows the typical production scan spectra and the proposed patterns of fragmentation of the analytes and the IS.

Figure 1.

Product ion scan mass spectra: imatinib (A); N-desmethyl imatinib (B); palonosetron (IS) (C).

Figure 1.

Product ion scan mass spectra: imatinib (A); N-desmethyl imatinib (B); palonosetron (IS) (C).

Stock solutions

Stock solutions of imatinib and N-desmethyl imatinib at concentrations of approximately 800,000 and 90,000 ng/mL, respectively (both as the free base), were prepared in methanol–water (1:1, v/v). Working solutions of imatinib and N-desmethyl imatinib were prepared by serial dilution with the same solvent in the range from 80 to 50,000 ng/mL for imatinib and 30 to 7,000 ng/mL for N-desmethyl imatinib. The stock and working solutions of palonosetron hydrochloride (IS) were prepared similarly at 200,000 and 2,000 ng/mL as the free base. All solutions were stored under refrigeration (4°C) when not in use.

Plasma sample pretreatment

An aliquot of a 0.4 mL plasma sample was spiked with 40 µL of the IS and 40 µL of methanol–water (1:1, v/v), or 40 µL of the corresponding working standard solutions, for the preparation of the calibration plasma standards and quality control samples, followed by protein precipitation with the addition of 1.2 mL methanol and vortex-mixing for 1 min and centrifuging at 10,000 × g for 10 min at 4°C. The supernatant was transferred into an autosampler vial for LC–MS-MS analysis with an injection volume of 10 µL.

The calibration plasma standards of imatinib and N-desmethyl imatinib were prepared and analyzed separately to avoid possible cross-talk interferences, although these were not found in this study. To prepare the plasma calibration standards, an aliquot of 40 µL of each working standard solution was mixed precisely with 0.4 mL of blank plasma to produce the calibration standard in the ranges of 8.0–5000 ng/mL for imatinib and 3.0–700 ng/mL for N-desmethyl imatinib. The quality control (QC) samples were made at 16, 400 and 3,200 ng/mL for imatinib and 9, 90 and 500 ng/mL for N-desmethyl imatinib in the same way.

Method validation

The analytical method was validated for specificity, matrix effects, linearity, lower limit of quantification (LLOQ), accuracy, precision and recovery of measurements. The specificity was evaluated by comparing the chromatograms of six different batches of blank plasma with the corresponding spiked plasma to investigate the potential interferences near the retention times of either the analytes or the IS. The linearity of the method was determined by the analysis of a series of standard samples with concentrations from 8.0 to 5,000 ng/mL for imatinib and 3.0 to 700 ng/mL for N-desmethyl imatinib. The calibration curves were established through weighted linear least-squares regression of the peak area ratios (Y) of the analytes to the IS obtained against the corresponding concentrations (C, in ng/mL). Coefficients of correlation (r) were required to be 0.99 or better. The acceptance criterion for each back-calculated standard concentration above the LLOQ was ± 15% deviation from the nominal value, except at LLOQ. The LLOQ was defined as the concentration of the sample that could be quantified with less than 20% variation in precision (n = 6) and provided a signal-to-noise ratio ≥ 10; this was established by using six independent samples. The intra-batch and inter-batch accuracy and precision were determined by analysis of five replicates at three QC concentration levels. The criteria for acceptability of the data included accuracy within ± 15% deviation from the nominal values and precision within 15%. Recoveries of imatinib and N-desmethyl imatinib from plasma with protein precipitation by methanol were determined by comparing their peak areas in spiked plasma samples at three QC concentrations with those in samples prepared by spiking the blank plasma post-preparation with the same amounts of imatinib and N-desmethyl imatinib. The recovery of the IS was evaluated at 2,000 ng/mL. Matrix effects were caused by ionization competition occurring among imatinib, N-desmethyl imatinib, IS and endogenous co-eluting components. To evaluate the matrix effects, chromatographic peaks of imatinib, N-desmethyl imatinib and IS from the spiked solution after preparation were compared with those obtained by direct injection of the standard solutions prepared in the mobile phase at the QC concentrations.

The stability of the analytes was assessed by using triplicate spiked plasma samples containing imatinib and N-desmethyl imatinib at two concentration levels (30 and 1,600 ng/mL for imatinib; 18 and 350 ng/mL for N-desmethyl imatinib), which were analyzed after subjection to various storage and handling conditions over time periods that exceeded those applied to the actual study samples. The spiked stability samples were analyzed against a calibration curve that was obtained from spiked calibration standards prepared from freshly made stock solutions; the obtained concentrations were compared to the nominal concentrations. The mean concentration at each level should be within ± 15% of the nominal concentration. For freeze–thaw stability, samples were stored at −20°C for 24 h and thawed unassisted at room temperature. After complete thawing, samples were refrozen again under the same conditions. The freeze–thaw cycle was repeated three times and analysis was conducted on the third cycle. Short-term temperature stability was assessed by analyzing samples thawed at room temperature and kept at this temperature for 8 h, and the stability of the post-preparative samples kept at room temperature for 8 h was also evaluated. The stability of the post-preparative samples in the autosampler was conducted by re-analyzing processed samples kept in the autosampler at 4°C for 24 h. Long-term stability was determined by storing at −20°C for 75 days.

The stabilities of stock solutions of imatinib, N-desmethyl imatinib and IS (with an appropriate dilution, taking into consideration the linearity and measuring range of the detector) were evaluated by comparing the response of the stock solutions kept at 4°C for 50 days with that of freshly prepared solutions.

Method application

The validated method was applied for the determination of imatinib and its primary metabolite N-desmethyl imatinib in plasma samples in a pharmacokinetic study. Twenty-four healthy male Chinese volunteers were selected as subjects after clinical screening procedures. Each subject was fasted and administered a single oral dose of 400 mg of imatinib mesylate capsules. Venous blood samples of approximately 4 mL were collected into heparinized polypropylene tubes at pre-dose and 0.5, 1.0, 1.5, 2.5, 4.0, 6.0, 8.0, 12, 24, 36, 48, 72, 96 and 120 h after administration. Plasma was separated by centrifugation at 4,000 g for 5 min and stored at −20°C until analysis. The study was conducted at Xijing Hospital (Xi'an, China) in accordance with the principles of the Declaration of Helsinki after receiving approval from the independent ethics committee in the hospital. All subjects gave written consent for their participation after having been informed by the medical supervisor about the aim, course and possible risks of the study.

Results

Method validation

Under the proposed LC–MS-MS conditions, retention times for imatinib, N-desmethyl imatinib and the IS were 2.8 ± 0.1, 2.9 ± 0.1 and 2.8 ± 0.1 min, respectively, and the total run time was approximately 3.8 min. No obvious interferences from endogenous substances were observed. Typical chromatograms are shown in Figure 2 for blank plasma, plasma spiked with LLOQ standards of imatinib, N-desmethyl imatinib and IS, and plasma samples collected from a subject 4 h after dosing.

Figure 2.

Representative MRM chromatograms: blank plasma (A); blank plasma spiked with imatinib [8 ng/mL, LLOQ, retention time (tR) = 2.9 min] and the IS, palonosetron (200 ng/mL, tR = 2.8 min) (B); N-desmethyl imatinib (3 ng/mL, LLOQ, tR = 2.9 min) and the IS, palonosetron (200 ng/mL, tR = 2.8 min) (C); plasma sample of a subject 4 h after single oral administration of imatinib mesylate capsules (400 mg of imatinib free base) (D).

Figure 2.

Representative MRM chromatograms: blank plasma (A); blank plasma spiked with imatinib [8 ng/mL, LLOQ, retention time (tR) = 2.9 min] and the IS, palonosetron (200 ng/mL, tR = 2.8 min) (B); N-desmethyl imatinib (3 ng/mL, LLOQ, tR = 2.9 min) and the IS, palonosetron (200 ng/mL, tR = 2.8 min) (C); plasma sample of a subject 4 h after single oral administration of imatinib mesylate capsules (400 mg of imatinib free base) (D).

Suitable weighting factors were selected for linear regression because the F-tests and homoscedasticity tests, conducted by plotting residuals versus concentration, demonstrated the heteroscedasticity. Empirical weights of 1/Y, 1/C, 1/Y2 and 1/C2 were evaluated. The best weighting factor was chosen according to the percentage relative error (RE), which compares the regressed concentration computed from the regression equation obtained for each weighting factor with the nominal standard concentration. Results showed that the weighting factor of 1/C gave the least sum of absolute RE across the whole concentration range; thus, it was selected as the weighting factor. Good linear relationships were obtained over the ranges of 8–5,000 and 3–700 ng/mL for imatinib and N-desmethyl imatinib, respectively. Typical equations were Y = (2.188 ± 0.006028) C + (0.1794 ± 0.01397) (r = 0.9997 ± 0.0001) (n = 3) for imatinib and Y = (1.258 ± 0.01908) C + (0.01479 ± 0.001946) (r = 0.9997 ± 0.0001) (n = 3) for N-desmethyl imatinib. The accuracy observed for the mean of back-calculated concentrations for three calibration curves was within 97.26–106.3% and 90.90–106.2% for imatinib and N-desmethyl imatinib, respectively, whereas the inter-validation precision [percentage relative standard deviation (RSD)] of the back-calculated calibration standards ranged from 0.28 to 2.15% for imatinib and 0.10 to 3.28% for N-desmethyl imatinib. The accuracy, precision (RSD) and regression parameters of slope, intercept and correlation coefficient (r) calculated by weight (1/C) linear regression are summarized in Tables I and II. The LLOQ values were found to be 8 and 3 ng/mL for imatinib and N-desmethyl imatinib, respectively; these values are in agreement with the requirement in human pharmacokinetic study.

Table I

Accuracy and Precision (RSD) of Calibration Curve Data and Regression Parameters Calculated by Weighted (1/C) Linear Regression for Imatinib

STD-1 STD-2 STD-3 STD-4 STD-5 STD-6 STD-7 STD-8 STD-9 STD-10 Slope Intercept r
Nominal (ng/mL) 8.000 16.00 32.00 80.00 160.0 400.0 800.0 1,600 3,200 5,000 
Back-calculated (ng/mL) Validation 1 7.707 16.23 31.27 82.26 154.2 428.6 825.3 1,513 3,187 5,046 2.194 0.1896 0.9996 
Validation 2 7.663 16.39 31.79 81.61 154.1 423.1 816.7 1,529 3,206 5,025 2.187 0.1635 0.9998 
Validation 3 7.972 15.89 31.63 80.96 153.3 424.2 825.4 1,524 3,213 5,014 2.182 0.1852 0.9997 
Overall mean 7.781 16.17 31.56 81.61 153.9 425.3 822.5 1,522 3,202 5,028 2.188 0.1794 0.9997 
SD 0.1671 0.2528 0.2645 0.6506 0.4902 2.907 4.979 8.363 13.30 15.84 0.006028 0.01397 0.0001 
RSD (%) 2.15 1.56 0.84 0.80 0.32 0.68 0.61 0.55 0.42 0.31 0.28 7.79 
Accuracy (%) 97.26 101.1 98.64 102.0 96.16 106.3 102.8 95.12 100.1 100.6 
Bias (%) –2.74 1.07 –1.36 2.01 –3.84 6.32 2.81 –4.88 0.06 0.56 
STD-1 STD-2 STD-3 STD-4 STD-5 STD-6 STD-7 STD-8 STD-9 STD-10 Slope Intercept r
Nominal (ng/mL) 8.000 16.00 32.00 80.00 160.0 400.0 800.0 1,600 3,200 5,000 
Back-calculated (ng/mL) Validation 1 7.707 16.23 31.27 82.26 154.2 428.6 825.3 1,513 3,187 5,046 2.194 0.1896 0.9996 
Validation 2 7.663 16.39 31.79 81.61 154.1 423.1 816.7 1,529 3,206 5,025 2.187 0.1635 0.9998 
Validation 3 7.972 15.89 31.63 80.96 153.3 424.2 825.4 1,524 3,213 5,014 2.182 0.1852 0.9997 
Overall mean 7.781 16.17 31.56 81.61 153.9 425.3 822.5 1,522 3,202 5,028 2.188 0.1794 0.9997 
SD 0.1671 0.2528 0.2645 0.6506 0.4902 2.907 4.979 8.363 13.30 15.84 0.006028 0.01397 0.0001 
RSD (%) 2.15 1.56 0.84 0.80 0.32 0.68 0.61 0.55 0.42 0.31 0.28 7.79 
Accuracy (%) 97.26 101.1 98.64 102.0 96.16 106.3 102.8 95.12 100.1 100.6 
Bias (%) –2.74 1.07 –1.36 2.01 –3.84 6.32 2.81 –4.88 0.06 0.56 

View Large

Table II

Accuracy and Precision (RSD) of Calibration Curve Data and Regression Parameters Calculated by Weighted (1/C) Linear Regression for N-Desmethyl Imatinib

STD-1 STD-2 STD-3 STD-4 STD-5 STD-6 STD-7 STD-8 STD-9 STD-10 Slope Intercept r
Nominal (ng/mL) 3.000 7.500 15.00 30.00 60.00 87.50 175.0 350.0 525.0 700.0 
Back-calculated (ng/mL) Validation 1 3.301 7.759 13.80 28.45 61.64 83.34 176.7 357.7 509.7 709.8 1.236 0.01535 0.9996 
Validation 2 3.154 7.539 13.53 31.42 61.95 83.42 176.4 357.1 509.7 707.2 1.270 0.01263 0.9997 
Validation 3 3.099 7.466 13.57 31.22 61.31 86.41 178.2 357.1 511.4 701.3 1.268 0.01640 0.9998 
Overall mean 3.184 7.588 13.64 30.36 61.64 84.39 182.2 357.3 524.8 726.3 1.258 0.01479 0.9997 
SD 0.1044 0.1527 0.1468 1.6602 0.3201 1.748 1.004 0.3404 1.021 4.376 0.01908 0.001946 0.0001 
RSD (%) 3.28 2.01 1.08 5.47 0.52 2.07 0.55 0.10 0.19 0.60 1.52 13.15 
Accuracy (%) 106.2 101.2 90.90 101.2 102.73 96.45 104.1 102.1 99.97 103.8 
Bias (%) 6.15 1.18 –9.10 1.20 2.73 –3.55 4.09 2.08 –0.03 3.75 
STD-1 STD-2 STD-3 STD-4 STD-5 STD-6 STD-7 STD-8 STD-9 STD-10 Slope Intercept r
Nominal (ng/mL) 3.000 7.500 15.00 30.00 60.00 87.50 175.0 350.0 525.0 700.0 
Back-calculated (ng/mL) Validation 1 3.301 7.759 13.80 28.45 61.64 83.34 176.7 357.7 509.7 709.8 1.236 0.01535 0.9996 
Validation 2 3.154 7.539 13.53 31.42 61.95 83.42 176.4 357.1 509.7 707.2 1.270 0.01263 0.9997 
Validation 3 3.099 7.466 13.57 31.22 61.31 86.41 178.2 357.1 511.4 701.3 1.268 0.01640 0.9998 
Overall mean 3.184 7.588 13.64 30.36 61.64 84.39 182.2 357.3 524.8 726.3 1.258 0.01479 0.9997 
SD 0.1044 0.1527 0.1468 1.6602 0.3201 1.748 1.004 0.3404 1.021 4.376 0.01908 0.001946 0.0001 
RSD (%) 3.28 2.01 1.08 5.47 0.52 2.07 0.55 0.10 0.19 0.60 1.52 13.15 
Accuracy (%) 106.2 101.2 90.90 101.2 102.73 96.45 104.1 102.1 99.97 103.8 
Bias (%) 6.15 1.18 –9.10 1.20 2.73 –3.55 4.09 2.08 –0.03 3.75 

View Large

The intra-batch and inter-batch precision values were ≤8.3% for imatinib and ≤ 5.1% for N-desmethyl imatinib. The accuracy, expressed as deviation percentage, was found to be within the acceptable range. The recovery data show that the sample preparation method was able to produce consistent, precise, reproducible and absolute recovery for the analytes and IS (>97%). No obvious matrix effects were found for the analytes and IS: the ratios of the peak responses ranged from 85 to 115%, which were within the acceptable limits. The results of accuracy, precision, recovery and matrix effect of imatinib and N

Leave a Comment

(0 Comments)

Your email address will not be published. Required fields are marked *