|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Lipid Research, Vol. 47, 2340-2345, October 2006
Copyright © 2006 by American Society for Biochemistry and Molecular Biology
Methods |


* Department of Metabolism and Pharmacokinetics, Ranbaxy Research Laboratories, Jamia Hamdard, Delhi, India
Department of Pharmacology, Jamia Hamdard, Delhi, India
Published, JLR Papers in Press, July 21, 2006.
1 To whom correspondence should be addressed. e-mail: jyoti.paliwal{at}ranbaxy.com
| ABSTRACT |
|---|
|
|
|---|
Supplementary key words statin biomarker liquid chromatography-tandem mass spectrometry
| INTRODUCTION |
|---|
|
|
|---|
|
The main challenge in developing and validating a method for determining MVA in human plasma was that MVA is a polar, endogenous moiety that circulates in the blood stream at nanogram levels. In most methods, the extraction of MVA from plasma was carried out using ion-exchange resins in the form of mevalonolactone (MVAL) (11, 12). Complicated procedures such as column switching and gradient flow with long run times were followed (11). In a modified assay procedure, a polar-end-capped reverse-phase liquid chromatography column was used for the quantification of plasma MVA over a calibration range of 0.550 ng/ml in human plasma (12). This assay had the advantages of shorter run time and isocratic flow.
These methods have reported recovery to be 5087%. The procedure followed does not capture the effect of any constant impurity/substance that may suppress ionization. The exact recovery can be obtained by comparing the response of processed spiked plasma with that of aqueous samples at the same concentration. The matrix effect can be evaluated by comparing spiked processed plasma blanks with aqueous samples at the same concentration. By knowing the recovery and matrix effect, the sensitivity of the method can be improved.
A specific and sufficiently sensitive method was required for the quantification of plasma MVA levels in clinical trials. The reported normal range of human plasma MVA levels is 1.511.8 ng/ml (13). After statin treatment even at the highest dose, the percentage decrease in plasma MVA levels has been shown to be 3050% (11, 14, 15). Therefore, the lowest limit of quantification was kept at 0.5 ng/ml and the upper limit was 50 ng/ml. The use of surrogate matrix (water) has been reported (11, 12) and is necessary to achieve lower levels of quantification.
In this study, a simple, robust, and reproducible method has been developed and validated to estimate MVA concentrations in human plasma. Equilibration time for the conversion of MVA to MVAL under acidic pH was further optimized. An approach to differentiate the matrix effect from recovery was performed under validation. The validated method was applied to quantify plasma MVA concentrations in rat to determine the effect of rosuvastatin on plasma MVA levels.
| EXPERIMENTAL PROCEDURES |
|---|
|
|
|---|
Column liquid chromatography
The column was a HyPurity Advance column (50 mm x 4.6 mm, 5 µm particle size; Thermo Electron Corp.). The column was kept at ambient temperature. The mobile phase consisted of ammonium formate buffer (10 mM, pH 8.0, adjusted with liquid ammonia) and acetonitrile (70:30, v/v). The flow rate was 0.8 ml/min, and the total run time was 3 min.
Mass spectrometry
The liquid chromatograph (Agilent 1100; Agilent Technologies, Inc., Palo Alto, CA) was coupled to a mass spectrometer with a turbo electrospray ion source (4000 Qtrap; Applied Biosystems, Foster City, CA) and was used in negative ionization mode with the following source settings. The turbo ion-spray interface was maintained at 530°C with zero air nebulization. The zero air was kept at a pressure of 70 p.s.i. The turbo ion-spray drying gas (zero air) was kept at a pressure of 70 p.s.i. The collision-activated dissociation gas pressure was 7 p.s.i., and the curtain gas pressure was 30 p.s.i.; turbo ion-spray voltage was 3,500 V. Declustering potential was 35 V; entrance potential was 10 V; collision energy was 20 V; collision cell exit potential was 1 V; and channel electron multiplier was 2,600 V. The multiple reaction monitoring pair monitored was m/z 147
59 for MVA and m/z 154
59 for D7-MVA, with a dwell time of 200 ms. The autosampler cooler was maintained at 10°C. Analyst software (version 1.4; Applied Biosystems) was used for data registration and calibration.
Sample collection
After validation, to study the effect of rosuvastatin on plasma MVA levels in rat (n = 5), the method was used to quantify plasma MVA concentrations after 10 mg/kg oral administration. Blood samples were collected in tubes containing lithium heparin before drug administration and 0.5, 1, 1.5, 6, 16, and 24 h thereafter. The samples were centrifuged at 2,500 g for 10 min. The separated plasma was stored at 70°C until analysis.
Sample preparation
Samples were thawed in water. Plasma aliquots of 500 µl were added to a glass tubes to which were also added IS (100 µl, 200 ng/ml), 0.1 N HCl (1 ml), and water (0.5 ml); the tubes were than vortex-mixed. The sample solution was allowed to equilibrate for 30 min to convert MVA to MVAL. Each sample solution was individually transferred to a solid-phase extraction cartridge (IST ENV+; 100 mg/3 ml) that had been preconditioned with methanol (1 ml) followed by 0.1 N HCl (1 ml). Each cartridge was washed with 0.1 N HCl (1 ml) followed by water (1 ml) and 15% methanol in water. The cartridges were allowed to dry. The analytes were eluted with 3 x 0.5 ml of methanol. The resulting methanol extract solutions were evaporated to dryness under a stream of nitrogen at 15 p.s.i. and 40°C bath temperature for 15 min. The residues were reconstituted in 0.2% ammonium hydroxide solution (100 µl) to convert MVAL to MVA. Aliquots of 10 µl were injected into the LC-MS/MS apparatus for analysis.
Standard curves
The calibration curve (CC) standards were prepared in water by adding known amounts of MVA. Lower limit of quantification (LLOQ) QC and low-quality control (LQC) samples were obtained by spiking MVA in water; the final concentrations were 0.5 and 1.9 ng/ml, respectively. Middle-quality control (MQC) and high-quality control (HQC) samples were obtained by spiking in plasma with concentrations of 21.9 and 41.3 ng/ml, respectively. The bulk-spiked CC and QC samples were stored at 70°C. The endogenous MVA level obtained in plasma (12.2 ng/ml) was added to spiked plasma samples to obtain corrected concentrations for MQC and HQC samples. All calibration curves consisted of one blank sample and eight calibration points in the concentration range of 0.550 ng/ml. The concentrations were corrected for potency and amount weighed. The resulting peak area ratios were plotted against the concentrations.
Validation
Specificity
The approach of using water for the preparation of CC and lower QC standards has been verified and reported previously (11, 12). A specificity exercise was performed for both water and plasma. Individual blank plasma samples, LLOQ QC samples, and water (blank) (n = 6) were prepared according to the sample preparation procedure described above and screened for interference.
Recovery The recovery exercise was performed at all QC levels by comparing the response (area) of processed QC samples with those of directly injected QC samples. The dilutions were made in 0.2% ammonium hydroxide solution to keep conditions the same.
Matrix effect To study the matrix effect, blank plasma samples were processed and spiked later to obtain MQC and HQC concentrations. The response (area) was compared with directly injected samples at MQC and HQC levels.
Inter-assay and intra-assay imprecision and accuracy Inter-assay and intra-assay imprecision and accuracy were evaluated by spiking known amounts of MVA and IS in plasma (n = 5). Four different concentrations were used, and samples were prepared according to the procedure mentioned above. Intra-assay imprecision and accuracy were assessed within one batch, whereas inter-assay imprecision and accuracy were assessed on three separate occasions.
Dilution integrity and partial volume analysis MQC and HQC samples were diluted two and four times with water for dilution integrity. Partial volume analysis was performed at one-half and one-fourth the processing volumes at MQC and HQC levels. The samples were processed according to the procedure mentioned above in five replicates.
Stability The stability of MVA was studied in human plasma and water at room temperature (bench top) for 6 h and in an autoinjector for 20 h. The bulk-spiked plasma and water samples stored at 70°C underwent three freeze-thaw cycles. Stock solution stability studies were performed at room temperature for 4 h and in refrigeration for 22 days. In addition, a long-term (28 days) stability study was done in human plasma stored at 70°C. The stability of D7-MVA was assessed in an autoinjector for 20 h. Stock solution stability studies were performed at room temperature for 4 h and in refrigeration for 112 days. Stock solution stability studies were carried out at the MQC level, and working IS was prepared from fresh and refrigerated stock solutions. The drug and IS response ratios of stored and fresh stocks were compared. In other stability studies, five replicates of LQC and HQC were analyzed.
| RESULTS |
|---|
|
|
|---|
|
The imprecision of the assay was measured by the percentage coefficient of variation over the concentration range of LLOQ QC, LQC, MQC, and HQC samples during the course of validation. The accuracy of the assay was defined as the absolute value of the ratio of the calculated mean values of the QC samples to their respective nominal values, expressed as percentages. Within-batch precision ranged from 1% to 18%, and within-batch accuracy ranged from 97% to 116%. Intra-assay precision ranged from 1% to 17%, and intra-assay accuracy ranged from 98% to 109%. Inter-assay precision ranged from 3% to 12%, and inter-assay accuracy ranged from 99% to 108%.
The room temperature stock stability at 4 h was 101% for MVA and 111% for IS. The refrigerated stock solution stability on the 22nd day for MVA was 105%, and that for IS on the 112th day was 100%. The autoinjector stability results demonstrate that MVA and IS are stable for 20 h. The mean stability ranged from 97% to 102% for MVA and from 98% to 102% for IS. The mean stability of MVA in human plasma ranged from 97% to 102% and 92% to 98% for one and three freeze-thaw cycles, respectively. During bench-top stability analysis, MVA was found to be stable up to 4 h, and the mean stability ranged from 94% to 102%. MVA was found to be stable for up to 28 days of storage (plasma) below 50°C, and the mean stability ranged from 100% to 107% (Table 1 ). The absolute recovery of MVA and IS was calculated for replicate spiked QC samples (MQC and HQC). Results indicate overall recoveries of 21% for MVA and 21% for IS. The percentage matrix effect was 46% for analyte and 73% for IS.
|
|
|
| DISCUSSION |
|---|
|
|
|---|
The extraction of MVA is highly pH-specific (11, 12). The sample was equilibrated with 0.1 M HCl to convert MVA to MVAL. The equilibration time was optimized to 30 min after evaluating response at 15, 30, 45, and 60 min. Both liquid-liquid and solid-phase extraction procedures were assessed initially for the extraction of MVA from plasma. Better sample cleanup and reproducibility were obtained using polystyrene-divinylbenzene cartridges. The method described is sensitive, selective, precise, and accurate for the determination of MVA in human plasma at very low concentrations (<1 ng/ml) over a concentration range extending up to 50 ng/ml. The method has been validated for a maximum batch size of 103 samples (with a total run time of 6.5 h). The use of water as a surrogate matrix allows the assay to be used down to the required lower limit of 0.5 ng/ml, which would not be possible if standards were prepared in plasma containing endogenous MVA.
In previously reported methods, recovery was determined by comparing processed plasma concentrations against spiked processed blank plasma samples (11, 12). This approach to determine recovery does not capture the effect of any constant endogenous substance that may reduce the response. Hence, the calculated recovery will not be the actual recovery. During validation, recovery was calculated by comparing processed plasma QC sample concentrations against the same aqueous concentrations for MVA and D7-MVA prepared in reconstitution solution (0.2% ammonium hydroxide).
The matrix effect was determined by comparing spiked processed blank plasma QC samples against the same aqueous concentrations for MVA and D7-MVA. A significant matrix effect was observed. The low recovery is attributable to the matrix effect. During method development, six different plasma samples were spiked to add 10 ng/ml to endogenous MVA. The response observed in spiked plasma was increased proportionately compared with that in blank plasma, and similar results were seen with QC samples. It can be concluded that the presence of some constant endogenous substance other than MVA or any reagent effect during sample processing can contribute to the matrix effect. Hence, this method can be used to obtain accurate plasma MVA concentrations.
The method was successfully applied to estimate rat plasma MVA levels after a single oral dose of rosuvastatin (10 mg/kg). Although the rat is not an adequate model for studying lipid metabolism, it has been used extensively to study the mevalonate pathway (2, 8, 16). No diurnal rhythm was observed, unlike in humans, and the normal range of plasma MVA in rats was found to be 2040 ng/ml, as reported previously (16). A significant decrease was observed in rat plasma MVA levels after rosuvastatin administration.
In conclusion, a simple, sufficiently sensitive, and reproducible method was developed for the quantification of MVA in plasma, and the developed method has been validated. The stability studies demonstrated that MVA was stable during normal assay procedures and in long-term frozen storage conditions (below 50°C). This should allow clinical samples to be stored and analyzed efficiently. The matrix effect and recovery should be differentiated to improve the sensitivity of the method and to capture the effect of any endogenous interference. The use of MVA as biomarker needs to be explored further.
| ACKNOWLEDGMENTS |
|---|
Manuscript received May 15, 2006 and in revised form June 16, 2006 and in re-revised form July 21, 2006.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Honda, Y. Mizokami, Y. Matsuzaki, T. Ikegami, M. Doy, and H. Miyazaki Highly sensitive assay of HMG-CoA reductase activity by LC-ESI-MS/MS J. Lipid Res., May 1, 2007; 48(5): 1212 - 1220. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Journal of Biological Chemistry |
| Molecular and Cellular Proteomics | ASBMB Today |