1. INTRODUCTION:
Cytochrome P450 3A (CYP3A) enzymes play an important role in the phase I metabolism of xenobiotics and contribute to the biotransformation of approximately 50% of the drugs currently in the market. Drug–drug interactions in humans based on CYP3A enzyme inhibition are generally considered to be undesirable1.
To predict such interactions at the early drug-discovery stage, CYP3A inhibition studies are conducted using probe drugs in animal models2. Among those animals, rats are widely used in non-clinical studies, such as toxicity and Absorption, Distribution, Metabolism and Excretion (ADME) and pharmacological studies and are considered to be the most convenient animal3.
There are two approaches to study the drug interactions and basal catalytic activity of these metabolic enzymes, viz. In-vitro and in-vivo approaches. Amongst these, an in-vivo approach so called non-invasive procedures utilizes readily available fluids such as plasma, saliva or excretions such as urine. These non-invasive measures form the basis for measuring in vivo metabolic activity and demands the use of model compounds or, as currently termed, in vivo probes, has been extensively applied for these purposes. 4
Midazolam (8-chloro-6-(2-fluorophenyl)-1-methyl-4H-imidazo-[1, 5-a] [1, 4] benzodiazepine, CAS [59467-70-8]) (Fig. 1A) is a short-acting benzodiazepine, commonly used in intravenous anesthesia induction, short-term sedation and oral hypnotic medication. In humans, midazolam is rapidly and almost completely metabolized to its primary l-hydroxy metabolite and, to a much lesser extent, to 4-hydroxymidazolam. Both of these pathways are selectively mediated by CYP3A. In addition, both intestinal and hepatic microsomes exhibit high midazolam hydroxylation activity, which in the case of the liver is significantly correlated with the drug’s systemic clearance5. Moreover, scale-up of such in vitro measures was found to provide an excellent prediction of the in vivo extraction ratios of the two organs6. A further characteristic of midazolam’s metabolism is that it is readily altered by administration of known CYP3A inhibitors (e. g., azole antifungal agents, certain macrolide antibiotics, HIV protease inhibitors, and grape fruit juice) and inducing agents (e. g., anticonvulsants and rifampin). Collectively, these characteristics fulfill most of the criteria usually accepted for validation of an in vivo probe7.
In spite of 70-fold higher amounts of CYP3A4 in the liver than in the intestine, it was found that the unbound intrinsic clearance of midazolam by liver and intestinal microsomes was similar. The Thomas Jefferson University (TJU) Clinical Research Unit has compared the oral midazolam pharmacokinetics to assess the effect of various agents on CYP3A4-mediated metabolism, including rifampin, ketoconazole, and mibefradil. These studies have demonstrated that oral midazolam pharmacokinetics is useful to examine drug interactions involving hepatic and intestinal CYP3A4. Moreover, the putative probe drug for CYP3A, midazolam belongs to BCS Class I, where transporters effects are generally unimportant in vivo.
Being the probe for CYP3A family midazolam is anticipated to use extensively in pharmacokinetic interaction studies and rats are reliable predictor of such pharmacokinetic interactions in humans. This complete system can therefore be used in carefully selected pharmacokinetic drug-drug interaction studies, from overall in vivo clearance measurements to subcellular or molecular mechanism studies, all in the same single integrated model. Such pre-clinical pharmacokinetic investigations require the support of a fast and reliable bio-analytical methodology for the measurement of the drug involved.
The pharmacokinetic studies of midazolam in rats shows the highest plasma level 120 ± 35. 38 ng ml-1 for single oral dose 5 mg kg-1. Taking into account this fact, a quantitation limit of 5 ng ml-1 is required for single dose design pharmacokinetic study of midazolam. To best of our knowledge none of the previously published high performance liquid chromatographic method (HPLC) with ultraviolet or fluorescence detection had such lower limit of quantitation. Also these methods are designed for human biological samples in relatively large volume, typically 0.5-1.0 ml (Table 1). Preclinical studies exclusively demands small amount of blood samples to be collected as blood contained in rats or mouse is less as compare to clinical study samples. This situation makes it impossible to get desired set of blood samples from one small animal to perform pharmacokinetic study.
A review of literature reveals number of assay methods have been reported over the period of time for midazolam. In most of these published methods the limit of quantification ranged from 0.10 to 10.0 ng ml-1. 0.10 ng ml-1 limit of quantification was possible from 1000 µL of plasma. A bioanalytical method for the quantitation of midazolam using UV detection has also been reported by Ha et al. Although the limit of quantitation (LOQ) of this method is 5 ng ml-1, a sample volume of 500 µl is needed for analysis and method has been developed in human plasma and urine8. Recently Jan et al has reported the LOQ of 7.8 ng ml-1 with a sample volume of 450 µl9. Elbarbry et al. recently published the HPLC-UV method with 200 µl plasma and achieved LOQ of 10.0 ng ml-1. But the method has been developed for human plasma samples and reveals 25% spiking which is not accepted by regulatory bodies10. Such high spiking volumes of drug constituents may cause protein precipitation of plasma proteins. Recently many methods using HPLC coupled with mass spectrometric detection also have been published which are more specific and sensitive11-14. However these methods are not affordable for most laboratories because of their specialty requirements and high equipment cost.
Under the scope of this view, the aim of our research work was to develop more precise, accurate, rugged and reliable method for midazolam determination in rat plasma that proves to be of immense use for carrying out the preclinical pharmacokinetic studies efficiently in terms of small sample volume, very low concentration and a relatively short run time. Taking into account the relative non-polar nature of the analyte (Log P 2.25), liquid-liquid extraction has been proven to be an efficient and highly preferred technique in most of the published methods.
We have developed a simple isocratic method. This research work exploits the high selectivity and sensitivity of ultraviolet (UV) detector to develop and validate a method for the determination of midazolam in rat plasma. The described HPLC assay method needs a plasma volume of 200µl, which is suitable for carrying out future pharmacokinetic studies in rats or other animals larger than rats. The lower solvent consumption with a short chromatographic run (<12. 5 min) indicates that this method can also be optimized for high throughput quantitative analysis, such as monitoring the plasma level of the compound in clinical patients. In addition to this, pharmacokinetics of midzolam was studied in a rat model that were administered a single oral dose of 5.0 mg midazolam per kg body weight, and single intra venous dose of 1.0 mg midazolam per kg body weight in order to validate the method with real samples.
2. EXPERIMENTAL:
2. 1 Reagents and materials:
Midazolam (8-chloro-6-(2-fluorophenyl)-1-methyl-4H-imidazo-[1, 5-a][1, 4] benzodiazepine, CAS [59467-70-8], (99. 50%, as is basis) was supplied by Wockhardt Limited (Aurangabad, India). Diazepam (99. 02%, as is basis) was supplied by IPCA Labs (Mumbai, India). HPLC-grade reagents such as acetonitrile, methanol, Methyl-ter-Butyl Ether (MTBE) were purchased from Rankem Chemicals (Delhi, India). Analytical grade reagents such as potassium dihydrogen phosphate, triethylamine, sodium hydroxide, potassium hydroxide were purchased from Merck (Mumbai, India). HPLC-grade water was produced in the laboratory by a Milli-Q purification system (Millipore Corp., Vienna, Austria).
2. 2 Apparatus:
The analysis was carried out on a Shimadzu LC-10 series chromatographic system (Shimadzu Corporation, Kyoto, Japan). More precisely, the system consisted of a model SCL-10A controller unit, a model DGU-2A degasser unit, two LC-10AT piston pumps, SIL10AD refrigerated autosampler and a Model SPD-10AVP UV- detector. System control, data acquisition and processing were performed with a PC-Pentium IV Processor personal computer operated with Microsoft Windows NT version 4. 0 and Shimadzu CLASS-VP version 6. 12 SP1 chromatography software with the system suitability option installed. Standard substances were weighed on AY 220 Shimadzu analytical balance. A glass vacuum-filtration apparatus (Alltech Associates) was employed for the filtration of buffer solution using 0. 45 µm filter obtained from Pall Life Sciences. Degassing of the mobile phase was performed by sonication in Oscar Micro clean-103 Ultrasonic bath. Sanyo ultra low temperature freezer (Leicestershire, UK) was used to store the study samples. A model Genie-2 Spinix vortex mixer, a REMI C24 refrigerated centrifuge (REMI, India) and TurboVap LV Evaporator (Caliper Life Sciences, Hopkinton, MA, USA) were employed for sample pretreatment.
2. 3 Chromatography:
The mobile phase consisting of a mixture of potassium dihydrogen phosphate buffer (containing 3.402g of potassium dihydrogen phosphate and 0.6 ml of triethylamine per liter, pH adjusted to 7.2 with potassium hydroxide) and acetonitrile in the ratio 50:50 (v/v) was delivered isocratically at a flow rate of 1.0 ml min-1. Following its preparation, the mobile phase was filtered under vacuum through a 0.45 μm membrane filter and ultrasonically degassed prior to use. The chromatographic separation was achieved by a Vydac C18 monomeric (250 × 4. 6 mm inner diameter × 5-µm particle size) column. Diazepam (Fig. 1B) was used as an internal standard (I. S.). The column temperature was maintained constant at 250 C using a thermostatically controlled column oven. Ultraviolet measurements were done at 245 nm wavelength. The chromatographic run time for each analysis was 12. 5 min.
Figure 1. (A) Midazolam and (B) Diazepam, an internal standard.
2. 4 Preparation of standard solutions:
Midazolam was dissolved in methanol to prepare a primary stock solution at a concentration of 1000 µg ml-1. Corrections to the theoretical concentration were performed according to the degree of standard substance impurities. Intermediate stock solutions (100 and 10 µg ml-1), in duplicate, were prepared from the primary stock solution. Primary and intermediate stock solutions were kept at 40 C and remained stable for at least 15 days. Working stock solutions of midazolam were prepared using a mixture of methanol–water, 50:50, v/v (diluent). The I. S. working solutions were prepared in a similar manner, providing finally a plasma concentration of 400 ng ml-1. The intermediate stock solutions were prepared weekly, while working stock solutions used for the calibration curves were prepared daily.
Calibration standards in plasma (5, 10, 25, 50, 100, 200, 400, 800, 1000, 2000 and 3000 ng ml-1) samples were prepared by spiking blank rat plasma with 10 µL of working stock solutions of midazolam. Quality control (QC) samples at four different levels were independently prepared at concentrations of 5 ng ml-1 (LOQ-QC), same concentration as the lowest non-zero standard), 15 ng ml-1 (LQC), 500 ng ml-1 (MQC) and 1500 ng ml-1 (HQC, High QC) of midazolam in the same manner. The quality control samples were prepared from a stock solution that was different from the one used to generate standard curve samples. These quality control samples were used to investigate intra- and inter-run variations.
2. 5 Sample preparation:
Ten microlitre of I. S. solution (8 μg ml-1) was added to 200 µL plasma standard or sample followed by 200 µL of 0. 5M sodium hydroxide. The sample was vortexed for 30s and 1. 8 ml of MTBE was added for drug extraction. The mixture was vortexed again and centrifuged at 2500×g for 10min. The organic phase was collected and evaporated to dryness in a water bath at 500 C under a gentle stream of nitrogen. The dried residue was reconstituted with 200 µL of diluent and vortexed for 1min. The supernatant was transferred entirely into a 250 µL vial-insert and a volume of 45 µL was injected into the HPLC system.
2. 6 Validation of method:
Validation of the developed method allowed us to conform the linearity over the tested concentration range and to assess precision (repeatability and intermediate precision), trueness, and selectivity of the analytical method. To validate these different criteria, two kinds of plasma samples were prepared: calibration and validation samples corresponding to QC samples used in the routine analysis. The validation range was selected on the basis of the preliminary experiments to cover the expected midazolam concentrations in the animal studies. These concentrations were between 5 and 3000 ng ml-1 in the rat plasma.
2. 6. 1 Selectivity:
Selectivity is generally defined as the lack of interfering peaks at the retention times of the assayed drug and the internal standard in the chromatograms. The selectivity of the assay was investigated by processing and analyzing (a) blanks prepared from six independent lots of control plasma and (b) a group of potentially co-administrated drugs with midazolam as CYP 3A4 inhibitors i. e ketoconazole, itroconazole, erythromycin etc and inducers i. e rifampin, phenytoin etc in a final concentration of 5 µg ml-1in plasma. The retention times for these drugs under the chromatographic conditions for the midazolam assay were determined. The method is selective if the response of interfering peaks at the retention time of drug is less than 20% of mean response of six extracted LOQ-QC samples.
2. 6. 2 Linearity:
The linearity of an analytical procedure is its ability (within a given range) to obtain test results which are directly proportional to the concentration (amount) of analyte in the sample. The linearity of assay for the test compounds was evaluated with a total of five calibration standards over the concentration range 5–3000 ng ml-1. Calibration curves were constructed by linear least-squares regression analysis plotting of peak-area ratios (midazolam /I. S.) versus the drug concentrations. The linearity of the calibration curve was tested and evaluated using both linear regression model of internal standard calibration (ISTD) curve. Coefficients of calibration equation and the correlation coefficient (r) were expressed. The calibration model was accepted, if (a) coefficient of correlation was greater than or equal to 0. 998; (b) residuals were within ±20% at the lower limit of quantification and within ±15 % at all other calibration levels; (c) no two consecutive calibration curve standards fail to meet the above acceptance criteria; and (d) at least 2/3 of the standards meet this criterion. Five calibration curves were prepared to ensure that the regression model was the most accurate for quantitative purposes. These standard curves were used to determine the concentrations of the QCs and to back calculate the relative error (residuals) for each point in the standard curve.
2. 6. 3 Detection and quantification limits (sensitivity):
The detection limit (LOD) was defined as the lowest concentration level resulting in a peak area of three times the baseline noise. The ratio of signal size to that of noise was calculated by equation 2H/h, where H is the height of the peak corresponding to the component concerned in a chromatogram obtained with the prescribed reference solution and h is the range of the background noise in the chromatogram obtained after the injection of a blank.
The limit of quantification (LOQ) was determined as the lowest concentration on the standard calibration curve that provided a peak area with a signal-to-noise ratio higher than 10 with a precision ≤20% and accuracy of 80–120% of its nominal value.
2. 6. 4 Assay precision and accuracy:
The precision of an analytical procedure expresses the closeness of agreement between a series of measurements obtained from multiple sampling of the same homogeneous sample, while, the accuracy of an analytical method describes the closeness of the test results obtained by the method to the true (nominal) value of the analyte. Accuracy, intra- and inter-day precisions of the method was determined for midazolam according to FDA guidance for bioanalytical method validation15. The run consisted of a calibration curve plus six replicates of each QC sample (LOQ-QC, LQC, MQC and HQC) on the same day. The inter-day accuracy and precision were assessed by the analysis of six samples (two samples per day at each level of four QC samples) on three different days. The accuracy (%bias) was calculated from the mean value of observed concentration (C) and the nominal concentration (Cnom) as follows accuracy (% bias)=[(Cobs)-(Cnom)/(Cnom)]. The percent coefficient of variation, C. V. %, was calculated from the observed concentrations as follows: precision C. V % =[standard deviation (S. D)/Cobs] × 100.
The accuracy determined at each concentration level must be within ±15 % of the respective nominal value except at LOQ-QC where it must be within ±20% of the nominal value. The precision around the mean value must not exceed 15% of the C. V. except for LOQ-QC where it must not exceed 20% of the C. V.
2. 6. 5. Recovery:
The trueness of an analytical procedure refers to the closeness of agreement between a conventionally accepted value and a mean experimental one. Trueness was expressed as percentage recovery of the target value and assessed by means of validation standards in the matrix at three independent concentration levels, LQC, MQC and HQC. Recovery was calculated by comparing the mean peak area of an extracted sample (n = 3) to the one obtained after the direct injection of a solution with the same drug concentration.
A
Absolute % Recovery = -----------× 100
B
Where, A= Mean peak area response of extracted samples at LQC, MQC or HQC level and B= Mean peak area response of non extracted samples at LQCMQC or HQC level
2. 6. 6 Stability tests:
The stock solution stability of midazolam and I. S. was tested at room temperature for 12 h and upon refrigeration (40 C) for 15 days period.
The drug stability in a biological fluid is a function of the storage conditions, the chemical properties of the drug and the matrix. The stability procedures should evaluate the stability of the analytes in biological fluids after long-term (frozen at the intended storage temperature and conditions) and short-term (bench top, room temperature and conditions) storage, and after going through freeze and thaw cycles and the analytical process. Twenty-four hours autosampler stability of plasma sample extracts at 40 C was determined for analyte as well as I. S.
The long-term stability of midazolam in rat plasma was assessed by performing the experiment after 30 days of storage at −200 C. Freeze–thaw stability of midazolam (LQC, MQC and HQC) in rat plasma samples was determined for three freeze–thaw cycles. The samples were thawed at the room temperature without any assistance, and then kept in the freezer (−20◦ C) for minimum of 12 h before taking out for the next thawing. The post-preparative short-term room temperature stability of midazolam in processed samples left at ambient temperature was followed for 24 h. The concentration of after each storage period was related to the initial concentration as determined for the samples that were freshly prepared. The criterion for an acceptable stability of compounds in plasma samples under different storage conditions is that the relative recovery of the drug should be at least 90% of the initial concentration.
2. 7 Pharmacokinetic application of the developed HPLC method:
To assess the applicability of the present work, the method was used to investigate the pharmacokinetics of midazolam in rats. The study was conducted in accordance with the ethical guidelines for investigations in laboratory animals and was approved by the Institutional Animal Ethics Committee (IAEC), Manipal University (MU). All procedures and care of the rats were in accordance with institutional guidelines for animal use in research. Male Wistar rats, 250 ± 20 g, fasted overnight with free access to water for at least 12 h, were dosed orally by gavages with 5.0 mg kg-1 body weight of Midazolam base suspended in 1. 0% (w/v) carboxyl methyl cellulose aqueous solution as vehicle with a blunt needle via the esophagus into stomach; or injected intravenously (1. 0 mg kg-1 body weight, in a physiological NaCl solution sterilized through a Millipore filter before use)
Rats were divided into six groups (n =6) based on the time of blood sampling, each group having five animals. The blood samples were collected in heparinized tubes at control and 0. 084, 0. 25, 0. 50, 0. 75, 1. 0, 1. 5, 2. 0, 3. 0, 5. 0, 7. 0 hrs after the administration. The samples were immediately placed on ice and centrifuged at 2500×g for 10min. The plasma was separated into clean tubes and frozen at −20◦C until assay.
2. 7. 1 Calculation of pharmacokinetic parameters:
The pharmacokinetic parameters were calculated using non-compartmental approach. The area under the plasma concentration-time curve after oral administration was calculated using the linear trapezoidal rule upto the last measured plasma concentration and extrapolated to infinity, AUC, 0 – ∞ using a correction term, namely Cp(last) / λz. Cp(last) denotes either the observed or predicted concentration at last sample time and λz is the first rate order constant associated with the terminal (log-linear) portion of the curve and is estimated via linear regression of time versus log concentration. The terminal half life (t1/2) was determined by dividing ln2 by λz. Total body clearance (CL) was calculated by D/AUC, where D represents the dose. The absolute bioavailability, F(%) of orally administered midazolam was calculated by
AUC oral Dose IV
F (%) = ×
× 100
AUC IV Dose oral
2. 7. 2 Statistics:
All values were expressed as mean ± S. D. unless otherwise stated. Statistical differences were assumed to be reproducible when p <0. 05 (two-tailed t -test).
3. RESULTS AND DISCUSSION:
3. 1 Selection of chromatographic parameters:
Midazolam is a weak base with primary amino group (Pka=6. 25) and having intermediate hydrophobicity indicated by its logP 2. 25. Midazolam chemical structure reveals unique property of benzodiazepine ring opening as pH changes from acidic side to basic side. In acidic pH range the benzodiazepine ring opens from the amino group, which contributes to increase its relative hydrophilicity. Due to increase in hydrophilicity, midazolam elutes faster in reversed phase columns. In intermediate pH range some molecules remains in ring form and some in open ring form, so retention increases in this case as compare to acidic pH. Towards the basic side of pH most of the form goes into closed ring form resulting into increase in relative hydrophobicity and hence increase in retention towards basic side. Consequently, optimizing chromatographic conditions to obtain desired retention and adequate resolution in a short time were critical factors in the method development process.
The effect of mobile phase pH was studied with phosphate buffer and the best results were obtained when slightly alkaline medium was assayed. This was achieved at a mobile phase pH 7.2 that provided optimum selectivity (α) with acceptable peak shapes. Use of triethylamine, a peak modifier, helps in reducing peak tailing by the suppression of silanols16. The presence of triethylamine in the mobile phase provided tighter and well-defined peaks. Among the several columns evaluated, a Vydac monomeric column was chosen because a good peak shape and acceptable retention times were obtained. Different mobile phases comprising several combinations of buffers (such as phosphate buffer, ammonium acetate buffer) and organic solvents (viz. methanol, acetonitrile, and tetrahydrofuran) were tested to provide sufficient resolution between analyte and I. S. The effective separation of the bands in the chromatogram was achieved when the mobile phase composition was 50% potassium dihydrogen phosphate buffer and 50% acetonitrile.
The chromatographic conditions and sample preparation for the proposed method were optimized to suit the preclinical pharmacokinetic studies. Interferences from biomatrix components like plasma proteins were observed, and therefore, separation of analytes from matrix components was a key issue during method development process.
The isocratic elution of mobile phase was preferred as midazolam and I. S. peaks were well resolved within acceptable retention range (1< k <10) with retention times of 8. 0 and 10. 9 min, respectively. Both, system pressure and column performance, remained stable after the analysis of a large number of validation samples.
3. 2 Sample preparation, injection volume and peak shape:
Although midazolam contains polar amino group, it is hydrophobic in basic condition on the virtue of its pH dependant ring opening property. Various extraction procedures including protein precipitation and liquid-liquid extraction were investigated17. Direct protein precipitation with acetonitrile and trichloroacetic acid resulted in interference at either midazolam or I. S. retention time with high background noise. The extraction efficiency of different solvents including ethyl acetate, n-hexane, diethyl ether, dichloromethane and chloroform each alone and in combination with different percentages of MTBE was compared and with regard to both recovery and interference with endogenous peaks. Amongst all MTBE gave good recoveries and no interference at midazolam and I. S retention times with 0. 5 M NaOH buffering. Buffering helps to keep midazolam in unionized form and hence increased its partition into extraction solvent thereby its recovery.
In reverse phase chromatography, diluent has a great effect on peak shape and sensitivity. Solutions containing midazolam and I. S. in different diluents were injected. A mixture of methanol–water, 50:50 (v/v) was finalized as a diluents keeping in mind the solubility and consequently the recovery of the analyte. The injection volumes of 10, 25 and 45 µL provided no significant change in peak shape other than slight increase in peak broadening with increasing injection volume, as expected. After use, the column was flushed with methanol and water (80:20, v/v) and stored in the same solvent. The use of other benzodiazepines with a similar molecular structure (such as alprazolam, diazepam and clonazepam) as an I. S. was considered. Diazepam was selected as an I. S. because of high recovery and suitable retention time. Furthermore, the assay as described gives satisfactory validation results with the use of diazepam as the I. S.
3. 3 Method performance:
The method demonstrated excellent chromatographic specificity with both analyte and I. S. , being well resolved from endogenous compounds within 12. 5min of isocratic elution. The representative chromatograms for rat blank plasma and rat plasma spiked with midazolam (500 ng ml −1) and the I. S. (400 ng ml−1) are shown in (Fig. 2A) and (Fig. 2C), respectively.
3. 4 Method validation:
3. 4. 1 Selectivity:
The assay was found to be selective for midazolam as no interfering peaks were observed in the extracts of the different blank plasma samples. The representative chromatogram for the blank plasma sample is shown in (Fig. 2A).
(A)
(B)
(C)
Figure 2. Chromatograms obtained from (A) a drug-free, blank rat plasma, (B) limit of quantification and (C) a plasma sample spiked with 500 ng ml-1 midazolam. Spiked concentration of diazepam (internal standard) is constant throughout, at 400 ng ml-1.