2015 年 38 巻 3 号 p. 380-388
A conventional, rapid and high throughput method for tissue extraction and accurate and selective LC-MS/MS quantification of 2′-C-methylguanosine triphosphate (2′-MeGTP) in mouse liver was developed and qualified. Trichloroacetic acid (TCA) was used as the tissue homogenization reagent that overcomes instability challenges of liver tissue nucleotide triphosphates due to instant ischemic degradation to mono- and diphosphate nucleotides. Degradation of 2′-MeGTP was also minimized by harvesting livers using in situ clamp-freezing or snap-freezing techniques. The assay also included a sample clean-up procedure using weak anion exchange solid phase extraction followed by ion exchange chromatography and tandem mass spectrometry detection. The linear assay range was from 50 to 10000 pmol/mL concentration in liver homogenate (250–50000 pmol/g in liver tissue). The method was qualified over three intraday batches for accuracy, precision, selectivity and specificity. The assay was successfully applied to pharmacokinetic studies of 2′-MeGTP in liver tissue samples after single oral doses of IDX184, a nucleotide prodrug inhibitor of the viral polymerase for the treatment of hepatitis C, to mice. The study results suggested that the clamp-freezing liver collection method was marginally more effective in preventing 2′-MeGTP degradation during liver tissue collection compared to the snap-freezing method.
Hepatitis C virus (HCV) infection is an important global healthcare concern with approximately 150 million individuals infected worldwide and an estimated 4 million newly infected patients added each year.1–3) For many years, pegylated α-interferon in combination with ribavirin (PegIFN/RBV) was the standard of care therapy for HCV patients. This combination therapy boosts patient’s immune response against the HCV infection to achieve a sustained virologic response (SVR).4–6) In 2011, protease inhibitors were the first direct-acting antivirals (DAAs) introduced in combination with PegIFN/RBV, resulting in increased SVR rates and shorter duration of therapy for the majority of HCV patients.7)
The development of nucleoside and nucleotide analogue molecules have also been key to the evolution of HCV treatment as inhibitors of the important viral replication step catalyzed by HCV NS5B RNA polymerase.8–12) For this class of compounds, the active entity is nucleoside triphosphate (NTP). The NTPs have poor cellular uptake and they undergo rapid enzymatic degradation. The conversion of a nucleoside to its monophosphate is a rate limiting first step towards the formation of the active NTP. To overcome these difficulties prodrug versions of these molecules have been developed to deliver and target nucleosides to the liver for HCV and address bioavailability and stability issues.7–9,11,13,14)
2′-C-Methylguanosine triphosphate or 2′-MeGTP is a nucleotide analog which effectively disrupts RNA formation by the viral polymerase.8) IDX184 (Fig. 1) is a liver-targeted prodrug that is metabolized to 2′-C-methylguanosine monophosphate (2′-MeGMP) and subsequently converted to the active triphosphate (2′-MeGTP) by cellular kinases following uptake by hepatocytes.15)
For the purpose of pharmacokinetic assessments in drug development, it was crucial to develop a rapid, selective and accurate method for quantifying 2′-MeGTP in liver samples. Previously a method for analyzing 2′-MeGTP in mouse liver was reported using a chelating reagent to stabilize the triphosphate homogenization step, which lacks sufficient detailed descriptions, and included a laborious pulverizing step on dry ice.16) Other poorly attractive extraction procedures and LC-MS methods for quantification of different nucleotides in tissues have been developed and reported.17–22)
A limitation in the measurement of 2′-MeGTP is that it undergoes rapid ischemic degradation in collected liver tissue. This instant degradation during sample collection and sample processing makes the quantification of the nucleotide very challenging. Without stabilizing the liver samples during sample collection and extraction, determination of the triphosphate nucleotide may not be possible since it quickly converts to its mono- and diphosphate forms.
Herein we provide a convenient, reliable high throughput assay for quantifying 2′-MeGTP in mouse liver samples. In this article specifically we introduce trichloroacetic acid (TCA) as a convenient reagent to overcome instability of 2′-MeGTP in collected liver samples during extraction and homogenization. The method described minimizes degradation of 2′-MeGTP during tissue excision and extraction by utilizing clamp-freezing and snap-freezing techniques for tissue harvest and by maintaining tissue in a frozen state during weighing before homogenization in the TCA solution. TCA effectively blocks rapid conversion of triphosphate (TP) to its diphosphate (DP) and monophosphate (MP) forms by deactivating cellular enzymes involved in that metabolism. TCA was found to be superior to other common reagents used for stabilizing TP in liver homogenates while speed of tissue extraction and method convenience was maintained. A weak anion exchange extraction procedure was employed for removing TCA from liver sample extracts, reducing matrix effect, enhancing mass spectrometry detection response and increasing chromatography column lifetime. The method was qualified and widely used during development PK assessments of the nucleotide prodrug IDX184.
Although not discussed in this article, TCA was readily applied to liver measurements of active moieties from other nucleotide analogue prodrugs and tissue samples from several animal species, and thus has found widespread bioanalytical application in our liver TP screening assays.
The analyte, 2′-MeGTP, and the internal standard (IS), 2′-MeATP (2′-C-methyladenosine-5′-triphoshate) (Fig. 1) were synthesized at Idenix Pharmaceuticals, Inc. 2′-MeGTP was provided in the tris(triethylammonium) salt form (molecular weight (MW) 840.78) and 2′-MeATP was provided in the trisodium salt form (MW 587.15). Deionized water was used from a Milli-Q® system (Milli-Q, Millipore Corp., Bedford, MA, U.S.A.). The chemicals were purchased from the following sources; HPLC grade methanol, HPLC grade acetonitrile (ACN) and formic acid from EMD (Rockland, MA, U.S.A.), ammonium acetate and ammonium hydroxide from J.T Baker (Center Valley, PA, U.S.A.), trichloroacetic acid (TCA) from VWR (West Chester, PA, U.S.A.), and perchloric acid from Sigma-Aldrich (St. Louis, MO, U.S.A.). All chemicals were analytical grade.
Preparation of Calibration Stock and Spiking SolutionsStock solutions of 2′-MeGTP for calibration standards and QC standards were prepared by adding accurately weighed amounts of the tris(triethylammonium) salt to 1.0 mL Milli-Q water and mixing on a vortex mixer to yield 3.43 mg/mL (6380 µM) solutions of 2′-MeGTP after correcting for purity and salt content. Equivalency of the two prepared stock solutions were checked by LC-MS injection of diluted solutions from the both stock solutions and was found to be less than 5% difference. These stock solutions were stored at −20°C and used to prepared intermediate calibration stock solution and intermediate QC stock solution that had a final concentration of 1000 µM.
The intermediate calibration stock solution was then used to prepare ten calibration standard spiking solutions (200, 160, 100, 50.0, 20.0, 10.0, 4.00, 2.00, 1.00, and 0.500 µM), and the intermediate QC stock solution was then used to prepare the High, Mid, Low and lower limit of quantification (LLOQ) quality control spiking solutions of 150, 60, 3, and 1.0 µM. The intermediate stock solutions and spiking solutions were stored at −20°C. Ten 2′-MeGTP calibration standards (ranging from 50 to 10000 pmol/mL), and QC samples (50, 150, 3000, and 7500 pmol/mL) were prepared on a daily basis by adding a 10 µL aliquot of the respective spiking solutions to 190 µL aliquots of liver blank homogenates.
Stock solutions of 2′-MeATP (IS) were prepared by combining accurately weighed 1 mg of the trisodium salt in 1 mL milli-Q water and determining the concentration of the free acid after correcting for purity and salt concentration. The IS working stock solution with the concentration of 10 µg/mL was also prepared by diluting of the IS stock solution in proportional volume of Milli-Q water. Both solutions were stored at −20°C.
Preparation of Homogenization and Extraction ReagentsThe tissue extraction reagent solution, 0.95 M trichloroacetic acid (TCA), was prepared by dissolving 500 g of solid TCA in 350 mL Milli-Q water to form a saturated solution and then diluting the saturated solution 5-fold with water (v/v). A 20% ammonium hydroxide solution was prepared by diluting concentrated (25–28%) ammonium hydroxide 5-fold with Milli-Q water (20 : 80, v/v). The 1.0% formic acid was prepared by diluting concentrated (98%) formic acid with water (v/v). A solution of methanol–water (75 : 25, v/v) was prepared and used to prepare 1% formic acid in methanol–water (75 : 25, v/v). Washing and eluting solutions 0.5% formic acid in methanol–water (50 : 50, v/v) and 4% ammonia in methanol–water (75 : 25, v/v) were prepared proportionally from the concentrated acid and base solutions, respectively.
Chromatographic and Mass Spectrometric ConditionsHPLC analysis was carried out on a Agilent 1100 series system (Agilent, Santa Clara, CA, U.S.A.) equipped with a binary pump, vacuum de-gasser unit, and a temperature controlled micro well plate autosampler. The reconstituted 96 well plate holding samples, standards and quality controls was covered with a pre-slit well cap (Thermo Scientific, Rochester, NY, U.S.A.) and placed in the refrigerated (4°C) autosampler. Samples were loaded onto and eluted from a Phenomenex Luna NH2, 5 μm microbore, 30×2.0 mm column (Phenomenex, Torrance, CA, U.S.A.) using a flow rate of 0.8 mL/min. The column was maintained at ambient temperature. The LC gradient started with 100% mobile phase A (MPA, 1 mM ammonium acetate in water–acetonitrile (70 : 30, v/v) pH 8.0) for the first 1.5 min of the run, then switched to 100% mobile phase B (MPB, 20 mM ammonium acetate in water–acetonitrile (70 : 30, v/v) pH 10.0) over the next 0.7 min. One hundred percent MPB was continued between 2.2 and 7.4 min into the run and then rapidly switched to 100% MPA over 0.1 min. One hundred percent MPA was continued for an additional 2.5 min to ensure the column re-equilibrated before the next sample was injected. Total run time was 10 min per sample.
Detection was carried out using a triple quadropole tandem mass spectrometer (Applied Biosystems API3200) equipped with a TurboIonSpray in the positive ion mode. The ion source temperature was set at 550°C, and the ion spray voltage was set at 5.5 kV. Ultra high purity (UHP) nitrogen was used as the curtain, nebulizer, collision and TuboIonSpray auxiliary gas. Detection of ions was performed in the multiple reaction monitoring (MRM) mode. Analyst 1.4.2® software was used to monitor the following transitions: m/z 537.8→m/z 152.0 for 2′-MeGTP and m/z 521.8→m/z 136.1 for the internal standard (IS). The declustering potential (DP) was set at 35 and 47 eV for 2′-MeGTP and the IS, respectively, and the collision energy (CE) was set at 35 and 66 eV for 2′-MeGTP and the IS, respectively. The dwell time for both 2′-MeGTP and the IS was set at 100 ms.
Calculated concentrations of 2′-MeGTP were based on peak area ratios of the analyte to the IS. Linear regression was processed by Applied BioSystem Analyst 1.4.2® with a weighting factor of 1/concentration2.
Animals Liver Sample CollectionsIDX184 was formulated in PEG200 and administered to 4 groups of 12 male CD-1 mice each by oral gavage at single doses of 11.4, 84.4, 168.6, and 84.4 mg/kg in a dose volume of 5 mL/kg at Xenometrics (Stilwell, KS, U.S.A.). The study protocol was reviewed and approved by the Xenometrics IACUC. Liver samples were obtained from 3 mice per group at 2, 4, 8, and 24 h post dose. Clamp-freezing was used to collect livers from groups 1–3 (doses of 11.4, 84.4, 168.6 mg/kg), and snap-freezing was used to collect livers from group 4 (dose of 84.4 mg/kg). In the clamp-freeze process, the liver while still being perfused with blood was clamped in situ between two liquid nitrogen chilled metal plates. The clamped liver was then excised and immediately frozen by immersion in liquid nitrogen. In the snap-freezing process, the tissues were excised first and then immediately frozen by immersion in liquid nitrogen. All liver samples were immediately placed on dry ice and then stored at −70°C freezer. Liver samples shipped on dry ice to Idenix Pharmaceuticals (Cambridge, MA, U.S.A.) where the samples were again stored at −70°C freezer until processing.
Since the clamp-freezing group 2 and snap-freezing group 4 were both dosed at 84.4 mg/kg, the analysis data from these two groups were compared to determine which sample collection method was better to keep the nucleotide stable in tissue during liver sample collections.
Sample PreparationsA 100–500 mg section of mouse liver (obtained by hammering the frozen liver to chunks) was weighed while still kept frozen on dry ice and immediately placed in a 50 mL medical grade Corning polypropylene conical bottom centrifuge tube (Corning, NY, U.S.A.) on dry ice. Ice-cold 0.95 M TCA was added to the weighed samples at a volume of 4 mL per gram of tissue and the samples were then immediately homogenized on wet ice for 1–2 min using a OMNI TH (Kennesaw, GA, U.S.A.) homogenizer with Omni-disposable RNase/DNase free tips. It was critical that care was taken to ensure liver samples did not thaw prior to homogenization. Two hundred microliter of sample homogenates from dosed animals were then transferred to 2.0 mL pre-labeled Eppendorf vial tubes with snap caps and placed in a rack on ice. Matrix blank samples were prepared by transferring 200 µL of the control liver homogenate to Eppendorf test tubes. Twenty microliter of the IS working solution was then added to all samples except the matrix blank samples and the samples were mixed by gentle vortexing. For the matrix blank samples, 20 µL of water was added in lieu of the IS working solution. After all samples were prepared, TCA was neutralized by the addition of 80 µL ice-cold 20% ammonium hydroxide followed by mixing. The pH of samples was adjusted within the range of 3–4 by adding 500 µL of 1% formic acid. The samples were then centrifuged in a Eppendorf 5415D centrifuge (Eppendorf, Hamburg, Germany) for 5 min at 13200 rpm (16100 rcf).
In the mean time, 96 well Waters Oasis WAX 30 mg cartridges (Waters, Milford, MA, U.S.A.) were prepared for weak anion exchange extraction of nucleotides by being conditioned sequentially with 1 mL each of a) methanol, b) 2% formic acid in methanol–water (75 : 25, v/v), and c) water. Between each step, the conditioning solutions were drawn through the cartridges under vacuum using a Biotage® (Charlotte, NC, U.S.A.) VacMaster-96 Sample Processing Manifold 96.
The entire supernatants from blanks, calibration standards, QC standards and tissue samples were loaded onto the conditioned cartridges and drawn into the cartridges slowly under vacuum. The loaded cartridges were then washed slowly under vacuum with 200 µL of water followed by 200 µL of 0.5% formic acid in methanol–water (50 : 50, v/v) and then eluted twice slowly under vacuum into a clean NunC Agilent 96 deep-well plate using 400 µL of 4% ammonium hydroxide in methanol–water (75 : 25, v/v) per elution step. The 96 deep well collection plate was then dried under nitrogen gas at 40–50°C for 45–60 min using a Caliper TurboVap 96 evaporator (Waltham, MA, U.S.A.). Immediately prior to analysis, each well was reconstituted using 100 µL of mobile phase A. A volume of 5–20 µL was injected into LC-MS.
Method QualificationTo determine intraday precision and accuracy of this assay, 3 batches were extracted on three separate days. Each batch included six replicates of LLOQ, Low, Mid and High QC samples that were prepared individually by spiking into liver homogenate yielding nominal concentrations of 50, 150, 3000, and 7500 pmol/mL, respectively.
For each batch duplicate sets of calibration standards were prepared and bracketed the QCs by spiking into liver homogenates yielding nominal concentrations of 50, 100, 200, 500, 1000, 2500, 5000, 8000, and 10000 pmol/mL.
The intraday precisions and accuracies were determined by calculating ratios of standard deviation of QCs measured concentrations to their mean at each QC level for each batch. The interday precision and accuracy were determined by pooling all individual assay QC results of replicate (n=6) quality control samples over the three separate batch runs.
Assay specificity was evaluated by preparing 6 lots of control (blank) mouse liver homogenates and analyzing a matrix blank and a matrix blank with IS from each lot for peaks co-eluting with 2′-MeGTP.
Assay interference selectivity was determined by assaying 6 Low QC replicates from each 6 individual lot of blank mouse liver homogenates (nominal concentrations at 150 pmol/mL).
Recovery of 2′-MeGTP was assessed by assaying Low, Mid, and High QCs in triplicate (nominal concentrations 150, 3000, and 7500 pmol/mL, respectively) and comparing peak area ratios following extraction of 2′-MeGTP with peak area ratio values obtained by spiking Low, Mid, and High QC concentration levels in triplicate into post extracted blank matrix. Recovery of 2′-MeATP was also assessed at one concentration.
The instability of NTPs in liver samples is generally known.22,23) Consequently, clamp-freezing or snap-freezing techniques were used during tissue excision as a first step to prevent degradation of the NTP to its mono- and diphosphate forms. Collected liver samples were stored at −70°C prior to tissue weighing and TP extraction. It was also essential not to allow tissues to thaw during the weighing procedure and hence tissues were maintained on dry ice during this process.
The importance of using ice-cold reagents to stabilize NTPs during extraction was also recognized in method development. We demonstrate how critical this step is by comparing yields of 2′-MeGTP in mouse liver samples homogenized on ice using either ice-cold or room temperature TCA. Both set of IDX184-dosed mouse liver study samples were extracted and assayed within one batch. Use of the room temperature TCA solution allowed the liver tissues to partially thaw and therefore did allow 2′-MeGTP to degrade within the tissue samples before homogenization. Obviously if the weighed liver samples are allowed to thaw the enzymes inside the non-frozen tissue start reacting on 2′-MeGTP and degrade it to 2′-MeGDP and 2′-MeGMP. It should be noted that this ischemic degradation occurs instantly inside thawed liver tissue.22,23) However, TCA reagent does not penetrate inside the tissue samples before homogenizing the samples. Therefore, it is essential to keep all samples in frozen condition before homogenization. Adding ice-cold TCA reagent to liver samples just before the homogenization step keeps the samples in frozen condition. Quick homogenization of the frozen liver samples with ice-cold TCA disables the enzymatic ischemic degradation. The data (see Table 1) show when room temperature TCA reagent was used all measured TP concentrations were below LLOQ (250 pmol/g) except sample number 28 whose measured TP concentration was very close to LLOQ (254 pmol/g). In this experiment the MRM transitions of monophosphate (2′-MeGMP) and diphosphate (2′-MeGDP) were also monitored. Disappearance of 2′-MeGTP and the internal standard 2′-MeATP as well as increasing in 2′-MeGDP and 2′-MeGMP peaks in the sample LC-MS/MS chromatograms are indications of the enzymatic degradation. It should be noted that 2′-MeGDP could further degrade to 2′-MeGMP if the enzymatic degradation is not completely inhibited. Therefore observing high peaks of 2′-MeGMP is good marker for the enzymatic degradation.
| Sample No. | Group No. | Time after dose (h) | Measured concentration of 2′-MeGTP (pmol/g liver) using ice-cold TCA homogenization reagent | Measured concentration of 2′-MeGTP (pmol/g liver) using room temp. TCA homogenization reagent |
|---|---|---|---|---|
| 25 | 3 | 2 | 7850 | BQLa) |
| 26 | 3 | 2 | 7100 | BQL |
| 27 | 3 | 2 | 6400 | BQL |
| 28 | 3 | 4 | 7600 | 254 |
| 29 | 3 | 4 | 6050 | BQL |
| 30 | 3 | 4 | 4290 | BQL |
| 31 | 3 | 8 | 9350 | BQL |
| 32 | 3 | 8 | 6400 | BQL |
| 33 | 3 | 8 | 4460 | BQL |
| 34 | 3 | 24 | 870 | BQL |
a) LLOQ: 250 pmol/g liver.
When the same set of samples were processed in frozen condition by adding ice-cold TCA the degradation of 2′-MeGTP was completely inhibited. 2′-MeGTP was quantifiable in very high levels (see Table 1). 2′-MeGDP and 2′-MeGMP were not highly present in LC-MS/MS chromatograms which indicate complete suppression of 2′-MeGTP degradation. The data illustrate that this procedure provides complete protection of liver 2′-MeGTP from ischemic degradation reactions (supplemental Figures 1A–C and supplemental Figures 2A–C).
The effectiveness and advantage of using TCA homogenization reagent was compared to five other solvents or reagents. Some of the solvents or reagents have been used elsewhere as protein precipitation methods to isolate TPs in liver, other tissues or cell cultures. These reagents included acetonitrile–water (70 : 30, v/v), methanol–water (70 : 30, v/v); acetonitrile; and 12% perchloric acid (w/v).17,22,24) The buffer reagent 10 mM ammonium acetate, pH 3.5, was also used to test whether a low pH buffer could stabilize 2′-MeGTP in liver homogenate. In each set of experiments, two IDX184-dosed mouse liver study samples were homogenized with ice-cold TCA reagent or with one of five other ice-cold solvents or reagents. All samples were extracted according to the aforementioned procedure and were quantified in one batch. In all cases, 2′-MeGTP was below the quantification level (BQL) of 250 pmol/g in the samples homogenized apart from those using the TCA reagent (Table 2). These results emphasize that using TCA reagent as well as maintaining cold temperature to preserve the tissue in frozen condition are both important requirements to prevent enzymatic degradation of 2′-MeGTP.
| Study sample No. | Group No. | Time after dose (h) | Measured 2′-MeGTP (pmol/g liver) using TCA homogenization reagent* | Measured 2′-MeGTP (pmol/g liver) using other homogenization reagents* | Other homogenization reagent |
|---|---|---|---|---|---|
| 13 | 2 | 2 | 6750 | BQLa) | Acetonitrile–water, 70 : 30 (v/v) |
| 15 | 2 | 2 | 3145 | BQL | |
| 17 | 2 | 2 | 4345 | BQL | Methanol–water, 70 : 30 (v/v) |
| 22 | 3 | 4 | 640 | BQL | |
| 8 | 1 | 4 | 340 | BQL | Acetonitrile |
| 16 | 2 | 4 | 3405 | BQL | |
| 27 | 3 | 8 | 6100 | BQL | 10 mM ammonium acetate, pH 3.5 |
| 31 | 3 | 8 | 9250 | BQL | |
| 31 | 3 | 8 | 9250 | BQL | 12% (w/v) perchloric acid |
| 46 | 4 | 24 | 585 | BQL |
a) LLOQ: 250 pmol/g liver. * TCA and other reagents were used in ice-cold condition.
In this experiment run in addition to 2′-MeGTP and 2′-MeATP MRM transitions, the MRM transition for monophosphate, 2′-MeGMP, was also monitored. The LC-MS/MS chromatogram for all samples showed that 2′-MeGTP and the IS in the samples were degraded when homogenized in all the reagents other than TCA. As a result no 2′-MeGTP and 2′-MeATP chromatogram peaks could be found for these samples. Moreover, peak height and peak area comparisons showed 2′-MeGMP levels, as a degradation product of 2′-MeGTP, were increased in these sample extracts compared to the same samples homogenized and extracted using the TCA reagent (Figs. 2A–C, Figs. 3A–C, supplemental Figures 3A–C and supplemental Figures 4A–C).
(A) 2′-MeGTP MRM transition (B) 2′-MeGMP MRM transition and (C) internal standard 2′-MeATP MRM transition.
(A) 2′-MeGTP MRM transition (B) 2′-MeGMP MRM transition and (C) internal standard 2′-MeATP MRM transition.
This indicates that TCA most likely inhibited the enzymes associated with conversion of 2′-MeGTP to 2′-MeGDP and 2′-MeGMP and also 2′-MeATP to its mono- and diphosphate nucleotides. More importantly TCA was able to stop enzymatic activity in the ice-cold temperature condition which was essential to keep the samples frozen. However, the other homogenization solvents or reagents could not stop the enzymatic degradation activity in the same cold condition. Also, no 2′-MeGMP peaks could be found in the chromatograms of standards and QCs which further indicates that TCA inhibited conversion of 2′-MeGTP to 2′-MeGDP and 2′-MeGMP in spiked TCA liver homogenates. This lack of 2′-MeGMP presence is also illustrated in the representative chromatogram of High QC (Figs. 4A–C).
(A) 2′-MeGTP MRM transition (B) 2′-MeGMP MRM transition and (C) internal standard 2′-MeATP MRM transition.
Although TCA extracts could be injected into LC-MS without further solid phase clean up, it was noticed during method development that TCA extracts caused mass spectrometer contamination and decreased column lifetime. Consequently, a weak anion exchange solid phase extraction procedure was developed to remove TCA from the liver sample extracts. After inclusion of a weak anion exchange procedure the bioanalytical method was progressed to qualification.
Accuracy, Precision, and LinearityThree intraday batches of calibration curve standards in control mouse liver homogenates were prepared and analyzed. These batches included six independent replicates of LLOQ, Low, Middle, and High QCs. Linear regressions were constructed by Analyst 1.4.2® weighted 1/concentration2 for best fit ratios of concentration/detector area response. Coefficients of determination (r2) for the three intraday runs were greater than 0.9976.
The accuracy was determined as relative error (%RE) by comparing the means of the measured concentrations with the nominal values using the formula %RE=(D×100/N) where D is the experimentally measured concentration and N is the nominal concentration. Precision was calculated as coefficient of variation (%CV) using the formula %CV=(Std Dev/M) (100) where M is the mean of the measured concentrations and Std Dev is the standard deviation of the measured concentrations. The acceptance criteria for intraday and interday precision were 20% or less variation for LLOQ and 15% or less for the other QC concentrations. The acceptance criteria for accuracy were 80–120% of the nominal concentrations at the LLOQ level (QC and standard) and 85–115% of the nominal concentration at the other levels of concentrations.
Measured accuracy values for 2′-MeGTP standards were within 98.2 to 102.1% of nominal values over the 50 to 10000 pmol/mL concentration range. Precision values (%CV) were lower than 7.7% for the standards.
Intraday precision and accuracy of 2′-MeGTP quality control samples in mouse liver homogenates was also determined in the three assay runs (see Table 3). In the three runs, the %RE ranged from 84.5 to 106.1% and the %CV were ≤8.7% for the four levels of QCs.
| Quality controls | LLOQ QC | Low QC | Mid QC | High QC |
|---|---|---|---|---|
| Nominal concentration (pmol/mL) | 50 | 150 | 3000 | 7500 |
| Intraday Batch 1 | ||||
| Mean | 47.4 | 157 | 2808 | 6900 |
| Std Dev | 4.13 | 10 | 133 | 285 |
| %CV | 8.7 | 6.6 | 4.7 | 4.1 |
| %RE | 94.8 | 104.5 | 93.6 | 92.0 |
| Intraday Batch 2 | ||||
| Mean | 47.0 | 159 | 3043 | 7557 |
| Std Dev | 4.1 | 7.3 | 123 | 155 |
| %CV | 8.7 | 4.6 | 4.0 | 2.1 |
| %RE | 94.1 | 106.1 | 101.4 | 100.8 |
| Intraday Batch 3 | ||||
| Mean Conc. | 42.3 | 155 | 3153 | 7380 |
| Std Dev | 2.7 | 11.2 | 96.5 | 81.7 |
| %CV | 6.5 | 7.2 | 3.1 | 1.1 |
| %RE | 84.5 | 103.4 | 105.1 | 98.4 |
| Interday Statistics | ||||
| Mean | 45.6 | 157 | 3002 | 7279 |
| Std Dev | 4.2 | 16.3 | 185 | 338 |
| %CV | 9.3 | 10.4 | 6.2 | 4.6 |
| %RE | 91.1 | 104.7 | 100.1 | 97.1 |
The results for interday validation of QC samples are also listed in Table 3. The mean accuracy values ranged from 91.1 to 104.7% and the %CVs were ≤10.4%.
Overall these results met the acceptance criteria. We conclude that the analytical method is reproducible in terms of intra- and interday accuracy and precision. Representative chromatogram of High QC is illustrated in Figs. 4A–C.
Selectivity and SpecificitySpecificity of the method was assessed by analyzing control mouse liver tissue samples from 6 different lots. For each lot, one replicate matrix blank and one replicate matrix blank spiked with IS were extracted and assayed. Backgrounds were negligible and no interfering peak was observed at retention times of 2′-MeGTP or IS for any of the six lots of blank or blank plus IS samples. Representative chromatographs are illustrated in Figs. 5A and B and supplemental Figures 5A and B.
(A) 2′-MeGTP MRM transition (B) internal standard 2′-MeATP MRM transition.
Selectivity was assessed by preparing six replicates of the low QC sample with each mouse liver lot. The Low QCs were also extracted, and assayed by LC-MS/MS in the same run. The %CVs for the six lots of Low QCs were within 15% acceptance criteria (see Table 4) which indicates the analytical method is selective for quantification of 2′-MeGTP in mouse liver.
| Nominal conc. (pmol/mL) | 150 | 150 | 150 | 150 | 150 | 150 |
|---|---|---|---|---|---|---|
| Liver lots | Lot 1 | Lot 2 | Lot 3 | Lot 4 | Lot 5 | Lot 6 |
| Mean measured concentration | 155 | 141 | 149 | 140 | 141 | 139 |
| Std Dev | 11.2 | 8.93 | 10.3 | 7.97 | 9.46 | 3.74 |
| %CV | 7.2 | 6.3 | 6.9 | 5.7 | 6.7 | 2.7 |
| %RE | 3.4 | −5.9 | −1.0 | −6.4 | −5.8 | −7.3 |
| n | 6 | 6 | 6 | 6 | 6 | 6 |
Recovery of 2′-MeGTP was assessed by preparing and extracting Low, Mid, and High QCs in triplicate and comparing peak area ratios following extraction of 2′-MeGTP with peak area ratio values obtained by spiking Low, Mid, and High QC concentration levels in triplicate into post extracted blank matrix. The recovery at each level of concentration was calculated using the following formula:
 |
| Nominal concentration (pmol/mL) | 150 | 3000 | 7500 |
|---|---|---|---|
| QC levels | Low | Middle | High |
| %Recovery | 43.4 | 46.5 | 47.6 |
| Mean % recovery | 45.8 | ||
The described method was applied to analyzing for “active” (2′-MeGTP) metabolite in liver tissue samples obtained from mice administered a single oral dose of the prodrug IDX184. Values were corrected for the dilution factor of liver homogenates (1 : 5) and reported as pmol/g. The measured concentration values were averaged for 3 mice per group per collection time after dose. The pharmacokinetic parameters were computed using Microsoft Excel 97–2003. The evaluated statistical data were Cmax (maximum observed concentration), AUC0–24 (area under the concentration–time curve measured to 24 h post dose using the linear trapezoidal rule) and Tmax (time of observed Cmax). The parameters for groups 2 and 4 are presented in Table 6. The AUC0–24 for groups 1 and 3 were 12700 h.pmol/g and 106000 h.pmol/g with dose levels of 11.4 and 168.8 mg/kg, respectively.
| Group | Tissue freezing method | Dose (mg/kg) | Tmax (h) | Cmax (pmol/g) | AUC(0–24) h.pmol/g |
|---|---|---|---|---|---|
| 2 | Clamp | 84.4 | 4 | 4780 | 71200 |
| 4 | Snap | 84.4 | 4 | 4130 | 59000 |
The mean liver concentrations versus time post-dose are plotted for groups 2 and 4 at the dose level of 84.4 mg/kg to compare measured concentrations of the two tissue collection methods (Fig. 6). The comparison between the measured exposures of these two groups indicates that measured concentrations for the clamp-freezing method were marginally higher than the exposures of the samples collected using the snap-freezing group. Therefore, the clamp-freezing liver collection method would be more optimal in preventing 2′-MeGTP degradation during liver tissue collection compared to the snap-freezing method. However, if animal variability and standard deviations are taken into consideration the relative differences are small. So the snap-freezing method could be still considered a suitable, effective and convenient method of liver collection for NTP analysis.
A reliable procedure was developed to overcome instability of 2′-MeGTP during liver tissue extraction. TCA was completely effective in stabilizing 2′-MeGTP during liver homogenization and sample extraction. Using TCA, a rapid and reproducible bioanalytical assay was developed and qualified to measure concentrations of 2′-MeGTP in mouse liver tissues. The method included sample clean up using weak anion exchange solid phase extraction followed by ion exchange chromatography and tandem mass spectrometry detection. The assay was linear over a 50–10000 pmol/mL concentration range in liver homogenate (or 250–50000 pmol/g in liver tissue) and had acceptable intraday and interday precision and accuracy.
Method performance was demonstrated for the analysis of liver samples from a mouse dosing study with IDX184. Within the study, two different liver collection methods were evaluated. The data indicated that a clamp-freezing liver collection method was marginally better than a snap-freezing method to prevent degradation of 2′-MeGTP in mouse liver. The bioanalytical method can be easily adapted for extraction of active moieties of other nucleotide analogues and tissue samples from several animal species.
This study was conducted at Idenix Pharmaceuticals. The authors would like to thank Xenometrics LLC for conducting the in-life section of this study and BethAnn Friedman for assistance to prepare this manuscript.
The authors declare no conflict of interest.
The online version of this article contains supplementary materials.
