Abstract
Atazanavir is a commonly prescribed protease inhibitor for treatment of HIV-1 infection. Thus far, only limited data are available on the in vivo metabolism of the drug. Three systemic circulating metabolites have been reported, but their chemical structures have not been released publicly. Atazanavir metabolites may contribute to its effectiveness but also to its toxicity and interactions. Thus, there is a need for extensive metabolic profiling of atazanavir. Our goals were to screen and identify previously unknown atazanavir metabolites and to develop a sensitive metabolite profiling method in plasma. Five atazanavir metabolites were detected and identified in patient samples using liquid chromatography coupled to linear ion trap mass spectrometry: one N-dealkylation product (M1), two metabolites resulting from carbamate hydrolysis (M2 and M3), a hydroxylated product (M4), and a keto-metabolite (M5). For sensitive semiquantitative analysis of the metabolites in plasma, the method was transferred to liquid chromatography coupled to triple quadrupole mass spectrometry. In 12 patient samples, all the metabolites could be detected, and possible other potential atazanavir keto-metabolites were found. Atazanavir metabolite levels were positively correlated with atazanavir levels, but interindividual variability was high. The developed atazanavir metabolic screening method can now be used for further clinical pharmacological research with this antiretroviral agent.
Atazanavir is a commonly used HIV protease inhibitor and is used in combination with other antiretroviral agents for the treatment of HIV infection. Atazanavir has been approved in the European Union for once-daily administration of 300 mg in combination with 100 mg of ritonavir. The protease inhibitor ritonavir is given as a pharmacokinetic booster and increases systemic atazanavir exposure by inhibiting cytochrome P450 enzyme 3A4 (CYP3A4) metabolism in the liver and intestines. In the United States, atazanavir has also been approved in a dose of 400 mg once daily without ritonavir in treatment-naive patients (Swainston Harrison and Scott, 2005).
CYP3A4 in the intestinal wall and liver is thought to be the major enzyme responsible for atazanavir biotransformation. Biliary excretion is the main route of excretion of atazanavir. Renal elimination plays a minor role in excreting atazanavir and its metabolites. The main systemically circulating atazanavir metabolites have been described to be mono-oxygenated and dioxygenated products of atazanavir, but their molecular structures have not been released publicly. Other metabolic routes have been proposed to be glucuronidation, N-dealkylation, hydrolysis, hydroxylation, and dehydrogenation (http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_Pharmr_P1A.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P1.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P2.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P3.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P4.pdf).
Adverse reactions related to atazanavir use are jaundice and QT prolongation. Atazanavir inhibits the UDP-glucuronosyltransferase isoenzyme 1A1 (UGT1A1), the enzyme responsible for bilirubin glucuronidation, causing decreased glucuronide excretion that may result in jaundice. Drug metabolites may contribute to effectiveness of the parent compound but also to its toxicity and interactions. Both atazanavir and its metabolites are known to exhibit electrophysiological effects (e.g., QT prolongation); however, no studies have been performed investigating the relationship between metabolite exposure and jaundice (http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_Pharmr_P1A.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P1.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P2.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P3.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P4.pdf).
Thus far, most work in metabolite detection and identification has been performed using liquid chromatography coupled to triple quadrupole mass spectrometers. These tandem mass spectrometers have excellent sensitivity and selectivity when operating in the multiple reaction monitoring (MRM) mode, allowing sensitive quantification of multiple analytes at once. However, in full-scan mode the identification sensitivity of triple quadrupole mass spectrometers is limited. In contrast, ion trap mass spectrometers allow sensitive full mass scans in combination with fast scanning time, and multistage mass spectral scans and monitoring of fragmentation cascades can be performed for identification of unknown compounds.
Because little is known about the identity and the amount of the circulating atazanavir metabolites and their role in treatment outcome, there is need for further identification and metabolic profiling of atazanavir. Our goals were to screen and to identify previously unknown atazanavir metabolites in plasma using liquid chromatography coupled to linear ion trap mass spectrometry and to develop a sensitive profiling method using liquid chromatography coupled with triple quadrupole mass spectrometry for purposes of clinical pharmacological research.
Materials and Methods
Chemicals and Reagents. Methanol, ethanol, and acetonitrile were high-performance liquid chromatography (HPLC) grade and obtained from Biosolve (Valkenswaard, The Netherlands). Ammonium acetate, tert-butyl methyl ether, and glacial acetic acid were obtained from Merck (Amsterdam, The Netherlands). Eppendorf tubes (2 ml) and autosampler vials with 250-μl inserts were obtained from VWR (Amsterdam, The Netherlands). Distilled water originated from Aqua B. Braun (Melsungen, Germany). Atazanavir sulfate was a gift from Bristol-Myers Squibb Co. (Stamford, CT), and the stable D5-atazanavir isotope was obtained from Toronto Research Chemicals Inc. (North York, ON, Canada). HPLC eluent A consisted of 10 mM ammonium acetate buffer in water, pH 5, and methanol (65:35, v/v), and HPLC eluent B consisted of methanol. The internal standard solution consisted of 500 ng/ml D5-atazanavir in methanol/water (25:75, v/v).
Sample Collection and Pretreatment. Patient samples were EDTA anti-coagulated plasma samples from HIV-infected patients at the outpatient clinic of the Slotervaart Hospital (Amsterdam, The Netherlands). All the patients were taking atazanavir for at least 4 weeks. Plasma samples from patients not taking atazanavir were screened for atazanavir and its metabolites as a control. The study design was observational. Ethics approval was obtained from our institutional ethics board, and all the patients gave written informed consent. For metabolite screening and identification, samples randomly collected during the dosing interval were used. For comparison of metabolic profiling samples, trough samples were used (collected 18–24 h after atazanavir intake).
For extraction of atazanavir and its metabolites from plasma, protein precipitation was performed using 200 μl of plasma and 500 μl of a mixture of methanol and acetonitrile (1:1, v/v). After centrifugation for 10 min at 23,100g, the supernatant was transferred to an autosampler vial. Furthermore, a previously described liquid-liquid extraction was tested, extracting atazanavir from plasma with tert-butyl methyl ether (Rezk et al., 2006). In brief, 1.5 ml of tert-butyl methyl ether was added to 200 μl of plasma in a 2-ml Eppendorf tube. After vortex-mixing for 2 min, the tubes were placed in a bath of ethanol with dry ice, and the aqueous layer was snap-frozen. The organic layer was transferred to a new Eppendorf tube and evaporated to dryness under a gentle stream of nitrogen at 40°C. The residue was reconstituted in 100 μl of eluent A and transferred to an autosampler vial. For metabolite screening and identification, no internal standard was added to the sample. For metabolic profiling, 50 μl of the internal standard solution containing D5-atazanavir was added to the plasma before extraction.
Liquid Chromatography. Eluent was delivered at a constant flow rate of 325 μl/min, and separation was performed using a Phenomenex (Utrecht, The Netherlands) Gemini C18 column (150 × 2.0 mm; particle size, 5 μm) with a Phenomenex SecurityGuard Gemini C18 precolumn (4.0 × 2.0 mm; particle size, 5 μm). Initially, chromatography consisted of a slow gradient starting with 100% eluent A and 0% B at time 0 min, linearly increasing to 10% A and 90% B in 60 min, followed by a 5-min re-equilibration step at 100% A and 0% B. The injection volume was set to 25 μl.
Ion Trap Mass Spectrometry. For metabolite screening and identification, an Accela HPLC coupled to an LTQ XL linear ion trap mass spectrometer with a heated electrospray ionization probe was used (Thermo Fisher Scientific, Waltham, MA). The mass spectrometer operated in positive ion mode with the spray voltage at 6 kV, the capillary and heated electrospray ionization vaporizer temperature set at 275°C, the tube lens voltage set at 110 V, the skimmer offset at 0 V, and the sheath and aux gas flow set at 40 and 35 arbitrary units (a.u.), respectively. The isolation width was set at 2.0, and normalized collision energy of 30% was used for collision-induced dissociation.
Triple Quadrupole Mass Spectrometry. For quantitation of atazanavir in plasma, a previously validated assay was used (ter Heine et al., 2007). For semiquantitative profiling of metabolites, an Agilent 1100 (Agilent Technologies, Santa Clara, CA) HPLC system was used, coupled to an API3000 triple quadrupole mass spectrometer (PE Sciex, Thornhill, ON, Canada) operating in MRM in positive ion mode, with Q1 and Q3 at low resolution, connected to the column with a turbo-ion spray source. The flow was introduced into the source after splitting 1:4. The ionization source parameters were nebulizer gas, 12 a.u.; curtain gas, 6 a.u.; ionspray voltage, 5500 V; heater gas, 350°C; and turbo gas, 7 l/min. Nitrogen was used as curtain gas, nebulizer gas, collision-activated dissociation gas, and turbo gas.
Systematic Workflow. The systematic workflow is schematically depicted in Fig. 1. First, both mass spectrometers were optimized for atazanavir and D5-atazanavir detection, and their fragmentation patterns were studied using an infusion pump connected directly to the electrospray source during continuous infusion of atazanavir and D5-atazanavir at a concentration of 1 μg/ml. Then, in step 2, patient samples were screened for knowledge-based predicted metabolites using linear ion trap mass spectrometry. The main atazanavir metabolic routes have been proposed to be glucuronidation, N-dealkylation, hydrolysis, and dehydrogenation (http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_Pharmr_P1A.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P1.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P2.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P3.pdf; http://www.fda.gov/cder/foi/nda/2003/21–567_Reyataz_BioPharmr_P4.pdf). Table 1 shows the molecular structures of possible atazanavir metabolites and the predicted mass-to-charge ratios (m/z) of the parent ions. If necessary, chromatography was adjusted to separate structurally similar metabolites, and sample pretreatment was adjusted for higher yields by concentration of the samples with liquid-liquid extraction. In step 3, possible metabolites were subjected to multistage mass spectral scans on the ion trap mass spectrometer for identification. In step 4, an MRM method was set up on the triple quadrupole mass spectrometer for semiquantitative metabolite screening based on the observed fragmentation patterns in step 3. If more sensitivity was needed for further metabolite identification on the ion trap mass spectrometer, patient samples containing relative high amounts of metabolites were screened using triple quadrupole mass spectrometry, and selected samples were subjected to extraction and further identification. In step 5, semiquantitative metabolic profiling was performed in patient samples. Metabolites were not available as pure standards; therefore, exact quantification was not possible. The metabolic ratio for each metabolite was determined dividing the area ratio of the metabolite by the area ratio of atazanavir in the same sample. The relative metabolite concentrations were then calculated with the formula: relative metabolite level = (metabolite molecular mass/atazanavir molecular mass) · metabolic ratio · atazanavir level.
Results
Method Optimization (Steps 1 and 2). After subjection of atazanavir and D5-atazanavir to collision-induced dissociation on the triple quadrupole mass spectrometer, four major fragments were found, as shown in Fig. 2, together with the proposed fragmentation. The optimal mass spectrometer settings for atazanavir and its corresponding fragments are shown in Table 2. The same optimal settings were found for the corresponding D5-atazanavir fragments. Similar fragments were found during multistage mass spectral scans on the linear ion trap mass spectrometer (see Fig. 3A).
During metabolite screening experiments on the ion trap mass spectrometer, a better signal/noise ratio for all the metabolites and atazanavir was observed in extracts obtained by liquid-liquid extraction. Less interference and higher signals were observed for all the compounds, most probably caused by cleaner extracts and sample concentration as a result of evaporation and reconstitution in a smaller volume. For further experiments, therefore, liquid-liquid extraction with tert-butyl methyl ether was used for sample pretreatment.
During initial metabolite screening (step 2), overlapping peaks were detected in chromatograms. To resolve this, a prolonged isocratic period was added before the start of the gradient elution. Final chromatographic conditions were as follows: eluent was delivered at a constant flow rate of 325 μl/min. At time 0, the flow consisted of 77.5% eluent A and 22.5% eluent B, and this composition was maintained for 45 min. Between time 45 and 125 min, the amount of eluent B was linearly increased from 22.5% to 60%. Thereafter, the column was flushed with 85% eluent B and subsequently reconditioned at the initial conditions of 77.5% A and 22.5% B for 5 min, resulting in a total run time of 135 min. The observed fragmentation patterns were used in the next step to identify the atazanavir metabolites.
Metabolite Identification Using Linear Ion Trap Mass Spectrometry (Step 3). Five potential atazanavir metabolites (M1–M5) were identified. No atazanavir glucuronide conjugates were found to circulate in plasma samples resulting from protein precipitation or liquid-liquid extraction. No metabolites or atazanavir was found in plasma extracts of patients not taking atazanavir. Proposed fragmentation pathways of M1 through M5 and atazanavir are depicted in Fig. 3. In most mass spectra, the fragment corresponding with the parent with a loss of a water molecule could be found, resulting in a loss of mass/charge ratio of 18.
Atazanavir. Atazanavir eluted after 96 min, and its proposed fragmentation pathway can be observed in Fig. 3A. The top panel shows the MS2 spectrum of the parent ion ([M+H]+ m/z 705). The middle panel shows the MS3 spectrum of the major fragment ([M+H]+ m/z 534) found in MS2 experiments, and the bottom panel shows the MS4 spectrum major fragment ([M+H]+ m/z 335) found in MS3 experiments.
M1. One N-dealkylation product was found to circulate in plasma ([M+H]+ m/z 538) (M1), eluting after 36.9 min (see Fig. 3B). The top panel of Fig. 3B shows the MS2 mass spectrum of M1. The panels below show the resulting MS3 spectra of the two major fragments ([M+H]+ m/z 335 and 367) found in MS2 experiments. The found fragmentation mass spectra corresponded with an N-dealkylation product resulting from cleavage of the (4-pyridin-2-ylphenyl)-methyl moiety from the atazanavir molecule.
M2 and M3. Two metabolites resulting from carbamate hydrolysis ([M+H]+ m/z 647) (M2 and M3) were detected to circulate in plasma, eluting after 38.7 and 45.1 min, respectively. Figure 3C shows the MS2 and the two respective MS3 mass spectra of the two major fragments found in MS2 experiments ([M+H]+ m/z 534 and 277) of M2.
Figure 3D shows the proposed fragmentation pathways of M3. The top panel shows the MS2 spectrum. The middle panel shows the MS3 spectrum of the fragment ([M+H]+ m/z 534) found in MS2 experiments, and the bottom panel shows the MS4 spectrum of the major fragment ([M+H]+ m/z 335) found in MS3 experiments. Found fragmentation patterns of M3 and M4 correlated well with atazanavir fragmentation (Fig. 3A) and confirmed their identity.
M4. An hydroxylated product ([M+H]+ m/z 721) (M4) was found, eluting after 75.3 min. The top panel of Fig. 3E shows the MS2 spectrum of the molecule. The middle panel shows the MS3 spectrum of the fragment ([M+H]+ m/z 550) found in MS2 experiments. Its fragmentation corresponded with the MS3 fragmentation of atazanavir ([M+H]+ m/z 534) but confirming the presence of an hydroxyl group on the fragment [M+H]+ m/z 379 in MS3 experiments. The bottom panel shows the MS4 fragmentation of the fragment containing the putative hydroxyl group ([M+H]+ m/z 379) found in MS3 experiments. As observed, the hydroxyl group on M4 was traced to be situated on the phenylbutane moiety of the molecule, as a result of either aliphatic or aromatic hydroxylation.
M5. The MS2 spectrum of the compound eluting at 101.4 min (Fig. 3F, top) corresponded with an atazanavir keto-metabolite ([M+H]+ m/z 719). The middle panel the MS3 spectrum of the fragment ([M+H]+ m/z 548) was found in MS2 experiments and corresponded well with MS3 spectra found for atazanavir and M1 through M4 but with the presence of a carbon-oxygen double bond. The bottom panel shows the MS4 spectrum of the major fragment found in MS3 experiments ([M+H]+ m/z 349) and shows that the oxygen atom was situated on either the tert-butyl moiety or the methyl-carbamate moiety.
Metabolic Profiling Using Triple Quadrupole Mass Spectrometry (Steps 4 and 5). Chosen mass transitions and compound-specific settings for MRM of the metabolites are shown in Table 2. As observed, for the metabolites either the fragment of m/z 144 (M1) or m/z 168 (M2–M5) was monitored in Q3. For calculation of the metabolic ratio for M1, the atazanavir response in the mass transition m/z 705→144 was monitored, and the mass transition m/z 705→168 was used for all the other metabolites.
Figure 4 shows the typical chromatograms for atazanavir and identified metabolites in a patient sample, recorded in MRM mode on the triple quadrupole mass spectrometer. Identity of the metabolites was confirmed with their retention time and specific mass transition. In the mass transition window of M5 (m/z 719→168), besides the main peak corresponding with M5, eluting at 101.4 min, other eluting compounds were observed, eluting at 78.2, 81.0, 82.8, 87.2, and 91.2 min. These possible atazanavir metabolites were not detected on the linear ion trap mass spectrometer during screening experiments. The exact identity of these compounds remains to be elucidated, but detector response in this specific mass transition and their absence in controls indicated that these compounds may be atazanavir metabolites.
Twelve patient trough samples were available for metabolic profiling. Eleven patients were taking 300 mg of atazanavir once daily, boosted with 100 mg of ritonavir. One patient used 400 mg of unboosted atazanavir once daily. No obvious differences in the metabolic profile were observed in the patient taking atazanavir without ritonavir. Figure 5A depicts the response for atazanavir (atazanavir peak area/D5-atazanavir peak area) versus atazanavir concentrations, showing a positive linear relationship between them. Figure 5B shows the observed relative metabolite concentrations of M5 versus the concentration of atazanavir. As shown, the metabolite concentrations correlate with atazanavir concentrations, but variability was high. All the measured atazanavir concentrations, arbitrary metabolite concentrations, their correlation with atazanavir concentrations, and interindividual variability expressed as the relative S.D. are shown in Table 3. As mentioned before, atazanavir trough concentrations are all positively correlated with all the other metabolite trough levels as well, and interindividual variability of the metabolic ratios was high.
Discussion
We have shown that the number of circulating atazanavir metabolites is higher than previously assumed. Using a systematic approach, we have screened for metabolites and identified five previously unknown atazanavir metabolites. No glucuronide conjugation products were found in plasma. Whether this finding is a result of limited glucuronide conjugation, fast glucuronide conjugate excretion, or limitations of the analytical capacity of the instrument remains to be elucidated. Because these glucuronide metabolites do not appear to abundantly circulate systemically, it is unlikely that these metabolites will influence treatment outcome. However, it may not be excluded that as a result of the liquid-liquid sample pretreatment glucuronide conjugates were not found because liquid-liquid extraction introduces an additional selectivity step. Furthermore, limitations in detector sensitivity and presence of current unknown metabolic pathways may cause certain metabolites not to be detected.
There are no data available on the influence of atazanavir metabolites on treatment outcome. However, they may significantly contribute to drug-drug interactions (e.g., as a result of inhibition of CYP3A4), atazanavir toxicity (because of inhibition of UGT1A1 or prolongation of the QT interval), or treatment efficacy. As shown in this small study with 12 patients, interindividual variability in atazanavir metabolism is high. Causes for variability in metabolic profiles (e.g., drug-drug interactions or pharmacogenetics) can be studied using the developed method. Furthermore, outlier metabolic profiles may be identified and correlated with treatment outcome.
The strengths of two different mass spectrometers were combined for identification, screening, and quantitation sensitivity. We have used the linear ion trap mass spectrometer for identification and screening purposes. By transferring the method to a triple quadrupole mass spectrometer, a sensitive MRM profiling method was developed allowing simultaneous determination of atazanavir and its metabolites in plasma, including possible keto-metabolites that could not be detected using the linear ion trap mass spectrometer. Linear ion trap mass spectrometers may not be routinely available in clinical laboratories. However, triple quadrupole mass spectrometers are becoming more available, and the developed atazanavir metabolic profiling method can be applied in other laboratories equipped with a triple quadrupole mass spectrometer after optimization for atazanavir and application of the proposed atazanavir metabolite fragmentation routes. Atazanavir metabolite trough levels were positively related with atazanavir trough levels, but interindividual variability was high.
The relative atazanavir metabolite concentrations were calculated to be low (<10% of the atazanavir concentrations). Measured metabolite levels in plasma depend on several factors. Not only does their formation depend on the amount of systemic circulating parent drug but also on the metabolite formation and excretion rate. The relative metabolite levels may be different from those reported, but their exact concentration remains unknown until reference substances are available for quantification. The developed profiling method can now be used for metabolic profiling of atazanavir in pharmacological studies.
Acknowledgments
We thank Robert Jansen for help with interpretation of the mass spectra.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.109.028258.
-
ABBREVIATIONS: MRM, multiple reaction monitoring; HPLC, high-performance liquid chromatography; a.u., arbitrary unit; MS2, double-stage mass spectrometry; MS3, triple-stage mass spectrometry; MS4, quadruple-stage mass spectrometry.
- Accepted June 18, 2009.
- Received April 28, 2009.
- The American Society for Pharmacology and Experimental Therapeutics