Anal. Chem. 2004, 76, 5127-5135
Confirmation and Quantification of Hemoglobin-Based Oxygen Carriers in Equine and Human Plasma by Hyphenated Liquid Chromatography Tandem Mass Spectrometry Fuyu Guan,† Cornelius E. Uboh,*,†,‡ Lawrence R. Soma,† Yi Luo,† Jonathan S. Jahr,§ and Bernd Driessen†,§
School of Veterinary Medicine, Department of Clinical Studies, University of Pennsylvania, New Bolton Center Campus, Kennett Square, Pennsylvania 19348, PA Equine Toxicology and Research Laboratory, Department of Chemistry, West Chester University, 220 East Rosedale Avenue, West Chester, Pennsylvania 19382, and Department of Anesthesiology, David Geffen School of Medicine at UCLA, Los Angeles, California 90095-1778
Oxyglobin (OXY) and Hemopure (HMP) are produced from bovine hemoglobin (Hb) and were developed for the treatment of anemia in animal and human patients, respectively. Hemolink (HML) is a blood substitute of human Hb origin under development. The ability of these agents to carry oxygen in circulating blood and their promise to improve oxygen delivery to tissues supports the potential for their abuse in equine and human athletes. To deter athletes from abuse of these agents, a method has been developed for the detection, confirmation and quantification of OXY, HMP, and HML in equine and human plasma. OXY, HMP, and HML were extracted from equine or human plasma by solid-phase extraction using Bond Elut ENV cartridges and were digested by trypsin at 37 °C for 3 h. The tryptic digests were analyzed by LC-MS/MS, and tryptic peptides specific for bovine and human Hbs were targeted. OXY and HMP were detected, quantified, and confirmed using the y14 ion and b8 ion of the tryptic peptide from bovine Hb r chain residues 69-90, and HML was quantified using the tryptic peptide from human Hb r chain residues 63-91. The limit of detection for OXY in equine plasma and HML in human and equine plasma was 50 and 250 µg/mL for HMP in human and equine plasma. The limit of confirmation was 250 µg/mL for OXY in equine plasma, 500 µg/ mL for HML in human and equine plasma, and 1000 µg/ mL for HMP in human and equine plasma. The linear range for quantification was 50-5000 µg/mL for OXY in equine plasma and for HML in human and equine plasma, and 250-5000 µg/mL for HMP in human and equine plasma. The intraday and interday CV were less than 17% for quantification of OXY in equine plasma with external calibration. OXY was stable for more than 30 days at -20 and -70 °C. OXY was detected and quantified in equine * Corresponding author. Tel: +01-610-436-3501. Fax: +01-610-436-3504. E-mail:
[email protected]. † University of Pennsylvania. ‡ West Chester University. § David Geffen School of Medicine at UCLA. 10.1021/ac035430x CCC: $27.50 Published on Web 07/24/2004
© 2004 American Chemical Society
plasma up to 24 h following administration of a very low dose of OXY (32.5 g in 2 × 125 mL per horse), and its presence in equine plasma was confirmed up to 12 h postadministration. Hemoglobin glutamer-200 (Oxyglobin (OXY), Biopure, Cambridge, MA) and hemoglobin glutamer-250 (Hemopure (HMP), Biopure) are solutions of glutaraldehyde-polymerized bovine hemoglobin (Hb), and hemoglobin raffimer (Hemolink (HML), Hemosol, Toronto, Canada), is a solution of human Hb crosslinked and polymerized with o-raffinose. These hemoglobin-based oxygen carriers (HBOCs) were developed for treatment of anemia in both animal and human patients, regardless of the cause of anemia (hemolysis, blood loss, or ineffective erythropoiesis). OXY is approved and marketed in the United States and Europe for treatment of anemia in dogs. HMP, while already approved in South Africa, is currently awaiting approval in the United States by the U.S. Food and Drug Administration (FDA). It is intended for the treatment of acute anemic conditions in adult patients during surgery and for eliminating, reducing, or delaying the need for allogenic red blood cell transfusion. HML is still in phase III clinical trials. Though developed primarily as temporary blood substitutes, these agents are excellent candidates for abuse in human and equine athletes, due to their potential to increase oxygen-carrying capacity of circulating blood and to improve tissue oxygenation.1 All three HBOCs have a higher oxygen partition pressure (P50) than native Hb and, thus, release oxygen easily to the tissues, particularly in situations of high tissue oxygen demand, such as during maximum athletic performance.2-5 The results of a study of human volunteers during a submaximal exercise test and of exercising normovolemic pigs seem to support this (1) Schumacher, Y. O.; Schmid, A.; Dinkelmann, S.; Berg, A.; Northoff, H. Int. J. Sports Med. 2001, 22, 566-571. (2) Standl, T.; Freitag, M.; Burmeister, M. A.; Horn, E. P.; Wilhelm, S.; Am Esch, J. S. J. Vasc. Surg. 2003, 37, 859-865. (3) Standl, T. G.; Reeker, W.; Redmann, G.; Kochs, E.; Werner, C.; Schulte am Esch, J. Intensive Care Med. 1997, 23, 865-872. (4) Mallick, A.; Bodenham, A. R. Br. J. Hosp. Med. 1996, 55, 443-448. (5) Palaparthy, R.; Wang, H.; Gulati, A. Adv. Drug Delivery Rev. 2000, 40, 185198.
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004 5127
concept.6,7 To deter athletes from blood doping, the International Olympic Committee (IOC) banned the administration of any products that enhance the uptake, transport, or delivery of oxygen. In this pursuit, HBOCs are included in the latest list of Prohibited Classes of Substances and Prohibited Methods released by IOC. Likewise, the Association of Racing Commissioners International has banned OXY and HMP by including them in its list of Prohibited Practices. In advance of the 2004 Olympic Games, the World Anti-Doping Agency (WADA, the anti-doping arm of IOC) lists Compounds and/or methods that enhance oxygen delivery as one of the research priorities. Similarly, the Racing Medication and Testing Consortium (U.S.A.) has listed OXY and HMP as one of its Research Priorities for prohibited substances and “difficult to detect foreign substances” in racehorses. Thus, it is not surprising that the abuse of HBOCs for blood doping purposes is one of the challenges the international human sports community, horse racing and horse show industries, and equine forensic chemists are facing. To date, there is no method available for the detection and confirmation of HBOC in equine athletes. However, soon after this study was completed, an independent study on LC-MS of HMP in human plasma was reported.8 Techniques available for quantification of HBOCs in blood are based on the appearance of free Hb in plasma and the application of readily available bedside hemometers (HemoCue)9 or cooximeters to measure plasma Hb content. However, these methods lack specificity in that they do not distinguish between native Hb (from hemolyzed red blood cells) and HBOCs. Biological macromolecules such as proteins can be determined in either intact form or digested segments such as peptides. LC separation of proteins and peptides10 and MS of proteins and peptides11,12 have been reported. A conventional yet powerful technique used in proteomics involves two-dimensional gel electrophoresis separation followed by digestion. The technique provides very good separation, but it is a tedious process and is not easily adaptable to large-scale screening of samples. Solidphase extraction (SPE) is widely used for extracting small organic compounds from biological samples, but it is rarely used for the extraction of proteins from blood or plasma samples. In this study, SPE was successfully used for the separation of the macromolecular analytes from equine and human plasma. The purpose of this study was to address the concerns of the Pennsylvania Horse and Harness Racing Commissions over the persistent lack of an analytical method for screening, quantification, and confirmation of the presence of HBOCs in postrace equine samples. As a result, a reliable method for detection, confirmation, and quantification of OXY, HMP, and HML in human and equine plasma using LC/Q-TOF-MS/MS was developed. (6) Hughes, G. S., Jr.; Yancey, E. P.; Albrecht, R.; Locker, P. K.; Francom, S. F.; Orringer, E. P.; Antal, E. J.; Jacobs, E. E., Jr. Clin. Pharmacol. Ther. 1995, 58, 434-443. (7) Crago, M. S.; West, S. D.; McKenzie, J. E. Heart Vessels 1999, 14, 1-8. (8) Thevis, M.; Loo, R. R. O.; Loo, J. A.; Schanzer, W. Anal. Chem. 2003, 75, 3287-3293. (9) Lurie, F.; Jahr, J. S.; Driessen, B. Anesth. Analg. 2002, 95, 870-873. (10) Hancock, W. S. Handbook of HPLC for the separation of amino acids, peptides, and proteins; CRC Press: Boca Raton, FL, 1985. (11) Snyder, A. P. Interpreting Protein Mass Spectra: A Comprehensive Resource; Oxford University Press: Oxford, 2000. (12) Kinter, M.; Sherman, N. E. Protein Sequencing and Identification Using Tandem Mass Spectrometry; Wiley-Interscience: New York, 2000.
5128
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
EXPERIMENTAL SECTION Reagents. OXY solution (13 g/dL) was purchased from Biopure (Cambridge, MA), HMP solution (13 g/dL) and HML solution (10 g/dL) were kindly provided by Biopure and Hemosol Corp. (Toronto, Canada), respectively. Lyophilized powders of horse heart myoglobin and trypsin from bovine pancreas (TPCK treated, Sigma Catalog No. T1426) were purchased from Sigma (St. Louis, MO). Acetonitrile (HPLC grade) and ammonium bicarbonate (certified) were purchased from Fisher Scientific (Pittsburgh, PA), formic acid was purchased from EM Science (Gibbstown, NJ), and HPLC grade methanol and water were purchased from J.T. Baker (Phillipsburg, NJ). Control equine plasma (citrate dextrose as anticoagulant) and urine were collected from horses. Human plasma (containing citrate dextrose as anticoagulant) was purchased from Bioreclamation Inc. (Hicksville, NY). Preparation of Reagents. OXY, HMP, and HML (10.0 and 1.0 mg/mL in H2O) were freshly prepared by dilution of the respective standard solutions using deionized H2O. Trypsin solution (400 µg/mL in H2O) was prepared by weighing 2.0 mg of the lyophilized powder of trypsin standard and dissolving it in 5.0 mL of H2O. All protein solutions used were freshly prepared shortly before use and discarded soon after use. Horse heart myoglobin stock solution (200 µg/mL) was prepared by weighing 1.0 mg of the lyophilized powder of the standard and dissolving it in 5.0 mL of acetonitrile/H2O/formic acid (50/50/0.2, v/v/v) and was stored at 4 °C for less than 15 days. Horse heart myoglobin solution (10 µg/mL) for calibration of the Q-TOF mass spectrometer was freshly prepared by dilution of the stock solution with acetonitrile/H2O/formic acid (50/50/0.2, v/v/v). Ammonium bicarbonate stock solution (1.0 M) was prepared by dissolving 7.9 g of ammonium bicarbonate in 100 mL of H2O and was stored at 4 °C for less than 90 days. Ammonium bicarbonate buffer (50 mM) was freshly prepared by dilution of the stock solution with H2O. Formic acid (10%, v/v) was prepared by adding 1.0 mL of formic acid to 9.0 mL of H2O. Drug Administration and Sample Collection. The University of Pennsylvania Institutional Animal Care and Use Committee approved the study protocol for the administration of OXY to horses. Six female horses ranging in age from 4 to 10 years with an average weight of 550 ( 49 kg were used. OXY was intravenously administered to horses (32.5-g dose in 2 × 125 mL per horse). Blood and urine samples were collected before drug administration (0 h) and at various time intervals postdrug administration (2, 5, 15, 30, and 45 min, 1, 2, 4, 6, 8, 10, 12, 16, 20, and 24 h). Blood samples were collected via a venous catheter placed into the contralateral jugular vein to ensure that circulation of the administered agent was completed prior to obtaining the samples. Tubes containing potassium oxalate as an anticoagulant and sodium fluoride as an inhibitor of plasma esterases were used to collect blood samples. The samples were centrifuged (25003000 rpm or 776-1318g) at 4 °C for 15 min to obtain plasma. Each plasma sample was transferred into four separate tubes; two tubes of the plasma samples were stored at 4 °C and the remaining two were frozen and stored at -20 °C until analysis was performed. Urine samples were collected via an indwelling catheter placed in the bladder and attached to a drainage bag. Urine samples were divided into 15-mL aliquots and stored at -20 °C. Each aliquot
was used once to avoid any effects of freeze-thaw cycles on the concentration of OXY. Preparation of OXY, HMP, or HML Calibrators in Equine Plasma and Urine and Human Plasma. Control equine plasma and urine and human plasma used in this study were previously confirmed to be free of OXY, HMP, and HML by the same LCMS method described in this study. Control plasma (1 mL) was transferred into different disposable glass tubes (16 × 100 mm). To each plasma sample, OXY, HMP, or HML (50-50 000 µg in 19-100 µL of H2O) was added. Each sample was briefly vortex mixed. The analyte standard spiked into plasma was used as calibrator, and the spiked plasma samples were freshly prepared immediately before use. Quality control (QC) plasma samples were independently prepared using the same procedure. OXY calibrators and QC samples in equine urine were similarly prepared. Safety Precautions. Plasma, especially of human origin, must be handled at all times as a potentially infectious source, and thus, precautions must be exercised by using protective gloves and goggles. Glass tubes designated for disposal were soaked in 80% Chlorox solution for 12-24 h prior to disposal. Solid-Phase Extraction of OXY, HMP, or HML from Equine Plasma and Urine and Human Plasma. Solid-phase extraction of OXY or HMP or HML from equine plasma and urine and human plasma was performed using Bond Elut ENV cartridges (100 mg/3 mL; Varian, Harbor City, CA) using a Speedisk 48 Pressure Processor (J.T. Baker, Philipsburg, NJ). Bond Elut ENV cartridges were sequentially conditioned with 1.0 mL of methanol and 2. 0 mL of H2O. One milliliter of each sample was loaded onto the cartridge and allowed to pass through. The cartridge was sequentially rinsed with 1.0 mL of H2O and 1.0 mL of acetonitrile/H2O (90/10, v/v). Analyte was eluted with 1.0 mL of acetonitrile/H2O/formic acid (80/20/0.2, v/v/v) into a fresh glass culture tube (12 × 75 mm). All SPE steps including conditioning, sample loading, rinsing, and eluting were conducted by gravity. The eluent was dried at 80 °C under a steady stream of nitrogen or air using a sample concentrator (Techne Dri-Block DB‚3, Duxford, Cambridge, U.K.). The dried extracts were dissolved in 500 µL of ammonium bicarbonate buffer (50 mM, pH 7.8) prior to digestion. Enzyme Digestion of OXY, HMP, and HML. Trypsin digestion of OXY, HMP, and HML was performed by modification of an existing procedure.13 A 25-µL sample of trypsin (400 µg/ mL in H2O) was added to OXY, HMP, or HML extract in 500 µL of ammonium bicarbonate buffer (50 mM, pH 7.8) and mixed. The mixture was incubated in a water bath (model 1245PC, VWR Scientific, Bridgeport, NJ) at 37 °C for 3 h. The digestion was stopped by adding 20 µL of formic acid (10%, v/v) to each of the mixtures and mixing. An aliquot of 100 µL of the digestion solution was transferred to a 250-µL vial insert for LC-MS analysis. Instrumentation and Operating Parameters. Analysis of protein digests was performed by LC-MS. The system comprised a Hewlett-Packard 1100 LC binary pump with an on-line vacuum degasser and an autosampler (Agilent, Wilmington, DE) and a Q-TOF mass spectrometer equipped with a z-spray electrospray ionization (ESI) source (Micromass, Manchester, U.K.). (13) Kinter, M.; Sherman, N. E. In Protein Sequencing and Identification Using Tandem Mass Spectrometry; Wiley-Interscience: New York, 2000; pp 160163.
Table 1. Gradient in LC Mobile-Phase Composition and Flow Rate time (min) 0 0.5 13.0 13.5 15.0 15.5 16.0 20.0 20.5 21.0 B% a 15 15 40 80 80 15 15 15 15 15 flow rate 0.2 0.2 0.2 0.2 0.2 0.2 0.4 0.4 0.2 0.2 (mL/min) a LC mobile phase A: acetonitrile/H O/formic acid (5/95/0.2, 2 v/v/v). LC mobile phase B: acetonitrile/H2O/formic acid (95/5/0.2, v/v/v).
LC separation of HBOCs digests was performed on a Zorbax 300SB-C8 column (2.1 × 50 mm, 3.5 µm) with a Zorbax 300SB-C8 guard column (2.1 × 12.5 mm) maintained at 27 °C. LC mobilephase gradient in composition and flow rate for resolution of tryptic peptides are shown in Table 1. Q-TOF mass spectrometry was operated in positive ion mode. The mass spectrometer was calibrated over mass range of m/z 500-1500 using horse heart myoglobin (10 µg/mL) in acetonitrile/H2O/formic acid (50/50/0.2, v/v/v). For LC-MS analysis of HBOC digests, the ESI source parameters were as follows: capillary, 3000 kV; cone, 35 V; extractor, 0 V; source block temperature, 120 °C; desolvation temperature, 400 °C. For MS/ MS experiments, the collision gas (argon) pressure was adjusted so that the analyzer vacuum readback was 2.0 × 10-5 mbar. Data acquisition and analysis were accomplished by MassLynx software version 3.5 (Micromass). RESULTS AND DISCUSSION Initial Attempt To Determine Intact OXY by LC-MS. Initially, the strategy was to determine intact OXY in equine plasma. Unlike many proteins, OXY showed one dominant mass spectrometric peak at m/z 1231.5 and two minor peaks at m/z 1253.5 and 1270.0. Changes in pH from below 3-4.7 to 6.8 resulted in only changes in the relative height of the mass spectrometric peaks of OXY but did not cause more mass spectrometric peaks to be generated from OXY. An LC-MS method was developed for the detection and quantification of intact OXY in equine plasma. The method was able to detect and quantify OXY spiked into the sample and was successful using unhemolyzed plasma in detecting OXY. However, the drawback in the method was interference by hemolyzed equine plasma that is commonly observed in racetrack equine samples. The interference came from a hemolysis product the retention time of which was quite close to that of OXY and the mass spectra was the same as that of OXY, and it could not be eliminated by means of mass spectrometry or LC. Thus, the method was considered unsuitable for confirmation of the presence of OXY in the sample. Nonetheless, the method was useful in evaluating the efficiency of SPE for the extraction of OXY from unhemolyzed equine plasma as described below. Extraction of OXY from Equine Plasma by SPE. Separation of target analyte proteins from plasma was quite challenging, due to the fact that plasma contains a large number of proteins and that the concentrations of these endogenous proteins may be considerably higher than those of the analyte proteins. Techniques available for separation of proteins from biological matrixes are conventional two-dimensional (2-D) gel electrophoresis, filtration based on the size of target protein molecules, and the latest 2-D Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5129
Figure 1. LC-MS/MS chromatograms of tryptic digests of blank equine plasma (a, c) and OXY (50 µg/mL) spiked into blank equine plasma (b, d). Graphs a-d are from bottom to top. The chromatographic peaks at retention time of 8.95 min in graph b and of 8.93 min in graph d indicated the presence of OXY. The chromatograms were reconstructed from the product ions of m/z 737.9 (a doubly charged y14 ion from the triply charged tryptic peptide from bovine Hb R chain residues 69-90) and of m/z 893.5 (a b8 ion from the same tryptic peptide). The ions used in reconstructing the chromatograms were “m/z 737.9 + 226.0” for graphs a and b, and “m/z 893.4 + 780.3” for graphs c and d. The collision energy was 30 × 2 eV for CID of m/z 737.9 and 35 × 1 eV for CID of m/z 893.5.
(ion-exchange and reversed-phase) liquid chromatography.14-17 These techniques are tedious and time-consuming and, thus, are not easily adaptable to high-throughput analysis. Although SPE is widely used for extraction of small organic molecules from plasma or blood, there are limited publications on the extraction of peptides18-21 and proteins22 from plasma or blood by SPE. In this study, SPE method was developed from preliminary experiments in which the retention of red-colored OXY on an Oasis HLB SPE cartridge (Waters Corp.) was visually noted. Further experiments showed that the SPE efficiency for OXY by the Oasis HLB cartridge was very low, probably due to small pore size (80 Å) of the cartridge sorbent. Proteins such as OXY are macromolecules (14) Bonneil, E.; Li, J.; Tremblay, T. L.; Bergeron, J. J.; Thibault, P. Electrophoresis 2002, 23, 3589-3598. (15) Liu, H.; Berger, S. J.; Chakraborty, A. B.; Plumb, R. S.; Cohen, S. A. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2002, 782, 267-289. (16) Gygi, S. P.; Rist, B.; Griffin, T. J.; Eng, J.; Aebersold, R. J. Proteome Res. 2002, 1, 47-54. (17) Clarke, N. J.; Crow, F. W.; Younkin, S.; Naylor, S. Anal. Biochem. 2001, 298, 32-39. (18) Darby, S. M.; Miller, M. L.; Allen, R. O.; LeBeau, M. J. Anal. Toxicol. 2001, 25, 8-14. (19) Marquez, C. D.; Lee, M. L.; Weintraub, S. T.; Smith, P. C. J. Chromatogr., B: Biomed. Sci. Appl. 1997, 700, 9-21. (20) Plumpton, C.; Haynes, W. G.; Webb, D. J.; Davenport, A. P. J. Cardiovasc. Pharmacol. 1995, 26, S34-S36. (21) O’Flynn, M. A.; Causon, R. C.; Brown, J.; Kageyama, S. J. Chromatogr., A 1988, 452, 469-476. (22) Ji, Q. C.; Gage, E. M.; Rodila, R.; Chang, M. S.; El-Shourbagy, T. A. Rapid Commun. Mass Spectrom. 2003, 17, 794-799.
5130 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
(for example, the diameter of hemoglobin molecule is 65 Å23), and separation by SPE requires a sorbent with large pore size (g300 Å) so that the macromolecules can move freely without any spacious restrictions inside pores of the sorbent. Unlike most SPE sorbents, sorbents in Bond Elut ENV cartridges have large pore size. The extraction of OXY from equine plasma by Bond Elut ENV was tested even though it is designed for small organic molecules from environmental samples. Bond Elut ENV resulted in higher efficiency and cleaner extracts in the recovery of OXY from plasma by SPE than Oasis HLB cartridges upon evaluation using the LC-MS method for intact OXY as previously described. The efficiency of extraction of intact OXY (100 µg/mL) by Bond Elut ENV was 13% from triplicate experiments. Even though the recovery of intact OXY from plasma was only 13%, the extraction efficiency was acceptable considering the significant quantities of endogenous proteins in plasma that could cause interference and yet the extract obtained was clean. In addition, given HBOCs concentration of ∼5 mg/mL in plasma predicted from the recommended dose,8 this extraction efficiency is sufficient. Digestion of OXY, HMP, and HML. Proteins can be digested by chemical reagents or enzymes to small peptides for elucidation of amino acid sequences. Trypsin is commonly used for digestion of proteins, and tryptic peptides from trypsin digestion of proteins are suitable for electrospray LC-MS study. Furthermore, tryptic (23) Stryer, L. In Biochemistry, 4th ed.; W. H. Freeman and Co.: New York, 1995; p 5.
Figure 2. LC-MS/MS chromatograms of tryptic digests of blank human plasma (a, c) and HMP (500 µg/mL) spiked into blank human plasma (b, d) showing detection of HMP by the chromatographic peaks at retention time of ∼8.9 min in graphs b and d. Graphs a-d are in the order from bottom to top. The ions used in reconstructing the chromatograms were “m/z 226 + 311 + 1136.7” (product ions of the doubly charged y14 ion at m/z 737.9) for graphs a and b and “m/z 893.4 + 780.3” for graphs c and d. The collision energy was 30 × 2 eV for CID of m/z 737.9 and 35 × 1 eV for CID of m/z 893.5.
peptides may form doubly, triply, or multiply charged ions in an ESI source, and collision-induced dissociation (CID) of these multiply charged ions may produce specific b-ion series and y-ion series unique to the precursor peptide and to the protein as well. Thus, trypsin was chosen for use in the digestion of OXY, HMP, and HML. Incubation time for digestion by trypsin was 3 h in this study compared with 24 h in a previous study,24 and the yield of tryptic peptides from OXY, as evaluated by LC-MS chromatographic peak heights of the peptides, was only slightly decreased. Thus, an incubation time of 3 h was chosen for this study to reduce lengthy digestion time and the total time of experiment without sacrificing the quality of the tryptic peptides formed. Detection, Confirmation, and Quantification of OXY, HMP, and HML. As concluded in another report related to this study,24 the tryptic peptide with amino acid sequence of ′AVEHLDDLPGALSELSDLHAHK′ from bovine Hb R chain residues 69-90 (Hb R T9) is specific for bovine Hb and thus to the identities of OXY and HMP since both preparations are of bovine Hb origin. The bovine Hb R T9 peptide formed triply charged ions in ESI source, and the triply charged peptide ion produced an abundant singly charged b8 ion at m/z 893.5 and a doubly charged y14 ion at m/z 737.9 via unintended in-source fragmentation. The b8 ion and the y14 ion were stronger in intensity than the triply charged precursor peptide ion, and thus, they were chosen for detection, confirmation, and quantification of OXY and (24) Guan, F.; Uboh, C.; Soma, L.; Luo, Y.; Driessen, B. Anal. Chem. 2004, 76, 5118-5126.
HMP. As shown in Figure 1, OXY spiked into equine plasma (50 µg/mL) was detected using the b8 ion at m/z 893.5 and y14 ion at m/z 737.9. HMP spiked to human plasma was also detected using the b8 ion and y14 ion (Figure 2). Similarly, the tryptic peptides with the amino acid sequences of ′VADALTNAVAHVDDMPNALSALSDLHAHK′ from human Hb R chain residues 63-91 (human Hb R T9) and ′FFESFGDLSTPDAVMGNPK′ from human Hb β chain residues 42-60 (abbreviated as human Hb β T5) are specific for human Hb and thus are the “fingerprints” of HML since it is of human Hb origin. The human Hb R T9 peptide formed triply charged ions at m/z 1000, and the human Hb β T5 peptide formed doubly charged ions at m/z 1030.7. Using the two specific peptides from human Hb R and β chains, HML spiked into human plasma was detected (Figure 3). Confirmation of the presence of OXY and HMP in the sample was conducted using the b-ion series and y-ion series from MS/ MS spectra of the doubly charged y14 ion at m/z 737.9 and singly charged b8 ion at m/z 893.5 of the triply charged bovine Hb R T9 peptide mentioned above. The presence of OXY (250 µg/mL) spiked into equine plasma was confirmed (Figure 4). Interpretation of b-ion series and y-ion series in the MS/MS spectra can be found in another related report.24 Confirmation of the presence of HML in the sample was accomplished by MS/MS spectra of the triply charged human Hb R T9 peptide mentioned above. The b-ion series and y-ion series of the peptide are listed in another related report.24 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5131
Figure 3. LC-MS/MS chromatograms of tryptic digests of blank human plasma (a, c) and HML (500 µg/mL) spiked into blank human plasma (b, d) showed detection of HML by the chromatographic peaks at retention time of ∼11.5 min in graph b and 9.6 min in graph d. Graphs a-d are in the order from bottom to top. The ions used in reconstructing the chromatograms were “m/z 999.9” (the triply charged tryptic peptide from human Hb R chain residues 63-91) for graphs a and b and “m/z 1030” (the doubly charged tryptic peptide from human Hb β chain residues 42-60) for graphs c and d. The collision energy was 35 × 3 eV for CID of m/z 999.9 and 30 × 2 eV for CID of m/z 1030.
Figure 4. MS/MS spectra of the y14 ion at m/z 737.9 (bottom panel) and the b8 ion at m/z 893.5 (top panel) of the tryptic peptide from bovine Hb R chain residues 69-90 showing that the presence of OXY (250 µg/mL) spiked into blank equine plasma was confirmed. The confirmation was based on the b-ion series and y-ion series of the y14 ion and b8 ion. Interpretation of the spectra can be found in another related report.24 5132 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
Table 2. LC Retention Time, Precursor Ions, and Product Ions for Detection, Quantitation, and Confirmation of OXY, HMP, and HML
specific tryptic peptidea
OXY
HMP
HML
bovine Hb R chain residues 69-90
bovine Hb R chain residues 69-90
human Hb R chain residues 63-90
human Hb β chain residues 42-60
11.52
9.55
retention time (min)
8.95
8.95
8.88
8.95
precursor peptide ions monitored
y14 at m/z 737.9 of the tryptic peptide ion m/z 737.9+ 226
b8 at m/z 893.5 of the tryptic peptide ion m/z 893.5+ 780.4
y14 at m/z 737.9 of the tryptic peptide ion m/z 226+311+ 1136.7b
b8 at m/z 893.5 of the tryptic peptide ion m/z 893.5+ 780.4
tryptic peptide ion at m/z 999.9 m/z 999.9
y-ion series, b-ion series
b-ion series
y-ion series, b-ion series
b-ion series
y-ion series, b-ion series
product ions for detection and quantitationa product ions for confirmationa
tryptic peptide ion at m/z 1030 m/z 1030
a See ref 24. b m/z 226, 311, and 1136.7 are the b3, a4, and y10 ions of the doubly charged peptide ion at m/z 737.9, respectively. For nomenclature of MS/MS product ions (such as b, a, and y series ions) of a peptide ion, see refs 26 and 27.
OXY spiked into equine plasma was quantified with external calibration using the product ions (m/z 226 + 737.9) of the doubly charged y14 ion at m/z 737.9 from bovine Hb R T9 mentioned above, HMP spiked into human plasma using the product ions of (m/z 226 + 311 + 1136.7) of the y14 ion at m/z 737.9, and HML spiked into human plasma using the ion at m/z 1000 (the triply charged human Hb R T9 peptide). All the information on detection, quantitation, and confirmation of OXY, HMP, and HML, such as the LC retention time, the precursor ions, and the product ions, are summarized in Table 2. It should be pointed out that the retention times for the specific peptides in this study are shorter than those in the other related study since the LC gradient used in this study was shorter to reduce the analysis time. Validation of the LC-MS/MS Method. The LC-MS method for detection, confirmation, and quantification of OXY in equine plasma was validated with respect to limits of detection, confirmation, quantification, and determination of accuracy and precision. The limit of detection (LOD) was 50 µg/mL at which concentration OXY resulted in an LC-MS/MS chromatographic peak with signal-to-noise ratio of >3, and the limit of confirmation (LOC) was 250 µg/mL at which concentration OXY resulted in MS/MS spectra with b-ion series and y-ion series that were reasonably and unambiguously interpreted. The linear quantification range was 50-5000 µg/mL. As summarized in Table 3, the intraday accuracy was 97-103% for OXY (100-2500 µg/mL); the interday accuracy was from 87 to 116%. The intraday and interday precision expressed as percent coefficient of variation (CV) was less than 18%, and thus, it was acceptable. OXY spiked into equine urine was also detected and confirmed. The LOD was 50 µg/mL, and the LOC was 250 µg/mL. The calibration curve for OX in equine urine was not linear. HMP and HML spiked into human and equine plasma were detected, confirmed, and quantified. The LOD was 250 µg/mL for HMP and 50 µg/mL for HML; the LOC was 1000 µg/mL for HMP and 500 µg/mL for HML, in human and equine plasma. The linear quantification range was 250-5000 µg/mL for HMP and 50-000 µg/mL for HML. Estimation of Measurement Uncertainty. The importance of providing a measurement of uncertainty (MU) estimate in
Table 3. Accuracya and Precision (CV)b for Quantification of OXY in Equine Plasma (n ) 6) intraday added (µg/mL)
detected (( SD, µg/mL)
accuracy (%)
100 500 2500
97 ( 16 489 ( 27 2573 ( 70
97 98 103
interday CV (%)
detected (( SD, µg/mL)
17 116 ( 14 5.6 433 ( 43 2.7 2547 ( 322
accuracy CV (%) (%) 116 87 102
12 10 13
a Accuracy ) quantified/added × 100. b Coefficient of variation (CV; %) ) standard deviation of the concentration quantified/mean of the concentration quantified × 100.
quantitative results has been described elsewhere by the authors.25 MU was estimated using laboratory control. MU for OXY in equine plasma was (36%. The MU is somewhat high but it is acceptable based on the small n (6) and the fact that we used a concentration (100 µg/mL) close to LOD (50 µg/mL). Stability of OXY under Different Storage Conditions. According to the manufacturer, OXY as a blood substitute should be stored at room temperature or refrigerated (2-30 °C) but not frozen. However, frozen samples can be stored usually for a longer period of time than unfrozen ones. Since postrace samples are not shipped to the laboratory immediately after the race and collection of the samples and because detection, confirmation, and quantification have to be demonstrated, the stability of OXY under simulation of the different storage conditions at the racetracks and in the laboratory was evaluated in this study. As shown in Table 4, OXY was stable for 11 days at 4 °C, for 35 days at -20 °C, and for 38 days at -70 °C. It should be noted that the stability experiments were discontinued right after 35 days at -20 °C and after 38 days at -70 °C and that OXY might be stable for a longer period of time at -20 and -70 °C than the time period for this (25) Luo, Y.; Rudy, J.; Uboh, C.; Soma, L.; Guan, F. J. Chromatogr., B: Biomed. Appl. 2004, 801, 173-184. (26) Kinter, M.; Sherman, N. E. Protein Sequencing and Identification using Tandem Mass Spectrometry; Wiley-Interscience: New York, 2000; pp 6877. (27) Roepstorff, P.; Fohlman, J. Biomed. Mass Spectrom. 1984, 11, 601.
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5133
Table 4. Stability of OXY under Different Storage Conditions 4 °C
-20 °C
-70 °C
concn days of days of days of (µg/mL) storage detected storage detected storage detected 100 100 100 100 100 100 100 500 500 500 500 500 500 500 2500 2500 2500 2500 2500 2500 2500
1 4 5 6 7 8 11 1 4 5 6 7 8 11 1 4 5 6 7 8 11
111 101 91 108 117 84 105 464 438 393 525 433 522 453 2282 2120 2188 2585 2335 2627 2571
7 15 21 28 35
110 101 101 142 99
8 15 21 28 38
55 106 142 109 94
7 15 21 28 35
491 614 503 520 455
8 15 21 28 38
466 432 445 452 421
7 15 21 28 35
2651 2956 2453 2255 2295
8 15 21 28 38
2934 2542 2518 2606 2551
evaluation suggested. Discontinuation after 38 days was necessary because storage of test samples beyond this time period before the final results are released to our commissions is unusual. The results on stability of OXY have important implications for storage of equine plasma samples for detection, confirmation, and quantification of OXY: the samples are stored at -20 or -70 °C for 35 days if analysis is not completed earlier than 30 days postsample collection and delivery to the laboratory for analysis. Analysis of Plasma Samples after OXY Administration. The LC-MS/MS method we have described was successfully applied to the analysis of equine plasma and urine samples obtained from six horses that were administered a very low dose of OXY (0.06 g/kg; ∼0.45 mL/kg of body weight ) total dose of 32.5 g in 2 × 125 mL per horse, iv, while the recommended dosage of OXY for dog by the manufacture is 10-30 mL/kg of body weight). OXY was detected in equine plasma up to 24 h postadministration and confirmed in equine plasma up to 12 h postadministration. OXY quantitation results for the administration plasma samples will be published elsewhere in a report on pharmacokinetics of Oxyglobin in the horse. However, OXY was not detected in all the urine samples collected from 2 min to 24 h postadministration of OXY, suggesting that urinary excretion of OXY is not the preferred route. The results show that the method is useful and applicable to the detection, confirmation, and quantification of OXY in equine plasma. Analysis of Racehorse Samples. The method was also applied to analysis of suspected equine plasma and urine samples submitted to our laboratory by the Pennsylvania Horse and Harness Racing Commissions. The samples were suspected to contain OXY because the urine samples were tinted red in color, and so they were analyzed for the presence of OXY. Results showed that all the suspected plasma and urine samples were negative for the presence of OXY. In a search for an explanation for the red coloration of the urine samples, the samples were further analyzed for equine Hb, and the specific tryptic peptide 5134 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
with amino acid sequence of ′VGDALTLAVGHLDDLPGALSNLSDLHAHK′ from equine Hb R chain residues 63-91, as reported in a related study,24 was targeted. Most of the red coloration in equine urine samples contained equine Hb, which accounted for the color of the samples. Compared to the LC-MS method for analysis of bovine hemoglobin-based oxygen therapeutics in human plasma recently reported,8 the LOD and LOC were different: 50 versus 250 µg/ mL8 and 250 µg/mL versus 1 mg/mL,8 respectively. Furthermore, SPE of HBOCs from equine and human plasma used in this study is more adaptable to sample preparation and automation than the filtration method used by the previous investigators.8 The use of an inexpensive form of trypsin in this method compared with the highly purified and expensive trypsin used by other investigators make the current method relatively inexpensive and easily adaptable to analysis of large-scale screening of test samples. CONCLUSION An LC-MS/MS method for the detection, confirmation, and quantification of OXY in equine plasma and urine, and of HMP and HML in human plasma, is presented. The method was validated regarding limit of detection, confirmation, quantification, accuracy, and precision. The stability of OXY in equine plasma under different storage conditions was evaluated, and the results showed that it was stable for more than 30 days at -20 and -70 °C. The method was successfully applied to the detection, confirmation, and quantification of OXY in equine plasma samples collected from horses to which OXY was administered. The method is useful for the detection, confirmation, and quantification of OXY, HMP, and HML in equine and human plasma. ACKNOWLEDGMENT This study was funded by the Pennsylvania Horse and Harness Racing Commissions for which we are very thankful. The authors are also thankful to Dr. Enrico Bucci of the University of Maryland for his useful suggestions. The authors used the opportunity to study the remainder samples of Hemopure and Hemolink solutions that were previously given by Biopure Corp. and Hemosol Corp,, respectively, to J.S.J. for other studies in his laboratories. The authors also express their appreciation to The Meadows Standardbred Owners Association at Meadows, Thoroughbred Horse Association at Philadelphia Park, The Horsemen Benevolent of Pennsylvania Association at Penn National, and The Horsemen Association at Pocono Downs for their cooperation, continuing support, and encouragement. SUPPORTING INFORMATION AVAILABLE (1) LC-MS/MS chromatograms of the tryptic digest of HMP spiked into blank equine plasma (500 µg/mL). (2) LC-MS/MS chromatograms of the tryptic digest of HML spiked into blank equine plasma (500 µg/mL). (3) MS/MS spectra showing that HMP was confirmed at 1000 µg/mL in human plasma and that HML was confirmed at 500 µg/mL in human plasma, (4) LCMS/MS chromatograms showing OXY was detectable up to 24 h postadministration of OXY to horses. (5) MS/MS spectra showing that OXY was confirmed up to 12 h postadministration of OXY to horses. (6) Calibration curve for quantitation of OXY spiked to
equine plasma by external standard calibration. (7) Calibration curve for quantitation of HMP spiked to human plasma by external standard calibration. (8) Calibration curve for quantitation of HML spiked to human plasma by external standard calibration. Response (y axis)/peak area for the chromatographic peak reconstructed using the ions m/z 999.9. x axis: concentration of
HML spiked to human plasma (in µg/mL). This material is available free of charge via the Internet at http://pubs.acs.org. Received for review December 4, 2003. Accepted June 10, 2004. AC035430X
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5135
