Introduction
Erythropoietin (EPO) is produced mainly in kidney cells and circulates to bone marrow where it stimulates erythroid progenitor cells to proliferate. The resulting increase in red cells and haemoglobin delivers more oxygen to muscles, and leads to enhanced athletic performance in humansReference Ekblom1 and horsesReference McKeever, Agans, Geiser, Lorimer and Maylin2. Recombinant human EPO (rHuEPO) became available in 1989 and almost immediately, athletes were rumoured to be doping with itReference Catlin, Hatton, Elliott, Foote and Molineux3. Twenty years later, human and equine sports are contending with many new rHuEPOs that are referred to as biosimilars or erythrocyte-stimulating agents (ESAs), such as dynepo, peg-epoetin β and rHuEPOs produced in various countriesReference Elliott4. All ESAs are prohibited in human and equine sports. Peg-epoetin β is the latest ESA to appear in doping. It was detected in 2008 in urine and plasma samples of Tour de France riders, then in 2009 in plasma samples from competitors in the 2008 Olympic GamesReference Lasne, Martin, Martin and de Ceaurriz5. It has not yet been reported in the urine or blood of racehorses.
Although sport persons were aware that rHuEPO was in use 20 years ago, there was no method for detecting abuse until 2000 when Lasne et al. Reference Lasne and de Ceaurriz6, Reference Lasne, Martin, Crepin and de Ceaurriz7 showed that it could be identified in human urine with a method based on isoelectric focusing (IEF) and a clever double-blotting (DB) techniqueReference Lasne8. We used the same technique in 2002 to detect darbepoetin α, the long-acting sequel to rHuEPO, in Olympic athletesReference Catlin, Breidbach, Elliott and Glaspy9. The IEF–DB method produces immunoblots that distinguish endogenous human EPO from ESAs based on characteristic band patterns. The test has survived numerous legal attacks virtually unscathed since it was introduced to human testing in 2000.
The IEF–DB test successfully detected epoetin α (Eprex®) and darbepoetin α (Aranesp®) in horse urine for several hours after administrationReference Lasne, Popot, Varlet-Marie, Martin, Martin and Bonnaire10. That work also showed that the bands arising from endogenous horse EPO were very faint. This was expected since the test uses an antibody directed at an epitope of human EPO, which would not be expected to completely cross-react with endogenous horse EPOReference Sato, Yamashita, Kugo, Hasegawa, Mitsui and Kijima-Suda11, Reference Kearns, Lenhart and McKeever12. Peg-epoetin β is difficult to detect in human urine with the IEF–DB test; however, it is readily detected in human plasmaReference Lasne, Martin, Martin and de Ceaurriz5. Herein, we describe how immunoaffinity followed by the IEF–DB method can detect it in horse plasma. We also show that an automated chemiluminescent immunometric assay designed to quantitate EPO in human plasma will quantitate peg-epoetin β equivalents in horse plasma.
Materials and methods
Animals
A healthy ~450 kg mature female Thoroughbred received a single intravenous bolus of 50 μg peg-epoetin β. Plasma was obtained from the contralateral jugular vein at times 0 (prior to administration), 2, 4, 6, 8, 24 and 48 h. Ten healthy Standardbred horses provided control plasma. All samples were stored at − 20°C before and after shipment on dry ice.
Drugs and standards
Pharmaceutical Mircera® (Hoffmann-La Roche, Roche Diagnostics GmbH, Mannheim, Germany; 50 μg/0.3 ml) was obtained from a pharmacy. Darbepoetin α (Aranesp®) was obtained from Amgen, Thousand Oaks, CA, USA. EPO biological reference preparation (BRP) was obtained from the European Directorate for the Quality of Medicines (Strasbourg, France). Human urinary EPO was obtained from the National Institute for Biological Standards and Controls (Potters Bar, Hertfordshire, UK). The monoclonal mouse antibody (anti-hEPO clone, AE7A5) was obtained from R&D Systems, Minneapolis, MN, USA. The secondary biotinylated polyclonal antibody (goat anti-mouse immunoglobulin G) was obtained from Thermo Fisher Scientific, Rockford, IL, USA. The streptavidin–horseradish peroxidase conjugate was obtained from Millipore, Temecula, CA, USA. The chemiluminescent substrate was obtained from Alpha Innotech, San Leandro, CA, USA.
Analytical methods
A rapid, automated and chemiluminescent immunometric assay (Immulite® 1000, Diagnostic Products, Inc., Los Angeles, CA, USA) designed to quantitate EPO in human plasma was used to quantitate peg-epoetin β in horse plasma. A standard curve was constructed from standards containing 0.0066, 0.0166, 0.0332, 0.1660 and 0.3320 μg ml− 1 of peg-epoetin β in 50 mM Tris–HCl, 1% bovine serum albumin, pH 7.4, diluted 1:100 in the Immulite® diluent.
Peg-epoetin β was extracted from plasma by immunoaffinity as described by Lasne et al. Reference Lasne, Martin, Martin and de Ceaurriz13. The immunoaffinity column utilized monoclonal anti-human EPO clone 9C21D11 antibodies (R&D Systems) conjugated to Affi-Gel® Hz gel (Bio-Rad Laboratories, Hercules, CA, USA). The column was conditioned with ~100 ml of phosphate-buffered saline (PBS), pH 7.4. The plasma samples were diluted 1:10 in PBS and applied to the column. The eluates were concentrated with a centrifugal filter with a cut-off of 30 kDa.
The immunoblot profile of peg-epoetin β was obtained with the IEF–DB test described by Lasne and de CeaurrizReference Lasne and de Ceaurriz6. IEF was conducted using a pH gradient of 2–6. The DB utilized a monoclonal anti-human antibody (AE7A5 from R&D Systems). The isoforms were revealed by chemiluminescence and the image captured with an imaging system (Fluor Chem CCD camera; Alpha Innotech).
Results
The immunoblot patterns for the standards and plasma samples collected from the treated horse at different times are shown in Fig. 1. The isoform pattern of the human urine EPO standard is shown in lane 1. Both lanes 2 and 11 show the pattern of darbepoetin α in the acidic region and BRP, which is a combination of epoetin α and β, in the basic region. As expected, lane 3 that contains the baseline plasma collected at time 0 is blank. The samples collected at 2, 4, 6, 8, 24 and 48 h after administration (lanes 4–9) all show the characteristic pattern of peg-epoetin β, which is shown for a standard in lane 10. For the plasma of five out of the ten control horses (results not shown), no isoforms were visible. The cathode and anode are indicated by a circled − and +.
Electropherogram showing isoelectric-focused patterns of standards and peg-epoetin β isolated from equine plasma with immunoaffinity chromatography. Lane 1, human urinary EPO standard; lanes 2 and 11, mixture of BRP standard (composed of epoetin α and epoetin β, appearing as the six bands in the basic region) and darbepoetin α standard (the four bands in the acidic region); lanes 3–9, plasma collected from horse treated with peg-epoetin β at 0 (baseline), 2, 4, 6, 8, 24 and 48 h; lane 10, peg-epoetin β standard. The most acidic band of peg-epoetin β is labelled band 1 (arrow) in lane 4
The peg-epoetin β standard has seven bands and the most acidic (band 1) is barely separated from band 2. The inter-band interval between bands 3 and 4 is greater than between bands 2 and 3. Band 6 is faint and band 7 is barely visible. The greatest inter-band intervals are between bands 5 and 6 and 6 and 7. At 24 and 48 h, the bands are fainter and only three or four of them are still discernible. This immunoblot of peg-epoetin β is distinct and unlike that of endogenous human or horse EPO, or any of the known ESAs.
The Immulite® assay was used to estimate the concentrations of peg-epoetin β in the horse plasma. (Fig. 2) The peak value was approximately equivalent to 800 pg ml− 1 in the 4-h sample. At 48 h, the value was equivalent to 200 pg ml− 1. The samples from all ten control horses were negative.
Plasma peg-epoetin β concentrations determined by the Immulite® 1000 assay following a single intravenous injection of 50 μg to a horse. The standard curve is shown in the inset
Discussion
We have shown that peg-epoetin β can be detected in horse plasma by the IEF–DB method and the Immulite® method for at least 48 h after intravenous administration of only 50 μg. The IEF–DB method requires at least 2 days, while the Immulite® assay takes minutes.
The IEF–DB test separates the isoforms or bands of peg-epoetin β by IEF. The second step (DB) prepares the bands for visualization by chemiluminescence. An isoform of EPO is a subset of the EPO molecules that has a defined charge at a given pH. Isoforms appear on the immunoblot as bands. The isoform pattern provides their number, position and density relative to each other. The number and position of the bands is a direct result of the structural characteristics of the molecules, and their relative density reflects relative abundance. The immunoblot of peg-epoetin β is distinct from that of any other known ESAs. The Immulite® test and the IEF–DB test are based on different principles yet both reveal the presence of peg-epoetin β. The Immulite® test result is numerical and indicates the presence of human EPO-like material, while the IEF–DB test provides a detailed visual pattern that can be compared to that of reference standards to identify the exact human EPO-like material present. The IEF–DB test has been extensively validated on human plasma and urine. The Immulite® test will require additional studies for complete validation.
Following the IEF–DB test, the time 0 plasma sample was blank, indicating, as expected, that no peg-epoetin β was present. The blank lane also indicates that no isoforms of endogenous equine EPO are detected. This is not surprising because the antibodies used during the immunoaffinity purification step and during the blotting process are anti-human EPO (clones 9C21D11 and AE7A5), and endogenous horse EPO is known to differ from human EPOReference Sato, Yamashita, Kugo, Hasegawa, Mitsui and Kijima-Suda11, Reference Kearns, Lenhart and McKeever12. In the only study where the IEF–DB test was applied to horse body fluids, the baseline urine did show faint bandsReference Lasne, Popot, Varlet-Marie, Martin, Martin and Bonnaire10.
Plasma samples from several athletes competing in cycling events and the Olympic Games have been reported to contain peg-epoetin βReference Lasne, Martin, Martin and de Ceaurriz5. This led the human doping control movement to develop an enzyme-linked immunosorbent assay (ELISA) screen for peg-epoetin β that uses a monoclonal anti-PEG antibody labelled with digoxigeninReference Lamon, Giraud, Egli, Smolander, Jarsch and Stubenrauch14. This method is rapid, inexpensive and partially validated. Lasne et al. Reference Lasne, Martin, Martin and de Ceaurriz13 also devised an immunoaffinity technique for extracting EPO from plasma. We used the same method on the horse plasma with good results.
The Immulite® 1000 test performed satisfactorily on the plasma from the study horse even though it is designed to detect human EPO in human plasma. This is somewhat surprising since the PEG groups might have shielded the relevant epitopes. In this single-dose study, the Immulite® assay was sufficient as a simple screening method to detect peg-epoetin β in horse plasma for 48 h following a dose of only 50 μg.
At this time, the optimal strategy for testing horses for peg-epoetin β is to screen their plasma with the Immulite® assay, and if the screen shows evidence of human EPO-like material, submit the sample to the IEF–DB procedure for confirmation of the immunoreactive material. In time, we expect that research studies with ELISA will provide new assays for EPO. The data herein pertain to ten control horses and one peg-epoetin-treated horse. If these results are repeatable in studies that are underway with other horses, the screen and confirmation test, if applied, will essentially eliminate the use of peg-epoetin β as a doping agent in the equine community.
Acknowledgements
The authors express their gratitude to the individuals and industry groups that provided funding to the Grayson–Jockey Club Research Foundation's Equine Drug Research Institute and to the New York State Racing and Wagering Board.
