Drug Testing and Analysis

Research article Received: 30 June 2013

Revised: 2 October 2013

Accepted: 2 October 2013

Published online in Wiley Online Library: 4 November 2013

(www.drugtestinganalysis.com) DOI 10.1002/dta.1579

Detection of recombinant EPO in blood and urine samples with EPO WGA MAIIA, IEF and SAR-PAGE after microdose injections Yvette Dehnes,* Alexandra Shalina and Linda Myrvold The misuse of microdoses of performance enhancing drugs like erythropoietin (EPO) constitutes a major challenge in doping analysis. When injected intravenously, the half-life of recombinant human EPO (rhEPO) like epoetin alfa, beta, and zeta is only a few hours and hence, the window for direct detection of rhEPO in urine is small. In order to investigate the detection window for rhEPO directly in blood and urine with a combined affinity chromatography and lateral flow immunoassay (EPO WGA MAIIA), we recruited nine healthy people who each received six intravenously injected microdoses (7.5 IU/kg) of NeoRecormon (epoetin beta) over a period of three weeks. Blood and urine samples were collected in the days following the injections and analyzed with EPO WGA MAIIA as well as the current validated methods for rhEPO; isoelectric focusing (IEF) and sarcosyl polyacrylamide gel electrophoresis (SAR-PAGE). For samples collected 18 h after a microdose, the sensitivity of the EPO WGA MAIIA assay was 100% in plasma and 87.5% in urine samples at the respective 98% specificity threshold levels. In comparison, the sensitivity in plasma and urine was 75% and 100%, respectively, with IEF, and 87.5% in plasma and 100% in urine when analyzed with SAR-PAGE. We conclude that EPO WGA MAIIA is a sensitive assay for the detection of rhEPO, with the potential of being a fast, supplemental screening assay for use in doping analysis. Copyright © 2013 John Wiley & Sons, Ltd. Keywords: erythropoietin; SAR-PAGE; isoelectric focusing; microdoses; doping control

Introduction

Drug Test. Analysis 2013, 5, 861–869

*

Correspondence to: Yvette Dehnes, Norwegian Doping Control Laboratory, Oslo University Hospital, Trondheimsveien 235, 0586 Oslo, Norway. E-mail: [email protected] Norwegian Doping Control Laboratory, Oslo University Hospital, Trondheimsveien 235, 0586 Oslo, Norway

Copyright © 2013 John Wiley & Sons, Ltd.

861

Erythropoietin (EPO) is a glycoprotein hormone produced mainly in the kidneys, which stimulates the proliferation and maturation of red blood cells in bone marrow.[1] Due to the increased oxygen transport capacity in blood, elevated erythropoiesis will increase an athlete’s performance, especially in endurance sports.[2–6] Recombinant human EPO (rhEPO) has been widely misused by athletes for many years, and its use in sports was prohibited by the International Olympic Committee (IOC) in 1990. Since 2000, misuse of rhEPO has been detected by the isoelectric focusing (IEF) and double-blotting procedure,[7] and later also SDS-or SAR-PAGE.[8–10] Due to the short half-life of rhEPO in the blood, it is only detectable in urine 3–7 days when using therapeutic doses (40–80 IU/kg, 3 x wk-1).[11–13] The physiological effect, on the other hand, lasts for several weeks.[2] After the launch of the EPO-test, cheaters started to inject small doses (microdosing) of rhEPO in order not to fail a doping test. Small doses of EPO (approximately 20–25 IU/kg) were reported to have a detection window with the IEF procedure of only 12–18 h[14] using the criteria of that time (TD2004EPO). However, the inclusion of SDS/SAR-PAGE as well as optimizations of the double blotting method over the years, have improved the sensitivity of the original IEF EPO-test. A new test for EPO, EPO WGA MAIIA (MAIIA Diagnostics), has been under development for some time.[15,16] The assay combines WGA affinity chromatography with lateral flow immunoassay, which takes place on a dipstick. Endogenous EPO and different EPO-analogues bind with different affinity to the WGA due to differences in terminal sialic acid residues as well as differences in poly-N-acetyl lactosamine (LacNAc) residues.[15,17–19] In this study we have used a research version of the assay, EPO

WGA MAIIA Isoform Distribution Kit. EPO-Fc fusion protein and Darbepoetin alfa have the highest affinities towards WGA of the tested EPO-analogues, followed by epoetin alfa and beta, while endogenous EPO has a weaker affinity for WGA. Pegylated epoetin beta, CERA (Continues Erythropoietin Receptor Activator), has an even lower WGA-affinity than endogenous EPO. WGA-bound EPO is eluted with the sugar N-acetyl glucosamine (GlcNAc). Separation of EPO-isoforms is achieved by using two different concentrations of GlcNAc for each sample, adjusted to give the best separation of epoetin alfa/beta and endogenous EPO. The elution buffer with a low concentration of GlcNAc releases only EPO-isoforms with lower affinity for WGA, while the elution buffer with a high concentration of GlcNAc releases all EPO-isoforms (total EPO). Released EPO will bind to anti-EPO in the capturing line, and the amount of bound EPO is detected by another anti-EPO labelled with carbon black nano-strings. The obtained blackness in the capturing line is quantified by an image scanner, and is proportional to the concentration of EPO in the range 10–600 ng/L. We have previously tested the assay on several different recombinant epoetins (e.g. epoetin alfa, beta, delta, omega, darbepoetin alfa, CERA) in both plasma and urine samples in cooperation with MAIIA Diagnostics (unpublished). Detection of CERA with the research kit is achieved by using an even lower concentration of GlcNAc in the elution buffer low than for rhEPO.

Drug Testing and Analysis

Y. Dehnes, A. Shalina and L. Myrvold

One obvious advantage with the MAIIA assay is that it is very fast; the assay itself takes less than 1 h, and about 3 h when the necessary affinity purification step that precedes the assay is included. In comparison, the current EPO-analysis methods, IEF and SDS/SAR-PAGE followed by double blotting both take 2–3 working days. If fit for purpose therefore, a screening assay for doping samples with EPO WGA MAIIA would mean a significant reduction in time spent. Recently, Ashenden et al. reported modest sensitivity of the EPO WGA MAIIA assay in blood samples collected 72 and 96 h after injection of 10–40 IU EPO/kg body weight.[20] We wanted to test the sensitivity of the MAIIA isoform assay on samples of both blood and urine collected in the first few days following intravenous administration of microdoses of rhEPO. Our aim was to keep the injected dose low enough not to cause a build up of rhEPO throughout the three weeks of injections, in order to keep a possible effect on haematological parameters to a minimum. In order to look at the applicability of EPO WGA MAIIA in routine doping analysis, the plasma and urine samples were also analyzed with IEF- and SAR-PAGE, using the criteria of the current technical document for EPO analysis (TD2013EPO).[21] We can report that EPO WGA MAIIA is a sensitive screening assay for the detection of microdoses of rhEPO (7.5 IU/kg) in doping samples, and has, similar to SAR-PAGE, a higher sensitivity than IEF for the detection of rhEPO in blood.

Experimental Materials Blood was collected in Vacuette K2E tubes from Greiner Bio-One GmbH (Kremsmünster, Austria). Devices for microfiltration; HPF Millex-HV filters (0.45 μm), Millex GV filters (0.22 μm) and polyvinylidene difluoride (PVDF) membranes (Durapore, Immobilon-P) were purchased from Millipore (Billerica, MA, USA). PBS-tablets were purchased from OXOID (Basingstoke, UK). Tris(hydroxymethyl) aminometane, glycine, glacial acetic acid and methanol (HPLCgrade) were obtained from Merck (Darmstadt, Germany). Acrylamide/bisacrylamide solution for IEF (PlusOne ReadySol IEF, 40% T, 3% C) and urea was from GE Healthcare (Uppsala, Sweden). Carrier ampholytes (Servalytes 2–4, 4–6, and 6–8) were purchased from Serva (Heidelberg, Germany). Dithiothreitol (DTT), ammonium peroxidisulfate (APS), 4-morpholinepropanesulfonic acid (MOPS), sodium dodecyl sulphate (SDS), sodium N-lauroylsarcosinate (Sarcosyl, SAR), N,N,N′,N′-tetramethylethylenediamine (TEMED) and chemiluminiscent peroxidase substrate (CPS 160) were from SigmaAldrich (St Louis, MO, USA). The primary antibody used, mouse anti-human EPO antibody (Clone AE7A5), was from R&D Systems (Oxford, UK), and the secondary antibody; a biotinylated polyclonal goat anti-mouse IgG (H + L) (ImmunoPure), was from Pierce (Rockford, IL, USA). Streptavidin-Horse Radish Peroxidase (HRP) complex was obtained from BioSpa (Milan, Italy). Sarcosyl polyacrylamide gel electrophoresis (SAR-PAGE) was performed using readymade BisTris-gels (NuPAGE, 10%, 1.5 mm), sample reducing agent (10x) and antioxidant from Invitrogen (Carlsbad, CA, USA). Molecular

weight standard (Precision Plus Protein Dual Color Standards) was purchased from BioRad (Hercules, CA, USA). NIBSC (human urinary erythropoietin, second international reference preparation) was from the National Institute for Biological Standards and Control (NIBSC, Potters Bar, UK) and BRP (2nd Biological Reference Preparation of erythropoietin (rhEPO)) was from the European Directorate for the Quality of Medicines (Strasbourg, France). The EPO Purification Kit containing anti-EPO monolith columns, UPD (Urine Precipitate Dissolvation) buffer, washing buffer, desorption buffer, Detergent Aid, Exposure Aid and adjustment buffer, as well as the EPO WGA MAIIA Isoform Distribution Kit and the EPO Quantification Kit were from MAIIA Diagnostics (Uppsala, Sweden). NeoRecormon was purchased from Roche (Mannheim, Germany).

Volunteers and samples Nine healthy and either recreationally active or well-trained, noncompeting students from the Norwegian School of Sports Sciences were recruited as volunteers for the study (S#1–S#9). The inclusion criteria were normal haematological parameters, including haemoglobin between 12.5–16.5 g/dl for men, 11.5–15.5 g/dl for women, blood pressure ≤ 90 (diastolic) and ≤ 140 (systolic) and no known medical condition (Table 1). The participants each received six intravenous injections of the recombinant EPO NeoRecormon (epoetin beta), given twice a week (Mondays and Thursdays) for three weeks. The injected dosages (approximately 7.5U/kg body weight) corresponded to 15U/kg/week, and was neither preceded nor followed by higher doses of rhEPO. One subject (S#9) was recruited later than the first eight, and did not receive the first injection. Another subject (S#2) received only four injections due to repeated problems with blood withdrawal during sample collection. After each injection, the participants were supervised for 30 min to make sure there were no adverse reactions to the injected drug. There was a medical doctor either present or nearby during the injections, which were given by registered nurses. This study was approved by the local data protection officer and the regional ethics committee. Informed consent was obtained in writing after informing all participants orally and in writing about the drugs and the study. Two control samples of blood taken two days apart were collected from each participant (apart from S#9 who only gave one control sample) prior to the first EPO injection. The first was a fasting blood sample for total haemoglobin mass measurement. Collection of the remaining blood was standardized as non-fasting samples and the participants stayed seated for 10 min prior to sampling. Blood was collected twice after each injection, apart from injection 3 (1 sample) and 6 (3 samples) (Table 2). After the first injection, blood was collected at approximately 17 and 72 h post injection. For the remaining five injections, blood was collected twice at 24, 48, 72, or 96 h post injection. Ten days after the sixth injection, a last blood sample was collected. Due to illness and travel only 7 and 6 participants were able to give a blood sample after the fifth (96 h) and sixth (48 and 72 h) injection, respectively. A stock of negative control sample was made by mixing the sera from ten different

Table 1. Volunteer characteristics

862

Women (4) Men (5)

Age

Weight (kg)

Hb (g/dl)

Hct (%)

BP (mmHg)

25.3 ± 2.4 25.0 ± 3.2

63.0 ± 8.9 76.2 ± 8.2

12.9 ± 0.7 14.9 ± 0.5

38.4 ± 1.9 43.0 ± 1.3

109/70 ± 7.8/6.6 126/72 ± 12.5/5.7

wileyonlinelibrary.com/journal/dta

Copyright © 2013 John Wiley & Sons, Ltd.

Drug Test. Analysis 2013, 5, 861–869

Drug Testing and Analysis

Screening for rhEPO in plasma and urine after microinjections

Table 2. Post-injection blood collection scheme. Blood samples from each volunteer were collected at 1-3 different time-points after each microdose Injection number Blood sample collection (h post last injection) Number of blood samples

1 18 8

2 92 8

24 9

employees from the laboratory, and this sample was analyzed in every MAIIA assay run. Blood was collected in 6-mL Vacuette® K2EDTA tubes and inverted 6–7 times according to the instructions of the manufacturer. The blood was homogenized on a roller for minimum 15 min prior to analysis on a Sysmex 2000i XT haematology analyzer (Sysmex, Skjetten, Norway) within 2–3 h of collection. The remaining blood was centrifuged (3200 rpm, 10 min), and the plasma transferred to separate tubes and stored at -30 °C until further analysis. Urine samples were collected at approx. 12, 24, or 36 h after some, but not all injections, as we had a greater focus on blood samples. The urine samples were stored at -30 °C until further analysis. EPO affinity purification Both the plasma and the urine samples were subjected to affinity purification with EPO Purification Kit (MAIIA Diagnostics) prior to analysis with the EPO MAIIA WGA assay, IEF or SAR-PAGE. Affinity purification with the EPO Purification Kit has been validated and described previously,[22] and was performed according to the manufacturer’s instructions (MAIIA Diagnostics). Briefly, 2 mL UPD-buffer was added to urine aliquots of 20 mL, the mixtures incubated for 10 min at room temperature (RT) to dissolve precipitates, followed by heating in a water bath (95 °C) for 9 min. After cooling, the samples were diluted with 20 mL dilution mix (1:20 Protection Aid, 1:20 Exposure Aid, 1:100 Detergent Aid in H2O) and filtered through HPF Millex-HV filters (0.45 μm). Urines were passed through the monolith columns at 0.8–1 mL/min using a vacuum manifold at a pressure of -5 kPa. After washing, the isolated EPO was eluted with 50 μL Desorption (low pH) buffer into 5 μL Adjustment buffer for pH-neutralization, with a resulting buffer composition of 0.1 M Bis-Tris pH 7.0, 0.1 M NaCl, 10 mM glycine, 0.1% Tween 20, 0.05% BSA and 0.02% NaN3. Samples were stored at -30 °C until analysis. The plasma samples (usually 700 μL) were pretreated with 1/11 part (70 μL) ethanol prior to affinity purification, in order to remove fibrinogen which can obstruct the pores of the affinity columns. After addition of ethanol, the samples were vortexed before centrifugation (10 000 g, 10 min). The supernatants were transferred to new tubes and diluted to 10% in buffer (20 mM Tris pH 7.5, 0.1 M NaCl, 0.1% Tween 20 and 0.02% NaN3). The diluted plasma was filtered through Millex GV filters (0.22 μm) and then affinity purified using the same columns and desorption buffers as for urine. Electrophoresis

Drug Test. Analysis 2013, 5, 861–869

48 9

24 9

4 48 9

5 72 9

24 8

6 96 7

48 7

72 6

240 9

incubation with primary antibody (1:1000), the bound EPO-antibodies were blotted in semi-dry, acidic conditions (0.7% acetic acid) to a second PVDF-membrane. After incubation with biotinylated secondary antibody (1:2000), strepdavidine-HRP complex (1:1500) and HRPsubstrate, EPO isoforms were visualized by chemiluminescence in a CCD (Charge Coupled Device) camera (Fusion FX7, Vilber Lourmat, Marne-la-Vallée, France and FUJIFILM LAS 1000 Plus, Science Imaging Scandinavia, Saltsjö-Boo, Sweden). We have tested several other HRPcomplexes which reportedly shall be of higher sensitivity (e.g. SuperSignal West Femto (Thermo) and WesternBright Sirius (Advansta)), in our laboratory; however, the best sensitivity has been achieved using CPS 160 (Sigma). EPO isoform profile analysis was performed using the GASepo software (v2.1, ARC Seibersdorf Research GmbH, Austria).[23] In the first round of analysis, we visualized our results with our LAS 1000 Plus camera. After the arrival of our new Fusion FX7 camera, 16 gels were reanalyzed. This resulted in the detection of rhEPO in 4 additional samples after SAR-PAGE analysis. For SAR-PAGE, 20 μL of eluate (or less, depending on measured EPO concentration) was mixed with 10x Reducing Agent (Invitrogen) and 4x SAR sample buffer, both to a final concentration of 1x. The samples were separated by SAR-PAGE at constant voltage (150 V) for 110 min on 10% TrisBis (1.5 mm) gels. After electrophoresis, the gels were incubated 3x5 min in Bjerrum buffer (48 mM Tris, 39 mM glycine, 1.3 mM SDS, 20% methanol) and the separated proteins were blotted semidry in Bjerrum buffer to PVDF-membranes (0.8 mA/cm2, 60 min, using a TransBlot SD Semi-Dry Transfer Cell, BioRad, Hercules, CA, USA). All subsequent steps were as described for the IEF gels. In doping analysis, a sample will be regarded as suspicious, and not positive, if it fulfils the criteria for the presence of, for example, rhEPO after the initial testing procedure (screening). Only when it also fulfils the identification criteria of the following confirmation procedure, will the sample be considered as an adverse analytical finding (positive). To avoid confusion, we will use this nomenclature here and identify samples only as suspicious after screening if criteria are met, even though rhEPO is clearly present. EPO WGA MAIIA The EPO WGA MAIIA Isoform Distribution Kit (Art. No.1310) was obtained from MAIIA Diagnostics and used in accordance with the instructions from the supplier. Briefly, the EPO WGA MAIIA strips, four for each sample for duplicate determination, were treated in five subsequent steps (one row per step, four wells per sample). In the first row the strips were placed in 25 μL of pre-washing buffer and incubated for 5 min. The second row was prepared by dispensing 4 x 25 μL of diluted eluate of each sample, and the third row by dispensing 25 μL of W elution buffer low (10 mM GlcNAc) to the first two wells and 25 μL of W elution buffer high (300 mM GlcNAc) to the remaining two wells per sample. 25 μL of anti-EPO-CBNS (carbon black nano-strings) was dispensed to the wells in the fourth row, and finally 25 μL

Copyright © 2013 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/dta

863

IEF and double-blotting was performed as described by Lasne et al.[7] Briefly, 20 μL of sample eluate (or less, depending on measured EPO concentration) was focused on an IEF-gel with a pH-range of 2–6. Focused proteins were electro blotted to a PVDF-membrane, incubated with 5 mM DTT in PBS (1 h, RT) and blocked with 5% skimmed milk powder in PBS (1 h, RT). After

3

Drug Testing and Analysis

Y. Dehnes, A. Shalina and L. Myrvold

washing solution to the last row. The last four steps required 10 min of incubation each. After incubation in elution buffer, the WGA zone was cut from the strip before the remaining strip was placed in the anti-EPO-CBNS well. After the washing the step, the strips were mounted on the provided template, dried and the obtained signal intensity in the anti-EPO zone of the strip was detected by an Epson Perfection V700 scanner (Epson, Sollentuna, Sweden) in accordance with the Scanner Detection Instruction (MAIIA Diagnostics). The MAIIA calc software calculated the intensity signal in the anti-EPO zone as delta blackness per pixel (dbpp).[24] Calculation of PMI and thresholds In order to obtain the best resolution between endogenous and recombinant EPO, the optimal concentration of W elution buffer low had to be determined prior to running the project samples. This was done in accordance with the instructions of the EPO WGA MAIIA kit by running a standard curve (epoetin beta, 10–600 ng/L) and varying concentrations (5–20 mM) of W elution low. Ten mM GlcNAc was used as W elution low in this study. The concentration values for the samples after treatment with W elution buffer low and high were calculated by using a four-parameter logistic curve fit program (ReaderFit from Hitachi Solutions America Ltd, South San Francisco, CA, USA). By calculating the ratio of the concentration values determined with elution buffer low and high, the Percentage of Migrated Isoforms (PMI) was established. All the samples were run in duplicate, and samples with a coefficient of variation over 15% were re-run. A screening assay should have high sensitivity, but can afford to have slightly lower specificity. As the EPO WGA MAIIA may have its greatest potential as a screening assay, this aspect was taken into account when calculating threshold values that would define suspicious samples, and PMI thresholds for urine and plasma samples were therefore calculated at a specificity level of 98%. Between subject (bs) and within subject (ws) standard deviations for plasma PMI were calculated using the nine individual non-fasting control samples and the six pairs of 0-control samples, respectively. For the urine samples, the standard deviation (SD) of previously measured athlete samples was used, due to lack of 0-control samples from the volunteers. The threshold PMI values for plasma and urine were calculated according to TPðPMIÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Mean  2:08 bs2 þ ws2 and TU(PMI) = Mean  2.08 SD, respectively. Graphing and statistics of the calculated PMI values was done in Prism5 software from GraphPad (La Jolla, CA, USA).

Results Assay performance A positive control (100 ng/L epoetin beta for the plasma samples, a rhEPO control from the manufacturer for the urine samples) and a

negative control (the plasma mix from laboratory personnel for plasma samples and an endogenous, urinary EPO control from the manufacturer for the urine samples) were analyzed with each run both for monitoring the run’s performance as well as the inter-assay reproducibility. After 14 assays, the PMI for the positive plasma control was 15.9 ± 3.6 (mean ± SD) and for the negative plasma control 68.4 ± 7.8 (Table 3). For the urinary controls, the mean PMI from seven assay runs was 20.4 ± 2.4 and 59.7 ± 3.7 for the positive and negative control, respectively. The PMI in plasma samples were generally slightly higher than the PMI in urine samples, both in assay controls and in samples from volunteers and athletes. The calculated CV for EPO-concentration duplicates from the low and high elution conditions are shown in Table 3. Samples with a duplicate CV > 15% were re-run, and their CV’s were excluded from the statistics. In most cases the high variation resulted from either a very low EPO-concentration in the sample or a too high concentration outside the linear range of the assay. In both cases adjustments were made by either diluting the sample less or more, respectively, in sample dilution buffer during assay setup. A total of 3.7% (16 of 437 duplicates) of the urine and plasma duplicates exceeded the 15% CV-limit (Table 3).

PMI before and after microdoses of rhEPO The PMI in the samples dropped significantly after an intravenously injected microdose rhEPO (7.5 IU/kg) (Figure 1). The mean PMI in plasma samples collected 18 h after the first microdose was 43.6 ± 6.7 (p < 0.001), and though the mean plasma PMI had increased to 58.8 ± 8.5 in the samples collected 24 h postinjection (of the second, third, and fifth microdose), it was still significantly lower than the samples collected prior to the EPO injections (p < 0.001). At 48, 72, and 96 h post-injection, there was no significant reduction in mean PMI. A similar reduction was seen for the urine samples, where the mean PMI was 37.0 ± 6.7 in samples collected 8–10 h after the first injection, and 47.4 ± 5.5 in the samples collected 6–9 h later (Figure 1B). At 23–25 h after a microdose injection, the mean PMI was still significantly lower (55.4 ± 7.9, p < 0.001) than the negative control (athlete samples, 68.3 ± 6.2). Ten days after the last injection the mean plasma PMI was 73.4 ± 5.7. There were individual differences in how quickly the plasma PMI bounced back towards pre-injection level after each microdose, probably reflecting differences in rhEPO excretion. This is illustrated in Figure 2 with the PMI-profiles for two of the volunteers. While only the plasma sample collected at the 18-h time-point had a PMI value below threshold for subject #4, also samples collected at 24, 48, and 72 h time-points were below threshold for S#5.

Table 3. EPO WGA MAIIA assay performance during the 14 runs of blood samples and 7 runs of urine samples. A positive and a negative control sample, either plasma or urine depending on sample matrix, was run with each assay for estimation of reproducibility Matrix

864

Plasma Urine

%CV, EPO concentration Low buffer

High Buffer

5.9 6.1

4.6 3.5

wileyonlinelibrary.com/journal/dta

Duplicates with CV > 15%

Mean PMI ± SD Control samples

Count

%

Pos. contr

Neg. contr.

12 of 336 4 of 101

3.6 4.0

15.9 ± 3.6 20.4 ± 2.4

68.4 ± 7.8 59.7 ± 3.7

Copyright © 2013 John Wiley & Sons, Ltd.

Drug Test. Analysis 2013, 5, 861–869

Drug Testing and Analysis

Screening for rhEPO in plasma and urine after microinjections

Figure 1. Distribution of PMI in plasma samples (A) and urine samples (B) before and after microinjections of epoetin beta. The samples collected at the same time points (e.g. 24 and 48 h), but after different injections, are grouped together according to time point. Number of samples (n) for each group is indicated, as well as group mean and standard deviation (SD). The stippled line in each graph indicates the calculated PMI threshold for plasma and urine, respectively.

Figure 2. Plasma and urine PMI profiles for two subjects prior to and during injections of rhEPO microdoses (7.5 IU/kg). A. The profile of subject #5, of which six plasma samples (filled circles) and four urine samples (open squares) are below the respective thresholds at 98% specificity. B. The profiles of subject #4, of which only one plasma sample (the 18 h time point) and two urine samples are below the respective threshold levels.

Calculation of threshold values

Drug Test. Analysis 2013, 5, 861–869

Copyright © 2013 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/dta

865

Prior to the EPO injections, the mean PMI for the plasma samples collected from the nine volunteers was 71.6 ± 4.4 (n = 15), slightly higher than the negative assay control (plasma mix, see above). Using this mean and the calculated within and between standard

deviations (ws = 3.6 and bs = 5.0), gave a plasma threshold of PMI = 59 at a 98% specificity level The mean PMI for 18 athlete urines (presumed negative from their IEF-profiles) was 68.3 ± 6.2, slightly higher than the negative urinary control from MAIIA Diagnostics. This gave rise to a threshold in urine of PMI = 55 at 98% specificity.

Drug Testing and Analysis

Y. Dehnes, A. Shalina and L. Myrvold

Sensitivity and specificity of EPO WGA MAIIA – plasma samples The assay showed a sensitivity of 100% for the 8 plasma samples collected 18 h after injection (Table 4). For the 26 plasma samples collected 24 h after a microdose (injections 2, 3, and 5; Table 2), the sensitivity was 50%, and after 48 h (25 samples in total) the sensitivity was reduced to 12%. At 72 and 96 h post-injection, 27% and 20%, respectively, of the 15 samples in each group had a PMI below the threshold of 59. Seven out of eight plasma samples collected after 18 h were confirmed positive for rhEPO with SAR-PAGE (Table 4). Of the 13 samples with PMI below threshold 24 h after an injection, 6 were confirmed positive. None of the plasma samples collected at later time points could be confirmed with SAR-PAGE. The observed specificity of the control plasma samples (n = 15) at the 98% specificity level was good with no samples below the threshold, and also the 9 samples collected 10 days after the sixth injection all had PMI above the threshold. Sensitivity of EPO WGA MAIIA – urine samples The sensitivity for urine samples collected 8–10 h after the first injection was 100%, and 15–19 h after injection the sensitivity in urine was down to 87.5% (Table 5). After 24 h, the sensitivity in urine was 50%, just as observed for the plasma samples. Amongst the 10 urine samples collected between 30 and 40 h after a microdose, one sample had PMI below 55; this sample fulfilled the criteria for both IEF and SAR-PAGE as well. At the later timepoints the sensitivity was 0% in urine. All the urine samples with PMI below threshold collected up to 19 h after injection were confirmed positive for rhEPO with SARPAGE (Table 5). Of the samples below threshold after 24 h and 30–40 h, 80% and 50%, respectively, were confirmed positive.

broadness, and several samples displayed a shift in mobility towards rhEPO (Figure 3A). For the blood samples collected 24 h after a microdose, the SAR-PAGE sensitivity was 60% (Figure 3A). The SAR-PAGE mobility of EPO in blood samples collected at later time-points (48, 72, and 96 h) was as for endogenous EPO (Figure 3B). Of the plasma samples identified suspicious for rhEPO with IEF, 75% of the samples collected 18 h after injection and 19% of those collected after 24 h were confirmed positive with SARPAGE (Table 4). Sensitivities of IEF and SAR-PAGE – urine samples All the urine samples collected up to 19 h after the first microdose fulfilled the criteria for the presence of rhEPO analyzed with IEF (Figure 4 A+B). After 24 h the sensitivity of IEF was 50%, and for the samples collected between 30 and 40 h post-injection 40% fulfilled the rhEPO criteria (Figure 4C). From 48 h and onwards the sensitivity for the urine samples with IEF was 0%. As for IEF, the SAR-PAGE sensitivity was 100% for all samples collected up to 19 h after the first microdose (Figure 5). For urine collected 24 h after the third, fifth, and sixth microdose 50% were considered suspicious, while the sensitivity was 20% for samples collected after 30–40 h. None of the samples from time-points 48 h and later, that either had a PMI below threshold or were found suspicious with IEF, were found suspicious when analyzed with SAR-PAGE. SAR-PAGE confirmed all the urine samples that were collected up to 19 h after injection and identified as suspicious for rhEPO with IEF (Table 5). Of the 11 suspicious samples detected with IEF 24 h after injection, 10 samples were confirmed positive with SAR-PAGE.

Discussion Sensitivities of IEF and SAR-PAGE – plasma samples The criteria used for analyzing samples for recombinant EPO with IEF and SAR-PAGE were according to TD2013EPO.[21] With IEF, 75% of the isoelectric EPO-profiles of the blood samples collected 18 h after the first microdose fulfilled the criteria and were considered suspicious for the presence of rhEPO (Table 4). The two 18-h samples that did not fulfil the criteria with IEF, were however, identified as suspicious by SAR-PAGE, and both had PMI below threshold with EPO MAIIA WGA. At the 24 h timepoint, the sensitivity of the IEF assay was reduced to 27%. No plasma samples collected at later time points (48, 72, and 96 h post-injection) fulfilled the criteria for rhEPO with IEF. When analyzed with SAR-PAGE, 87.5% of the plasma samples collected 18 h post-injection displayed an electrophoretic mobility differing from endogenous EPO, and fulfilled the criteria for rhEPO (Table 4). The samples had a diffuse recombinant EPO band above the endogenous EPO band of varying strength and

In this study we have analyzed 98 plasma samples and 56 urine samples with EPO WGA MAIIA, IEF, and SAR-PAGE, potentially positive for rhEPO following microdose injections. We find that EPO WGA MAIIA is a sensitive and very rapid assay for the detection of rhEPO in both blood and urine, which has the potential to be a valuable supplemental screening method in doping analysis. It has an assay principle that complements the current electrophoretic methods IEF and SDS/SAR-PAGE, as the chromatographic separation of epoetins and endogenous EPO in EPO WGA MAIIA is based on differences in lectin (WGA) affinity. We wanted to look at the detection window for sub-performanceenhancing doses of rhEPO, without preceding boosting doses that could affect our results. To avoid accumulative effect of the microdoses, intravenous doses were given far apart (72 and 96 h, respectively, every week) so that most of the rhEPO should be washed out before next injection. The rationale for this setup was that if we can detect – for a reasonable time post-injection – rhEPO

Table 4. Plasma results. Assay sensitivities and % suspicious samples confirmed positive, for blood samples collected at different times (given in hours) after microdose injections of epoetin beta (7.5 IU/kg) Sensitivity (% suspicious samples, initial testing) Time p.inj.

866

MAIIA IEF SAR-PAGE

18 h 100 % 75 % 87.5 %

wileyonlinelibrary.com/journal/dta

Samples confirmed positive (with SAR-PAGE)

24 h

48 h

18 h

24 h

50 % 27 % 60 %

12 % 0% 0%

87.5 % 75 %

46 % 19 %

Copyright © 2013 John Wiley & Sons, Ltd.

48 h 0% -

Drug Test. Analysis 2013, 5, 861–869

Drug Testing and Analysis

Screening for rhEPO in plasma and urine after microinjections

Table 5. Urine results. Assay sensitivities and % suspicious samples confirmed positive, for urine samples collected at different times after microdose injections of epoetin beta (7.5 IU/kg). The sensitivity at two different calculated threshold levels (98% and 95% specificity) are shown for EPO WGA MAIIA Sensitivity (% suspicious samples, initial testing) Time p.inj.

8-10 h

MAIIA_98 MAIIA_95 IEF SAR-PAGE

100 % 100 % 100 % 100 %

15-19 h 87.5 % 100 % 100 % 100 %

24 h

Samples confirmed positive (with SAR-PAGE)

30-40 h

50 % 75 % 50 % 50 %

10 % 20 % 40 % 20 %

8-10 h

15-19 h

24 h

100 % 100 % 100 %

100 % 100 % 100 %

80 % 55 % 91 %

30-40 h 50 % 50 % 50 %

A

A

B

B

Figure 3. SAR-PAGE of blood samples collected after rhEO microdoses. A. Samples collected 18 h (lanes 3 and 4, subjects #2 and #7) and 24 h (lanes 5-9, subjects #1, #9, #8, #6, #5) after the first microdose of 7.5 IU/ kg epoetin beta. After 24-h, the diffuse bands of rhEPO vary from clearly visible (lane 5), weaker but detectable (lanes 6-8), to not detectable (lane 9). D= Dynepo (epoetin delta). B. Samples collected after 48 h (lanes 4 and 5, subjects #5 and #1) and after 72 h (lanes 6 and 7, subjects #8 and #9). Note that the diffuse rhEPO band no longer is visible above the endogenous EPO band in any of the samples. NC= negative control sample, D= Dynepo, NESP (darbepoetin alfa); BRP (2nd Biological Reference Preparation of erythropoietin (rhEPO)). The contrast of the bands has been individually optimized.

Drug Test. Analysis 2013, 5, 861–869

Figure 4. IEF of urine samples collected (A) 8-10 h, (B) 22-26 h and (C) 37-39 h after a rhEPO microdose. A. 8-10 h after injection, the presence of rhEPO is clearly detectable in the basic area of the gel (bands 1-5) in addition to urinary EPO bands in the endogenous (α-ε) and acidic (A-D) areas. B. 22-26 h after the 5th injection, over 50 % of the samples still fulfill the detection criteria for rhEPO, here exemplified with subjects #2, #5, #6 and #9. C. Of the samples collected 37-39 h after the 2nd microdose, only two are suspicious for rhEPO (S#3 and S#6). B= BRP, N= NESP.

Copyright © 2013 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/dta

867

doses too small to have any measureable effect, this would be encouraging news as athletes will run a high risk failing a doping control using larger, effective doses. In previous studies looking at the detection of rhEPO microdoses, the subjects have either been given larger (boosting) doses prior to the microdoses or the doses of rhEPO have increased over time.[20,25] While this is a desired scheme from a performance point of view, the larger doses of rhEPO can be detected for several days after an injection with the current EPOanalyses,[9,11,13,26–28] and for weeks by the athlete biological passport (ABP)[29–31] due to the larger doses’ subsequent effect on haematological parameters. Twenty-four hours after a microdose, SAR-PAGE and EPO WGA MAIIA detected about twice as many suspicious blood samples (15 and 13, respectively) as did IEF (7 samples). Of these, 12 suspicious samples from MAIIA and 5 from IEF could be confirmed with SAR-PAGE (Table 4). The lesser sensitivity of IEF for detection of rhEPO in blood is well known and due to the more basic isoelectric profile of blood EPO compared to urinary

C

Drug Testing and Analysis

Y. Dehnes, A. Shalina and L. Myrvold

Figure 5. SAR-PAGE of urine samples from six subjects collected 8-10 h after the first microdose. Note the shift in apparent molecular weight towards that of BRP, the broader band width as well as the presence of two bands due to the presence of both endogenous and recombinant EPO in the urine samples. NC = negative urine control. The contrast of the bands has been individually optimized.

868

EPO. With EPO WGA MAIIA, the slight difference in WGA affinity of blood and urinary EPO goes in favour of blood EPO, which displays slightly higher PMI and hence, greater separation from rhEPO (Table 3). Since the EPO WGA MAIIA assay works well for plasma samples, it can easily be applied to the samples analyzed for haematological parameters. Even without a B-sample, suspicious results can be used for additional or targeted testing. The MAIIA-assay can analyze 15 samples (including controls) in duplicates per run, which only takes about 1 h (including preparations) if the samples are already affinity purified. Also, the similar sensitivity for rhEPO in blood indicates that EPO WGA MAIIA and SAR-PAGE could make up a compatible assay pair, combining high sensitivity and throughput in the screening assay (MAIIA), with high sensitivity and specificity in the confirmation assay (SAR-PAGE). This way, suspicious samples can be quickly identified for confirmatory analysis with SDS/SAR-PAGE. According to TD2013EPO, a confirmation procedure for rhEPO in urine includes the SDS/SAR-PAGE method, either alone or in combination with IEF, depending on the initial testing procedure.[21] Confirmation of rhEPO in blood samples is done with SDS/SAR-PAGE, independent of initial testing procedure. As all samples were affinity purified prior to initial testing, SAR-PAGE was used as the single confirmation procedure. The samples were only analyzed with each of the three methods once, so when looking at the number of suspicious samples that could be confirmed positive with SAR-PAGE, only MAIIA and IEF were compared. The new criteria for IEF and SAR/SDS-PAGE combined with the improvement in methods of later years (i.e. affinity purification, loading in wells, western blotting reagents with increased sensitivity) meant that the assumed non-performanceenhancing intravenous microdoses given in this study could be confirmed with existing methods for 18–24 h, some even longer. In urine, the MAIIA assay performed with slightly lower sensitivity at the set 98% specificity compared to IEF; one 15–19 h sample and three 30–40 h samples with PMI above threshold fulfilled the criteria for rhEPO with IEF (Table 5). When adjusting the threshold of EPO WGA MAIIA in urine to a specificity of 95% (TU (PMI) = Mean  1.64 SD = 58), the sensitivity increased to 100% for the samples collected 15–19 h after injection, and after 24 h the sensitivity was 75%. After 24 h, three of the additional five urine samples with PMI below threshold at 95% specificity could be confirmed with SAR-PAGE. We had however, only a limited number of samples for calculation of within and between subject variations. More data is needed, and preferably from several laboratories, in order to establish suitable decision limits for EPO WGA MAIIA in urine and plasma. Also, the level of sensitivity for a screening procedure must be weighed against the cost of too many confirmation procedures of negative samples. In this

wileyonlinelibrary.com/journal/dta

study, a total of ten blood samples collected between 48–96 h after injection that had PMI below the 98% threshold level for plasma, could not be confirmed positive with SAR-PAGE. Since we cannot confirm the results, these samples will be considered negated and negative in the context of doping analysis. At present, however, we do not really know whether they represent increased sensitivity or lower specificity of EPO WGA MAIIA. Recently Mørkeberg et al.[25] published a paper where they assessed the detection of microdoses using the same EPO WGA MAIIA assay as we have here, the only difference being the assay instructions for incubation times, which were slightly longer in the more recent lots that we used for this study. They collected blood and urine at 2, 6, 12, and 72 h after an injection of 10 IU/kg body weight in seven subjects. At an absolute threshold of 99% specificity, the sensitivity was 57% in both blood and urine at the 72 h time-point. In our study with 25% lower rhEPO doses (7.5 IU/kg), the sensitivity at 72 h was less than half of that (27%), and none of those samples could be confirmed with SAR-PAGE. However, the facts that six normal doses of rhEPO (50 IU/kg) were given prior to the microdoses, and that a confirmation procedure (SDS/SAR-PAGE) was not performed on the samples below threshold in Mørkeberg’s study, make comparison difficult.

Conclusion EPO WGA MAIIA is able to detect microdoses of EPO in blood and urine samples with sensitivity comparable to that of SAR-PAGE. It is a very rapid assay which is simple to use and its method principle complements that of IEF and SDS/SAR-PAGE. After proper validation and threshold establishment it could represent a fast and valuable supplemental screening tool in EPO analysis. Acknowledgements We wish to sincerely thank the students at the Norwegian School of Sports for participating in this study. We gratefully acknowledge the assistance from employees at the Section of Endocrinology and the Norwegian Doping Control Laboratory at Oslo University Hospital, Aker; in particular Lillian Broderstad and Mette Istad for technical assistance and Peter Hemmersbach for helpful comments on the manuscript.

References [1] G.L. Israels, E.D. Israels, in Erythropoietins and Erythropoiesis, (Eds: G. Molineux, M.A. Foote, S.G. Elliott), Birkhäuser Verlag, Basel, 2003, pp. 3. [2] K.I. Birkeland, J. Stray Gundersen, P. Hemmersbach, J. Hallén, E. Haug, R. Bahr. Effect of rhEPO administration on serum levels of sTfR and cycling performance. Med. Sci. Sports Exerc. 2000, 32, 1238. [3] B. Ekblom, B. Berglund. Effect of erythropoietin administration on maximal aerobic power. Scand. J. Med. Sci. Sports 1991, 1, 88. [4] M. Audran, R. Gareau, S. Matecki, F. Durand, C. Chenard, M.T. Sicart, B. Marion, F. Bressolle. Effects of erythropoietin administration in training athletes and possible indirect detection in doping control. Med. Sci. Sports Exerc. 1999, 31, 639. [5] P. Connes, S. Perrey, A. Varray, C. Prefaut, C. Caillaud. Faster oxygen uptake kinetics at the onset of submaximal cycling exercise following 4 weeks recombinant human erythropoietin (r-HuEPO) treatment. Pflügers Archiv - Eur. J. Physiol. 2003, 447, 231. [6] G. Ninot, P. Connes, C. Caillaud. Effects of recombinant human erythropoietin injections on physical self in endurance athletes. J. Sports Sci. 2006, 24, 383.

Copyright © 2013 John Wiley & Sons, Ltd.

Drug Test. Analysis 2013, 5, 861–869

Drug Testing and Analysis

Screening for rhEPO in plasma and urine after microinjections [7] F. Lasne, L. Martin, N. Crepin, J. de Ceaurriz. Detection of isoelectric profiles of erythropoietin in urine: differentiation of natural and administered recombinant hormones. Anal. Biochem. 2002, 311, 119. [8] C. Reichel, F. Abzieher, T. Geisendorfer. SARCOSYL-PAGE: a new method for the detection of MIRCERA-and EPO-doping in blood. Drug Test. Anal. 2009, 1, 494. [9] C. Reichel, R. Kulovics, V. Jordan, M. Watzinger, T. Geisendorfer. SDSPAGE of recombinant and endogenous erythropoietins: benefits and limitations of the method for application in doping control. Drug Test. Anal. 2009, 1, 43. [10] M. Kohler, C. Ayotte, P. Desharnais, U. Flenker, S. Lüdke, M. Thevis, E. Völker-Schänzer, W. Schänzer. Discrimination of Recombinant and Endogenous Urinary Erythropoietin by Calculating Relative Mobility Values from SDS Gels. Int. J. Sports Med. 2008, 29, 1. [11] S. Lamon, N. Robinson, P. Mangin, M. Saugy. Detection window of Darbepoetin-[alpha] following one single subcutaneous injection. Clin. Chim. Acta 2007, 379, 145. [12] J. Morkeberg, C. Lundby, G.R.O. Nissen-Lie, T.K. Nielsen, P. Hemmersbach, R. Damsgaard. Detection of Darbepoetin Alfa Misuse in Urine and Blood: A Preliminary Investigation. Med. Sci. Sports Exerc. 2007, 39, 1742. [13] G.R.O. Nissen-Lie, K. Birkeland, P. Hemmersbach, V. Skibeli. Serum sTfR Levels May Indicate Charge Profiling of Urinary r-hEPO in Doping Control. Med. Sci. Sports Exerc. 2004, 36, 588. [14] M. Ashenden, E. Varlet-Marie, F. Lasne, M. Audran. The effects of microdose recombinant human erythropoietin regimens in athletes. Haematologica 2006, 91, 1143. [15] M. Lönnberg, M. Andrén, G. Birgegård, M. Drevin, M. Garle, J. Carlsson. Rapid detection of erythropoiesis-stimulating agents in urine and serum. Anal. Biochem. 2012, 420, 101. [16] L. Franco Fraguas, J. Carlsson, M. Lönnberg. Lectin affinity chromatography as a tool to differentiate endogenous and recombinant erythropoietins. J. Chromatogr. A 2008, 1212, 82. [17] M. Takeuchi, S. Takasaki, H. Miyazaki, T. Kato, S. Hoshi, N. Kochibe. Comparative study of the asparagine-linked sugar chains of human erythropoietins purified from urine and the culture medium of recombinant chinese hamster ovary cells. J. Biol. Chem. 1988, 263, 3657. [18] E. Llop, R. Gutiérrez-Gallego, J. Segura, J. Mallorquí, J.A. Pascual. Structural analysis of the glycosylation of gene-activated erythropoietin (epoetin delta, Dynepo). Anal. Biochem. 2008, 383, 243. [19] J. Bones, N. McLoughlin, M. Hilliard, K. Wynne, B.L. Karger, P.M. Rudd. 2D-LC Analysis of BRP 3 Erythropoietin N-Glycosylation using Anion

[20] [21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29] [30] [31]

Exchange Fractionation and Hydrophilic Interaction UPLC Reveals Long Poly-N-Acetyl Lactosamine Extensions. Anal. Chem. 2011, 83, 4154. M. Ashenden, K. Sharpe, A. Garnham, C.J. Gore. Evaluation of the MAIIA dipstick test to detect recombinant human erythropoietin in plasma. J. Pharmaceut. Biomed. 2012, 67/68, 123. World Anti-Doping Agency. Available at: http://www.wada-ama.org/ Documents/World_Anti-Doping_Program/WADP-IS-Laboratories/ Technical_Documents/WADA-TD2013EPO-Harmonization-Analysisof-Recombinant-Erythropoietins-EN.pdf [28 June 2013]. Y. Dehnes, S. Lamon, M. Lönnberg. Erythropoietin (EPO) immunoaffinity columns - A powerful tool for purifying EPO and its recombinant analogues. J. Pharmaceut. Biomed. 2010, 53, 1028. I. Bajla, I. Holländer, M. Minichmayr, G. Gmeiner, C. Reichel. GASepo - a software solution for quantitative analysis of digital images in Epo doping control. Comput. Methods Programs Biomed. 2005, 80, 246. M. Lönnberg, M. Drevin, J. Carlsson. Ultra-sensitive immunochromatographic assay for quantitative determination of erythropoietin. J. Immunol. Methods 2008, 339, 236. J. Mørkeberg, K. Sharpe, K. Karstoft, M.J. Ashenden. Detection of microdoses of rhEPO with the MAIIA test. Scand. J. Med. Sci. Sports 2013, DOI: 10.1111/sms.12049. F. Lasne, J. Thioulouse, L. Martin, J. de Ceaurriz. Detection of recombinant human erythropoietin in urine for doping analysis: Interpretation of isoelectric profiles by discriminant analysis. Electrophoresis 2007, 28, 1875. Y. Dehnes, P. Hemmersbach. Effect of single doses of methoxypolyethylene glycol-epoetin beta (CERA, Mircera™) and epoetin delta (Dynepo™) on isoelectric erythropoietin profiles and haematological parameters. Drug Test. Anal. 2011, 3, 291. S. Lamon, J. Boccard, P.-E. Sottas, N. Glatz, G. Wuerzner, N. Robinson, M. Saugy. IEF pattern classification-derived criteria for the identification of epoetin-δ in urine. Electrophoresis 2010, 31, 1918. P.-E. Sottas, N. Robinson, O. Rabin, M. Saugy. The Athlete Biological Passport. Clin. Chem. 2011, 57, 969. K. Sharpe, M. Ashenden, Y. Schumacher. A third generation approach to detect erythropoietin abuse in athletes. Haematologica 2006, 91, 356. M. Zorzoli, F. Rossi. Case studies on ESA-doping as revealed by the Biological Passport. Drug Test. Anal. 2012, 4, 854.

869

Drug Test. Analysis 2013, 5, 861–869

Copyright © 2013 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/dta