Printer Friendly

Detection of recombinant human erythropoietin in urine by isoelectric focusing.

Recombinant human erythropoietin (rHuEPO) [4] is a glycoprotein with 165 amino acids and three N-linked and one O-linked glycans that is expressed in Chinese hamster ovary cells transfected with DNA encoding HuEPO. HuEPO is the main factor responsible for the proliferation of erythrocytes in the human body. Approximately 90% of HuEPO production takes place in the kidneys whenever a tissue oxygen sensor detects oxygen depletion (1, 2). For patients with kidney failure, the loss of the main HuEPO production site leads to severe anemia, which was treated by blood transfusion until 1988 when rHuEPO became available. Subsequently, rHuEPO has been approved for the treatment of anemia attributable to renal failure, cancer, and other types of anemia. Doping with erythropoietic proteins such as rHuEPO and darbepoetin alfa is perhaps the most complex and serious issue facing sport authorities today. A prodigious amount of anecdotal data indicates that such drugs are widely used in some sports (3).

In endurance sports, the main performance-limiting factor is the oxygen-carrying capacity of the blood. rHuEPO enhances athletic performance by increasing the number of erythrocytes (4). The International Olympic Committee added rHuEPO to its "List of Prohibited Substances" in 1990, although at that time no method existed to detect it in body fluids. A method for detecting rHuEPO in urine by electrophoresis was first described in 1995 (5), and in 2000, Lasne and de Ceaurriz (6) introduced an isoelectric focusing (IEF) method coupled with a technique that reduced the nonspecific binding that accompanies immunoblotting. The technique, called "double-blotting", led to a practical method that is used to detect rHuEPO and darbepoetin alfa (a long-acting analog of rHuEPO) in the urine of athletes (7, 8).

Very little is known about the time course of urinary rHuEPO after its administration. An increase in immunoreactive EPO has been reported during rHuEPO administration but not in samples collected 4 days after the last dose (9), and Wide et al. (5) reported detecting rHuEPO in urine by electrophoresis up to 48 h after the last dose. In this study, our objective was to determine the time course of detectable rHuEPO in urine, using the improved IEF method (6), after nine doses of rHuEPO. We also describe numerical and visual criteria for determining whether rHuEPO is present.

Materials and Methods


rHuEPO (Epogen[TM]) was obtained from Amgen Inc. The sources of all other drugs, reagents, and chemicals have been described elsewhere (7).


We studied 13 men (9 Caucasians, 3 African Americans, and 1 Asian) and 12 women (11 Caucasians, 1 African American) between 23 and 40 years of age. The study participants were screened by medical history, physical examination, and blood and urine tests. Hematology screening tests included hematocrit, hemoglobin, macrocytes, and percentage reticulocytes. Serum screening tests were creatinine, ferritin, EPO, and soluble transferrin receptor (sTfr). The inclusion criteria included hematocrit of 36-45%, serum ferritin >30 [micro]/L, and regular physical exercise. Individuals were excluded if the history revealed major organ system disease, substance abuse, participation in any sport drug-testing program, or medications that affect the hematologic, hepatic, or renal systems. Females were screened for pregnancy and African Americans for sickle cell trait. The study was approved by the Institutional Review Board of the University of California at Los Angeles, and all study participants gave written informed consent.


We collected one urine from each of 59 men and 37 women between 18 and 44 years of age. The ethnicity of the individuals was 10 African Americans, 7 Asian, 48 Caucasians, 18 Hispanics, and 13 undesignated. Forty-nine individuals lived in Los Angeles, CA, and 47 lived in Salt Lake City, UT. Medical history examination determined that the individuals had no active disease and were not taking any medications affecting the hematologic, hepatic, or renal system. The minimum hematocrits for males and females were 40% and 35%, respectively. The study was approved by the Institutional Review Board of the University of California at Los Angeles, and all study participants gave written informed consent.


After 25 individuals qualified for the study, they were paired according to ethnicity, gender, and similar hematologic weighted erythropoietic index (10). One individual in each pair was randomly assigned to the epoetin alfa treatment group and the other to the placebo group. Five individuals who could not be paired were assigned to the epoetin alfa treatment group. The participants were blinded with regard to group. The 15 individuals in the epoetin alfa group (50 IU/kg) and the 10 in the placebo group (1 mL of 9 g/L NaCI) received epoetin alfa or saline subcutaneously between 0800 and 1100 on days 0, 2, 4, 7, 9,11,14,16, and 18. All participants received 325 mg/ day of ferrous sulfate (65 mg of elemental iron) by mouth from day -7 to day 25.

Blood and untimed urine samples were collected during the morning of study days -3, 0, 2, 4, 7, 9, 11, 14, 16, 18, 20, 21, 22, and 25. The last 4 study days corresponded to days 2, 3, 4, and 7 postadministration (washout). Blood and urines were collected immediately before administration of saline or epoetin alfa. Blood pressure, pulse, and queries regarding potential side effects were monitored on each visit. If the hematocrit exceeded 48%, the dose for that day was withheld; if the hematocrit exceeded 48% on two or more consecutive visits, the individual was excluded from the study. For all individuals, the red cell indices were measured within 8 h of collection. Serum and urine were stored at -70[degrees]C until analysis. The methods for determining the erythropoietic index have been described (10). In brief, hematocrit, percentage macrocytes, and percentage reticulocytes were determined with an ADVIA120 Hematology Analyzer (Bayer Diagnostics). The serum EPO concentrations were determined with an automated immunometric chemiluminescent assay (Immulite EPO; Diagnostic Products Corporation). The sTfr concentrations were determined by an automated immunonephelometric assay (Dade Behring GmbH). The performance characteristics of the serum EPO and sTfr assays have been described (11, 12).


IEF was performed as described previously (7). In brief, 20 mL of urine was subjected to ultrafiltration, and the retentate containing EPO was focused in an IEF gel (pH 2-6). The isoforms were double-blotted (13) and visualized by chemiluminescence. The emitted light was captured with a F1uorChem 8000 CCD camera (Alpha Innotech Corp.).


A representative example of an IEF gel with two lanes containing markers (lanes S) and four lanes of unknowns (lanes 1-4) is shown in Fig. 1. For the lanes representing samples whose rHuEPO content was unknown, band zero (0) is identified as the band with the same isoelectric point (pI) as the most acidic band in the rHuEPO marker. Starting from band 0, the bands toward the cathode are considered "basic" and the bands toward the anode are considered "acidic". Ideally, all bands with the same pI would line up perfectly horizontally across all lanes. In reality, the pH gradient created by ampholytes is subject to disturbances by sample constituents that lead to deflections of the lines of equal pH (iso-pH) and produce bands that have been characterized as "frowns" or "smiles". Smiles are bands with the left and right ends (tips) curved upward, such as band 0 in lane 4 of Fig. 1. Frowns are bands with the tips curved downward (not shown in Fig. 1). As seen in Fig. 1, the horizontal line connecting the 0 bands in the two lanes containing markers (lanes S) also passes through the outer tips of the band labeled 0 in lane 4, leading to the assignment of this band as band 0.



Three main criteria for identifying rHuEPO were developed. The first criterion was that bands that focus in the basic area of the lane, as determined by the location of the rHuEPO marker, must be darker than other bands in the same lane. The second criterion was that these bands must have the same pI values as the bands in the nearest lane containing a rHuEPO marker. The third criterion was that band 0 and the adjacent two bands in the direction of the cathode must be present. If all three of these visual criteria were met, the sample was considered to contain rHuEPO. If band 0 and the two adjacent basic bands were incompletely resolved or associated with any light or dark spots or imperfections, the sample was reanalyzed. If any of these criteria failed, the sample was categorized "rHuEPO not detectable". Four experienced readers followed the above three criteria to evaluate 48 images of lanes obtained from 24 individuals on days 3 and 7 postadministration. The images were randomized, and the readers were blinded. The instructions were to assign each image to the epoetin alfa or the placebo group.


Lane densitometry was carried out with F1uorChem AlphaEase (Ver. 2.01) software (Alpha Innotech Corp), using the 1D-MultiAnalysis, Auto Grid option. The lane template was positioned such that it included two lanes containing markers with a maximum of four unknown lanes in between. For baseline adjustment, the default "Auto Base" method was used, in which the left boundary of the respective lane was broken into 16 regions. The mean background of each of those 16 regions was then used to construct the baseline. The resulting peaks were auto-integrated, using the default settings. Peaks that were not integrated by the software were manually integrated. We defined a peak as one in which the baseline-subtracted height was >30% of its total peak height, i.e., the corresponding band was at least 30% darker than its immediate surroundings. This process is illustrated in Fig. 2, which shows the densitogram of lane 3 of Fig. 1. For each lane in the gel, we calculated a ratio of peak areas. We referred to this ratio as the two-band ratio (TBR) where:


[SIGMA]P[A.sub.2-BASIC] is the combined peak area of the two bands immediately adjacent to band 0 on the basic side, and [SIGMA]P[A.sub.2-ACIDIC] is the combined peak area of the two bands adjacent to band 0 on the acidic side. To avoid division by 0, if the latter two bands failed the 30% test, the denominator in the TBR was assigned the value of 1.



All computations were performed on a PC with a Pentium[TM] chip utilizing the statistical software suite "R" (14). We used the "survival" package with procedure "survreg" to calculate the geometric mean, x-axis scale, and various percentiles. Zero values were considered to be left-censored and the distribution to be log-normal. The Wilcoxon rank-sum test was used to determine differences between the epoetin alfa- and placebo-treated groups. P values were two-sided, and values <0.05 were considered statistically significant.



The baseline characteristics of the placebo and epoetin alfa individuals are shown in Table 1. All individuals were between the ages of 23 and 40 and were well matched for age, weight, height, exercise, and erythropoietic index. The participants completed 347 of the possible 350 visits. Of 135 possible rHuEPO injections, 3 doses were withheld from one individual who was then excluded from the IEF analyses. Two other individuals missed one dose of epoetin alfa but remained in the study.

The efficacy of EPO was established by calculating an erythropoietic index (on-score), which is a weighted index of rHuEPO activity (10). The five variables in the index are hematocrit, reticulocyte hematocrit, percentage macrocytes, serum concentration of EPO, and sTfr. The onscore was calculated using blood samples obtained 3 and 0 days before first drug administration (baseline); samples obtained on days 2, 4, 7, 9, 11, 14, 16, and 18 of treatment; and on days 2, 3, 4, and 7 after the last dose of epoetin alfa or placebo. The mean (SD) on-scores for all females and males in the placebo group on all days were 2.02 (0.045) and 2.27 (0.063), respectively, and there was no change over the 28 days of the study. For the females treated with epoetin alfa, the mean on-score steadily increased to a maximum of 2.77 on day 20 (day 2 postadministration) and then steadily decreased. For the male epoetin alfa group, the mean on-score steadily increased to a maximum of 2.91 on day 18 (last day of epoetin alfa administration) and then steadily decreased. For the male epoetin alfa group, the mean on-score was significantly higher than that of the placebo group for all days after day 2. For the female epoetin alfa group, the mean on-score was significantly higher than that of the placebo group on days 2 and 7-25.


Shown in Fig. 3 are representative examples of an IEF image obtained from the analysis of markers (lanes S), a negative quality-control urine (lane QCN), a clinical positive quality-control urine (lane QCP), and urines obtained from two epoetin alfa- and one placebo-treated individuals. The Lanes S (markers) contain 2 fmoles each of rHuEPO (the five bands closest to the cathode) and darbepoetin (the four bands closest to the anode).

Also shown in Fig. 3 are the isoform patterns of samples obtained on days 2, 3, and 4 postadministration from individual A in the placebo group (Fig. 3, section A). The placebo-group patterns were very similar to the patterns of the QCN. They were also similar to the patterns observed in the other 9 placebo-treated individuals and the day -3 and 0 samples from the 14 remaining individuals in the epoetin alfa group (not shown). Four samples obtained on days 2, 3, 4, and 7 postadministration from individual B in the epoetin alfa-treated group had patterns very similar to that of the QCP for all days (Fig. 3, section B). Four samples from individual C in the epoetin alfa-treated group, obtained on the same days, had a pattern that changed over time (Fig. 3, section C): close to QCP on day 2, close to QCN on day 7, and intermediate on days 3 and 4. Results for individual C were included in Fig. 3 to show such changes in patterns over time.


The four experienced readers properly identified the day 3 postadministration samples as being from the epoetin alfa group (n = 14) or the placebo group (n = 10). For the day 7 postadministration samples, the readers properly assigned the 10 placebo samples to the placebo group; however, for the 14 epoetin alfa samples, some were classified as from the placebo group (Fig. 4).


For 19 of the 96 controls, the TBR could not be calculated because of absent peaks (n = 15) and interfering spots (n = 4). Of the remaining 77 samples, 12 (16%) showed a TBR of 0. The 99th percentile was calculated to be 1.19. For the epoetin alfa group, the TBR values ranged from 0.42 to 91740 on the four postadministration days. None of the samples from the placebo group exceeded a TBR of 1.19 (range, 0-0.77). Of 95 samples analyzed from the epoetin alfa- and placebo-treated individuals on the 4 post-epoetin alfa days, only 3 samples (3%) contained too little EPO for evaluation. Those three samples came from three epoetin alfa-treated individuals on days 4 (n = 1) and 7 (n = 2) postadministration.




Shown in Fig. 4 are the differences between the TBR and the visual method of data analysis; Fig, 4 also provides data on the time course of detection of rHuEPO in urine. On post-epoetin alfa days 2, 3, 4, and 7, the TBR properly identified 13, 14 (all), 12, and 7 of the 14 samples, respectively. The visual readers properly identified all the samples on post-epoetin alfa day 3 and between 6 and 11 of the 14 samples on day 7. The percentages of positive results for the two methods of data analysis were identical for the post-epoetin alfa day 3 data and very similar on day 7.


The pattern of urinary isoforms of rHuEPO differs from that of endogenous EPO. The former are clustered into four or five bands in the most basic portion of the gel, whereas the latter, which include as many as 14 bands, overlap with and are parallel to the rHuEPO bands in the basic region but are also found in the more acidic region of the gel. Furthermore, after several days of epoetin alfa administration, the endogenous bands become less dense and eventually disappear. These results confirm the findings of Lasne and de Ceaurriz (6) and further establish that the method is suitable for detecting rHuEPO in urine.

Our results pertain to epoetin alfa, the drug administered; however, the results would be very similar if epoetin beta were administered. Epoetin beta has been shown to migrate like epoetin alfa, and it has one more band in the basic region (6). It is not possible to comment on the applicability of the method to epoetin omega (15) because its pattern of urinary isoforms has not been published.

When rHuEPO is administered to healthy or diseased individuals, it increases the hematocrit, reticulocytes, macrocytes, serum EPO concentration, and sTfr concentration (16). In this study, the increase in the weighted erythropoietic index had reached statistical significance vs the placebo group by day 2, and it continued to increase until it plateaued about day 12. The weighted erythropoietic index did not begin to decrease until the second day after rHuEPO was discontinued and thereafter decreased rapidly but was still greater than the index for the placebo group on day 7 postadministration. This confirms that an indirect index of erythropoietic activity, such as the onscore (10), is a valid indicator of rHuEPO activity.

The presence of rHuEPO in urine followed a similar course. Although endogenous HuEPO contains isoforms that focus in the same area as rHuEPO, there is a significant difference between epoetin alfa and placebo groups in the urinary EPO isoform patterns with respect to the density of the bands within one lane (6). The main difference is where the darkest bands are located. The time course of detection of urinary rHuEPO (Fig. 4) revealed that nearly all of the samples collected on days 2, 3, and 4 after epoetin alfa was discontinued contained rHuEPO. Thus, rHuEPO was highly detectable in urine for up to 4 days after we discontinued epoetin alfa that had been given nine times in 18 days. Between days 3 and 7 after the discontinuation of epoetin alfa, the number of cases detected decreased to -50%. Therefore, as with many performance-enhancing drugs, the most effective use of this test is in out-of-competition testing. In the urines of individuals B and C, there was hardly any detectable endogenous EPO. The reason for that could be down-regulation of endogenous EPO production (17,18). It could also be explained by the presence of overwhelming amounts of urinary rHuEPO, which led to the need to dilute the retentate, which in turn would make endogenous EPO harder to detect.

In 1995, Wide et al. (5) first reported detecting rHuEPO in urine by electrophoresis. Their assay was sufficient to detect rHuEPO in the urine of 11 of 11 individuals in samples collected up to 24 h after rHuEPO was administered (20 IU/kg three times per week) for 7-9 weeks. The detection rate decreased to 75% at 48 h. Urinary EPO has also been detected by RIA during rHuEPO administration (200 IU/kg on 5 of 10 days) but not 4 days after the last dose (8). In the current study, the last collection day was day 7 postadministration, and approximately one-half of the participants still had detectable urinary rHuEPO.

The visual and numerical methods for evaluating IEF data produced very similar results (Fig. 4). Four different data readers, using the three visual criteria, correctly identified the 14 individuals in the epoetin alfa group and the 10 individuals in the placebo group on samples collected 3 days after the last dose of epoetin alfa. On day 7 postadministration, when the amount of rHuEPO was expected to be decreasing and the serum concentrations of EPO had decreased (16), the readers still detected rHuEPO in -50% of the samples, although with a lesser degree of concordance.

Densitometry converts band patterns into numerical values that can be readily combined into various scores or ratios. Because the isoforms of HuEPO and rHuEPO focus in overlapping areas, whatever scores or ratios are used must account for the usual variation of HuEPO isoform patterns. We selected the TBR as our preferred method of numerical analysis after evaluating various other ratios that could be used to characterize urinary rHuEPO. Compared with numerical methods that require analyzing all of the bands, the advantages of the TBR method include simplicity (only four bands need to be analyzed), speed (the analysis and calculation times are considerably reduced), fewer lanes declared invalid because of interfering spots, and easier accommodation of lanes that are not in perfect vertical alignment. In addition, there is no ambiguity regarding assignment of band 0.

In our procedure, the isoform pattern is visually inspected and the three criteria described here are applied. Samples that do not meet these criteria are declared "no detectable rHuEPO". If the criteria are met but some bands overlap or have interfering spots, the samples are reanalyzed. The TBR is then determined by lane densitometry. If both the visual and numerical criteria are met, it is highly likely that the sample contains rHuEPO. From the perspective of statistical limitations, at this time our numerical data are sufficient to determine the 99th percentile. The 99.9th percentile or higher margins of safety require a larger number of samples. Accordingly, we believe it is advisable to accumulate more data and to participate in interlaboratory studies designed to enhance the criteria for stating that a sample contains rHuEPO.

We are grateful to David Shaskey, MD, for assistance in managing participants, Annie Ramseyer for technical work, Laure Dulce for assistance with sample collection, and C.K. Hatton, PhD, for providing continuing advice and assistance in evaluating the data and preparing the manuscript. We thank Amgen, Inc. (Thousand Oaks, CA) for providing rHuEPO. This work was supported by a grant funded jointly by the International Olympic Committee and the United States Anti-Doping Agency of the United States Olympic Committee.

Received October 31, 2002; accepted March 27, 2003.


(1.) Roberts D, Smith DJ. Erythropoietin: induction of synthesis to signal transduction. J Mol Endocrinol 1994;12:131-48.

(2.) Daghman NA, Elder GE, Savage GA, Winter PC, Maxwell AP, Lappin TR. Erythropoietin production: evidence for multiple oxygen sensing pathways. Ann Hematol 1999;78:275-8.

(3.) Cazzola M. Further concerns about the medical risks of blood doping. Haematologica 2002;87:232.

(4.) Berglund B, Ekblom B. Effect of recombinant human erythropoietin treatment on blood pressure and some haematological parameters in healthy men. J Intern Med 1991;229:125-30.

(5.) Wide L, Bengtsson C, Berglund B, Ekblom B. Detection in blood and urine of recombinant erythropoietin administered to healthy men. Med Sci Sports Exerc 1995;27:1569-76.

(6.) Lasne F, de Ceaurriz J. Recombinant erythropoietin in urine. Nature 2000;405:635.

(7.) Catlin DH, Breidbach A, Elliott S, Glaspy J. Comparison of isoelectric focusing patterns of darbepoetin alfa, recombinant human erythropoietin, and endogenous erythropoietin from human urine. Clin Chem 2002;48:2057-9.

(8.) Lasne F, Martin L, Crepin N, de Ceaurriz J. Detection of isoelectric profiles of erythropoietin in urine: differentiation of natural and administered recombinant hormones. Anal Biochem 2002;311: 119-26.

(9.) Souillard A, Audran M, Bressolle F, Gareau R, Duvallet A, Chanal JL. Pharmacokinetics and pharmacodynamics of recombinant human erythropoietin in athletes. Blood sampling and doping control. Br J Clin Pharmacol 1996;42:355-64.

(10.) Parisotto R, Wu M, Ashenden MJ, Emslie KR, Gore CJ, Howe C, et al. Detection of recombinant human erythropoietin abuse in athletes utilizing markers of altered erythropoiesis. Haematologica 2001;86:128-37.

(11.) Wians FH Jr, Urban JE, Kroft SH, Keffer JH. Soluble transferrin receptor (sTfR) concentration quantified using two sTfR kits: analytical and clinical performance characteristics. Clin Chim Acta 2001;303:75-81.

(12.) Benson EW, Hardy R, Chaffin C, Robinson CA, Konrad RJ. New automated chemiluminescent assay for erythropoietin. J Clin Lab Anal 2000;14:271-3.

(13.) Lasne F. Double-blotting: a solution to the problem of non-specific binding of secondary antibodies in immunoblotting procedures. J Immunol Methods 2001;253:125-31.

(14.) R Development Core Team. The Comprehensive R Archive Network, Ver. 1.6.0. (Accessed October 2002).

(15.) Sikole A, Spasovski G, Zafirov D, Polenakovic M. Epoetin omega for treatment of anemia in maintenance hemodialysis patients. Clin Nephrol 2002;57:237-45.

(16.) Parisotto R, Gore CJ, Emslie KR, Ashenden MJ, Brugnara C, Howe C, et al. A novel method utilising markers of altered erythropoiesis for the detection of recombinant human erythropoietin abuse in athletes. Haematologica 2000;85:564-72.

(17.) Koury ST, Koury MJ, Bondurant MC, Caro J, Graber SE. Quantitation of erythropoietin-producing cells in kidneys of mice by in situ hybridization: correlation with hematocrit, renal erythropoietin mRNA, and serum erythropoietin concentration. Blood 1989;74: 645-51.

(18.) Cazzola M, Mercuriali F, Brugnara C. Use of recombinant human erythropoietin outside the setting of uremia. Blood 1997;89: 4248-67.


[1] UCLA Olympic Analytical Laboratory, [2] Department of Family Medicine, Division of Sports Medicine, and [3)] Department of Molecular and Medical Pharmacology, University of California at Los Angeles, Los Angeles, CA 90025.

* Address correspondence to this author at: UCLA Olympic Analytical Laboratory, 2122 Granville Ave., Los Angeles, CA 90025. Fax 310-205-9177; e-mail

[4] Nonstandard abbreviations: rHuEPO, recombinant human erythropoietin; IEF, isoelectric focusing; sTfr, soluble serum transferrin receptor; TBR, two-band ratio; QCN, negative quality control; and QCP, positive quality control.
Table 1. Demographic characteristics and baseline data of
the individuals who received placebo and epoetin alfa.

 Placebo group Epoetin alfa group

 Mean Range Mean Range

Age, years 31.6 23-40 30 23-40
Weight, kg 71.1 55.9-95.9 68.7 53.6-86.4
Height, cm 174 163-190 174 160-190
Exercise, h/week 11.7 5.0-22.5 10.2 5.0-25.0
Creatinine, mg/L 9 08-12 10 08-12
Hematocrit, % 41.8 35.8-44.9 41.2 35.8-44.8
On-score 2.1 1.9-2.3 2.1 1.81-2.38

Male, n 5 8
Female, n 5 7

Caucasian, n 9 11
African-American, n 1 3
Asian, n 0 1
COPYRIGHT 2003 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Drug Monitoring and Toxicology
Author:Breidbach, Andreas; Catlin, Don H.; Green, Gary A.; Tregub, Inna; Truong, Henry; Gorzek, Jeffrey
Publication:Clinical Chemistry
Date:Jun 1, 2003
Previous Article:Plasma total cysteine, mortality, and cardiovascular disease hospitalizations: the Hordaland Homocysteine Study.
Next Article:Time-resolved fluorometric assay for detection of autoantibodies to glutamic acid decarboxylase (GAD65).

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters