Papers in Press. First published November 21, 2003 as doi:10.1373/clinchem.2003.026583
Clinical Chemistry 50:2 000 – 000 (2004)
Automation and Analytical Techniques
Detection of Hemoglobin-Based Oxygen Carriers in Human Serum for Doping Analysis: Screening by Electrophoresis Franc¸oise Lasne,1* Nathalie Crepin,1 Michael Ashenden,2 Michel Audran,3 and Jacques de Ceaurriz1
Background: Hemoglobin-based oxygen carriers (HBOCs) have recently been included in the International Olympic Committee and World Anti-Doping Agency lists of substances and methods prohibited in sports. To enforce this rule and deter abuse of HBOCs in elite sport, it is necessary to develop HBOC-specific screening and confirmation tests that are the usual steps in antidoping control analysis. Methods: We developed a screening method based on electrophoresis of serum samples cleared of haptoglobin (Hp). Four successive steps (immunoprecipitation of Hp, electrophoresis of the cleared serum, Western blotting of the separated proteins, and detection of hemoglobin-related molecules based on the peroxidase properties of the heme moiety), provided electropherograms that could be easily interpreted in terms of the presence of HBOCs. This method was tested with serum samples enriched with various types of HBOCs: polymerized, conjugated, and cross-linked hemoglobins. It was also applied to blood samples collected from 12 healthy volunteers who had been infused with either 30 or 45 g of Hemopure, a glutaraldehyde-polymerized bovine hemoglobin. Results: The method clearly detected the presence in serum of the various types of HBOCs tested and demonstrated no possible confusion with endogenous hemoglobin that may be present in cases of hemolysis. The
test was able to detect Hemopure for 4 –5 days after administration of 45 g to healthy individuals. Conclusions: The electrophoretic method is a simple, fast, and sensitive procedure that appears to fulfill the criteria of a screening test for the presence of HBOCs in antidoping control samples. © 2004 American Association for Clinical Chemistry
For several decades, endurance-oriented sports have been linked with the illegal practice of blood doping, the goal of which is to increase the amount of hemoglobin (Hb)4 available to transport oxygen to muscle and thereby improve aerobic performance (1 ). It is also clear that athletes will adopt novel doping strategies to avoid detection (2 ). Faced with the continued refinement of a urine-based test to detect the use of recombinant human erythropoietin (which has been the drug of choice during the past decade) (3 ) and the recent commercial release of a hemoglobin-based oxygen carrier (HBOC) used to substitute for the oxygen-carrying functions of erythrocytes (4 ), athletes may be tempted to experiment with the latter product in an attempt to obtain a surreptitious performance advantage. This possibility must be considered, because this HBOC reportedly provides better tissue oxygenation and exercise capacity than the equivalent Hb supplied by an erythrocyte infusion (5, 6 ). Although the current International Olympic Committee and World Anti-Doping Agency lists of prohibited substances have anticipated this type of doping and specifically prohibit the use of “modified hemoglobin products”, there is currently no test to detect the abuse of these agents. The short half-life of HBOCs in circulation implies that athletes would have to infuse HBOCs at or near the time of competition to obtain a meaningful
1
National Antidoping Laboratory, Chaˆtenay-Malabry, France. Science and Industry Against Blood Doping (SIAB) Research Consortium, Gold Coast, Australia. 3 Biophysical & Bioanalysis Laboratory, Faculty of Pharmacy, University Montpellier I, Montpellier, France. *Address correspondence to this author at: Laboratoire National de De´pistage du dopage, 143 avenue Roger Salengro, 92290 Chaˆtenay-Malabry, France. Fax 33-1-46603017; e-mail
[email protected]. Received August 28, 2003; accepted November 3, 2003. Previously published online at DOI: 10.1373/clinchem.2003.026583 2
4 Nonstandard abbreviations: Hb, hemoglobin; HBOC, hemoglobin-based oxygen carrier; and Hp, haptoglobin.
1 Copyright (C) 2003 by The American Association for Clinical Chemistry
2
Lasne et al.: Method for Detecting HBOCs in Doping Control Screening
performance benefit. Some HBOCS are not excreted by the kidney; therefore, urinary concentrations are too low and variable to be considered of any value in developing an effective test for HBOC use. This necessitates a bloodbased detection methodology and the requisite collection of a blood sample from the athlete at the competition site. A striking characteristic associated with HBOCs, which circulate in the plasma and have a product-dependent half-life in the range of 8 –24 h (7 ), is the pinkish-red discoloration of serum or plasma samples containing HBOC. Although this visual identification is certainly helpful in identifying the presence of HBOCs, it is insufficient evidence for a doping control sanction. To satisfy legal precedents that demand detection of a banned substance by two different technologies, we undertook to develop both a test that could serve as a rapid and simple screen to identify the presence of HBOCs and a separate confirmation method to provide irrefutable evidence of doping. Both methods had to meet two special requirements: (a) to be able to detect each of the various HBOCs currently under development and (b) to specifically differentiate them from natural Hb because low concentrations of Hb (⬍50 mg/L) are typically encountered in serum and much higher concentrations may be reached in the case of in vivo (pathologic conditions) or in vitro (after blood collection) hemolysis. We present here a method based on investigation of the electrophoretic patterns of HBOCs to fulfill the criteria of an efficient screening test.
Materials and Methods analytical samples After the establishment of appropriate Material Transfer and Confidentiality Agreements, the following HBOC products were provided by the respective pharmaceutical companies responsible for developing each product: Oxyglobin and Hemopure (glutaraldehyde-polymerized bovine Hb; Biopure Corporation), PolyHeme (glutaraldehyde-cross-linked pyridoxalated human Hb; Northfield Laboratories), PHP (pyridoxalated human Hb polyoxyethylene conjugate; Apex Bioscience Inc.), Hemospan (maleimide-polyethylene-glycol-conjugated human Hb; Sangart Inc.), and HemAssist (diaspirin-cross-linked human Hb; Baxter Healthcare Co.). Hemolink (Hemosol Inc) was not available for our studies. Natural Hb was obtained from red blood cells washed three times in a solution containing 9 g/L NaCl and lysed in water. Myoglobin from human heart was obtained from Sigma. Natural Hb, myoglobin, and the different HBOCs were all analyzed at 3 g/L in water and serum.
serum samples from individuals infused with hemopure Blood samples were obtained from 12 males (age range, 20 –35 years) infused with either 30 or 45 g of Hemopure. All were healthy and engaged in aerobic exercise for at
least 20 min three times per week, and all signed an informed consent form that was approved by the University of Montpellier Ethics Committee. Hemopure is available as a sterile solution of polymerized bovine Hb in a solution similar to Ringer’s lactate. The nominal (SD) Hb concentration of Hemopure is 130 (10) g/L, and the percentage of methemoglobin is ⬍15%. Hemopure was infused at a rate of 0.33 or 0.50 g/min over a 90-min period (for 30- and 45-g doses, respectively). Venous blood samples were collected into 8-mL serum tubes immediately at the end of the infusion procedure and on days 1, 2, 3, 5, and 8 postinfusion. Samples were also collected from six individuals on day 4. Serum samples were frozen and kept at ⫺20 °C until they were analyzed.
sample analysis Immunoprecipitation of haptoglobin (Hp). Polyclonal rabbit anti-human Hp antibodies from Dako (antibody titer, 5.1 g/L) were used either directly or concentrated fourfold by ultrafiltration in a Centricon YM30 (Mr cutoff, 30 000) from Millipore as indicated by the manufacturer. We mixed 30 L of undiluted serum samples with an equal volume of antibody solution. After 1 h of incubation at room temperature, the immune complexes were removed by centrifugation at 1800g for 10 min, and the supernatant was subjected to electrophoresis. Electrophoresis. We applied 2 L of the supernatant directly to agarose Hydragel Hemoglobin gels (Sebia). Electrophoresis was performed at 165 V constant voltage for 12 min in the Tris-barbital buffer (pH 9.2) purchased with the gels. For demonstration, serum samples were analyzed in the same electrophoretic system without immunoprecipitation of Hp. In these cases, 2 L of serum was applied directly to the gels. Western blotting. Because an agarose matrix was used for electrophoresis, a simple capillary blot was sufficient for transferring the separated proteins. At the end of the run, Immobilon P polyvinylidene fluoride membranes (Millipore) with a 0.45 m pore size were soaked in methanol, rinsed in distilled water, and equilibrated in 25 mmol/L Tris–192 mmol/L glycine buffer. The membranes were layered on the gels and covered with one sheet of filter paper soaked in the same buffer and three dry sheets. This assembly was placed between two glass plates and secured by two clamps. The contact was maintained for 45 min. Chemiluminescent detection of the heme moiety. The membranes were then rinsed in Tris– glycine buffer and covered with a chemiluminescent solution (Covalight; Covalab) prepared with four times more oxidant than indicated by the manufacturer. This modification was necessary because heme was used as catalyzer of the oxidation reaction. This preparation was layered onto the
3
Clinical Chemistry 50, No. 2, 2004
surface of the membrane and covered with a plastic sheet. Luminescence was detected by exposure in a Luminescent Image Analyzer LAS-1000 Plus CCD camera (Fuji Film).
Results electropherograms of human serum As shown in Fig. 1A, physiologic serum samples analyzed without removal of Hp (lane b) gave a faint luminescent band in a more anodal position than pure Hb (lane a). When increasing amounts of natural Hb were added to serum, this band was intensified (lane c), and beyond a certain concentration (depending on the analyzed serum sample), a second band in the position of pure Hb was observed (lanes d–f). If Hp was removed from serum by immunoprecipitation (Fig. 1B), only the band corresponding to pure Hb was recovered. This indicated that the band in the more anodal position corresponded to Hb–Hp complexes and the other band to free Hb. In serum, the tetrameric molecules of natural Hb spontaneously dissociate into dimeric ␣- subunits that are bound to Hp (8 ). Beyond a certain concentration of Hb, according to its concentration and phenotype, serum Hp becomes saturated and free Hb appears. In cases of high concentrations of free Hb, two additional minor bands were detected in a more cathodal position (lanes e and f). The most cathodal band corresponded to HbA2, a typical minor fraction of Hb that was recovered from high concentrations of pure Hb (data not shown). The second band, which has not yet been identified, was not present in pure Hb but only when Hb was diluted in serum. To facilitate the interpretation of the electropherograms, natural Hb in both its complexed (Hb–Hp) and free forms was used as markers. For this, a reference preparation was obtained by adding natural Hb to a serum sample at an approximate concentration of 6 g/L, which saturated the Hp and gave rise to the two marker positions, designated as m1 and m2 for Hb–Hp complexes and free Hb, respectively.
electropherograms of hboc products As shown by Fig. 2, pure solutions of the different HBOCs gave rise to broad bands located in the area of the m1 marker for Hemopure (lane b), Oxyglobin (lane c) and HemAssist (lane d). To respect the confidential nature of ongoing research and development of these novel products, electropherograms from HBOCs still undergoing clinical trials have not been included in this report, although the remaining products were also distinguishable based on their migration patterns (our unpublished observations). The concentration of myoglobin, another heminic protein, is typically too low to be detected on an electropherogram (⬍100 g/L) (9 ). Nevertheless, the position of pure myoglobin was established and found to be clearly distinct from those of Hb and the tested HBOCs (Fig. 2, lane e). When added to serum, these products gave results very similar to those described above (Fig. 3). The immunoprecipitation of Hp was necessary to differentiate some of the tested HBOCs from the Hb–Hp complexes because some of them (as was the case for Hemopure, Oxyglobin, and HemAssist) migrated to a position close to the m1 marker associated with the Hb–Hp complex. The immunoprecipitation conditions were established by use of various serum samples enriched with natural Hb at ⬃6 g/L and checked by electrophoresis for disappearance of the Hb–Hp complexes. Incubation for 1 h (other tested times were 2 and 16h) at room temperature (other tested temperatures were 4 and 37 °C) efficiently removed all of the Hp present in serum samples presenting normal concentrations of this protein (normal range obtained by the IFCC, 0.27–2.14 g/L). However, to ensure that immunoprecipitation was complete in the presence of abnormally increased concentrations of Hp (most increased concentration tested was 5.75 g/L), we finally settled on the use of concentrated antibodies. In all cases, a very simple interpretation could be drawn from the results obtained after removal of Hp. None of the investigated HBOCs migrated in the position
Fig. 1. Electrophoresis of natural Hb at increasing concentrations in human serum. Lane a, pure natural Hb (not in serum); lanes b–f, serum samples enriched with increasing concentrations (0.5, 1.0, 2.0, 4.0, and 8.0 g/L) of natural Hb and submitted to electrophoresis before (A) and after (B) immunoprecipitation of Hp. The anode is at the top of the gel. Samples were applied at the cathodal (bottom) edge of the gel. ? indicates the unidentified band.
4
Lasne et al.: Method for Detecting HBOCs in Doping Control Screening
Fig. 2. Electrophoresis of different pure HBOCs compared with natural Hb and myoglobin. Lane a, natural Hb; lane b, Hemopure; lane c, Oxyglobin; lane d, HemAssist; lane e, myoglobin. Lane m is natural Hb in serum, illustrating the positions of the m1 and m2 markers corresponding with Hb–Hp complexes and free Hb, respectively. The anode is at the top of the gel. Samples were applied at the cathodal (bottom) edge of the gel.
(lane P⬘). In contrast, some samples demonstrated a less intense band in the same position (lane D3) that persisted after precipitation of Hp and was found to be Hemopure (lane D⬘3). This clearly corroborated that immunoprecipitation of Hp was essential for clear interpretation of the results. In a few samples, this method clearly established the presence of free endogenous Hb in addition to Hemopure (as was the case for the sample in lane D2). Most likely the free Hb in serum emanated from in vitro hemolysis associated with blood collection. Immediately after infusion of either 30 or 45 g of Hemopure, all 12 samples were highly positive for this product. Despite an apparent substantial decrease in the serum concentration of Hemopure, the product was still easily detectable in all samples on the first and second days after infusion. Although much less intense, the band corresponding to Hemopure was still present in all 12 samples on the third day and in each of the 6 samples taken on the fourth day. On the fifth day, 11 of 12 samples were still weakly positive. No Hemopure was detectable 8 days after the infusion.
Discussion of free Hb (band m2), which is the only possible form of endogenous Hb after a serum sample has been cleared of Hb–Hp complexes. In other words, the presence of any bands in positions different from that of the m2 marker was highly suggestive of the presence of an HBOC in the analyzed sample.
electropherogram of serum from individuals infused with hemopure The electrophoretic method was applied to blood samples derived from individuals infused with Hemopure, before and after removal of Hp (Fig. 4). In some samples an intense band in position m1 (lane P), was confirmed as natural Hb because it was removed by precipitation of Hp
Fig. 3. Electrophoresis of different HBOCs in serum. Lanes a– d correspond with the same lanes in Fig. 2. Lanes a⬘–d⬘ correspond with the same entities after immunoprecipitation of Hp by polyclonal rabbit anti-human Hp antibodies. The anode is at the top of the gel. Samples were applied at the cathodal (bottom) edge of the gel.
In native-gel electrophoresis, proteins are separated according to a combination of several factors, e.g., net charge, molecular size, shape, and eventual interactions with the matrix used. When submitted to agarose gel electrophoresis at pH 9.2, the different types of HBOCs tested all migrated in positions different from that of free Hb. However, some products were positioned near the complexes formed by Hp and Hb dimers derived from dissociated endogenous Hb. Because such complexes are present physiologically (at low concentrations) in serum, it was necessary to eliminate these before electrophoresis of the serum samples. In the context of doping controls conducted on athletes, it is pertinent to consider the potential increase of such complexes in the case of hemolysis associated with exercise (10 ) as well as sample collection (11 ). The use of concentrated antibodies is highly advised to ensure efficient removal of the Hb–Hp complexes in the cases of abnormally high Hp concentrations. Although the polyclonal antibodies used are reactive against the three common Hp phenotypes (1-1, 2-1, and 2-2), some differences in immune reaction related to phenotype must be envisaged (12, 13 ). Nevertheless, the efficiency of Hp removal can be validated by checking that the electropherogram is not changed by use of more concentrated antibodies. After removal of Hb–Hp complexes, the only possible residual form of natural Hb in the sample is free Hb (for example, in cases of substantial hemolysis), which is clearly identified by electrophoresis. The only conceivable situation in which this approach might be confounded would be the presence in serum containing an abnormal natural Hb (i.e., an amino acid mutation without clinical significance) in its free form that would present an abnormal electrophoretic migration.
5
Clinical Chemistry 50, No. 2, 2004
Fig. 4. Electrophoresis of serum samples collected from a healthy individual infused with 30 g of Hemopure. The m1 (Hb–Hp complexes) and m2 (free Hb) markers are illustrated in lane a, which contains natural Hb added to serum. Lane P is a serum sample taken from the individual before the infusion; the intense component at the m1 position in lane P corresponds to Hb–Hp complexes rather than to Hemopure, as demonstrated by its disappearance after immunoprecipitation of Hp (lane P⬘). Lanes D0–D3 are serum samples containing Hemopure (the subscript indicates the number of days postinfusion). Lanes a⬘, D0⬘–D3⬘, and P⬘ depict the same samples after immunoprecipitation of Hp. The presence of free endogenous Hb (unbound to Hp) in the samples in lanes D2 and D2⬘ are indicated by the intense band at the m2 position. Anode is at the top of the gel. Samples were applied at the cathodal (bottom) edge of the gel.
However, any doubt could be easily investigated by analyzing the endogenous Hb from the red blood cells of the athlete. Myoglobin, another heme-containing protein, is typically present at such low concentrations in serum (⬍100 g/L) that it is not detected by electrophoresis (9 ). However, in cases of intensive physical exercise, increased myoglobinemia attributable to transit rhabdomyolysis may be observed, which would give rise to increased serum concentrations (14 ). These concentrations are still insignificant (⬍1 mg/L) compared with the concentrations of endogenous Hb, and especially HBOC if it were used at the dosages considered sufficient to provide a meaningful performance advantage. However, to exclude any possible confusion in extreme case of rhabdomyolysis that could produce detectable concentrations of serum myoglobin, the position of myoglobin was shown to be clearly different from those of the investigated HBOCs. Because serum samples were analyzed without any additional purification process other than Hp removal, a general staining method could not be applied, requiring the specific detection of the hemoglobinic molecules. Both endogenous Hb and the modified Hb molecules found in HBOCs include heme prosthetic groups that are directly implicated in the binding of oxygen. It was therefore possible to use the peroxidase properties of the heme moiety to detect the hemoglobinic molecules without any interference from serum proteins. The use of a chemiluminescent substrate for peroxidase enabled electropherograms specific for hemoglobinic compounds to be obtained. This method was quite sensitive: even physiologic concentrations of serum Hb (⬍50 mg/L) were detected. Moreover, it was highly specific for hemoglobinic molecules; no interference from any other molecule with peroxidase properties was observed during the investigation of various serum samples. In conclusion, native-gel electrophoresis at basic pH of serum after immunoprecipitation of Hp appears to be a
convenient, rapid, and sensitive method able to separate natural human Hb from each of the studied HBOCs. The speed and low cost of electrophoresis make it well suited to the investigation of a large number of samples. For these reasons, we propose the above method as an efficient screening method in antidoping controls for the presence of HBOC.
We thank Maria Gawryl, Ted Jacobs, and Biopure Corporation (Boston, MA), without whose cooperation the administration trial would not have been possible. We sincerely thank each pharmaceutical companies that altruistically provided access to product samples still under development. We acknowledge the expert technical assistance of Melissa Arkinstall, the staff at Hoˆ pital Arnaud de Villeneuve, and Professor Christian Prefaut. We are deeply indebted to the volunteers for their commitment and compliance to all aspects of the protocol. This project was carried out with the support of the World AntiDoping Agency.
References 1. Gaudard A, Varlet-Marie E, Bressolle F, Audran M. Drugs for increasing oxygen transport and their potential use in doping. Sports Med 2003;33:187–212. 2. Ashenden MJ. A strategy to deter blood doping in sport. Haematologica 2002;87:225–32. 3. Lasne F, de Ceaurriz J. Recombinant erythropoietin in urine, an artificial hormone taken to boost athletic performance can now be detected. Nature 2000;405:635. 4. Lok C. Blood product from cattle wins approval for use in humans. Nature 2001;410:855. 5. Standl T, Freitag M, Burmeister MA, Horn EP, Wilhelm S, Am Esch JS. Hemoglobin-based oxygen carrier HBOC-201 provides higher and faster increase in oxygen tension in skeletal muscle of anemic dogs than do stored red blood cells. J Vasc Surg 2003;37:859 – 65. 6. Hugues GS Jr, Yancey EP, Albrecht R, Locker PK, Francom SF, Orringer EP, et al. Hemoglobin-based oxygen carrier preserves submaximal exercise capacity in humans. Clin Pharmacol Ther 1995;58:434 – 43.
6
Lasne et al.: Method for Detecting HBOCs in Doping Control Screening
7. Chang TM. Red blood cell substitutes. Baillieres Best Pract Res Clin Haematol 2000;13:651– 67. 8. Putnam FW. Haptoglobin. In: Putnam FW, ed. The plasma proteins, 2nd ed. New-York: Academic Press, 1975:2–51. 9. Wu AHB, Laios I, Green S, Gornet TG, Wong SS, Parmley L, et al. Immunoassays for serum and urine myoglobin: myoglobin clearance assessed as a risk factor for acute renal failure. Clin Chem 1994;40:796 – 802. 10. Telford RD, Sly GJ, Hahn AG, Cunningham RB, Bryant C, Smith JA. Footstrike is the major cause of hemolysis during running. J Appl Physiol 2003;94:38 – 42. 11. Grant MS. The effect of blood drawing techniques and equipment
on the hemolysis of ED laboratory blood samples. J Emerg Nurs 2003;29:116 –21. 12. Valette I, Pointis J, Rondeau Y, Waks M, Engler R. Is immunochemical determination of haptoglobin phenotype dependent? Clin Chim Acta 1979;99:1– 6. 13. Van Rijn HJM, van der Wilt W, Stroes JW, Schrijver J. Is the turbidimetric immunoassay of haptoglobin phenotype-dependent? Clin Biochem 1987;20:245– 8. 14. Thomas BD, Motley CP. Myoglobinemia and endurance exercise: a study of twenty-five participants in a triathlon competition. Am J Sports Med 1984;12:113–9.
