J. Delanghe, MD, PhD
Professor of clinical chemistry
De Pintelaan 185
B 9000 Gent
Ms. Genevieve Jeanson tested positive for recombinant erythropoeitin in both an A and a B urine sample which was prelevated on July 25th 2005. However, the athlete denied epo abuse. An independent investigation was therefore initiated to check the possibility that in the particular case of Ms Jeanson the results of the positive epo test could be explained by false positivity.
The pre-analytical phase
It is well known in clinical chemistry that the vast majority of so-called ”laboratory errors” originate in the pre-analytical phase (the phase prior to the analysis of the sample).
The urine sample No. 491880 has been prelevated shortly following a strenuous physical exercise. It is well known that in some athletes, this induces an exercise – induced proteinuria, which occurs after exercise and which is characterized by a half life of about 1 hour.
Despite the fact that the sample 491880 had to be analysed for brittle glycoprotein structures like recombinant erythropoeitin and despite the high ambient temperature (a warm summer night), the urine sample 491880 has not been cooled during the transport to the USADA lab. The care for the sample quality shows a marked difference before and after July 27: before July 27, no attention has been paid on the ambient temperature, afterwards, very careful handling (cooling, addition of cocktails of antiproteases to protect the quality of the sample and avoid proteolysis (degradation of the proteins like epo) despite the fact that a huge proteolysis already had taken place between July 25 and July 27!)
The transport was time consuming:
Sampling: July 25 2005 22.19 PM
Arrival in the lab: July 27 10.07 AM
This high ambient temperature has major consequences for the quality of the
specimen: the bacterial growth in the urine specimen is promoted. In urine specimens, bacterial growth is strongly dependent on the temperature (ref: European Guidelines for urinalysis, Scand J Clin Lab invest 2000; 231:1-86).
Since the urine specimen contained an important amount of protein (as evidenced by the USADA lab), this specimen contained a lot of nutrients promoting bacterial growth.
Since the specimen was not prelevated under sterile circumstances and given the fact that no “mid-stream catch” technique has been used for urine prelevation, it is certain that already during the time of prelevation, a female athlete’s urine did contain an important number of bacteria.
The high ambient temperature and the long transportation time (the sample arrived in Los Angeles on July 29 2005 at a.m.) undoubtedly have significantly increased the bacterial count of the urine specimen. Furthermore, the pH value of the specimen (pH 6.0 measured by the lab in Los Angeles) was optimal for promoting bacterial growth (a more acid pH value would have inhibited the growth). Also the high density of the sample (1020 kg/m3) (which is due to the pronounced dehydratation of the athlete following the strenuous physical exercise at a high ambient temperature) indicates that there was a high concentration of nutrients in the specimen which is in favour of bacterial growth. It can be estimated that Ms. Jeanson's urine sample may already have contained over 100,000,000 bacteria/ml (vs. the reference interval ± 1,000 - ±
100,000/ml) upon arrival in the USADA lab.
In human urine, Escherichia Coli (E. Coli) is the most abundant microorganism. Recently, scientific evidence has been published that certain E. Coli proteins do cross react with the AE7A5 monoclonal antibodies used in the official epo test in use (reference: Franke W, Heid H, Clin Chim Acta 2006, 373:189-190)
The monoclonal antibody AE7A5 used in the current epo test is definitely not specific for recombinant EPO, not even for EPO in general. This is true for native and denatured proteins, for urea- or for SDS-treated proteins. In fact, mAb AE7A5 can cross-react with a remarkably high number of other proteins (e.g., E. coli is known to occur in human urine and is widely adherent to human skin in the uro–ano–genital region), extracellular proteins as well as intracellular ones which, e.g., may be released into urine upon urothelial damage, by no means a rare happening. In this context it is also relevant that the antigenic peptide used for the generation of mAb AE7A5 contained a central tetrapeptide with an amino acid sequence that occurs in the positive bacterial thioredoxin reductase as well as in more than 500 human and almost 185 E.coli proteins.
PRELEVATED ON JULY 25 2005
WAS INSUFFICIENT TO OBTAIN
A RELIABLE TEST RESULT
FOR FURTHER READING: marked cross-reactivity was demonstrated in the epo test with bacterial antigens (reference: Franke W , Heid H, Clin Chim Acta 2006, 373:189-190).
In particular, E. coli, known to occur in human urine and is widely adherent to human skin in the uro–ano–genital region) produces thioredoxin reductase which cross reacts with the antibody of the epo test. The antigenic peptide used for the generation of mAb AE7A5 contained a central tetrapeptide with an amino acid sequence that occurs in the positive bacterial thioredoxin reductase as well as in more than 500 human and almost 185 E. coli proteins.
Effect of strenuous physical exercise on proteinuria in the case of Genevieve Jeanson
Since the urine sample No. 491880 was prelevated following strenuous physical exercise, the effect of physical exercise on proteinuria has been tested in detail by us in the Jeanson case.
Ms Genevieve Jeanson was asked to carry out the same type of exercise that as the one carried on on July 25th 2005 prior to the prelevation of sample 491880. Tests were carried out in Arizona in February 2006. Post exercise urine specimens were collected within the same time frame as had been the case for sample 491880. The urinary epo test is hampered by the presence of cross reactivity following strenuous physical exercise.
Assesment of exercise –induced proteinuria
Conclusions: a marked exercise –induced proteinuria is found in the samples (the upper reference limit for urinary protein is 150 mg/l). The high alpha 1 microglobulin indicates an important tubular component of the proteinuria. The urinary epo concentration is low. Moreover, a high leukocyte count is observed. Which is accompanied by a release of proteases carrying out a number of structural modifications to urinary proteins.
Electrophoresis of the proteins
Following a period of intense cycling activity, a typical mixed glomerular-tubular proteinuria is observed due to temporary renal ischemia (comparable to the case described in Blood 2006)
Figure: Agarose gel electrophoresis electrophoresis (followed by Coomassie Blue staining) of urine samples taken following training sessions of Genevieve Jeanson (from left to right samples of January 30, February 5, and February 6 2006, and a control sample (1:50 dilution of human serum). It is clear that a marked proteinuria can be observed on the three occasions, the urinary protein spectrum shows a typical mixed glomerular-tubular pattern.
Test of cross reacting substances
The sample obtained on Feb 11 was further analysed. When the monoclonal antibody 9C21D11 was used for quantitative epo analysis (in contrast to the WADA antibody this antibody is approved by the FDA for use in humans for diagnostic use).
We measured epo values in the four samples with a well validated (and FDA approved) monoclonal antibody (clone 9C21D11): only a small amount of urinary erythropoetin could be detected. The non-significant correlation between the FDA approved epo test and the staining with the AE7A5 monoclonal antibody was disappointing.
An advertisement of the double immunoblots technique (as used in the WADA approved epo test). According to the leaflet, the final protein concentration can get as high as 200 g/l. However, it is clear that in Jeanson’ case, final concentration in sample #491880 following a 700 to 1000 fold concentration must have been in the magnitude 300-500 g/l since the method used in the WADA lab is based on the so-called indicator principle which is only sensitive for albumin. In the case of Ms Jeanson, the relative amount of albumin is lower than normal due to the mixed character of the proteinuria (containg a lot of tubular proteins which are not detected by the method).
Post-exercise proteinuria is an important caveat for the epo test
As early as March 2003, Dr. G. Peltre and professor W. Thormann predicted these kinds of analytical problems with the epo test. This was mentioned in their official WADA report (which was removed from the WADA website when we started our investigations). On page 11 of their report (copy enclosed), one can read some prophetic statements regarding the effect of strenuous physical exercise on the epo test.
Specific investigations on cross-reactivity
The adopted epo test is based on a combination of isoelectric focusing and double immunoblotting, and distinguishes between endogenous and recombinant human Epo. It was demonstrated that this test can occasionally lead to the false-positive detection of rhEpo (epoetin-B) in postexercise, protein-rich urine, probably because the adopted monoclonal anti-Epo antibodies are not monospecific (Beullens et al Blood 2006).
FOR FURTHER READING
M Beullens, J. Delanghe, and M Bollen False-positive detection of recombinant human erythropoietin in urine following strenuous physical exercise. Blood 2006 107: 4711-4713 and Blood, 2006; 108: 1778 – 1780).
Immunoblotting visualized a band of 42 kDa (a higher mass than the epo isoforms (32kDa-39 kDa)). Removal of N-linked carbohydrates decreased the mass of epoetin- isoforms, from between 32 kDa and 39 kDa to 18 kDa, but the treatment did not cause a band shift in a postexercise urine sample. Immunoblotting before and after glyco-sidase treatment confirmed that the major urinary protein reacting with the anti-Epo antibodies was not Epo. None of these bands were detected in the absence of the primary antibodies, showing that they did not result from the interaction with the secondary antibodies. Some of these proteins have a similar isoelectric point as the epoetin- isoforms, which possibly accounts for the false-positive detection of epo. The nonspecific binding of these anti-epo antibodies to several proteins in the urine of a nonathletic volunteer has been reported (Khan 2005).
False-positive detection of epo may be linked to the extent and type of proteinuria. The extent of proteinuria correlates more with the intensity than the duration of exercise and has a half-time decay of about 1 hour. In mixed glomerular-tubular proteinuria, some of the urinary proteins show some structural homology with epoetin-, which possibly accounts for their cross-reactivity with the anti-Epo antibodies. In a WADA report, the possible existence of such interferences was already predicted (Peltre).
Lasne F, de Ceaurriz J. Recombinant erythropoietin in urine. Nature. 2000;405: 635.
Lasne F. Double-blotting: a solution to the problem of non-specific binding of secondary antibodies in immunoblotting procedures. J Immunol Methods. 2003;276: 223-226.
Khan A, Grinyer J, Truong ST, Breen EJ, Packer NH. New urinary EPO drug testing method using two-dimensional gel electrophoresis. Clin Chim Acta. 2005;358: 119-130.
Peltre G, Thormann W. Evaluation report of the urine EPO test. Paris, France, and Bern, Switzerland: Council of the World Anti-Doping Agency (WADA); 2003.
As the half-life of Epo is about 8 hours, this indicates that the detected signals at 0 hours are not derived from Epo itself.
The lack of specificity of the anti-Epo antibody (clone AE7A5) lies at the heart of the false-positive detection of rhEpo. This issue has also been raised by Khan et al, and is not unexpected since this antibody is sold for research use only. We note that Lasne is the first WADA-supported investigator admitting to the nonspecificity of this antibody (Blood 2006). It is very surprising that this cross-reactivity has not been noted in the thousands of Epo tests that have been performed so far. The diagnostic use of an antibody that is not monospecific and the evident use of ad-hoc interpretation criteria by the WADA-accredited laboratories are worrisome and difficult to reconcile with the claim that the rhEpo test is infallible.
(LIKE IN G. JEANSON’S CASE),
THERE IS A CONCERN ABOUT
THE VALIDITY OF THE EPO TEST
About the choice of monoclonal antibodies
To the surprise of many biochemists, WADA laboratories are using the monoclonal antibody AE7A5. Monoclonal antibodies do offer the advantage being specific. However, it is surprising why AE7A5 has been chosen. The FDA has approved a number of diagnostic test kits which make use of the antibody 9C21D11, which is recognizing the entire epo molecule. This antibody is able to provide correct measurements even in presence of hige quantities of other antibodies. In contrast, the AE7A5 clone has only been tested against sheep, baboon and dog epo. The documentation of the manufacturer is enclosed.
Figure: manufacturers’ documentation of the monoclonal antibodies used by WADA laboratories. The information regarding cross-reactivity is extremely scanty. The manufacturer clearly states that this product is not intended for use in humans.
A thorough literature review on effect of pre-analytical variables on epo testing was started. To our surprise, very few peer-reviewed articles were found in Pubmed (the world’ s largest medical database containing over 11,000,000 records). Only one relevant publication can be retrieved (Robinson et al. Clin Lab 2003;49: 57 - 62).
Remarkably, the only WADA approved epo test is not mentioned in this paper. It can be concluded that the present knowledge on pre-analytical errors for the actual epo-test is very limited indeed.
Figure. A literature search concerning the pre-analytical phase for the epo-test yields a very poor result (early 2006) (print screen of Pubmed public database)
Some quantitative calculations.
Knowing the mean concentration of erythropoietin in Ms Jeanson’s specimens (mean value 6.5 mIU/ml), this corresponds to a mass concentration of 0.05 microgram/l or 0.00005 mg/l (specific activity 1.2 *105 IU/mg). For a lot of proteins which are virtually absent in urine specimens of individuals at rest, concentration in urine may go as high as in the magnitude of 1 mg/l. Since these values exceed the epo values by a factor of 20,000, a specificity of 99.99% will pick up the signals of these proteins. In the example(depicted in the table), the post-exercise proteinuria will give raise to a signal twice as high as the epo signal.
Two urine samples taken with a short interval offer a unique quality control of the sample 4918880
Further evidence of an abnormal test result
Urine samples 491880 and 493301 were sampled with an interval of only 60 hours and investigated by the same accredidated lab. In the first sample we see a very high concentration of protein in the rhEPO zone (>97 % of the staining intensity). With the current knowledge of rhEPO pharmacokinetics and the test (as described in the paper xxx) one would expect still a traceable amount of rHepo in the second sample 493301.
The first urine sample 491880 is taken following a streneous physical training and was stored initially stored under non-optimal conditions and arrived at UCLA 36 hours later. The second one was taken at rest in the morning and the transportation time was significantly shorter.
TABLE : COMPARISON OF THE PRE-ANALYTICAL PHASE OF Ms JEANSONS SAMPLES
In the second sample, we do not see any trace of endogenous or recombinant epo! In contrast we clearly see a cross-reacting substance (also the UCLA lab did observe this mysterious unidentified fraction). This unidentified fraction is nothing but a cross- reacting urinary protein which has nothing to do with epo. This finding of cross-reactivity (on itself a remarkable finding because it was never reported in the publications describing the method) was not interpreted as positivity (this is no matter of debate) but clearly illustrates the presence of cross-reacting substances.
Careful analysis of the banding pattern of sample 491880 (figure 2 p.25) shows that some rungs of the ladder of the middle lane (G. Jeanson's sample) do not fit very well with the epo standards. It is clear that the distance between the rungs corresponds with the effect of 1 net charge/mol mass of about 30,000 Da, but the index rung at the very bottom of the middle lane is different.
Sample 491880 does not show a sample that is compatible with the standards rHuEPO/NESP nor with the positive Quality control sample: the chemical nature of the proteins is not identical. The “rungs of the ladder” in Ms Jeanson’s pherogram (indicating the various glycosylated forms of the proteins believed to be erythropoietin) (MS Jeanson’s specimen 491880) do not fit with the recombinant epo standard, indicating a different nature of the protein investigated. These data are compatible with the presence of cross-reacting substances in the urine sample.
The findings of cross - reacting proteins in the urinary epo test have already been predicted or described in literature (Khan A et al. New urinary EPO drug testing method using two-dimensional gel electrophoresis. Clin Chim Acta 2005 ;358:119-30) and the WADA report written by J. Peltre and W. Thormann on March 11th 2003 (which mysteriously disappeared from the WADA website some months ago when some scientists were using this documents to prove that the epo test was not 100% reliable in all cases). A very similar observation was made in the case of Rutger Beke in Summer 2005 (appearance of cross-reacting substances in urine, proteins with a molecular mass different from epo appeared in the rhEPO area of the gel.
Just like in the Jeanson case, the cross-reacting urinary proteins disappeared quickly in the Beke case. Postexercise proteinuria is directly related to the intensity of exercise rather than to its duration. This excretion of excess proteins is a transient state with a half-time decay of about 1 hour characterised by an increased glomerular permeability and a partial inhibition of tubular reabsorption. Exercise decreases the glomerular electrostatic barrier and facilitates transfer of macromolecules.(ref. Poortmans Jr, Vanderstraeten J. Kidney function during exercise in healthy and diseased humans. An update. Sports Med. 1994 ;18:419-37)
A mixed glomerular-tubular type of proteinuria is characterised by a broad spectrum of urinary proteins. Some of them show some structural homology with epo. As in case of proteinuria the concentration range of these specific urinary proteins largely exceeds those of epo by several orders of magnitude, there is a potential risk for cross- reactivity of the urinary proteins towards the anti-epo antibodies commonly used in epo abuse testing.
In other words it means that Genevieve Jeanson produces specific urinary proteins which are cross reacting with monoclonal antibody AE7A5 as used in the epo test by UCLA. This monoclonal antibody has only been tested against baboon epo, goat epo and dog epo (and are not intended for use in humans according to the manufacturer). The unexpected huge difference between the two test results can be explained by the present of transient exercise induced proteinuria in sample 491880.
The faint fractions in the pH 6 zone in sample 493301 can be the less glycosylated remnants of low molecular mass plasma proteins which caused the false positive results of USADA 491880.
In order to clear the situation, additional mass spectrometry of the urinary proteins in the leftover of the sample (if still existing) could be very helpful.
TAKEN SHORTLY AFTER THE FIRST ONE
QUESTION THE POSITIVITY OF THE FIRST SAMPLE.
FURTHERMORE, A DETAILED ANALYSIS OF THE PHEROGRAM
IS SUGGESTIVE FOR THE PRESENCE OF
CROSS-REACTING (EPO-LIKE) PROTEINS
False positive epo testing, a rare finding
As pointed out by Dr Catlin (Blood 2006), positive cases for the urinary epo test are rare. So, per definition, false positive cases are very rare.
Knowing that cross-reacting substances reacting with the monoclonal are occurring in a small minority of athletes investigated, a medical investigation was carried out: a few people might be more prone to develop a false positive test than others.
So far, we have studied 4 athletes who were accused of epo abuse and who denied having taken recombinant epo (Genevieve Jeanson being one of them)
For privacy reasons, we cannot provide the athletes’ names. Striking clinical features were found: between Ms Jeanson and in the three other athletes. These striking clinical similarities cannot be explained by coinicidence but may point a common overexpression of certain proteins in the four individuals. These investigations are still ungoing.
Molecular modelling of the problem
In order to understand the phenomenon of cross-reactivity (leading to false positivity) at a molecular level, we have carried out computer-based calculations using protein sequence databases. Candidate proteins were compared with the sequence of the 26 amino acids of the N terminal of human epo (known to be the epitope of the monoclonal AE7A5 used in the WADA approved test). For some known proteins, we obtained striking structural similarities with the first 26 amino acids of the N terminus of human erythropoietin (epitope of the AE7A5 monoclonal antibody of the epo test).
Next to the bacterial proteins, similar cross reactivities towards the monoclonal antibody AE7A5 have been observed for other human proteins as well by our researchers (e.g the ones depicted in Blood, Jun 2006; 107: 4711 - 4713. and Blood, Sep 2006; 108: 1778 – 1780).
It is clear that given the potential amount of certain of these proteins in human urine following exercise, that the cross reactivity may lead to false positivity of the epo test.
At this moment (September 2006) these results cannot be distributed because the data so far have not been published in peer reviewed scientific journals.
In the present case, there are numerous reasons to doubt the interpretation of the test result as positive for rHuEPO. The sampling following a strenuous effort has created a pronounced proteinuria which exceeds the in-built capacity of the epo test to neutralise this kinds of effects. It should be noted that the double blotting technique used in the epo test does not protect against the cross-reactivity type of error.
The very poor conditions of transportation of the urine sample (containing a abnormally high amount of protein) have facilitated destruction of the brittle epo glycoprotein structures and have promoted bacterial growth. Despite its poor pre- analytical quality, the specimen was still analysed. The atypical patterns observed in sample 491880 support the presence of cross-reacting proteins. This is not surprising in view of the conditions of sampling and the storage.
The prelevation of a second sample (No 493301) shortly after the first one yields an entirely different pattern than the one observed for sample 491880. This unprecedented unique situation allows to confirm our previous findings: one would have expected a kind of “B” sample but just the opposite is observed. The striking difference between the two test results apparently seems to be a violation of our current knowledge about the erythropoietin pharmacokinetics. Since no evidence exists for an abnormal handling of erythropoetin by Ms Jeanson, the only scientifically possible explanation is again the presence of cross-reacting substances in sample # 491880.
The body of evidence from the literature is growing that this kind of sample cannot be interpreted with certainty.
Similar cases have recently been found in Belgium (n = 1) and in Spain (n = 2). All cases hitherto showed strong resemblances: the urine specimen was obtained following a streneous physical exercise in endurance athletes which are particularly prone to exercise-induced proteinuria. In all investigated cases, conditions of sample transportation were poor (unprotected against the heat) and a false positive test result was obtained for epo.
Le 20 septembre 2007 on a pu voir à l'émission Enquête à la télévisison de Radio-Canada Le Secret de Geneviève Jeanson au cours de laquelle le Dr Joris Delanghe est interviewé (transcription disponible en cliquant ici)
L'animateur Alain Gravel commente ensuite ainsi : Les plus grands spécialistes du dopage sanguin contredisent le docteur Delanghe. Ils affirment que la durée de vie de l’EPO dans l’organisme, soit la période durant laquelle on peut épingler les athlètes dopés, est très courte. Parmi ces spécialistes il y a Michel Audran de Montpellier, dans le sud de la France, qui affirme qu’il est parfaitement possible que l’EPO détectée dans le corps de Geneviève Jeanson lors du premier test ait disparu 60 heures plus tard lors du deuxième.
On peut également y voir la Dr Christiane Ayotte (directrice du laboratoire de contrôle du dopage INRS) qui commente ainsi le rapport que vous avez lu ci-haut :
Ça, ici, ce que M. Delanghe nous a mis dans la littérature, c’est de la crotte ! Sincèrement, il n’y a pas un biologiste... et ça c’est peut être difficile pour quelqu’un qui en fait pas, mais il n’y a pas un biologiste moléculaire qui va regarder des gels de cette nature-là, qui étaient censés être la preuve que le test était un faux positif, et qui va dire...
(Pour visionner l'émission cliquez ici pour accéder à la page de l'émission, puis sur la date du 20 septembre au calendrier).