identification of blood erythroid markers useful in revealing erythropoietin abuse in athletes

uranceOlympiciderations,ers uponcantlyand an

of rEpo.ormone isvell of thet, soluble

tricf Thus, them experi-m

Magnani et al. Blood Cells, Molecules, and Diseases (2001)27(3) May/June: 559–571

doi:10.1006/bcmd.2001.0419, available online at http://www.idealibrary.com on

Identification of Blood Erythroid Markers Usefulin Revealing Erythropoietin Abuse in AthletesSubmitted 04/03/01(Communicated by E. Beutler, M.D., 04/04/01)

Mauro Magnani,1 Dario Corsi,1 Marzia Bianchi,1 Mirko Paiardini,1 Luca Galluzzi,1 Elisa Gargiullo,1

Attilio Parisi,2 and Fabio Pigozzi2

ABSTRACT: Recombinant human erythropoietin (rEpo) is being used with increasing frequency by endathletes to improve aerobic potential. Although rEpo administration has been banned by the InternationalCommittee, no methods are available to unequivocally detect its abuse in sports. Prompted by these conswe evaluated the main hematological and biochemical modifications measured in the blood of 18 volunterEpo administration. Different rEpo regimens, iron, folic acid, and vitamin B12 administration did not signifimodify the percentage increase in hematocrit. However, a significant decrease in circulating ferritin (fr)increase in the soluble transferrin receptor (sTfr) were not found in athletes receiving low (30 IU/kg) dosesThus, an increase in the sTfr/fr ratio cannot be used as an indicator of rEpo abuse, at least when the hadministered at low concentrations. In contrast, the amounts ofb-globin mRNA detected by quantitaticompetitive (RT)-PCR in whole blood samples significantly increased above the threshold levels in atreatments investigated. Taken together, these data suggest that hematocrit value, reticulocyte countransferrin receptor content, and concentration ofb-globin mRNA, when included in a new multiparameormula, can detect rEpo abuse in 57.5% of the samples examined with a confidence interval of 99.99%.ethod reported in this paper could significantly improve the tests currently available, which in similarents allowed the detection of rEpo abuse in only 7.6% of the samples examined.© 2001 Academic Press

Key Words:erythropoietin; doping; erythroid markers; soluble transferrin receptor; ferritin.

izegi-ortseryr-

singpur-ionan

nec-ry-n

antoidifictionth-

tan-ion,and

ed-icail-ectug-life

f Biolog -Ur

INTRODUCTION

To improve their performance, athletes utilspecific methods which optimize the physiolocal characteristics needed for their sport. In sprequiring endurance, intense effort and recovduring training and competition, it is very impotant to increase aerobic potential by increaoxygen transport. The methods used for thispose include altitude training, self-transfusand, in recent years, administration of humrEpo. Erythropoietin is a glycoprotein hormoproduced primarily by the kidney, with a moleular weight of 30.4 kDa (1), that stimulates ethroid cell proliferation (2, 3) and differentiatio

Correspondence and reprint requests to: Mauro Magnani, Institute oItaly. Fax139-0722-320188. E-mail: [email protected]
Institute of Biological Chemistry “G. Fornaini,” University of Urbino, 61022 University Institute for Motor Sciences, 00194-Rome, Italy.
559

(4, 5). Like endogenous Epo, the recombinhormone interacts with the precursor erythrcells (BFU-E, CFU-E) by means of a specmembrane receptor (6, 7), causing proliferaand differentiation of these cells in mature eryrocytes (8–11). The use of injections, as in sdard medical practice, is simpler than transfuswhich can instead cause various problemsinvolves greater risks.

Although rEpo has been banned by the mical commission of the International OlympCommittee, the antidoping tests currently avable cannot detect it with confidence. A dirdetection of rEpo in urine has recently been sgested (12). However, while the plasma half-

ical Chemistry “G. Fornaini,” University of Urbino, Via Saffi 2, 61029bino,

9-Urbino, Italy.

1079-9796/01 $35.00Copyright© 2001 by Academic Press

All rights of reproduction in any form reserved

, itseat-mes

innotneen-andnsau e(17)alue,ubleoidfer-dy)e of: theses

sesil-si-nd

28eseter ariaane-ndetesheentro-of

es

ntoin

Thecrit

be-10,indsclot

, toes,d foricu-ap-of

: alls ofible

Thethen

ndanre-io-ere-,

,

eachminday

Blood Cells, Molecules, and Diseases (2001)27(3) May/June: 559–571 Magnani et al.doi:10.1006/bcmd.2001.0419, available online at http://www.idealibrary.com on

of rEpo varies between 4 and 13 h (13, 14)biological effects occur several days after trment and thus the erythropoietic effect becoevident when rEpo is no longer detectablecirculation. Moreover, Epo concentrations areonly very low but also vary considerably from operson to another (15) and are influenced byvironmental factors such as fatigue, stressbody hydration. To overcome these limitatioand make the abuse of rEpo detectable, Gareal. (16) and subsequently Bressolle et al.suggested that, besides the hematocrit (Ht) vthe ratio between the concentration of the soltransferrin receptor (sTfr; a marker for erythractivity) (18) and the concentration of serumritin (fr; a measure of the iron stores in the bo(19) should also be evaluated. In fact, the usrEpo also causes a change in iron metabolismlevel of the soluble transferrin receptor increa(20), while the level of serum ferritin decrea(21). The main indirect method currently avaable for the detection of rEpo abuse utilizesmultaneously multiple indirect hematological abiochemical markers (22).

MATERIALS AND METHODS

Study Participants

Eighteen male athletes ranging from 19 toyears of age were included in this study. Thathletes were selected from 24 volunteers affull medical examination. The exclusion critewere age lower than 18 years, hypertension,mia, iron, vitamin B12 or folic acid deficiency ahematocrit greater than 50%. None of the athlparticipated in official competitions during tperiod of treatment. Written informed conswas obtained from all participants. The study ptocol was approved by the Ethics Committeethe University Institute for Motor Scienc(Rome, Italy).

Study Design

The 18 athletes were divided randomly ithree different groups (A, B, C): 5 in group A, 5
group B, and 8 in group C. The treatments used for
560

t

each of the three groups are shown in Table 1.administration of rEpo was interrupted if hematovalues greater than 50% were observed.

Sampling Procedures

Blood samples were taken in the morning,fore administration of drugs, on days 0, 2, 4, 8,14, 18, 24, 28, and 31 of the treatment. Three kof vacutainers containing heparin, SST gel andactivator, and EDTA were utilized, respectivelydetermine full hemograms with reticulocyte valuto separate serum and to collect the blood useRNA extraction. Complete hemograms and retlocyte counts were obtained using cell counterparatus (Vega Retix ABX) 3–5 h after drawingblood. Each sample was analyzed five timeshematological values reported are the meanthese five analyses, in order to minimize posserrors due to the variability of the apparatus.serum was separated at room temperature andstored at220°C until use to measure sTfr aferritin concentrations. The Quantikine IVD humsTfr kit (R & D System, Inc.) was utilized foquantification of sTfr, while serum ferritin was dtected using the Elegance Ferritin ELISA kit (Bclone). Blood samples for RNA extraction wdiluted 1:4 (v/v) in TRIZOL LS reagent (GIBCOBRL), mixed, and stored at270°C until processing

TABLE 1

Treatment Protocols for the Evaluation of Erythropoietin-Modified Parameters in Athletes

Treatmentcondition

Group A(5 subjects)

Group B(5 subjects)

Group C(8 subjects)

rEpo (U/kg) 200 200 30Days of

administration0, 2, 4, 8, 10 0, 2, 4, 8, 10 0, 2, 4, 8, 10,

12, 14, 16, 1821, 24, 28

Iron(mg/subject)

— 25 25

Folic acid(mg/subject)

— 25 25

Vitamin B12(mcg/subject)

— 2500 2500

Note.The amount of drugs administered to the athletes ingroup is shown. In groups B and C, iron, folic acid, and vitaB12 were administered intravenously at 4-day intervals from0 to day 28.

as described below.

ined

theicalheint,ri-

ata-

s

R

odu s-s uc-t e-l red ep 2.4m frome riesc endo

Asor-ple,tal

fesif-ot to

chof

aR

p lr

Tfrre30C., 5

5

ve ansefirst

in

m on1 iumb na ass Bio-R rew on-cc om-p

eder-s”

d as-

T ofcG ofc ct.

nc-micns

hu-

c r,”t ilars be-l R

r 31atol-r.ctsted



Statistical Analysis

All data are presented as the means obtafrom subjects of each group6 SE or SD. Themean of values obtained at time 0 was withinnormal range and was taken as physiologvalue. Student’st test was utilized to compare tbiological data for each group, at each time powith those obtained at time 0. Analysis of vaance (ANOVA) was utilized to compare the dfor the three groups. AP value,0.01 was conidered statistically significant.

NA Extraction from Whole Blood

Total RNA was extracted from whole blosing the TRIZOL LS reagent (GIBCO-BRL) eentially according to the manufacturer’s instrions, with the slight modifications detailed bow. Briefly, 2-ml aliquots of whole blood weiluted in 6 ml of TRIZOL LS reagent. After thhase separation step, fixed volumes equal tol of the upper aqueous phase were takenach sample in order to make all RNA recoveomparable. The RNA pellets obtained at thef the procedure were all dissolved in 200ml of

DEPC-treated water. Quantification of RNyields was performed by measuring the abbance at 260 nm of 1:60 dilutions of each samagainst appropriate blanks. The yield of toRNA from all samples ranged from 9 to 32mgRNA per milliliter of whole blood. Extraction ototal RNA in triplicate samples provided valuwithin 10% of the mean, indicating that the dferences are due to the sample analyzed and nthe procedure used.

Reverse Transcription Polymerase ChainReaction (RT-PCR) and QuantitativeCompetitive (RT)-PCR Assays

Total RNA isolated from whole blood at easampling time was used for RT-PCR analysisselected erythroid gene markers, such asb-globin

nd Tfr.Tfr mRNA content was analyzed by RT-PC

erformed with 0.5ml of total RNA in a fina
eaction volume of 50ml, using “Ready-To-Go
561

RT-PCR Beads” (Pharmacia Biotech). ForcDNA amplification the cycling conditions we30 s at 95°C, 15 s at 65°C, 25 s at 72°C forcycles with a final extension of 10 min at 72°The sequences of primers used were TFR 19-ACTCAGCAAAGTCTGGCGTGAT-39 (sensebases 626 – 647 of cDNA); and TFR 2,9-TGAAGGAAGGGAATCCAGGTGT-39 (anti-sense bases 1044–1023 of cDNA) which gaPCR product of 419 bp in length. The antiseoligonucleotide was also used to prime thestrand cDNA synthesis in the RT reaction (25 mat 42°C). RT-PCR products, in 10ml of reaction

ixture, were resolved by electrophoresis.5% agarose gels and quantified by ethidromide staining, in comparison with knowmounts of DNA molecular weight markertandard, using the Gel Doc 1000 apparatus (ad). All determinations, in triplicate, weithin the linear range of the standard DNA centrations used (1–100 ng).b-Globin mRNAontent was analyzed using a quantitative cetitive (RT)-PCR assay. In this case 10ml of

1:250 dilutions of total RNA samples were usin the one-step protocol for QC (RT)-PCR pformed with the “Ready-To-Go RT-PCR Beaddescribed above. The oligonucleotides usePCR primers were BETA 1, 59-GTCTGCCGT

ACTGCCCTGTGG-39 (sense bases 27–48DNA) and BETA 4, 59-ACGTTGCCCAG-AGCCTGAAG-39 (antisense bases 329–309

DNA), which gave a 303-bp-long PCR produb-Globin primers spanning adjacent exon jutions were selected in order to exclude genoDNA amplification and complementary to regionot usually susceptible to polymorphisms inmans (23). In this assayb-globin cDNA wasoamplified in the presence of a “competitohat is, an internal standard with a closely simequence but carrying a 29-bp deletion (seeow). b-globin cDNA and competitor DNA PCconditions were: 30 s at 95°C, 25 s at 72°C focycles followed by a final extension of 10 min72°C; the reaction was performed in a final vume of 50 ml with 25 pmol of each primeb-Globin and competitor amplification produ(303 and 274 bp, respectively) were discrimina
on 3% (3:1) Nusieve–agarose gels and target

om-

leBE-ThededT4ther

gpe

t n int id

ithAofe

s ofand

lysisy oftitorT7

calan

turalntl

m

w ncet teda ther

lyam-pet-thintoal

t thee

tu-r ithi tormpc dw1 fe re-s itieso tedb ra-t tra-t d ast ter-m icht toc oftt ep be2 h

250-

t ntso

efer-f

l .

R

HM

me


bands were densitometrically quantitated by cputer imaging as described below.

Generation of theb-Globin Competitor andReference Targets

Briefly, total RNA from a whole blood sampwas reverse-transcribed and amplified with theTA1–BETA4 primer pair, as described above.303-bp PCR product was then purified, blunt-enwith Klenow enzyme, phosphorylated bypolynucleotide kinase and finally ligated inBlueScript SK2 vector cloning box (“reference ta-et”). To generate the competitorb-globin cDNA,

the SK2 plasmid containing the 303-bp wild-tysequence was digested withDraII and StyI restric-ion enzymes, which produced a 29-bp deletiohe b-globin cDNA insert. The linearized plasmbearing the deleted fragment was then filled in wKlenow polymerase and reclosed with T4 DNligase (“competitor target”). Upon transformationEscherichia colistrain XL1-Blue, bacteria from onpositive colony were grown and large amountreference and competitor plasmids were purifiedcarefully quantitated by spectrophotometric anaand gel electrophoresis. To confirm the specificitthe 29-bp deletion both the reference and compeplasmids were bidirectionally sequenced withand T3 primers.

Validation of the RT-PCR Assay

The QC (RT)-PCR assays forb-globinmRNA were analyzed with the mathematimodel described by Vu et al. (24). Briefly, intypical QC (RT)-PCR experiment two dilutioseries were set up, one with a constant natarget amount (T0)1 and the other with a constareference target amount (T0)2. The mathematica

odel is based on the equation

XEP 5 (MEP)1/(T0)1 5 (MEP)2/(T0)2,

here (T0)2 is a constant amount of the referearget; (MEP)2 is the amount of competitor detect the equivalence point of coamplification witheference target itself; (MEP)1 is the amount of com-

petitor at the equivalence point of coamplification (

562

with the natural target;XEP is a constant, typicalnot 1, since it depends on the usually differentplification efficiencies of the target and the comitor DNA (a correction for the difference in lengbetween target and competitor must be takenaccount); and (T0)1 refers to the unknown naturarget amount, which can be calculated fromquation reported above.

Under our experimental conditions both naal and reference targets were coamplified wncreasing copy numbers of DNA competi

olecules (ranging from 106 to 109 molecules). Inarticular, 160 amol (corresponding to 108 mole-ules) of reference target (T0)2 were coamplifieith four different amounts of competitor (5307, 108, 2.53 108, 53 108 molecules). The re-rence and competitor PCR products wereolved on agarose gels and the relative intensf the bands were densitometrically quantitay computer imaging: the logarithms of their

ios were plotted as a function of the concenion of the competitor added. This plot, selectehe line best fitting our data, was used to deine the equivalence point, i.e., the point at wh

he logarithm of the ratio of reference targetompetitor is equal to the logarithm of the ratioheir relative sizes (target/competitor5 1.106; logarget/competitor5 0.044). At the equivalencoint the amount of competitor was found to13.8 amol (MEP)2. Total RNA extracted at eac

sampling time (natural targets) was diluted 1:in DEPC-treated water and 10ml of these diluions were coamplified with increasing amouf the competitor (ranging from 106 to 109 mol-

ecules) and analyzed as described for the rence target to calculate (MEP)1 (the amount ocompetitor at the equivalence point). Since (T0)2

is known in advance and (MEP)2 and (MEP)1 areexperimentally determined, (T0)1 can be calcu-ated easily from the equation reported above

ESULTS

ematological and Biochemical Valuesodified by rEpo Administration

The values for hematocrit, mean cell volu
MCV), reticulocytes, ferritin, soluble transferrin

fors Ajectse ofor-is-

cedfr,ed.up,

. 1.dur-criteivecrit

ac-the

meC

values,



receptor, and sTfr/fr ratio were determinedeach group at each sampling time. In groupand C the means were obtained excluding subnumber 23 (group A) and 22 (group C) becaustransferrin and folic acid values lower than nmal at time 0 and of intolerance to iron admintration, respectively. The treatments indumodifications in Ht, MCV, reticulocytes, fr, sTand sTrf/fr ratio in the three groups examinThese modifications were different in each gro

FIG. 1. Evaluation of MCV (a), hematocrit (b), reticfor each group at each sampling time, as an index of thC, Œ. Values are means for each group6 SE.

depending on the type of treatment used. These

563

differences are graphically represented in FigTwo athletes (numbers 10 and 12, group C)ing treatment were found to have a hematohigher than 50% and consequently did not recthe dose of erythropoietin until the hematovalue returned below this level.

During the period of treatment, a general mrocytosis was observed in all subjects, butvalues were never statistically different from ti0 (Student’st test analysis). In groups A, B, and

tes (c), ferritin (d), and soluble transferrin receptor (e)ect of different rEpo treatments. Group A,■; group B,F; group

ulocye eff

the MCV reached the highest value on day 10 of

thevelythethe

andnifi-es

thein-t ahedtheat-eon

thened4,ceesreeent,ri-in-upseat-luesevelluetial

tin. Intly

18.onlyC,in

lessofndnC,tond

ialin-

ilarum

aterec-n-axi-

ithn-in

asenttheaterasnd

11.3tiolue,atertheatthefr/fre of

if-els

-e ndi oie-s tor-i nt

poedandand. At

ofth-

, as

ay


the treatment and the increase, compared tovalue at time 0, was 1.4, 2.7, and 2.6, respecti(Fig. 1a). The greatest hematocrit increase inathletes of group A, a 7.5% increase overinitial value, was reached on day 10 (Fig. 1b)on days 10 and 14 the hematocrit was sigcantly different from that at time 0. At later timof the treatment, the hematocrit decreased toinitial values. In group B, the hematocritcreased slightly more than in group A and alater time. In fact, the greatest value was reacon day 14 (with a 10% increase compared tovalue at time 0). On days 14 and 18, the hemocrit was significantly different from that at tim0. In group C, the highest value was reachedday 10, with an increase of 7% compared tovalue at time 0. The hematocrit values remaisignificantly different from time 0 on days 10, 1and 24 of the treatment. Analysis of varian(ANOVA) showed that the hematocrit valuwere never significantly different among the thgroups throughout the entire period of treatmconfirming a similar effect of the three expemental protocols in producing a hematocritcrease. As shown in Fig. 1c, in the three grothe reticulocyte count also increased during trment. Groups A and B showed the highest vaon day 14 (respectively 2.9 and 3.3 times the lat time 0). Group C showed the maximum vaon day 18 with a 2.4-fold increase over the inivalue.

During erythropoiesis, the amount of ferridecreases due to the reduction of free irongroup A, the decrease in ferritin was significandifferent from time 0 on days 4, 8, 10, 14, andThe lowest value, reached on day 14, was20% that found at time 0. In groups B andwhich also received iron, folic acid and vitamB12 supplementation, this reduction wasmarked (Fig. 1d). In group B, the amountferritin decreased to 29% of the initial value awas statistically different from time 0 only odays 8, 10, and 14 of the treatment. In grouphowever, the value of ferritin decreased only64% of the value found at time 0 (on day 10) awas never significantly different from the initvalue. In our study, we confirmed that sTfr
creases greatly upon rEpo administration. As
564

shown in Fig. 1e, groups A and B showed simtrends. In both groups sTfr reached the maximon day 18 with values 4.3 and 4.4 times grethan the initial values for groups A and B, resptively. In group C, sTfr showed a lower but costant increase during treatment, reaching a mmal value on day 24 (twice the initial value). Wregard to the sTfr/fr ratio, we essentially cofirmed the results of Gareau et al. (16) onlygroup A (Table 2). In fact, in this case the increin the sTfr/fr ratio was statistically significathroughout the entire period of treatment, withhighest ratio found on day 14 (18.1 times grethan the initial value). In group B, this ratio wsignificantly different on days 4, 8, 10, 14, 18, a24 and reached the maximal value on day 10 (times the initial value). In group C the sTfr/fr rachanged much less drastically: the maximal vareached on day 10, was only 3.2 times grethan the initial value and the other values ofsTfr/fr ratio were statistically different from thfound at time 0 only on days 18 and 24 oftreatment. Furthermore, in this group the sTratio was always lower than the threshold valu403 established by Bressolle et al. (17).

Detection ofb-Globin mRNA Content, uponrEpo Administration, by QuantitativeCompetitive (RT)-PCR

In this study we evaluated the effects of dferent rEpo administration protocols on the levof b-globin mRNA. In fact,b-globin is consid

red a selective marker of erythroid activity ats expression is up-regulated during erythropis stimulation. The study was aimed at moning b-globin mRNA expression in three differegroups of athletes (A, B, and C), receiving rEwith different regimens. In particular we analyzthree athletes from group A (numbers 3, 17,18), three from group B (numbers 4, 5, and 8),three from group C (numbers 1, 11, and 12)each sampling time during treatment, 2 mlwhole blood were drawn from the selected aletes and processed for total RNA extractiondescribed under Materials and Methods.

A quantitative competitive (RT)-PCR ass
was developed to obtain an accurate detection of

ess ofas acu-ive, auateresstan-ith

am-

m-thes

w befi te

9di-

rew d ona pet-i ely)b re-s bles ffi-c

si-b r-e co-a argetD l totT then cans n: inf en-t rala QC( setu unt( nce

ina

eti-

Ma-

TA

BLE

2

sTfr

/frR

atio

Val

ues

ay0

24

810

1418

2428

31

upA

141.

606

70.0

720

6.13

693

.18

850.

3671

9.31

1260

.5665

6.45

1977

.4640

2.8

2557

.5611

67.7

1993

.2611

27.4

537.

2638

4.8

275.

7667

.35

287.

1613

8

upB

84.5

620

.52

81.5

3623

.24

237.

1610

4.79

882.

666

180.

2295

7.73

670

5.84

854.

176

570.

4987

8.66

1116

.717

5.59

648

.15

137.

5631

.22

127.

576

25.5

8

upC

77.4

4617

.22

109.

336

63.6

515

3.36

684

.03

159.

226

126.

7325

0.55

623

8.16

170.

666

116.

217

9.81

688

.44

212.

396

133.

2414

9.46

64.3

712

1.33

623

.25

ote

.Val

ues

are

mea

ns6S

Dob

tain

eddu

ring

trea

tmen

tin

grou

psA

,B

,an

dC

.



b-globin mRNA content in the RNA samplextracted at each time point. Since the kineticPCR are very complex and hence its usequantitative method is not usually exact, an acrate quantitation of nucleic acids by competit(RT)-PCR requires a valid internal standardreference for data normalization and an adeqmathematical model for data analysis. To addthese points we first generated an internal sdard, referred to herein as the “competitor,” wa sequence very similar to that of the targetplified fragment, naturalb-globin cDNA, but witha different size due to a 29-bp deletion. A coparison between the amplification kinetics oftarget and competitorb-globin DNA fragment

as performed by amplifying in the same tuxed amounts (107 molecules) of each templa

with the specificb-globin primer set. After 2amplification cycles and after each of six adtional cycles, 10-ml aliquots of each PCR mixtu

ere removed and the products were resolvegarose gels. Quantification of target and com

tor specific bands (303 and 274 bp, respectivy a gel analyzer (Gel Doc 1000, Bio Rad)ulted in two exponential curves with comparalopes, indicating a similar amplification eiency (data not shown).

Moreover, to completely eliminate the posility of errors arising even from minimal diffences in the amplification efficiencies of themplified sequences, we also synthesized a tNA (indicated as reference target) identica

he amplified fragment of naturalb-globin cDNA.his internal target DNA can exactly simulateatural target during (RT)-PCR and, hence,erve as a reference for absolute quantificatioact, the validity of the method relies on the idical amplification characteristics of the natund reference targets. Briefly, for a typicalRT)-PCR experiment two dilution series arep, one with a constant natural target amoT0)1 and the other with a constant refere

target amount (T0)2. Since the latter is knowndvance and the ratio R between (T0)1 and (T0)2

can be determined from the amounts of comptor at the apparent equimolar points, (T0)1 can becalculated from the equation reported under
terials and Methods.
D

Gro

Gro

Gro N

565

,ssti-r

o oft e

theroupnreefor

sal.69rvedB,dayints

p Cd

b

Mt

ticm ina hed-u

t hatc nt.T Ht),rT

e atersB),ectsblel toon-ineWeenceo.enth-oat-79

hanIn-al-mesh-al.

d inthela,imelue.tsig.

geses

ftervi-ca-Epo

asex-

0, areThe

b


The basal value forb-globin mRNA contentcalculated as the mean6 SD of the single valueat time 0 obtained from the nine subjects invegated, was equal to 3966 77 amol per microlite

f whole blood. Moreover, the quantificationhe increase inb-globin mRNA content at thdifferent sampling times is reported in Fig. 2:results, expressed as mean values for each gclearly indicate thatb-globin mRNA expressioincreased during rEpo administration in the thgroups, reaching the highest value on day 14groups A and B (5.78- and 5.29-fold the bavalue), while for group C the highest value (2times higher than the basal value) was obseon day 18. In the athletes in groups A andb-globin expression drastically decreased on18, reaching the basal value at the last time po(days 28 and 31), while in the athletes in grouthe induction ofb-globin mRNA was less marke

ut long-lasting during rEpo treatment.

ultiparametric Formula: A New Approach tohe Detection of rEpo Abuse

In an attempt to find a reliable diagnosethod with which to detect rEpo abusethletes, regardless of the administration sc

FIG. 2. Quantitative competitive (RT)-PCR ofb-glo-bin mRNA. b-Globin mRNA amounts were quantifieddescribed under Materials and Methods. All values,pressed as fold-increase respect to the value at timemeans6 SE for the three groups at each sampling time.

asal value forb-globin mRNA was equal to 3966 77amol per microliter of whole blood. Group A,■; group B,F; group C,Œ.

le, we developed a mathematical formula that

566

,

akes into account all the biological values thanged most significantly during treatmehe values considered were: hematocrit (eticulocytes (Ret), sTfr andb-globin mRNA.he mathematical formula is

[Ht(%) 3 Ret(%)3 sTfr (nmol/L)3 b-globin mRNA (amol/ml blood)]

100,000

To create a baseline we calculated this valutime 0 for nine subjects identified with numb3, 17, and 18 (group A), 4, 5, and 8 (groupand 1, 11, and 12 (group C). These subjwere the only ones with all parameters availaat time 0. The mean value plus a value equathree times the SD gave a baseline with a cfidence interval of 99.99% (25). The baselvalue established with our data was 5.023.assume that values greater than this referlimit may indicate a probable intake of rEpThe efficacy of this method is evident whcompared with the data obtained with the meods currently utilized for detection of rEpabuse (Fig. 3): in fact, considering the hemocrit values, only 6 time points among theconsidered after time 0 (7.6%) were higher t50% (the threshold limit established by theternational Cycling Federation) (Fig. 3a). Anysis of the sTfr/fr ratio shows values at 28 tipoints out of 79 (35.4%) higher than the threold value of 403 established by Bressolle et(17). In this case, rEpo use was not detecteany of the athletes in group C (Fig. 3b). Oncontrary, using our multiparametric formuthe values calculated at 42 of 73 (57.5%) tpoints were higher than the baseline vaMoreover, in group C, 12 of 24 time poin(50%) after time 0 revealed rEpo abuse (F3c). It is worth noting that these percentawere calculated considering all of the valuavailable, including those obtained 2 days athe beginning of treatment (although it is obous that no significant hematological modifition can be expected 2 days after a single r
injection).

se-ct,

smaTo

oto-TfreaseCRlingrs

aph-thee-

. InAalueilar

p B,he

oupNAble

ues,

oread-ofcela.intageimeob-alue

letesin

w-letesop-as

ideech-theane of

eivinge 50% forh ld valuesw babi



Multiparametric Formula Including Tfr mRNAExpression Detected by RT-PCR

Erythropoietin induces striking changes inrum-soluble transferrin receptor content: in fathe transferrin receptor is released in the plapredominantly by hematopoietic progenitors.evaluate whether the rEpo administration prcols investigated have different effects onmRNA expression, as suggested by the incrin plasma sTfr content, we performed RT-Panalysis of total RNA extracted at each samptime, using Tfr-specific oligonucleotide prime(see Materials and Methods). The results, grically represented in Fig. 4a, revealed thatincrease in Tfr mRNA expression is strictly dpendent on the rEpo administration protocolthe athletes in group A, a peak of Tfr mRNexpression was detected on day 8 and the vthen rapidly decreased to basal levels. A simpattern was observed in the athletes of groubut with a lower increase at each time point. Trelatively low-dose r-Epo treatment used in grC resulted in a less marked increase in Tfr mRexpression, which was however detectathroughout the period of treatment. These val

FIG. 3. Values of hematocrit (a), sTfr/fr ratio (b),rythropoietin with protocols A, B, and C described inematocrit, 403 for sTfr/fr ratio, and 5.023 for multiparaere considered an indication of rEpo intake with a pro

although probably less accurate than those deter-

567

mined by QC (RT)-PCR, can be obtained mquickly and easily and therefore represent anditional parameter for use in the investigationblood samples within the limits of the confideninterval defined by the multiparametric formuIn fact, the inclusion of Tfr mRNA expressionthe multiparametric formula raises the percenof rEpo abuse detection to 68.1% (at 47 tpoints out of the 69 investigated the valuestained were higher than the new threshold vof 48,048) (Fig. 4b).

DISCUSSION

Indirect evidence suggests that some athutilize specific substances, legitimately usedmedicine, to improve their performance. Hoever, in recent years the percentage of athfound positive for substance abuse in the antiding tests performed by different federations wonly about 1%, showing that it is possible to hthis practice due to the inadequate analytical tniques currently available (26). Furthermore,considerable availability of recombinant humerythropoietin has allowed the widespread us

ultiparametric values (c) in the individual athletes recle 1. Continuous lines represent the threshold limits:ic analysis. Values greater than the respective thresholity greater than 99.99%. Group A,■; group B,F; group C,Œ.

and mTabmetr

this drug in aerobic sports to increase oxygen

ests

etect9).tectveriner-

ingure

eth-ectsEpoof

au-f the

). PCRp to knowna of PCRp . (b)

reshold


transfer capacity (27). At present, antidoping tcannot reliably detect rEpo in circulation.

Several methods have been explored to dillicit drug use in competitive sports (28, 2Recently, a direct approach with which to derEpo in urine has been proposed (12). Howethis method should find its principal applicationout-of-competition testing since it does not ov

FIG. 4. (a) RT-PCR detection of Tfr mRNA in athroducts were resolved by electrophoresis and targetmounts of DNA of similar size (DNA molecular weigroduct per microliter of whole blood, are the mean6

Multiparametric values, including Tfr mRNA, of the athllimit of 48,048: greater values indicate intake of rEpo wC, Œ.

come the kinetic constraints of the short half-life

568

of rEpo (due to its rapid clearance), thus makits detection reliable only after a recent exposto the hormone. Several alternative indirect mods addressing the delayed erythropoietic effof rEpo have also been developed. Since radministration also induces the redistributioniron from stores to erythroid elements, somethors (16, 17, 30) have suggested the use o

receiving rEpo with different regimens (see Table 1s quantified by ethidium bromide staining with respectrker VI, Roche): all values, expressed as attomolesbtained for the three groups at each sampling timexamined in Fig. 3c. The continuous line indicates the throbability greater than 99.99%. Group A,■; group B,F; group

letesbandht masSE oetes eith a p

ratio between the soluble transferrin receptor and

bleormrrin–beIronrrinn ses ach isendandneu-lesstiocan

on-nlyanden

iatedcalon-ath-ug-lly

notsi-ap-ali-

res.tectinis-ul-

theam-po

nfi-stab-wasimitsy ofthethewnectned,

nifi-ctedluesula,f theasediontalsly

dlyhehe

esres-mestopa-therncela.

a-Epoob-

forb-

notporag-

nta-iblybys,

efi me-t trica hode sev-e

b-e nds , Dr.D ndd iana( .A.



serum ferritin to detect rEpo abuse. The soluTfr present in human plasma is a truncated fof the tissue receptor and exists as a transfereceptor complex (18, 31); serum ferritin canused as a measure of iron stores in the body.deprivation causes prompt induction of transfereceptor synthesis (18, 32) and a decrease irum ferritin levels: the sTfr/fr ratio increases aconsequence of these events. This approainteresting because in principle it does not depon body hydration (as the hematocrit does)could reveal rEpo abuse as well as other mavers that accelerate erythropoiesis. Neverthein this study we demonstrate that the sTfr/fr radepends on the administration schedule andvary upon iron supplementation. Our results cfirm the conclusion of Gareau et al. (17) owhen rEpo is administered at high doseswithout iron supplementation. In contrast, whrEpo is administered at low doses and assocwith an iron supplement, as is common in clinipractice (33, 34), the sTfr/fr ratio cannot be csidered a reliable marker of rEpo abuse inletes. The main method currently available, sgested by Parisotto et al. (22), it is potentiaeffective to identify users of rEpo but it doesexclude the possibility of registering false potives. Moreover, in this method there are nopropriate internal standards for inter-assay cbration laboratory and quality control procedu

We propose a new technique able to derEpo abuse in athletes regardless of the admtration protocol. Our approach is based on a mtiparametric formula that takes into account allhematological, biochemical and molecular pareters that change most significantly after rEintake: hematocrit, reticulocytes, sTfr andb-glo-bin mRNA. The reference ranges (99.99% codence interval) for these parameters were elished and the threshold value with our data5.023. Values greater than these reference lindicate a probable intake of rEpo. The efficacthis method is evident when compared withcurrently utilized antidoping tests based onhematocrit readings and sTfr/fr index. As shoin Fig. 3, our multiparametric formula can detrEpo abuse in 57.5% of the samples exami
considering all the values available, including (
569

-

,

those obtained at 2–4 days, although no sigcant hematological modification can be expeat the beginning of the treatment. If these vaare not considered, our multiparametric formcan detect rEpo abuse higher than 99.99% osamples examined. The screening methods bon hematocrit and sTfr/fr ratio allow the detectof rEpo intake in only 7.6 and 35.4% of tosamples, respectively. Moreover, we previoureported that Tfr mRNA expression is markeaffected by rEpo administration: in fact, while tconcentration of immunoreactive sTfr in tplasma increased to a maximal value five timgreater than the basal value, Tfr mRNA expsion detected by RT-PCR increased up to 40 ti(35). Therefore, Tfr mRNA content may helpreveal rEpo intake and thus is an additionalrameter that can be considered when the omarker values are within the reference confideintervals defined by the multiparametric formuThe inclusion of Tfr mRNA value in the multiprametric analysis raises the percentage of rabuse detection to 68.1%. These data weretained considering only the nine subjectswhom all the marker values were available. Oviously, the number of subjects considered issufficient to validate a method to detect rEabuse, but the results obtained are very encouing and prompt us to continue this experimetion on a greater number of athletes and possunder different physiological conditions (i.e.,varying body hydration). In preliminary studiewe found that 50ml of blood sampled from thngertip are enough to determine all the paraers included in the proposed multiparamenalysis. This approach would make this metven more acceptable to athletes involved inral-week-long competitions.

ACKNOWLEDGMENTS

We gratefully thank Drs. Ovidio Stocchi and Rorto Guazzolini of the Urbino Hospital for support auggestions regarding hematological determinationsaniele Ribelli (ABX Rome) for reticulocyte counts aiscussion, the Federazione Medico Sportiva ItalRome, Italy), and Pacific Electronic Instrument S
RSM) for partial support of this study.

gy

O.tic

ptor.etinslo-in

y-othn.

ng,ing,

D.,2)

theA

of

.,n-

anpa-

1 Yi,K2

ty-mu-

1 Lo-thero-A

1 inant

1 re-

1 J.,H.

en.

1 . E.,u-

nant.

1 ers,.

1 R.,tiontes.

ijk,ifi-.

ey,ytenous

ne,ofpa-ns-.lism

den,ut,ingn oftes.

ndetic

an

ell,an-byata

2 ical

22 nant

ce.

2 for


REFERENCES

1. Porter, D. L., and Goldberg, M. A. (1994) Physioloof erythropoietin production.Semin. Hematol.31,112–121.

2. Krantz, S. B. (1991) Erythropoietin.Blood.77, 419–434.

3. Kirby, S., Cook, D., Walton, W., and Smithies,(1996) Proliferation of multipotent hematopoiecells controlled by truncated erythropoietin recetransgene.Proc. Natl. Acad. Sci. USA93,9402–9407

4. Penta, K., and Sawyer, S. T. (1995) Erythropoiinduces the tyrosine phosphorylation, nuclear trancation, and DNA binding of STAT1 and STAT5erythroid cells.J. Biol. Chem.270,31282–31287.

5. Liboi, E., Carroll, M., D’Andrea, A., and MathePrevot, B. (1993) Erythropoietin receptor signals bproliferation and erythroid-specific differentiatioProc. Natl. Acad. Sci. USA90, 11351–11355.

6. Jones, S. S., D’Andrea, A. D., Hines, L. L., and WoG. G. (1990) Human erythropoietin receptor: Clonexpression, and biological characterization.Blood.76,31–35.

7. Watowich, S. S., Yoshimura, A., Longmore, G.Hilton, D. J., Yoshimura, Y., and Lodish, H. F. (199Homodimerization and constitutive activation oferythropoietin receptor.Proc. Natl. Acad. Sci. US89, 2140–2144.

8. Lacombe, C., and Mayeux, P. (1998) Biologyerythropoietin.Haematologica83, 724–732.

9. Biljanovic-Paunovic, L., Djukanovic, L., Lezaic, VStojanovic, N., Marisavljevic, D., and Pavlovic-Ketera, V. (1998)In vivo effects of recombinant humerythropoietin on bone marrow hematopoiesis intients with chronic renal failure.Eur. J. Med. Res.3,564–570.

0. Witthuhn, B. A., Quelle, F. W., Silvennoinen, O.,T., Tang, B., Miura, O., and Ihle, J. N. (1993) JAassociates with the erythropoietin receptor and isrosine phosphorylated and activated following stilation with erythropoietin.Cell 74, 227–236.

1. Klingmuller, U., Bergelson, S., Hsiao, J. G., anddish, H. F. (1996) Multiple tyrosine residues incytosolic domain of the erythropoietin receptor pmote activation of STAT5.Proc. Natl. Acad. Sci. US93, 8324–8328.

2. Lasne, F., and de Ceaurriz, J. (2000) Recomberythropoietin in urine.Nature405,635.

3. Flaharty, K. K. (1990) Clinical pharmacology ofcombinant human erythropoietin (rHuEpo).Pharma-cotherapy10, 9S–14S.

4. Flaharty, K. K., Caro, J., Erslev, A., Whalen, J.Morris, E. M., Bjornsson, T. D., and Vlasses, P.
(1990) Pharmacokinetics and erythropoietic response
570

to human recombinant erythropoietin in healthy mClin. Pharmacol. Ther.47, 557–564.

5. Egrie, J. C., Cotes, P. M., Lane, J., Gaines Das, Rand Tam, R. C. (1987) Development of a radioimmnoassay for human erythropoietin using recombierythropoietin as tracer and immunogen.J. ImmunolMethods99, 235–241.

6. Gareau, R., Audran, M., Baynes, R. D., FlowC. H., Duvallet, A., Sene´cal, L., and Brisson, G. R(1996) Erythropoietin abuse in athletes.Nature 380,113.

7. Bressolle, F., Audran, M., Guidicelli, C., Gareau,Baynes, R. D., and Gomeni, R. (1997) Populapharmacodynamics for monitoring epoetin in athleClin. Drug Invest.14, 233–242.

18. Feelders, R. A., Kuiper-Kramer EPA, and van EH. G. (1999) Structure, function and clinical signcance of transferrin receptors.Clin. Chem. Lab. Med37, 1–10.

19. Bhandari, S., Norfolk, D., Brownjohn, A., and TurnJ. (1997) Evaluation of RBC ferritin and reticulocmeasurements in monitoring response to intraveiron therapy.Am. J. Kidney Dis.30, 814–821.

20. Beguin, Y., Loo, M., R’Zik, S., Sautois, B., LejeuF., Rorive, G., and Fillet, G. (1993) Early predictionresponse to recombinant human erythropoietin intients with the anemia of renal failure by serum traferrin receptor and fibrinogen.Blood 82, 2010–2016

21. Adamson, J. (1997) Erythropoietin, iron metaboand red blood cell production.Semin. Hematol.33(2,Suppl. 2), 5–7.

22. Parisotto, R., Gore, J. C., Emslie, K. R., AshenM. J., Brugnara, C., Howe, C., Martin, D. T., TroG. J., and Hahn, A. G. (2000) A novel method utilismarkers of altered erythropoiesis for the detectiorecombinant human erythropoietin abuse in athleHaematologica85, 564–572.

23. Fullerton, S. M., Harding, R. M., Boyce, A. J., aClegg, J. B. (1994) Molecular and population genanalysis of allelic sequence diversity at the humb-globin locus.Proc. Natl. Acad. Sci. USA91, 1805–1809.

24. Vu, H. L., Troubetzkoy, S., Nguyen, H. H., RussM. W., and Mestecky, J. (2000) A method for qutification of absolute amounts of nucleic acids(RT)-PCR and a new mathematical model for danalysis.Nucleic Acids Res.28, 1–9.

5. Dacie, J. V., and Lewis, S. M. (1975) PractHaematology. Churchill Livingstone.

6. Unknown (1991)Le Mond Cycliste23, 10–12.7. Adamson, J. W., and Vapnek, D. (1991) Recombi

erythropoietin to improve athletic performanN. Engl. J. Med.324,698–699.

8. Cayla, J. L., and Duvallet, A. (1999) Screening
exogenous erythropoietin.Presse Med.28, 992–996.

ingal-

3 F.,lle,inin

.3 uble

ptor

poi-

yth-

.,ni--



29. Lippi, G., and Guidi, G. (2000) Laboratory screenfor erythropoietin abuse in sport: An emerging chlenge.Clin. Chem. Lab. Med.38, 13–19.

0. Audran, M., Gareau, R., Matecki, S., Durand,Chenard, C., Sicart, M. T., Marion, B., and BressoF. (1999) Effects of erythropoietin administrationtraining athletes and possible indirect detectiondoping control.Med. Sci. Sports Exerc.31, 639–645

1. Ahn, J., and Johnstone, R. M. (1993) Origin of a sol
truncated transferrin receptor.Blood81, 2442–2451.
571

32. Skikne, B. S. (1998) Circulating transferrin receassay—Coming of age.Clin. Chem.44, 7–9.

33. Spivak, J. L. (1994) Recombinant human erythroetin and the anemia of cancer.Blood 84, 997–1004.

34. Sunder-Plassmann, G., and Horl, W. H. (1997) Erropoietin and iron.Clin. Nephrol.47, 141–157.

35. Magnani, M., Corsi, D., Bianchi, M., Paiardini, MGalluzzi, L., Parisi, A., and Pigozzi, F. (1999) Motoring erythropoietin abuse in athletes.Br. J. Haema
tol. 106,260–261.

identification of blood erythroid markers useful in revealing erythropoietin abuse in athletes

Documents