The presence of a significant associationbetween elevated PRV-1 mRNA expressionand low plasma erythropoietin concentrationin essential thrombocythaemia

It is well agreed that serum or plasma erythropoie-tin (EPO) concentrations are subnormal in mostpatients with polycythaemia vera (PV) (1–6).Indeed, in newly diagnosed and untreated PVpatients, by using a highly sensitive assay for theidentification of low EPO concentrations, thevalues for plasma EPO were consistently shownto be either not measurable or below the referencelimit for normal adults (5, 7, 8). Additionally, it wasdemonstrated that 46–50% of newly diagnosed anduntreated patients with essential thrombocythae-mia (ET) also presented with subnormal plasmaEPO concentrations (5, 9). As to ET several othergroups of workers have reached similar conclusions

(6, 10, 11). Further, there is good reason to suggestthat a subnormal EPO concentration in a newlydiagnosed ET patient heralds that the patient ismore likely to experience occlusive vascular eventsthan ET patients who at diagnosis demonstrate anormal plasma EPO (12). Thus, in a retrospectivework of Andreasson et al. (9) the reason forinitiation of myelosuppressive treatment to ETpatients with subnormal EPO concentrations werevascular events in 11 of 13 patients compared withnine of 18 with normal plasma EPO values. Thisdifference did not quite reach statistical significancewhen microvascular events were included in thestudy. However, when the reason for treatment was

Johansson P, Andreasson B, Safai-Kutti S, Wennstrom L, Palmqvist L,Ricksten A, Lindstedt G, Kutti J. The presence of a significantassociation between elevated PRV-1 mRNA expression and low plasmaerythropoietin concentration in essential thrombocythaemia.Eur J Haematol 2003: 70: 358–362. � Blackwell Munksgaard 2003.

Abstract: Approximately 45% of newly diagnosed patients with essen-tial thrombocythaemia (ET) demonstrate subnormal plasma erythro-poietin (EPO) concentrations, which constitutes a risk factor forocclusive vascular events. In 58 ET patients, a possible associationbetween polycythaemia rubra vera-1 (PRV-1) overexpression and sub-normal plasma EPO was investigated, which was always measured priorto the institution of platelet lowering agents. At the time when PRV-1expression was measured, 28 of 58 (48%) ET patients had receivedplatelet lowering treatment. PRV-1 expression was measured by quan-titative real-time reverse transcription–polymerase chain reaction assayof mRNA extracted from purified peripheral blood buffy coat. The cyclethreshold (CT) value of PRV-1 was determined and was divided with theCT value for the housekeeping GAPDH gene transcript. A quotient<0.93 was defined as PRV-1 positive. Of the ET patients 12 of 58 (21%)were PRV-1 positive and 19 of 58 (33%) demonstrated subnormalplasma EPO. In the 58 ET patients there was a significant associationbetween low plasma EPO and PRV-1 positive results (P ¼ 0.001). The30 ET patients who had not received any platelet lowering treatmentshowed a significant (P ¼ 0.005) relation between PRV-1 positivity andsubnormal plasma EPO. No such relationship was present in the 28 ETpatients who had received prior treatment with the above drugs(P ¼ 0.147).

Peter Johansson1,2, Bj�rnAndr�asson1,2, Soodabeh Safai-Kutti1, Lovisa Wennstr�m1, LarsPalmqvist3, Anne Ricksten3, G�ranLindstedt3, Jack Kutti11Haematology and Coagulation Section, Departmentof Medicine, Sahlgrenska University Hospital, Universityof G�teborg, G�teborg; 2Department of Medicine,Uddevalla Hospital, Uddevalla, and 3Department ofClinical Chemistry and Transfusion Medicine,Sahlgrenska University Hospital, University of G�teborg,G�teborg, Sweden

Key words: PRV-1; erythropoietin; essentialthrombocythaemia; thrombosis; myelosuppression

Correspondence: Dr Peter Johansson, Haematology andCoagulation Section, Department of Medicine,Sahlgrenska University Hospital, S-413 45 G�teborg,SwedenTel: +46 522 93212Fax: +46 522 93232e-mail: peter.l.johansson@vgregion.se

Accepted for publication 26 March 2003

Eur J Haematol 2003: 70: 358–362Printed in UK. All rights reserved

Copyright � Blackwell Munksgaard 2003

EUROPEANJOURNAL OF HAEMATOLOGY

ISSN 0902-4441

358

restricted to cerebrovascular and other macrovas-cular events the difference became statisticallysignificant (P < 0.02). Similar observations wererecently made by Messinezy et al. (6). In theirwork, 11 of 16 ET patients with low EPO had ahistory of thrombosis or ischaemia, compared withonly 16 of 33 patients without a low EPO.Over recent years, by using clonality assays,

evidence has accumulated that ET is an heteroge-nous disorder. In view of available informationapproximately 50% of evaluable female ETpatients demonstrated monoclonal haemopoiesis(13–15). Additionally, it was also demonstratedthat thrombotic complications were significantlymore common in monoclonal than polyclonalpatients. The results of recent work carried outdid not, however, show any significant associationbetween monoclonality and low plasma EPO con-centration in ET (16).The cloning of a novel gene named polycythae-

mia rubra vera-1 (PRV-1) which codes for a cellsurface receptor was recently described by Temer-inac et al. (17). They found that mRNA from thePRV-1 gene was overexpressed in granulocytesfrom patients with PV, and occasionally in patientswith ET and idiopathic myelofibrosis. The presentwork was undertaken in order to investigate theincidence of increased PRV-1 mRNA expression inET, and also to study a possible associationbetween subnormal plasma EPO concentrationand PRV-1 overexpression in this disorder.

Material and methods

Patients and controls

The present study was approved by the EthicsCommittee of Goteborg University and all patientssigned informed consent forms in accordance withthe Declaration of Helsinki.A total of 58 patients (20 males and 38 females)

with ET aged 26–89 (mean 60) yr were includedfrom two haematology centres. The diagnosis ofET was established according to the PolycythemiaVera Study Group criteria (18). Additionally, itshould be stated that a bone marrow biopsy fromthe posterior iliac crest was obtained from each ofthe patients and that they also satisfied the revisedcriteria for the diagnosis of ET as proposed (19).Thirty of the patients had never received anyplatelet lowering treatment. Of the remaining 28patients, 20 were treated with hydroxyurea, fivewith anagrelide and three had received radiophos-phorous. Their disease duration ranged 0–20(median 1.8; mean 3.4) yr. Thirty of the 58 ETpatients had been diagnosed during the two recentyears. The control group consists of 40 healthy

volunteers; there were 19 males and 21 females aged29–65 (mean 48) yr.

Methods

Plasma EPO measurement. All blood samples werecollected between 08.00 and 11.30 hours. Blood cellcounts were analysed with Technicon H2equipment (Bayer Diagnostics Tarrytown NY,USA). Determination of blood haemoglobin,platelet count, total white cell count and totalgranulocyte count was performed in each patient.For the measurement of plasma EPO concentration,blood was collected into ethylenediaminetetraaceticacid (EDTA) vacuum tubes containing K2-EDTA.Prior to 1998 were all measurements of plasmaEPO concentrations carried out by using animmunoradiometric assay (125I-EPOCOATRIA;bioMerieux SA, Marcy-l�Etoile, France) with anormal reference interval in adults of 3.7–16 IU/L.Herein are included 23 ET patients of the presentseries. A plasma EPO<3.7 IU/L was considered tobe subnormal (5). In the remaining 35 ET patients,measurements of the plasma EPO concentrationwere carried out by using an immunoenzymaticassay (Quantikine IVDHumanEPODEP 00;R&DSystems, Minneapolis, MN, USA) with a normalrange of 3.1–14.9 IU/L in adults and<3.1 IU/L forsubnormal EPO concentrations. The reason forchange of method was the sudden withdrawal ofthe former assay from the market by the producer,no information being given to customers before-hand by bioMerieux SA or its local representative.The withdrawal was the result of a policy change asregards future delivery of immunoassays withradioactive ligands, favouring those with non-radioactive ligands. In retrospective analysis, withthe R &D assay, of samples previously assayed withthe bioMerieux assay, a good correlation betweenassays was found between the two methodsconfirming that they both are high-detectabilityassays although the degree of matrix effect, albeitlow, might differ.In most of the 58 ET patients, blood sampling for

measurement of the plasma EPO concentration wascarried out at diagnosis when untreated and alwaysbefore the institution of any myelosuppressive/platelet lowering treatment.

Buffy coat PRV-1 mRNA measurement. Ten milliliter of EDTA-anticoagulated venous bloodsamples were collected from the patients andblood nuclear cells were isolated and counted.mRNA was extracted from 105 cells using theGenoM-48 Robotic workstation (GenoVision,Norway). Complementary DNA was generated by

PRV-1 mRNA expression and low plasma erythropoietin concentration

359

reverse transcription (RT) with random primers(Hexanucleotidmix, Roche, Sweden) using theSuperscript II enzyme (InVitrogen, Stockholm,Sweden). The cDNA was stored at )20�C.PRV-1 expression was quantified by real-time

PCR with primers deduced from the human PRV-1mRNA sequence (NM 020406) and detected withfluorescent dye SYBR Green in combination withmelt curve analysis to ensure the specificity of thePCR product. The cycle threshold (CT) value ofPRV-1 was determined and was then divided withthe CT value for the housekeeping glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene tran-script, quantified with the TaqMan probe system,to standardise the result. Consequently, a lowerPRV-1/GAPDH ratio indicates a higher PRV-1mRNA expression. All patient samples were ana-lysed in duplicates together with reference sampleswith known PRV-1 expression levels.The day-to-day coefficient of variation (CV) for the assay wasdetermined to be approximately 3.1%.PRV-1 expression was detected in all the healthy

volunteers and showed a normal distribution, and a95% limit reference interval for the PRV-1/GAP-DH ratio was determined to 0.93–1.25. A PRV-1/GAPDH ratio below 0.93 was considered to be apositive result indicating increased expression ofPRV-1 in peripheral blood nucleated cells and isrelated to as PRV-1 positive.

Statistics

Standard statistical methods were employed. Thedifference between mean values was tested withStudent’s t-test. Fisher’s exact test was used to testpossible differences in significance in PRV-1expression between ET patients with subnormaland normal/elevated plasma EPO concentrations,respectively.

Results

The results for plasma EPO and PRV-1 in the totalgroup of ET patients are summarised in Table 1. Inthe present series 12 of 58 ET patients (21%)demonstrated overexpression of PRV-1 mRNA.Among the 58 ET patients, 19 (33%) had subnor-

mal plasma EPO concentrations, whereas theremaining 39 had EPO concentrations above thelower reference limits. Nine of the patients withsubnormal EPO (47%) were PRV-1 positive com-pared with the three subjects with normal/elevatedplasma EPO (8%), and the difference betweenthese two groups of patients was statisticallysignificant (P ¼ 0.001). Four patients had plasmaEPO concentrations above the upper referenceranges for the two assays, but they were all PRV-1negative.Table 2 shows the results for plasma EPO and

PRV-1 for those 30 ET patients who at the time forPRV-1 sampling had never been exposed to anymyelosuppressive/platelet lowering treatment. Theresults for the 28 patients who at the time of thestudy were receiving, or previously had receivedmyelosuppressive/platelet lowering therapy, alsoappear in Table 2. In the former group of subjectsit is seen that five of nine (56%) patients withsubnormal plasma EPO concentrations werePRV-1 positive. For the remaining 21 subjects withnormal/elevated plasma EPO concentrations wasonly one (5%) subject PRV-1 positive, and thedifference between those two groups was statisti-cally significant (P ¼ 0.005).In the group of ET receiving myelosuppressive/

platelet lowering therapy four of 10 (40%) patientswith subnormal plasma EPO concentration hadPRV-1 overexpression whereas two of 18 (11%)patients with normal/elevated plasma EPO concen-tration were positive for PRV-1; this difference did,however, not reach statistical significance(P ¼ 0.147).Total granulocyte counts were simultaneously

determined in each patient. There was no signifi-cant difference in total granulocyte counts betweenpatients with subnormal or normal/elevated plasmaEPO concentrations (P ¼ 0.90). Nor was there anydifference between patients who were PRV-1 pos-itive or PRV-1 negative (P ¼ 0.85).

Table 2. The results for plasma erythropoietin (p-EPO) and PRV-1 expression in 30patients with ET who prior to study had not received any myelosuppressive/plateletlowering therapy, and in 28 patients who were receiving such treatment

PRV-1 positive* PRV-1 negative**

No myelosuppressive therapySubnormal p-EPO (n ¼ 9) 5 4Normal/elevated p-EPO (n ¼ 21) 1 20Total (n ¼ 30) 6 24

Myelosuppressive therapySubnormal p-EPO (n ¼ 10) 4 6Normal/elevated p-EPO (n ¼ 18) 2 16Total (n ¼ 28) 6 22

*Quotient CT PRV-1/CT GAPDH < 0.93.**Quotient CT PRV-1/CT GAPDH ‡ 0.93.

Table 1. The results for plasma erythropoietin (p-EPO) and PRV-1 expression in 58patients with ET

PRV-1 positive* PRV-1 negative**

Subnormal p-EPO (n ¼ 19) 9 10Normal/elevated p-EPO (n ¼ 39) 3 36Total (n ¼ 58) 12 46

*Quotient CT PRV-1/CT GAPDH < 0.93.**Quotient CT PRV-1/CT GAPDH ‡ 0.93.

Johansson et al.

360

Discussion

Although the present study demonstrated that therewas a significant relationship between overexpres-sion of PRV-1 mRNA and subnormal plasma EPOconcentration in patients with ET the biologicalimplication of this observation remains obscure. Itwas shown that the plasma EPO concentrationswere subnormal in 33% of the 58 ET patients. Thisfigure is slightly lower than was reported in our twoprevious studies on ET where subnormal EPOvalues were recorded in 46–50% of the patients (5,9). However, our previous studies were carried outby using the immunoradiometric assay provided bybioMerieux whereas only 23 of the 58 ET patientsin the current study were assayed by using thismethod. In the remaining 35 patients the immuno-enzymatic assay provided by R & D Systems wasemployed. Indeed, in the recent work by Messinezyet al. (6) also the assay provided by R & D Systemswas employed; herein 33% of the ET patients wereconsidered to have low EPO values. It could,therefore, be that the slightly lower figure forsubnormal EPO concentrations in the current studyas compared with our previous reports could partlybe accounted to differences in assays employed.However, at the time when EPO analyses werechanged from bioMerieux to R & D Systems, verycareful comparative studies were carried outbetween the two assays and a very close correlation,also in the low range of EPO concentrations, wasshown to be present (20). Indeed, a subnormal EPOconcentration in PV and in ET is known to increaserapidly when myelosuppressive therapy is intro-duced (5, 7, 9).In view of what is known previously it thus

appears that a subnormal EPO concentration as wellas the presence of monoclonal disease in ET are riskfactors for the development of occlusive vasculardisease (9, 12–15). On the basis of recent workcarried out, however, it appears that these two riskfactors are likely to operate independently (16). Thedemonstration that there is a significant associationbetween PRV-1 positivity and subnormal EPO inET requires further studies. As to PRV-1 expressionseveral issues are presently unknown, e.g. it iscurrently not established whether PRV-1 expressionis affected by myelosuppressive treatment. In thepresent work we, therefore, separately studied therelationship between PRV-1 and subnormal plasmaEPO in 30 ET patients who had never been exposedto platelet lowering therapy and in 28 ET patientswho had received or were receiving chemotherapy.The results showed that the relation between thesetwo variables was statistically significant in subjectswho had never been exposed to chemotherapy or

platelet lowering therapy. However, in the group of28 ET patients who had been exposed to chemo-therapy no such significant relationship was present.This finding is likely to indicate that chemotherapycould affect PRV-1 expression. However, at whichlevel this would occur is not yet understood. On thebasis of the present results it also appears that anuneven distribution of the total granulocyte countbetween the groups could not possibly account forthe differences observed.Recently, by using a qualitative RT–polymerase

chain reaction (RT–PCR), Teofili et al. (21) studiedPRV-1 expression in granulocytes isolated frompatients with PV and ET. It was demonstrated thatall ET patients (37 of 37) and most PV patients (35of 37) in contrast to normal healthy controlsexpressed PRV-1. These findings are consequentlynot fully in accord with the present results. In ourexperience the vast majority of PV patients over-express PRV-1 (results to be published) but only21% of the ET patients of the present workdemonstrated overexpression of PRV-1, when com-pared with our reference interval for the PRV-1/GAPDH ratio. The difference between the results ofthese two studies could at least partly be accountedto differences in the assays employed. Thus, thepresent results were in contrast to the work ofTeofili et al. (21) based upon a quantitative RT–PCR and mRNA extraction from buffy coat cells.It could be speculated that PRV-1 overexpression

and a subnormal plasma EPO concentration metwith in approximately one-third of the ET patientsstudied heralds transition to a subsequent state ofPV. Indeed, it might well be that a fraction of ETpatients who at the time of diagnosis fulfil allpossible diagnostic criteria for ET rather representearly cases of PV. This hypothesis cannot beexcluded at present and requires careful prospectivefollow-up studies. It should be stated that mucheffort was conducted to ascertain that the patientsof the present work were true cases of ET inasmuchas they had all to satisfy not only the PVSG criteriafor the diagnosis of ET (18) but also the revisedcriteria for ET as proposed by us (19). It is evidentthat prospective studies have to solve severalpresently unknown issues including if and to whichextent PRV-1 expression is influenced by chemo-therapy and whether PRV-1 overexpression in ETrepresents a risk factor for the development ofthrombotic vascular complications.

Acknowledgements

The authors take the opportunity to thank Professor HeikePahl, Freiburg, Germany, for advice and most fruitful discus-sions. We gratefully acknowledge the expert measurements of

PRV-1 mRNA expression and low plasma erythropoietin concentration

361

plasma EPO by Per-Arne Lundberg. We also wish to thankCarina Wasslavik for expert technical assistance. This studywas supported by grants from �Stiftelsen JubileumsklinikensForskningsfond�, �FoU Vastra Gotaland�, �Assar GabrielssonsFond� and �FoU-radet Fyrbodal�.

References

1. Erslev AJ, Caro J. Pure erythrocytosis classified accordingto erythropoietin titers. Am J Med 1984;76:57–61.

2. Cotes PM, Dore CJ, Yin JA, Lewis SM, Messinezy M,Pearson TC, Reid C. Determination of serum immuno-reactive erythropoietin in the investigation of erythrocyto-sis. N Engl J Med 1986;315:283–287.

3. Birgegard G, Wide L. Serum erythropoietin in the diag-nosis of polycythaemia and after phlebotomy treatment. BrJ Haematol 1992;81:603–606.

4. Messinezy M, Westwood NB, Woodcock SP, Strong

RM, Pearson TC. Low serum erythropoietin – a strongdiagnostic criterion of primary polycythaemia even atnormal haemoglobin levels. Clin Lab Haematol1995;17:217–220.

5. Carneskog J, Kutti J, Wadenvik H, Lundberg PA,Lindstedt G. Plasma erythropoietin by high-detectabilityimmunoradiometric assay in untreated and treated patientswith polycythaemia vera and essential thrombocythaemia.Eur J Haematol 1998;60:278–282.

6. Messinezy M, Westwood NB, El-Hemaidi I, Marsden

JT, Sherwood RS, Pearson TC. Serum erythropoietinvalues in erythrocytoses and in primary thrombocythaemia.Br J Haematol 2002;117:47–53.

7. Andreasson B, Carneskog J, Lindstedt G, Lundberg

PA, Swolin B, Wadenvik H, Kutti J. Plasma erythro-poietin concentrations in polycythaemia vera with specialreference to myelosuppressive therapy. Leuk Lymphoma2000;37:189–195.

8. Johansson P, Safai-Kutti S, Lindstedt G, Suurkula M,Kutti J. Red cell mass, spleen size and plasma erythro-poietin in polycythaemia vera and apparent polycythaemia.Acta Haematol 2002;108:1–7.

9. Andreasson B, Lindstedt G, Kutti J. Plasma erythro-poietin in essential thrombocythaemia: at diagnosis and inresponse to myelosuppressive treatment. Leuk Lymphoma2000;38:113–120.

10. Viala JJ, Ville D, Sebban C, Assouline D, Devaux Y,Hanss M. Plasma erythropoietin in essential thrombo-cythaemia. Nouv Rev Fr Hematol 1993;35:423–424.

11. Griesshammer M, Kubanek B, Beneke H, Heimpel H,Bangerter M, Bergmann L, Schrezenmeier H. Serum

erythropoietin and thrombopoietin levels in patients withessential thrombocythaemia. Leuk Lymphoma 2000;36:533–538.

12. Andreasson B. Clinical and Experimental Studies inPolycythemia Vera and Essential Thrombocythemia (Dis-sertation). Vasastadens Bokbinderi AB: Goteborg, 2000.

13. Harrison CN, Gale RE, Machin SJ, Linch DC. A largeproportion of patients with a diagnosis of essentialthrombocythemia do not have a clonal disorder and may beat lower risk of thrombotic complications. Blood1999;93:417–424.

14. Chiusolo P, La Barbera EO, Laurenti L, et al. Clonalhemopoiesis and risk of thrombosis in young female pa-tients with essential thrombocythemia. Exp Hematol2001;29:670–676.

15. Shih LY, Lin TL, Lai CL, Dunn P, Wu JH, Wang PN,Kuo MC, Lee LC. Predictive values of X-chromosomeinactivation patterns and clinicohematologic parametersfor vascular complications in female patients with essentialthrombocythemia. Blood 2002;100:1596–1601.

16. Andreasson B, Harrison C, Lindstedt G, Linch D,Kutti J. Monoclonal myelopoiesis and subnormal ery-thropoietin concentration are independent risk factors forthromboembolic complications in essential thrombocythe-mia. Blood 2003;101:783–784.

17. Temerinac S, Klippel S, Strunck E, Roder S, Lubbert

M, Lange W, Azemar M, Meinhardt G, Schaefer HE,Pahl HL. Cloning of PRV-1, a novel member of the uPARreceptor superfamily, which is overexpressed in polycythe-mia rubra vera. Blood 2000;95:2569–2576.

18. Murphy S, Peterson P, Iland H, Laszlo J. Experience ofthe Polycythemia Vera Study Group with essentialthrombocythemia: a final report on diagnostic criteria,survival, and leukemic transition by treatment. SeminHematol 1997;34:29–39.

19. Kutti J, Wadenvik H. Diagnostic and differential criteriaof essential thrombocythemia and reactive thrombocytosis.Leuk Lymphoma 1996;22 (Suppl. 1):41–45.

20. Lindstedt G, Lundberg PA. Are current methods ofmeasurement of erythropoietin (EPO) in human plasma orserum adequate for the diagnosis of polycythaemia veraand the assessment of EPO deficiency? Scand J Clin LabInvest 1998;58:441–458.

21. Teofili L, Martini M, Luongo M, Di Mario A, Leone

G, De Stefano V, Larocca LH. Overexpression of thepolycythemia rubra vera-1 gene in essential thrombocyt-hemia. J Clin Oncol 2002;20:4249–4254.

Johansson et al.

362