Experimental Hematology 32 (2004) 520–525

Human endogenous retrovirus(HERV-K) particles in megakaryocytes cultured fromessential thrombocythemia peripheral blood stem cells

Doris Morgan and Isadore BrodskyDrexel University College of Medicine, Department of Medicine, Division of Hematology and Oncology, Philadelphia, Pa., USA

(Received 12 December 2003; revised 1 March 2004; accepted 2 March 2004)

Objective. The aim of this study was to determine the extent of human endogenous retrovirus(HERV) gene translation in megakaryocytes cultured from peripheral blood stem cells ofpatients with essential thrombocythemia previously reported with platelet-associated HERVsequences and reverse transcriptase activity.

Materials and Methods. Terminally differentiated megakaryocytes derived from circulatingstem cells in serum-free medium supplemented with stem cell factor and thrombopoietin wereprocessed for electron microscopic immunostaining using a monoclonal antibody againstthe gag protein of HERV-K10 and an electron dense gold-labeled secondary antibody.

Results. We found that HERV-K gag protein was detected as clusters in the cytoplasm as wellas associated with viral particles budding from the cell membrane and into intracellularvacuoles in megakaryocytes from two patients with essential thrombocythemia. None of thesestructures was observed in megakaryocytes from a normal control or from a patient withchronic myelocytic leukemia.

Conclusion. This is the first evidence of HERV-K protein synthesis (gene translation) in humantissue other than seminomas, placenta, or fetal tissue. Translation of the HERV-K gag genewith subsequent packaging of the protein product into viral particles adds a new and importantdimension to future studies on the role of HERVs in the myeloproliferative diseases. � 2004International Society for Experimental Hematology. Published by Elsevier Inc.

The potential role of viruses in the etiology of human myelo-proliferative diseases is not a recent hypothesis. Electronmicroscopic detection of viruslike particles in platelets frompatients with myeloproliferative disorders was reported asearly as 1975 [1] but was not sufficient evidence to implicatea role for viruses in human disease. With the discovery ofendogenous retroviruses in the human genome as reportedby Martin et al. [2] in 1981, virologists focused on thesevestigial evolutionary remnants by characterizing the geneticand molecular properties of human endogenous retroviruses(HERVs) normally present in human teratocarcinoma celllines. These cell lines actively translate viral genes andrelease viral particles from the cell membrane [3]. Itwas assumed HERV genes were not expressed in human

Offprint requests to: Doris A. Morgan, Ph.D., Drexel University Collegeof Medicine, MS 412, 245 N.15th Street, Philadelphia, PA 19102;E-mail: doris.morgan@drexel.edu

Presented in part at the 32nd Annual Meeting of the International Societyfor Experimental Hematology, July 5–8, 2003, Paris, France.

0301-472X/04 $–see front matter. Copyright � 2004 International Society fordoi: 10.1016/j .exphem.2004.03.003

tissues other than placenta [4] and therefore were of littleor no significance to the well-being of humans. HERVswere considered “junk” DNA. With the evolution of molecu-lar technology came the methodology to detect HERV geneexpression by reverse transcription-polymerase chain reac-tion (RT-PCR) and more stringent assays for RT (reversetranscriptase), the pol gene product, and a renewed interestin HERV and human diseases. Of the 360 HERV-relatedpublications, 222 have appeared within the past 5 years [5].

Historically, not all researchers had dismissed the poten-tial role of retroviruses in human disease. During 1960 to1980, hematologists were focusing on hematologic diseasesin mice induced by Friend leukemia virus, Rauscher virus[6], and myeloproliferative sarcoma virus [7]. The similarit-ies between the virus-induced abnormal hematology of miceand myeloproliferative diseases in humans did not go unno-ticed. Due to the significance of platelets and megakary-ocytes in the oncornavirus-induced murine and felineleukemias, Brodsky et al. [1,8] focused on the platelets of

Experimental Hematology. Published by Elsevier Inc.

D. Morgan and I. Brodsky/Experimental Hematology 32 (2004) 520–525 521

patients with myeloproliferative diseases. In 1975, they re-ported RNA-dependent DNA polymerase (reverse tran-scriptase) and C-type viruslike particles budding into thevacuoles of platelets from patients with thrombocythemiaand from polycythemia vera patients in 1977. Their obser-vations were corroborated by Boyd et al. [9], who confirmedthat the RT activity in platelets of patients with myeloprolif-erative disease wasvirus particle associated andnot of cellularorigin. The detection of RT activity in cells from hematologicdiseases reflected the apparent activation of HERV pol gene.Subsequently, one hypothesis was that activation of HERVsin the human genome may play a role in the etiology ofpolycythemia vera, essential thrombocythemia (ET), andchronic myelogenous leukemia [10].

Recently numerous publications have reported HERVgene sequences in human autoimmune diseases [11], schizo-phrenia [12], melanoma [13], and breast cancer [14,15].Lower et al. [16] stated that RT-PCR methods are verysensitive and the significance of overexpressed RNA tran-scripts is not clear; however, the demonstration of proteinsynthesis would have biologic relevance [16,17]. We pre-viously reported the copurification of HERV-K RT tran-scripts and RT activity associated with particles purifiedon sucrose gradients from platelet lysates of three patientswith ET [18]. Transcript analyses for other HERV-Kgenes were not included. With the availability of a mono-clonal antibody against HERV-K gag protein [19], we as-sayed megakaryocytes cultured from peripheral blood stemcells from two of these patients by electron microscopy.In this report, we present evidence for the first time thatHERV-K activation is reflected not only in gene transcrip-tion but also in translation of at least one viral gene intoprotein product.

Materials and methods

PatientsThe clinical data from patients UPN07 and UPN09 have beenpreviously reported [18]. Prior to peripheral blood collections forthis report, patient and normal control consent was obtained ac-cording to the rules of Drexel University Internal Review Boardand the tenets of the Helsinki protocol.

Megakaryocyte culturesMononuclear cells from 150 mL heparinized peripheral blood frompatients and the normal control were obtained as previously de-scribed [20]. CD34� cells were acquired by magnetic column sepa-ration (Miltenyi Biotec Inc., Auburn, CA, USA) according to themanufacturer’s specifications and cultured in serum-free mediumcontaining 50 ng/mL stem cell factor and 20 ng/mL thrombopoietin(R&D Research, Minneapolis, MN, USA) for 12 days. The cellswere harvested and morphology examined on Wright-Giemsastained cytocentrifuge preparations [21].

Electron microscopyCells were fixed in a mixture of 0.1% glutaraldehyde/2% paraform-aldehyde with 0.05% saponin for 20 minutes, washed, and trans-ferred to blocking buffer for 15 minutes. After an overnight

incubation with the HERMA-1 anti-gag antibody, cells werewashed, fixed in 1% glutaraldehyde/2% paraformaldehyde for 30minutes, and processed by standard methodology for epon embed-ding. Thin sections of 80 nm were deposited on 200-mesh gold grids,etched with aqueous saturated sodium meta-per-iodate, andquenched with sodium borohydride [22]. After blocking with 0.2%cold water fish skin gelatin/1% ovalbumin in phosphate-bufferedsaline at pH 7.4, sections were incubated with 12-nm gold-tagged secondary antibody for 1 hour. Grids were washed in deion-ized water and stained in alcoholic uranyl acetate for 5 minutesfor contrast [23]. Images were taken using the JEOL Jem 1010equipped with an AMT 12-HR system and Hamamatsu CCDcamera.

Results

Cultured megakaryocytesWith new innovations in serum-free stem cell culturing andthe availability of thrombopoietin, relatively pure popula-tions of megakaryocytes are now available. As observed bylight microscopy, well-differentiated megakaryocytes fromET patients (Fig. 1A) and normal control (not shown) exhib-ited distinctive size and well-differentiated morphology after12 to 14 days of culture. Dysfunctional megakaryocyto-poiesis was apparent when stem cells from a patient withchronic myelocytic leukemia (CML) were induced withthrombopoietin. The megakaryocytes were smaller and hadnot achieved terminal maturation as evidenced by the agranu-lar blue cytoplasm (Fig. 1B). Preliminary examination bytransmission electron microscopy did not yield conclusiveevidence of HERV-K viruslike particles in any of the mega-karyocytes. This was due in part to the complexity of meg-akaryocyte ultrastructure with the characteristic membraneblebbing and numerous intracellular bodies resemblingviruses.

Electron microscopic immunolabelingOur resolve to continue the search for virus particles inmegakaryocytes cultured from stem cells was enhancedby the availability of HERMA-1, the monoclonal antibodyagainst the gag protein of HERV-K. Visual localization ofcellular HERV-K gag was achieved by the electron densegold “marker” of the gag-antigag complex. Single immuno-labeled particles are seen budding from the cell membraneof megakaryocytes (Figs. 2 and 3A) and were observed in5% to 10% of the cells. Cells were considered positive onlyif viral structures were present. The condensed core of theviral particles is consistent with C-type retroviral morphol-ogy. Unlike C-type viruses, however, location of the HERV-K particles was not restricted to the cell membrane butincluded vacuolar spaces (Fig. 3B), as described previouslyfor ET platelets [1]. This would not be unusual for megakar-yocytes considering the extensive demarcation membranesystem and invagination of the cell membrane in preparationfor proplatelet formation and eventual platelet release.

D. Morgan and I. Brodsky /Experimental Hematology 32 (2004) 520–525522

Figure 1. Megakaryocytes cultured for 12 days from peripheral blood stem cells of patients with (A) essential thrombocythemia (ET) or (B) chronicmyeloid leukemia (CML). Cytocentrifuged cells were stained with Wright-Giemsa and examined at 400× magnification.

“Empty” virus particles with no discernible cores were ob-served in extracellular and membrane-associated structures(Fig. 4A and B). Not all gag protein observed had beenpackaged into particles but was clustered in the cytoplasm(Fig. 4C–E). None of the structures in Figures 2 to 4 wasdetected in megakaryocytes from the CML patient or normalcontrol [24].

DiscussionEndogenous retroviral-like particles and viral proteins havebeen reported in placenta [4,25], fetal tissue [25,26], andseminomas [27], but not in cells from the diseases withoverexpression of HERV transcripts. Endogenous viral pro-tein synthesis in human non-germline tissue would reflecta level of virus activation that is biologically significant and

Figure 2. (A) Electron immunolabeling of HERV-K gag protein in a megakaryocyte cultured for 12 days from peripheral blood stem cells of a patientwith essential thrombocythemia. The sites of primary monoclonal HERV-K gag antibody localization on the virus particle are visualized by an electrondense gold-labeled secondary antibody. (B) Enhanced view and higher magnification of the budding viruslike particle.

D. Morgan and I. Brodsky/Experimental Hematology 32 (2004) 520–525 523

Figure 3. Cultured essential thrombocythemia megakaryocytes with immunolabeled HERV-K gag protein associated with a single virus particle buddingfrom the cell membrane (A) and into a vacuolar space (B).

would provide information regarding the biology of the virusas well as of the host cell. Apart from our laboratory[1,6,8,10], there have been relatively few reports focusingon the activation of endogenous retroviruses in hematologicdiseases [17,28], even though the notion of the viral etiologyof leukemia is not new [29,30]. In this report, we haveprovided evidence of HERV-K gene translation with thedetection of gag protein in well-defined virus particles in

megakaryocytes from patients with ET. There was no evi-dence of HERV-K protein in megakaryocytes from a normalcontrol or CML, even though we previously had reportedthe highest level of RT activity in CML platelets. Thereare two possible explanations for negative results in ourstudy: 1) viral particles are not present, or 2) if viral particlesare present, they are below the level of detection. These dataare preliminary and require a more comprehensive study

Figure 4. Essential thrombocythemia megakaryocytes with immunolabeled HERV-K gag protein associated with an extracellular viruslike particle (A), anempty cell membrane-bound particle (B), and clustered in the cytoplasm (C—E).

D. Morgan and I. Brodsky /Experimental Hematology 32 (2004) 520–525524

before one may consider whether HERV-K plays any role inthe initiation or persistence of ET or other myeloproliferativediseases. First of all, there are technical issues that mustbe resolved. Due to the low frequency of megakaryocytescontaining definitive HERV-K viral structures, RT-PCRwould be the method of choice for screening a cohort ofpatients and normal controls. Analysis of HERV-K gagand env RNA transcripts in extracts of mononuclear cells andplatelets from patients with myeloproliferative disorderswould provide the first level of subgrouping these patientsbased on diagnosis and presence/absence of geneticallyactive HERV-K. Because there is no evidence that HERV-K activity is restricted to the megakaryocytic lineage, exvivo expansion and RNA analyses of megakaryocytic, ery-throid, and myeloid lineages derived from circulating stemcells should be included in the screening protocol whenpossible. The second level of analysis would determinewhether HERV-K transcripts are translated into viral pro-teins, gag and env, as detected by immunoblotting. If proteinlevels are too low to be detected on Western blots, immuo-precipitation may be required to concentrate the protein.Other protocols for up-regulating virus expression may needto be used, e.g., induction with dexamethasone or 5-azacyti-dine. The third level of HERV-K activity would be the pack-aging of viral proteins and RNA into well-defined viralstructures that can only be visualized and identified as HERV-K by electron microscopic immunostaining. It is unlikelythat complete, infectious viral particles will be rescued fromthe platelets or blood cells from the myeloproliferative pa-tients. Viral particles from the teratocarcinoma cell linesappear to be incomplete [19], and none of the HERVs thus fardescribed produce infectious virus [31]. Attempts to rescueinfectious virus even in the absence of visualized viral parti-cles may be worth the effort because cocultivation with apermissive cell may up-regulate or amplify virus production.Two candidate cell lines come to mind: a human megakaryo-blastic cell line HU-3 established in this laboratory [32] andHEK, human embryonic kidney, known to be permissivefor porcine ERV [33].

Even though the data presented in this report were gener-ated from few patients, these observations provide the ratio-nale for hematologists to reexamine the potential impact ofendogenous retrovirus activation in hematologic disorders.It is premature to reflect on the potential role of viralproteins in the disease process, but this study provides credi-ble evidence that endogenous retroviruses are not merelypassengers in megakaryocytes of ET patients.

AcknowledgmentsThis research was supported by the I. Brodsky Institute for Cancerand Blood Diseases. The authors express their appreciation toDr. Klaus Boller, Institut-Paul-Ehrlich, Langen, Germany, for theHERMA-1 antibody and to Neelima Shah and the University ofPennsylvania Biomedical Imaging Core Facility for excellence inelectron microscopy.

References1. Brodsky I, Fuscaldo A, Erlick B, Kingsbury E, Schultz G, Fuscaldo

K. Analysis of platelets from patients with thrombocythemia for reversetranscriptase and virus-like particles. J Natl Cancer Inst. 1975;55:1069–1074.

2. Martin MA, Bryan T, Rasheed S, Khan A. Identification and cloningof endogenous retroviral sequences present in human DNA. Proc NatlAcad Sci U S A. 1981;78:4892–4896.

3. Boller K, Konig H, Sauter M, et al. Evidence that HERV-K is theendogenous retrovirus sequence that codes for the human teratocarci-noma-derived retrovirus HTDV. Virology. 1993;196:349–353.

4. Simpson GR, Patience C, Lower R, et al. Endogenous D-type (HERV-K) related sequences are packaged into retroviral particles in theplacenta and possess open reading frames for reverse transcriptase.Virology. 1996;222:451–456.

5. MEDLINE search performed in March 2004, using text wordHERV.

6. Brodsky I. Role of the megakaryocyte and platelet in the leukemicprocess in mice and men: a review and hypothesis. J Natl Cancer Inst.1973;51:329–335.

7. Le Bousse-Kerdiles MC, Jasmin C, Smadja-Joffe F, Klein B, Prag-nell I, Ostertag W. Myeloproliferative sarcoma virus (MPSV), a newtool for the study of hematopoiesis. Blood Cells. 1981;7:105–123.

8. Brodsky I, Fuscaldo A, Erlick B, Fuscaldo K. Effect of busulfan ononcornavirus-like activity in platelets and chromosomes in polycythe-mia vera and essential thrombocythemia. J Natl Cancer Inst. 1977;59:61–67.

9. Boyd MT, Maclean N, Oscier DG. Detection of retrovirus in patientswith myeloproliferative disease. Lancet. 1989;1:814–817.

10. Brodsky I, Foley B, Haines D, Johnston J, Cuddy K, Gillespie D. Ex-pression of HERV-K proviruses in human leukocytes. Blood. 1993;81:2369–2374.

11. Perron H, Seigneurin J. Review: human retroviral sequences associatedwith extracellular particles in autoimmune diseases: epiphenomenon orpossible role in aetiopathogenesis? Microbes Infect. 1999;1:309–322.

12. Karlsson H, Bachmann S, Schroder J, McArthur J, Torrey E, YolkenR. Retroviral RNA identified in the cerebrospinal fluids and brains ofindividuals with schizophrenia. Proc Natl Acad Sci U S A. 2001;98:4634–4639.

13. Schiavetti F, Thonnard J, Colau D, Boon T, Coulie P. A human endoge-nous retroviral sequence encoding an antigen recognized on melanomaby cytolytic T lymphocytes. Cancer Res. 2002;62:5510–5516.

14. Liu B, Wang Y, Melana S, et al. Identification of a proviral structurein human breast cancer. Cancer Res. 2001;61:1754–1759.

15. Wang-Johanning F, Frost A, Johanning G, et al. Expression of humanendogenous retrovirus K envelope transcripts in human breast cancer.Clin Cancer Res. 2001;7:1553–1560.

16. Lower R, Lower J, Kurth R. The viruses in all of us: Characteristicsand biological significance of human endogenous retrovirus sequences.Proc Natl Acad Sci U S A. 1996;93:5177–5186.

17. Depil S, Roche C, Dussart P, Prin L. Expression of a human endogenousretrovirus, HERV-K , in the blood cells of leukemia patients. Leukemia.2002;16:254–259.

18. Boyd MT, Foley B, Brodsky I. Evidence for copurification of HERV-K-related transcripts and a reverse transcriptase activity in humanplatelets from patients with essential thrombocythemia. Blood. 1997;90:4022–4030.

19. Bieda K, Hoffmann D, Boller K. Phenotypic heterogeneity of humanendogenous retrovirus particles produced by teratocarcinoma cell lines.J Gen Virol. 2001;82:591–596.

20. Morgan DA, Gumucio DL, Brodsky I. Granulocyte-macrophagecolony-stimulating factor-dependent growth and erythropoietin-in-duced differentiation of a human cell line MB-02. Blood. 1991;78:2860–2871.

D. Morgan and I. Brodsky/Experimental Hematology 32 (2004) 520–525 525

21 Lower R, Lower J, Frank H, Harzmann R, Kurth R. Human teratocarci-nomas cultured in vitro produce unique retrovirus-like viruses. J. GenVirol. 1984;65:887–898.

22. Skepper JN. Immunocytochemical strategies for electron microscopy:choice or compromise. J Microsc. 2000;199:1–36.

23. Smith RM, Charron MJ, Shah N, Lodish HF, Jarett L. Immunoelectronmicroscopic demonstration of insulin-stimulated translocation of glu-cose transporters to the plasma membrane of isolated rat adipocytesand masking of the carboxyl-terminal epitope of intracellular GLUT4.Proc Natl Acad Sci U S A. 1991;88:6893–6897.

24. Brodsky I, Foley B, Gillespie D. Expression of human endogenousretrovirus (HERV-K) in chronic myeloid leukemia. Leuk Lymphoma.1993;11(suppl 1):119–123.

25. Kjellman C, Sjogren H, Salford L, Widegren B. HERV-F (XA34) isa full-length human endogenous retrovirus expressed in placental andfetal tissues. Gene. 1999;239:99–107.

26. Andersson A-C, Venables P, Tonjes R, Scherer J, Eriksson L, Lars-son E. Developmental expression of HERV-R (ERV3) and HERV-Kin human tissue. Virology. 2002;297:220–225.

27. Sauter M, Schommer S, Kremmer E, et al. Human endogenous retrovi-rus K10: Expression of gag protein and detection of antibodies inpatients with seminomas. J Virol. 1995;69:414–421.

28. Orenstein J. Human endogenous retroviral expression by leukocytes.Ultrastruct Pathol. 2000;24:123–124.

29. Sarngadharan M, Sarin P, Reitz M, Gallo R. Reverse transcriptaseactivity of human acute leukemia cells: purification of the enzyme,response to AMV, 70S RNA, and characterization of the DNA product.Nat New Biol. 1972;240:67–72.

30. Viola M, Frazier M, Wiernik P, McCredie K, Spiegelman S. Reversetranscriptase in leukocytes of leukemic patients in remission. N EnglJ Med. 1976;294:75–80.

31. Mayer J. Status of HERV in human cells: expression and coding capac-ity of human proviruses. Dev Biol (Basel). 2001;106:439–441.

32. Morgan DA, Class R, Soslau G, Brodsky I. Cytokine-mediated ery-throid maturation in megakaryoblastic human cell line HU-3. ExpHematol. 1997;25:1378–1385.

33. Ritzhaupt A, Van Der Laan L, Salomon D, Wilson C. Porcine endoge-nous retrovirus infects but does not replicate in nonhuman primate pri-mary cells and cell lines. J Virol. 2002;76:11312–11320.