In vitro EBV-infected subline of KMH2, derived from Hodgkin lymphoma, expressesonly EBNA-1, while CD40 ligand and IL-4 induceLMP-1 but not EBNA-2Lorand L. Kis1*, Jun Nishikawa1, Miki Takahara1, Noemi Nagy1, Liudmila Matskova1, Kenzo Takada2,P. Goran Elmberger3, Ann Ohlsson3, George Klein1 and Eva Klein1

1Microbiology and Tumor Biology Center, Karolinska Institutet, Stockholm, Sweden2Department of Tumor Virology, Institute for Genetic Medicine, Hokkaido University, Sapporo, Japan3Department of Pathology and Cytopathology, Karolinska University Hospital, Stockholm, Sweden

In about 50% of classical Hodgkin lymphomas, the Hodgkin/ReedSternberg (H/RS) cells carry Epstein-Barr virus (EBV). The viralgene expression in these cells is restricted to EBNA-1, EBERs,LMP-1 and LMP-2 (type II latency). The origin of H/RS cells wasdefined as crippled germinal center B cells that escaped apoptosis.In spite of numerous attempts, only few typical Hodgkin lym-phoma (HL) lines have been established. This suggests that thecells require survival factors that they receive in the in vivomicroenvironment. If EBV is expected to drive the cells for growthin culture, the absence of EBNA-2 may explain the incapacity ofH/RS cells for in vitro proliferation. In EBV carrying B lympho-cytes, functional EBNA-2 and LMP-1 proteins are required for invitro growth. For analysis of the interaction between EBV and theH/RS cells, we infected the CD21-positive HL line KMH2 with theB958 and Akata viral strains. Only EBNA-1 expression was de-tected in a few cells in spite of the fact that all cells could beinfected. Using a neomycin-resistance-tagged recombinant EBVstrain (Akata-Neo) we established an EBV-positive subline thatwas carried on selective medium. In contrast to the type II EBVexpression pattern of H/RS cells in vivo, the KMH2 EBV cells didnot express LMP-1. The EBV expression pattern could be modi-fied in this type I subline. LMP-1 could be induced by the histonedeacetylase inhibitors TSA and n-butyrate, by 5-AzaC, a demethy-lating agent, and by phorbol ester. None of these treatmentsinduced EBNA-2. Importantly, exposure to CD40 ligand and IL-4induced LMP-1 without EBNA-2 expression and lytic replication.The KMH2 EBV cells expressed LMP-2A, but not LMP-2BmRNAs. This result is highly relevant for the type II expressionpattern of H/RS cells in vivo, since these stimuli can be provided bythe surrounding activated T lymphocytes.© 2004 Wiley-Liss, Inc.

Key words: Epstein-Barr virus; Hodgkin lymphoma; latent membraneprotein 1

Classical Hodgkin lymphoma (HL) tissue contains the charac-teristic Hodgkin/Reed-Sternberg (H/RS) cells, about 1% in thepopulation of nonneoplastic inflammatory cells.1 The majority ofH/RS cells are of B-cell origin, as shown by their immunoglobulin(Ig) gene rearrangements.1,2 They derive from preapoptotic germi-nal-center B cells with nonfunctional Ig rearrangements.3 Severalmechanisms have been proposed that enable these cells to evadeapoptosis. Constitutive NF-�B activation is considered to be crit-ical.4 Reciprocal effects between the H/RS cells and the inflam-matory cells in the tissue through cytokines and adhesion mole-cules and autocrine signaling seem to be important for the survivalof these cells.4

The Epstein-Barr virus (EBV)-encoded genes expressed in non-virus-producing cells were defined in the virus-transformed lym-phoblastoid cell lines (LCLs). In such cells, 1 of the 2 alternativeviral promoters (Wp or Cp) is used to generate a giant message, outof which 6 nuclear protein-encoding mRNAs (EBNA1-6) arespliced. In addition, 3 membrane proteins, LMP-1, -2A and -2B,are expressed.5 Among the nuclear proteins, EBNA-2, -3, -5 and -6are required for transformation. In B cells, EBNA-2 activates theLMP-1, LMP-2 and the Cp viral promoters and induces the ex-pression of CD21, CD23, c-myc and c-fgr. LMP-1 imposes phe-notypic changes and has a key role in the transformation. ThisEBV latency pattern is referred to as type III.

The frequency of EBV-positive HL cases varies in differentgeographical areas, between 40% and 100%.6 Defined by in situanalysis, the EBV expression pattern in the HR/S cells is type II;they express Epstein-Barr virus early RNA (EBER), EBNA-1 andLMPs.7,8 LMP-1 and LMP-2A are believed to provide survivalsignals to B lymphocytes by mimicking the externally activatedCD40 and BCR pathways, respectively.9 In the EBV-positivecases, expression of these viral proteins could be responsible forthe survival of surface Ig-negative germinal-center B cells, theputative precursors of H/RS cells.2

The type II latency pattern was first seen in EBV carryingnasopharyngeal carcinoma (NPC).10 In addition to HL, it occursalso in T and NK lymphomas. In Burkitt lymphomas (BLs) and inthe EBV-carrying memory B cells of healthy individuals, only oneviral protein, EBNA-1, is expressed. This pattern is designatedtype I latency, or EBNA-1-only program. EBNA-1 is responsiblefor the maintenance of the virus genome in the latent virus-cellinteraction.

In adolescents and adults, primary EBV infection often causesinfectious mononucleosis (IM).11 Whether or not the infectioncauses specific symptoms, it is followed by a lifelong virus carrierstate. The virus genome persists in the memory B-cell compart-ment.

An association between EBV and HL is suggested by theelevated frequency of HL in individuals whose primary infectionwas accompanied by IM. In such EBV-positive HL cases, theEBV-encoded genes LMP-1 and LMP-2A are thought to preventapoptosis. Type II latency does not seem to impose cell prolifer-ation. Among the immunosuppression-associated EBV-positivelymphoproliferations, HL is relatively rare. The EBV-driven B-immunoblastomas (diffuse large B-cell lymphomas, or DLBCLs)express the proliferation-inducing and immunogenic virally en-coded proteins. Therefore, the immunodeficiency state allows sur-vival and expansion in the affected patients. On the other hand, itcan be envisaged that H/RS-like cells do not meet their require-ments from the inflammatory microenvironment that seem to con-tribute to the HL pathophysiology. The accumulated knowledge

Abbreviations: 5-AzaC, 5-azacytidine; BL, Burkitt lymphoma; DLBCL,diffuse large B-cell lymphoma; EBNA, EBV-encoded nuclear antigen;EBV, Epstein-Barr virus; GC, germinal center; HL, Hodgkin lymphoma;H/RS, Hodgkin/Reed Sternberg; IM, infectious mononucleosis; LMP, la-tent membrane protein; n-but, sodium-butyrate; NPC, nasopharyngeal car-cinoma; PMA, phorbol-12-myristate-13-acetate; TSA, trichostatin A.

Grant sponsor: the Swedish Cancer Society, Sweden; Grant sponsor:Cancer Research Fellowship of the Cancer Research Institute/ConcernFoundation.

Dr. Nishikawa’s current address is: Department of Gastroenterology andHepatology, Yamaguchi University, School of Medicine, Yamaguchi, Ja-pan.

*Correspondence to: Microbiology and Tumor Biology Center, Karo-linska Institute, S-171 77 Stockholm, Sweden. Fax: �46-8-33-04-98.E-mail: kis.lorand.levente@mtc.ki.se

Received 9 March 2004; Accepted after revision 20 July 2004DOI 10.1002/ijc.20654Published online 28 October 2004 in Wiley InterScience (www.

interscience.wiley.com).

Int. J. Cancer: 113, 937–945 (2005)© 2004 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

about the functions of the EBV-encoded genes in B lymphocytesmotivated us to investigate the interaction of the virus with H/RScells, which lack several phenotypic traits of B cells.

Material and methodsCell lines

We received the HL cell lines L428, L1236, L540, HDLM-2 andKMH2 from Magnus Bjorkholm (Department of Medicine, Divi-sion of Hematology, Radiumhemmet, Karolinska Hospital, Karo-linska Institutet, Stockholm, Sweden). Their characteristics areknown from previous publications.12–14 They were all EBV-neg-ative. The EBV-positive type I BL lines Akata and Rael, the typeIII Raji, the EBNA-2-deficient P3HR1 and Daudi, Akata LCL andCBMI LCL were used as controls in various assays throughout thestudy. The Akata LCL was made in our laboratory by infectingpurified tonsillar B cells with Akata-Neo and was 4-month-oldduring the time of study. They were grown in RPMI mediumsupplemented with 10% inactivated fetal calf serum (FCS) and 100U/ml penicillin and 100 �g/ml streptomycin. CD40L-expressingmouse fibrosarcoma L cells (CD40L-L) and the control L cellswere obtained from Pierre Garrone (Schering-Plough, Dardilly,France) and were maintained in IMDM medium.

EBV production and infectionThe following EBV strains were used: B95-8, Akata-Neo from

Akata cells and Akata-Neo-GFP from AGS (gastric carcinoma)cell line. For virus production and infection, we followed previouspublications.15,16 The Akata-Neo-GFP virus was received fromL.M. Hutt-Fletcher (Department of Microbiology and Immunol-ogy, LSU Health Sciences Center, Shreveport) through VictorLevitsky (MTC, Karolinska Institutet, Stockholm, Sweden).

To establish the EBV-positive subline, KMH2 cells were incu-bated for 1 hr with Akata-Neo viral supernatant, washed andcloned at 0.5 cell/well in round bottom 96-well-plate in 200 �lmedium containing 1 mg/ml active G418 (Calbiochem-Novabio-chem, San Diego, CA; 5 plates in total). Three weeks after cloning,5 wells contained growing cells and 1 clone could be culturedcontinuously.

Flow cytometry and immunoblottingStaining for FACS analysis was done according to standard

procedures. The surface molecules were stained by indirect immu-nofluorescence with the following mouse monoclonal antibodies:F18 (anti-CD21; Dako, Copenhagen, Denmark), J5 (anti-CD10;gift from L. Nadler, Dana Farber Cancer Institute, Boston, MA),MHM-6 (anti-CD23; gift from M. Rowe, University of Wales,Cardiff, Wales) and L243 (anti-HLA-DR). RPE-conjugated rabbitantimouse Ig F(ab)2 (Dako) was used as secondary antibody. Forthe detection of proteins by immunoblotting, the following anti-bodies were used: S12 supernatant (anti-LMP-1), PE-2 (anti-EBNA-2; Dako), OT1x (anti-EBNA-1; gift from J.M. Middel-dorp), the mixture of 7.1 and 13.1 (anti-GFP; Roche MolecularBiochemicals, Basel, Switzerland), AC-15 (anti-�-actin; Sigma,St. Louis, MO), clone 124 (anti-Bcl-2; Dako) and rabbit anti-SH2D1A serum (received from J. Sumegi, Nebraska Health Cen-ter, Omaha, NE). For detection of LMP-2A, the blots were incu-bated with the rat anti-LMP-2A mAb 14B7 (2 �g/ml; ITN,Neuherberg, Germany) and HRP-conjugated goat antirat IgG(H�L; diluted 1:3,000; Zymed, San Francisco, CA).

Detection of EBV and EBV-encoded genesEBNA was detected by anticomplement immunofluorescence

(ACIF) using the serum of a previously characterized seropositivehuman donor (K.F.).17 For EBER visualization, 2 in situ hybrid-ization (ISH) methods were employed. In the first approach, cellswere pelleted in the Shandon Cytocentrifuge 4 — and then oncewashed in phosphate-buffered saline (pH 7.4; S). The material wasprefixed in nonbuffered 4% formalin for 15 min. Cell blocks werethen prepared with the Cytoblock cell block preparation system

(Shandon). Cell blocks were postfixed in formalin for a totalfixation time of 24 hr. After conventional paraffin blocking, 5 �mthick sections were cut and mounted on Superfrost objectiveglasses. EBER detection was performed by ISH technique utilizingthe Inform kit (Ventana Medical Systems) on the automatic stain-ing platform BenchMark XT (Ventana Medical Systems, Tucson,AZ). EBER was visualized with the IVIEW assay and slides werefinally counterstained with the ISH red counterstain as part of theautostainer protocol. The recommendations provided by the man-ufacturer was followed in detail. Pictures of the slides were re-corded with digital microphotography using an Olympus BX 50microscope equipped with Olympus DP70 camera and finallycontrast was optimized through the Olympus DP-soft program.

EBER in situ hybridization coupled with FACS analysis wasdone with 5-carboxy-fluorescein-labeled EBER PNA probe (Da-kocytomation, Glostrup, Denmark) according to previously pub-lished protocol.18 LMP-1 immunofluorescence staining was doneas previously described.19 EBNA-2 staining followed the proce-dure for LMP-1 staining, with the difference that slides wereincubated with the anti-EBNA-2 PE-2 antibody (diluted 1:100).

For LMP-1 and EA double staining, the slides were first stained forLMP-1, visualized with 1:200 diluted Texas red-conjugated horseantimouse (Vector, Burlingame, CA) secondary antibody, then theslides were incubated with a previously characterized EA-reactiveantiserum conjugated with FITC (F6-esther; dilution 1:60).20

DNA extraction and DNA PCRGenomic DNA was extracted by cetyltrimethylammonium-

bromide (CTAB) method. PCR amplification of genomic frag-ments of EBER-1 and EBNA-2 was done in glass capillariesusing Rapid Cycler PCR (Idaho Technologies, Salt Lake City,UT). The reactions were carried out in 10 �l volume using 2mM MgCl2, 1 �M primers, 200 nM dNTP mixture and 0.1 Uplatinum Taq polymerase (Invitrogen, Carlsbad, CA). The ini-tial 60-sec denaturation at 94°C was followed by 35 cycles ofdenaturation at 94°C for 0 sec, annealing at 55°C for 0 sec andelongation at 72°C for 20 sec. The primer sequences wereEBER-1 sense, 5�-AGGACCTACGCTGCCCTAGA-3�;EBER-1 antisense, 5�-AAAACATGCGGACCACCAGC-3� (167bp); EBNA-2 sense, 5�-GGATGAAGACTAAGTCACAGGC-3�;EBNA-2 antisense, 5�-TGACGGGTTTCCAAGACTATCC-3� (615bp); �-actin sense, 5�-ATGGGTCAGAAGGATTCCTATGT-3�;�-actin antisense, 5�-TCAGGAGGAGCAATGATCTTGA-3� (862bp). The PCR products were fractionated on 2.0% agarose gels andvisualized with ethidium bromide.

Trichostatin A (TSA), sodium-butyrate (n-but), 5-azacytidine(5-azac) and phorbol-12-myristate-13-acetate (PMA) treatment

A total of 1 � 106 KMH2 EBV and Rael cells were cultured in3 ml medium containing 500 ng/ml TSA, 5 mM n-but, 10 �M5-AzaC, or 20 ng/ml PMA (all chemicals were obtained fromSigma). Forty-eight hours after treatment, the cells were harvestedand protein extracts and slides were prepared.

CD40L/IL-4 activationCD40L-transfected L cells or control L cells were irradiated

with 15,000 Rad and incubated overnight in 12-well plate (0.5 �106 cells/well). Thereafter, 0.5 � 106 KMH2 EBV and 2-day-infected KMH2 cells were plated on the CD40L-L and control Lcells in 3 ml FCS supplemented RPMI medium containing 50ng/ml IL-4 (PeproTech, London, U.K.).

cDNA synthesis and RT-PCRPoly(A) � RNA was isolated from 2 � 106 cells using oligo-dT

Dynabeads (Dynal, Oslo, Norway) according to the manufactur-er’s recommendation. cDNA was synthesized from poly(A) �RNA by oligo(dT12–18) priming using Superscript II (Invitrogen,La Jolla, CA). PCR reactions were carried out in glass capillarieswith Rapidcycler (Idaho Technology). The PCR was performed ina 10 �l volume [50 mM TRIS-HCl, pH 8.6, 50 �g/ml BSA

938 KIS ET AL.

(Sigma-Aldrich Sweden, Stockholm, Sweden), 3 mM MgCl2, 200�M of each dNTP, 0.4 U platinum Taq polymerase (Invitrogen)]using 1 �M of the appropriate primers. After an initial 15-secdenaturation at 94°C, the following cycles of 94°C for 0 sec, 55°C for0 sec and 72°C for 30 sec were repeated 35 times. The final extensionwas lengthened to 5 min. The PCR products were separated on a 2%agarose gel in 1 � TAE buffer in the presence of 10 �g/ml ethidiumbromide. The bands were visualized in an Intelligent Dark Box II(FujiFilm, Elmsford, NY). The sequences of primers used in our studyare (the expected size of the PCR products in parenthesis): LMP-2Asense, 5�-GGACGATGGCGGAAACAACTC-3�; LMP-2A anti-sense, 5�-TGAGGCCGTGAAACACGAGG-3� (376 bp); LMP-2Bsense, 5�-CAGGAAATGGAAAGGCAGTG-3�; LMP-2B antisense,

5�-GGCAGGCATACTGGATTCAT-3� (170 bp); EBNA-2 sense,�-AGAGGAGGTGGTAAGCGGTTC-3�; EBNA-2 antisense, 5�-TGACGGGTTTCCAAGACTATCC-3� (394 bp); GAPDH sense,5�-ACCACAGTCCATGCCATCAC-3�; GAPDH antisense, 5�-TCCACCACCCTGTTGCTGTA-3� (452 bp).

ResultsExpression of EBV-encoded genes in in vitro infected KMH2cells

Among the 5 classical HL-derived cell lines tested, only KMH2cells expressed the EBV receptor, CD21 (Fig. 1a). They all ex-pressed HLA-DR (Fig. 1a). When the KMH2 culture was infected

FIGURE 1 – (a) Expression of CD21 and HLA-DR in the cHD-derived cell lines assessed by FACS analysis stained by indirect immunoflu-orescence. Raji cells are included as positive control while Jurkat cells as negative control. The filled curves denote the fluorescence intensityof staining with specific primary antibody, while the black lines the background fluorescence (logarithmic amplification). (b) FACS analysis ofGFP expression of Akata-GFP-infected KMH2 cells at 24, 48 and 72 hr after infection. The filled curves represent the FL1 fluorescence intensityof the Akata EBV-infected KMH2 cells, while the black lines denote the FL1 fluorescence intensity of the uninfected cells (logarithmicamplification). The FACS curves shown are representative for 4 different experiments.

939IN VITRO EBV-INFECTED SUBLINE OF KMH2

with the B95-8 EBV strain, only 1% of the cells expressedEBNA-1 (not shown). After multiple reinfections, the fraction ofEBNA-1-positive cells increased to 10%. No EBNA-2 or LMP-1expressor cells were detected (not shown).

The question arose whether the low percentage of the cellsreflects the efficiency of viral infection. Therefore, we used alsothe GFP-tagged recombinant Akata EBV strain obtained from theepithelial cell line AGS, in which the GFP expression is drivenfrom a cytomegalovirus promoter and was reported to infect Bcells more efficiently.16 With this virus, all the cells were infectedas indicated by green fluorescence (Fig. 1b). Thus, the viral recep-tor was functional on the cells. After prolonged culture, the per-centage of GFP-positive cells gradually declined; we followed theculture up to 25 days and green cells were still detectable (notshown).

Protein extracts were prepared at different times after infection andanalyzed by Western blotting. They showed only low amounts ofEBNA-1 (Fig. 2). This result was in line with the visual examination,where only 5% of the cells were EBNA-1-positive 72 hr after infec-tion (not shown). Forty-eight hours after infection, 70% of the cellsexpressed EBERs, as visualized by ISH of paraffin-embedded sec-tions (Fig. 3a). At this time, all cells were GFP-positive according tothe result of FACS analysis. Thus, the viral DNA was present in thecells, but the EBNA genes were not expressed. The in vitro infectedHL-derived cells expressed the type I latency pattern, in contrast to thetype II latency pattern (EBNA-1, LMP) characteristic for EBV car-rying H/RS cells in vivo.

Establishment of an EBV-positive subline of KMH2In order to establish an EBV-converted subline of KMH-2, we

used the recombinant Akata EBV strain that carries a neomycin-resistance gene (Akata-Neo). Forty-eight hours after infection, theculture was cloned in medium containing G418. Three weeks later,colonies emerged in 5 wells and these were expanded. Only cellsderived from one colony could be kept in continuous culture. Thisyielded the KMH2 EBV subline. Analysis of the EBV-encodedgenes revealed that all cells expressed EBERs, as seen by ISH onparaffin sections or cells in suspension (Fig. 3). EBER expression

was considerably higher in the established subline than in the cells48 hr after infection with the GFP-Akata virus.

EBNA-1, but not EBNA-2 or LMP-1, was detected by Westernblotting. This gene expression pattern corresponds to that seen atthe early time point after infection. The line was stable during 1year of continuous culture in medium containing G418. Uponremoval of the selective medium, the virus was lost, the sublinebecame EBV-negative after 3 months. The EBV-positive and-negative cultures showed no difference in their immunoprofile;the expression levels of CD20, CD25, CD30, CD40, CD54(ICAM-1) and CD150 (SLAM) were identical as assessed byFACS analysis (not shown).

Induction of LMP-1 in KMH2 EBV cellsLMP-1 expression was induced in KMH2 EBV cells by histone

deacetylase inhibitors (TSA and n-but), by demethylation (5-AzaC) and by phorbol ester (PMA; Fig. 4a). The highest LMP-1level was seen in the cultures treated with n-but or PMA. Thesetreatments are known to induce the expression of EBNA-2 and theother high molecular EBNAs in type I BL lines.21 They also inducethe lytic cycle.20,21 In the KMH2 EBV cells, these treatments didnot induce EBNA-2 expression, however (Fig. 4a).

Double immunofluorescence staining for early lytic cycle anti-gens (EAs) and LMP-1 showed lytic replication only in the phor-bol ester-treated cultures. The percentage of LMP-1-positive cellswas 50%, but only 5% of these cells expressed EA (not shown).Lack of EBNA-2 expression was not due to the deletion of theDNA coding region of EBNA-2 during conversion of the KMH2cells. The DNA sequences of a part of EBNA-2 coding regionwere detected in the KMH2 EBV cells (Fig. 4b). DNA PCR withEBER-1-specific primers confirmed that EBV was present in theEBV-positive subline, but not in the uninfected KMH2 cells.

Induction of LMP-1 in KMH2 EBV cells by physiologic stimuliSince CD4-positive T cells with a predominantly Th-2 pheno-

type that express CD40L closely surround the H/RS cells in thetumor tissues,22,23 we posed the question whether these T cellsmay provide signals for the induction and/or maintenance ofLMP-1 expression. Therefore, we exposed the KMH2 EBV cellsto CD40 ligand and IL-4 and found that this treatment induced theexpression of LMP-1 (Fig. 5). No EBNA-2 protein and no lyticreplication were detected, and the EBNA-1 expression level didnot change. LMP-1 expression disappeared 48–72 hr after theremoval of CD40L and IL-4 (not shown).

Similarly, LMP-1 was induced in KMH-2 cultures freshly infectedwith Akata-GFP virus (not shown). The expression of Bcl-2, anantiapoptotic protein that was shown to be induced by LMP-1, wasnot changed irrespective of LMP-1 expression of the KMH2 EBVcells. We have previously found that HL-derived cell lines expressSH2D1A (SAP)14 and LMP-1 downregulates the expression ofSH2D1A in BL lines.24 Interestingly, SH2D1A expression was notaffected when LMP-1 was induced by CD40L and IL-4.

Expression of LMP-2 in KMH2 EBV cellsSince LMP-2 is an integral part of type II EBV latency and

EBV-infected H/RS cells express LMP-2A in vivo, we tested theexpression of LMP-2A protein by immunoblotting.8 In all 4 ex-periments performed, the 14B7 antibody reacted with a protein ofalmost identical molecular weight as LMP-2A (54 kDa) in theEBV-negative parental KMH2 cells (Fig. 6a). The same band,though more intense, could be seen in the KMH2 EBV sample.The band was absent from the total cell extracts of Raji. Forverification of the reactivity seen at the protein level, we examinedthe expression of LMP-2A and LMP-2B mRNAs by RT-PCR. TheLMP-2A and LMP-2B mRNAs are transcribed from 2 differentpromoters and the first exons are unique to each of them, thereforeprimer pairs can be designed to amplify LMP-2A or LMP-2Bspecifically. The KMH2 EBV cells expressed the LMP-2A, but notLMP-2B, transcripts (Fig. 6b). In accordance with the Westernblotting results, EBNA-2 message was not detected in the KMH2

FIGURE 2 – Expression of EBNA-1, GFP, LMP-1, EBNA-2 and�-actin proteins in the GFP-Akata-infected KMH2 cells. Total celllysates prepared at different time points after infection were examinedby SDS-PAGE and immunoblotting with specific monoclonal antibod-ies. The 200–220 kDa band recognized by the anti-EBNA-1 antibodyis nonspecific. Raji cells are shown as a positive control. Similarresults were obtained in 3 separate experiments.

940 KIS ET AL.

EBV cells. The control LCLs (CBMI LCL and Akata LCL) ex-pressed both LMP-2A and EBNA-2 mRNAs.

We concluded therefore that a protein of the same size as LMP-2Ais expressed in the KMH2 cells. The results with the LMP-2Amessage substantiate that the immunoblots showed LMP-2A proteinexpression in the EBV-carrying KMH2 subline, though its detectionis masked by the cross-reactivity of the antibody.

Discussion

During the long history of EBV-human coexistence, evolu-tion of the virus and of the host lead to a harmonious equilib-

rium. The virus carrier state can be easily revealed by EBV-specific antibodies in the serum, EBV-specific cellularimmunity and a low number of latently infected lymphocytes inthe B-lymphocyte population.5 In healthy individuals, persis-tence of the virus in face of the EBV-specific immunity and theavoidance of EBV-induced B-cell proliferation are due to thevariable cell-virus interactions. In the healthy individuals, vi-rus-carrying B lymphocytes express only one protein, EBNA-1,that does not drive the cells for proliferation and does notdecorate the cells with immunogenic peptides.5 Thus, they arenot detected by the immune system. This viral expressionpattern is designated as type I, EBNA-1 only.

FIGURE 3 – (a) EBER expression of uninfected KMH2 (a), GFP-Akata infected KMH2, 48 hr after infection (b) and KMH2 EBV subline (c)assessed by ISH on paraffin-embedded sections. EBER was visualized with the IVIEW assay and slides were finally counterstained with the ISHred counterstain. (b) EBER expression of KMH2 EBV cells assessed by ISH coupled with FACS analysis. The black lines represent the FL1fluorescence intensity after hybridization with the EBER FITC probe. Gray lines represent the FL1 fluorescence intensity of the unstained cells(logarithmic amplification). Daudi cells are included as positive control.

941IN VITRO EBV-INFECTED SUBLINE OF KMH2

EBV exploits a differentiation marker of B-lymphocytes CD21,the receptor for the complement component C3d, as its receptor.B-lymphocyte populations can be infected with high efficiency invitro. Infection induces B-cell activation and proliferation, whichcan be inhibited by NK and T cells. Therefore, the virus carrierstate in healthy individuals is harmless. The transforming potentialof the virus is controlled in vivo by the immune response. Theimmune system recognizes EBV-encoded antigens and costimula-tory molecules expressed in the transformed cells. Therefore, onlyin conditions with severe immunosuppression can EBV-infected Blymphocytes expand to malignancy.

Technical improvements led to the detection of EBV in a varietyof malignancies of hematopoietic and epithelial origin. Except forthe immunoblastomas (DLBCLs) in immunosuppressed patients,the role of EBV in the etiology of the tumors is not known. Weattribute significance to the fact that there are no B-cell lines witha type II latency. This suggests that the EBV genes expressed intype II latency do not impose in vitro proliferation without further

contributing factors. In spite of considerable efforts, only fewHL-derived cell lines were established and they are all EBV-negative except one, L591. This line contained initially more than60% EBNA-negative cells.25 The presently used line became pos-sibly superinfected in vitro by contaminating EBV-producing lym-phoblastoid cells. In any case, its EBV expression pattern is typeIII26 and therefore it is not representative for an EBV-positive HL.It cannot be excluded that similarly to the in vitro occurring typeI-type III latency shift seen in some BL lines, the L591 linechanged during prolonged culture.

The results with the KMH2 cells substantiate the notion that theEBV expression pattern is determined by the maturation/differen-tiation window of B lymphocytes. The GFP expression proved thatall cells of the KMH2 culture were infected with the virus. Im-portantly, practically all cells were EBER-positive, but they ex-pressed only EBNA-1.

The type II pattern of EBV-positive H/RS cells is explainableby the known molecular interactions involved in the EBV-encoded gene expression. In type II latent cells, promoters forEBNA-1 (Qp) and the LMPs are active, while Wp and Cp thatgenerate the giant message of the EBNA-2-6 are silent. Theinactivity of Wp/Cp promoter could be due to the lack ofB-cell-specific transcriptional factors in the H/RS cells. Theyexpress low levels of BSAP and they lack Oct-2.27–29 These 2transcription factors bind to and activate the Wp and the Cp,respectively.30,31 Although further experiments are needed toprovide direct evidence, it is tempting to speculate that the lackof BSAP and Oct-2 might be the reason why type III latency

FIGURE 4 – (a) EBNA-1, LMP-1, EBNA-2 and actin expression ofKMH2 EBV cells treated with TSA (500 ng/ml), n-but (5 mM),5-AzaC (10 �M), 5-AzaC/TSA, PMA (20 ng/ml) for 48 hr. Rael cellstreated with 5-AzaC (10 �M) for 48 hr are shown as control for theeffect of 5-AzaC. Scanned images of immunoblots are shown obtainedfrom 1 representative experiment out of 3. (b) DNA PCR analysis ofB95-8, P3HR-1 (EBNA-2-deficient), KMH2, KMH2 EBV and Akatacells with specific primers for EBNA-2, EBER-1 and �-actin genes.

FIGURE 5 – Expression of EBNA-1, LMP-1, �-actin, Bcl-2 andSH2D1A proteins in the KMH2 EBV total cell extracts prepared at 24,48 and 72 hr after exposure to CD40L and IL-4. Total cell extractsfrom one experiment were examined by SDS-PAGE and immunoblot-ting. Scanned images of immunoblots obtained examining the sameprotein extracts are shown (representative for 4 experiments).

942 KIS ET AL.

with EBNA-2-6 expression is not induced in the KMH2 EBVcells. It is likely that the same condition determines the silenceof Wp/Cp in other EBV-carrying lymphomas with a postger-minal center (post-GC) phenotype, such as cells in post-transplant lymphoproliferative disease (PTLD),32,33 the rareEBV-positive plasmacytomas,34 pleural effusion lymphomas(body cavity lymphomas)35,36 and plasmablastic lymphomas ofthe oral cavity.37

Requirement for the B-cell phenotype for type III EBV latencywas first shown in experiments with somatic cell hybrids, sinceonly those hybrids that maintained the B-cell phenotype, providedby one of the partners, maintained the type III latency.38 Furtherevidence for the finely tuned requirement for phenotype- anddifferentiation-related expression of the EBV-encoded genes canalso be seen in such LCLs that can be induced for plasma celldifferentiation by IL-6. These cells concomitantly downregulateEBNA-1, EBNA-2 and LMP-1.39 Furthermore, normally a smallfraction of cells in LCLs express cytoplasmic Ig, lack EBNA-2 andcease to proliferate.40

While the EBV expression type of the KMH2 EBV sublinedid not correspond to that of in vivo H/RS cells, LMP-1 ex-pression was induced by exposure to IL-4 and CD40 ligand bothin the EBV-positive established line and in freshly infectedcells. These results substantiate the notion that the inflammatorycells in the affected HL tissue contribute to the fate of H/RScells. In addition to other factors, for the EBV-positive H/RScells, the infiltrating CD4-positive T cells that express CD40ligand probably contribute to the expression of LMP-1. LMP-1and LMP-2A have already been proposed to contribute to thesurvival of sIg-negative GC B cells. Our results suggest that atleast at certain point during the transformation process of H/RSprecursors, external stimuli could induce LMP-1 and by thiscontribute to the survival of preapoptotic GC B cells. Similarly,T-cell-derived signals might contribute to the induction ofLMP-1 in B cells of PTLDs,41 angioimmunoblastic T-cell lym-phomas42 and MALTomas.43

Signals from CD40 and LMP-1 and activated NF-�B are knownto inhibit the entry into lytic cycle.44,45 Since NF-�B is constitu-tively activated in H/RS cells and in the HL-derived cell lines aswell,46 this may explain why lytic replication was not induced inthe KMH2 EBV cells.

The detection of H/RS-like cells in IM tonsils and lymph nodesmight be highly significant for the etiology of EBV-positiveHL.47,48 The risk for HL is elevated in individuals whose primaryEBV infection lead to clinical symptoms. Importantly, the LMP-1-positive, EBNA-2-negative cells were localized to the T-cell-rich interfollicular areas, where T-cell-derived signals could con-tribute to the induction of LMP-1. Interestingly, the EBV-positivecells seen in the germinal centers expressed a different, unusuallatency type, being EBNA-2-positive and LMP-1-negative.48 Thislatency type is regularly exhibited by the in vitro infected Bchronic lymphocytic leukemia cells.49

We thus showed that, one, the viral receptor-positive “crip-pled” B-cell HL line KMH2 can be infected, but it does notmaintain EBV unless the virus episome is required for cellsurvival (when the drug-resistance gene is needed). Two, theinfected cells express EBNA-1 that is instrumental for mainte-nance of the viral episome. Three, EBNA-2 could not be in-duced by any manipulation. Four, LMP-1 expression could beinduced in the KHM2 EBV cells by providing external stimuli.These stimuli can be provided in vivo by the inflammatory cellpopulation that surrounds the H/RS cells. Through our findingwe provide a direct proof for the factors that contribute to thephenotype of the EBV-carrying H/RS cells.

Acknowledgements

The authors thank Victor Levitsky and Andre Ortlieb Cacais(Microbiology and Tumor Biology Center, Karolinska Institutet)for the viral preparations and helpful discussions, and Britt-MarieLarsson for skilful technical help with the EBER ISH. Dr. J.M.Middeldorp (Department of Pathology, Vrije Universiteit MedicalCenter, Amsterdam, The Netherlands) is acknowledged for pro-viding the OT1x antibody. L.L.K. and N.N. are recipients of theCancer Research Fellowship of Cancer Research Institute/ConcernFoundation.

FIGURE 6 – (a) Scanned images of immunoblots probed withanti-LMP-2A-specific monoclonal antibody 14B7 of total cell ex-tracts prepared from CBMI LCL, KMH2, KMH2 EBV and Rajicells. Numbers on the right denote the molecular weight obtainedby running in a parallel well 10 �l PageRuler Prestained ProteinLadder (Fermentas, Vilnius, Lithuania). It is worth noting that Rajidoes not express LMP-2 because of a deletion in the gene’s codingregion. (b) RT-PCR results of LMP-2A, LMP-2B, EBNA-2 andGAPDH expression. PCR products were visualized in ethidiumbromide-stained agarose gels. The first lane contains 2 �l of Gene-Ruler 100 bp DNA ladder (Fermentas) with the 500 bp more intenseband. The expected size of PCR products is indicated on the right.The schematic diagram shows the transcription of the LMP-2 genes(open squares denote untranslated regions, hatched squares thecoding exons) and the location of PCR primers used for amplifi-cation of LMP-2A and LMP-2B genes.

943IN VITRO EBV-INFECTED SUBLINE OF KMH2

References

1. Kuppers R. Molecular biology of Hodgkin’s lymphoma. Adv CancerRes 2002;84:277–312.

2. Kuppers R. B cells under influence: transformation of B cells byEpstein-Barr virus. Nat Rev Immunol 2003;3:801–12.

3. Kanzler H, Kuppers R, Hansmann ML, Rajewsky K. Hodgkin andReed-Sternberg cells in Hodgkin’s disease represent the outgrowth ofa dominant tumor clone derived from (crippled) germinal center Bcells. J Exp Med 1996;184:1495–505.

4. Skinnider BF, Mak TW. The role of cytokines in classical Hodgkinlymphoma. Blood 2002;99:4283–97.

5. Middeldorp JM, Brink AA, van den Brule AJ, Meijer CJ. Pathogenicroles for Epstein-Barr virus (EBV) gene products in EBV-associatedproliferative disorders. Crit Rev Oncol Hematol 2003;45:1–36.

6. Herbst H. Epstein-Barr virus in Hodgkin’s disease. Semin Cancer Biol1996;7:183–9.

7. Pallesen G, Hamilton-Dutoit SJ, Rowe M, Young LS. Expression ofEpstein-Barr virus latent gene products in tumour cells of Hodgkin’sdisease. Lancet 1991;337:320–2.

8. Niedobitek G, Kremmer E, Herbst H, Whitehead L, Dawson CW,Niedobitek E, von Ostau C, Rooney N, Grasser FA, Young LS.Immunohistochemical detection of the Epstein-Barr virus-encodedlatent membrane protein 2A in Hodgkin’s disease and infectiousmononucleosis. Blood 1997;90:1664–72.

9. Thorley-Lawson DA. Epstein-Barr virus: exploiting the immune sys-tem. Nat Rev Immunol 2001;1:75–82.

10. Fahraeus R, Fu HL, Ernberg I, Finke J, Rowe M, Klein G, Falk K,Nilsson E, Yadav M, Busson P, Tursz T, Kallin B. Expression ofEpstein-Barr virus-encoded proteins in nasopharyngeal carcinoma. IntJ Cancer 1988;42:329–38.

11. Hjalgrim H, Askling J, Rostgaard K, Hamilton-Dutoit S, Frisch M,Zhang JS, Madsen M, Rosdahl N, Konradsen HB, Storm HH, MelbyeM. Characteristics of Hodgkin’s lymphoma after infectious mononu-cleosis. N Engl J Med 2003;349:1324–32.

12. Drexler HG. The leukemia-lymphoma cell lines facts book. SanDiego, CA: Academic Press, 2000.

13. Staratschek-Jox A, Wolf J, Diehl V. Hodgkin’s disease. In: MastersJRW, Palsson B, eds. Human cell culture. vol. 3. Lymphoma andleukemia cell lines. Dordrecht: Kluwer Academic Press, 2000. 339–53.

14. Kis LL, Nagy N, Klein G, Klein E. Expression of SH2D1A in fiveclassical Hodgkin’s disease-derived cell lines. Int J Cancer 2003;104:658–61.

15. Maeda A, Kiss C, Chen F, Ehlin-Henriksson B, Nagy N, Szekely L,Takada K, Klein E, Klein G. EBNA promoter usage in EBV-negativeBurkitt lymphoma cell lines converted with a neomycin-resistant EBVstrain. Int J Cancer 2001;93:714–9.

16. Borza CM, Hutt-Fletcher LM. Alternate replication in B cells andepithelial cells switches tropism of Epstein-Barr virus. Nat Med2002;8:594–9.

17. Reedman BM, Klein G. Cellular localization of an Epstein-Barr virus(EBV)-associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines. Int J Cancer 1973;11:499–520.

18. Just T, Burgwald H, Broe MK. Flow cytometric detection of EBV(EBER snRNA) using peptide nucleic acid probes. J Virol Methods1998;73:163–74.

19. Nishikawa J, Kis LL, Liu A, Zhang X, Takahara M, Bandobashi K,Kiss C, Nagy N, Okita K, Klein G, Klein E. Upregulation of LMP1expression by histone deacetylase inhibitors in an EBV carrying NPCcell line. Virus Genes 2004;28:121–8.

20. Di Renzo L, Avila-Carino J, Klein E. Induction of the lytic viral cyclein Epstein Barr virus carrying Burkitt lymphoma lines is accompaniedby increased expression of major histocompatibility complex mole-cules. Immunol Lett 1993;38:207–14.

21. Cuomo L, Trivedi P, de Campos-Lima PO, Zhang QJ, Ragnar E,Klein G, Masucci MG. Selective induction of allostimulatory capacityafter 5-azaC treatment of EBV carrying but not EBV negative Burkittlymphoma cell lines. Mol Immunol 1993;30:441–50.

22. Carbone A, Gloghini A, Gruss HJ, Pinto A. CD40 ligand is consti-tutively expressed in a subset of T cell lymphomas and on themicroenvironmental reactive T cells of follicular lymphomas andHodgkin’s disease. Am J Pathol 1995;147:912–22.

23. Poppema S, Potters M, Emmens R, Visser L, van den Berg A.Immune reactions in classical Hodgkin’s lymphoma. Semin Hematol1999;36:253–9.

24. Nagy N, Maeda A, Bandobashi K, Kis LL, Nishikawa J, Trivedi P,Faggioni A, Klein G, Klein E. SH2D1A expression in Burkitt lym-phoma cells is restricted to EBV positive group I lines and is down-regulated in parallel with immunoblastic transformation. Int J Cancer2002;100:433–40.

25. Diehl V, Kirchner HH, Burrichter H, Stein H, Fonatsch C, Gerdes J,Schaadt M, Heit W, Uchanska-Ziegler B, Ziegler A, Heintz F, Sueno

K. Characteristics of Hodgkin’s disease-derived cell lines. CancerTreat Rep 1982;66:615–32.

26. Vockerodt M, Belge G, Kube D, Irsch J, Siebert R, Tesch H, Diehl V,Wolf J, Bullerdiek J, Staratschek-Jox A. An unbalanced translocationinvolving chromosome 14 is the probable cause for loss of potentiallyfunctional rearranged immunoglobulin heavy chain genes in the Ep-stein-Barr virus-positive Hodgkin’s lymphoma-derived cell line L591.Br J Haematol 2002;119:640–6.

27. Hertel CB, Zhou XG, Hamilton-Dutoit SJ, Junker S. Loss of B cellidentity correlates with loss of B cell-specific transcription factors inHodgkin/Reed-Sternberg cells of classical Hodgkin lymphoma. On-cogene 2002;21:4908–20.

28. Arguello M, Sgarbanti M, Hernandez E, Mamane Y, Sharma S,Servant M, Lin R, Hiscott J. Disruption of the B-cell specific tran-scriptional program in HHV-8 associated primary effusion lymphomacell lines. Oncogene 2003;22:964–73.

29. Loddenkemper C, Anagnostopoulos I, Hummel M, Johrens-Leder K,Foss HD, Jundt F, Wirth T, Dorken B, Stein H. Differential Emuenhancer activity and expression of BOB.1/OBF.1, Oct2, PU.1, andimmunoglobulin in reactive B-cell populations, B-cell non-Hodgkinlymphomas, and Hodgkin lymphomas. J Pathol 2004;202:60–9.

30. Tierney R, Kirby H, Nagra J, Rickinson A, Bell A. The Epstein-Barrvirus promoter initiating B-cell transformation is activated by RFXproteins and the B-cell-specific activator protein BSAP/Pax5. J Virol2000;74:10458–67.

31. Contreras-Brodin B, Karlsson A, Nilsson T, Rymo L, Klein G. Bcell-specific activation of the Epstein-Barr virus-encoded C promotercompared with the wide-range activation of the W promoter. J GenVirol 1996;77:1159–62.

32. Cen H, Williams PA, McWilliams HP, Breinig MC, Ho M, McKnightJL. Evidence for restricted Epstein-Barr virus latent gene expressionand anti-EBNA antibody response in solid organ transplant recipientswith posttransplant lymphoproliferative disorders. Blood 1993;81:1393–403.

33. Oudejans JJ, Jiwa M, van den Brule AJ, Grasser FA, Horstman A, VosW, Kluin PM, van der Valk P, Walboomers JM, Meijer CJ. Detectionof heterogeneous Epstein-Barr virus gene expression patterns withinindividual post-transplantation lymphoproliferative disorders. Am JPathol 1995;147:923–33.

34. Zettl A, Lee SS, Rudiger T, Starostik P, Marino M, Kirchner T, Ott M,Muller-Hermelink HK, Ott G. Epstein-Barr virus-associated B-celllymphoproliferative disorders in angioimmunoblastic T-cell lym-phoma and peripheral T-cell lymphoma, unspecified. Am J ClinPathol 2002;117:368–79.

35. Horenstein MG, Nador RG, Chadburn A, Hyjek EM, Inghirami G,Knowles DM, Cesarman E. Epstein-Barr virus latent gene expres-sion in primary effusion lymphomas containing Kaposi’s sarcoma-associated herpesvirus/human herpesvirus-8. Blood 1997;90:1186 –91.

36. Trivedi P, Takazawa K, Zompetta C, Cuomo L, Anastasiadou E,Carbone A, Uccini S, Belardelli F, Takada K, Frati L, Faggioni A.Infection of HHV-8� primary effusion lymphoma cells with a re-combinant Epstein-Barr virus leads to restricted EBV latency, alteredphenotype, and increased tumorigenicity without affecting TCL1 ex-pression. Blood 2004;103:313–6.

37. Delecluse HJ, Anagnostopoulos I, Dallenbach F, Hummel M, Mara-fioti T, Schneider U, Huhn D, Schmidt-Westhausen A, Reichart PA,Gross U, Stein H. Plasmablastic lymphomas of the oral cavity: a newentity associated with the human immunodeficiency virus infection.Blood 1997;89:1413–20.

38. Contreras-Brodin BA, Anvret M, Imreh S, Altiok E, Klein G, MasucciMG. B cell phenotype-dependent expression of the Epstein-Barr virusnuclear antigens EBNA-2 to EBNA-6: studies with somatic cellhybrids. J Gen Virol 1991;72:3025–33.

39. Altmeyer A, Simmons RC, Krajewski S, Reed JC, Bornkamm GW,Chen-Kiang S. Reversal of EBV immortalization precedes apoptosisin IL-6-induced human B cell terminal differentiation. Immunity1997;7:667–77.

40. Wendel-Hansen V, Rosen A, Klein G. EBV-transformed lymphoblas-toid cell lines down-regulate EBNA in parallel with secretory differ-entiation. Int J Cancer 1987;39:404–8.

41. Brauninger A, Spieker T, Mottok A, Baur AS, Kuppers R, HansmannML. Epstein-Barr virus (EBV)-positive lymphoproliferations in post-transplant patients show immunoglobulin V gene mutation patternssuggesting interference of EBV with normal B cell differentiationprocesses. Eur J Immunol 2003;33:1593–602.

42. Brauninger A, Spieker T, Willenbrock K, Gaulard P, Wacker HH,Rajewsky K, Hansmann ML, Kuppers R. Survival and clonal expan-sion of mutating “forbidden” (immunoglobulin receptor-deficient)Epstein-Barr virus-infected B cells in angioimmunoblastic T celllymphoma. J Exp Med 2001;194:927–40.

944 KIS ET AL.

43. Xu WS, Chan AC, Lee JM, Liang RH, Ho FC, Srivastava G.Epstein-Barr virus infection and its gene expression in gastriclymphoma of mucosa-associated lymphoid tissue. J Med Virol1998;56:342–50.

44. Adler B, Schaadt E, Kempkes B, Zimber-Strobl U, Baier B,Bornkamm GW. Control of Epstein-Barr virus reactivation by acti-vated CD40 and viral latent membrane protein 1. Proc Natl Acad SciUSA 2002;99:437–42.

45. Brown HJ, Song MJ, Deng H, Wu TT, Cheng G, Sun R. NF-kappaBinhibits gammaherpesvirus lytic replication. J Virol 2003;77:8532–40.

46. Bargou RC, Emmerich F, Krappmann D, Bommert K, Mapara MY,Arnold W, Royer HD, Grinstein E, Greiner A, Scheidereit C, DorkenB. Constitutive nuclear factor-kappaB-RelA activation is required for

proliferation and survival of Hodgkin’s disease tumor cells. J ClinInvest 1997;100:2961–9.

47. Niedobitek G, Agathanggelou A, Herbst H, Whitehead L, Wright DH,Young LS. Epstein-Barr virus (EBV) infection in infectious mononu-cleosis: virus latency, replication and phenotype of EBV-infectedcells. J Pathol 1997;182:151–9.

48. Kurth J, Hansmann ML, Rajewsky K, Kuppers R. Epstein-Barr virus-infected B cells expanding in germinal centers of infectious mononu-cleosis patients do not participate in the germinal center reaction. ProcNatl Acad Sci USA 2003;100:4730–5.

49. Teramoto N, Gogolak P, Nagy N, Maeda A, Kvarnung K, BjorkholmT, Klein E. Epstein-Barr virus-infected B-chronic lymphocyte leuke-mia cells express the virally encoded nuclear proteins but they do notenter the cell cycle. J Hum Virol 2000;3:125–36.

945IN VITRO EBV-INFECTED SUBLINE OF KMH2