JOURNAL OF VIROLOGY, Jan. 2011, p. 165–177 Vol. 85, No. 10022-538X/11/$12.00 doi:10.1128/JVI.01512-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

A New Model of Epstein-Barr Virus Infection Reveals an ImportantRole for Early Lytic Viral Protein Expression in

the Development of Lymphomas�

Shi-Dong Ma,1 Subramanya Hegde,2 Ken H. Young,3 Ruth Sullivan,4 Deepika Rajesh,5 Ying Zhou,5

Ewa Jankowska-Gan,5 William J. Burlingham,5 Xiaoping Sun,1 Margaret L. Gulley,6Weihua Tang,6 Jenny E. Gumperz,2 and Shannon C. Kenney1,7*

Departments of Oncology1 and Medicine,7 McArdle Laboratory for Cancer Research, School of Medicine and Public Health,University of Wisconsin, Madison, Wisconsin; Department of Medical Microbiology, School of Medicine and Public Health,University of Wisconsin, Madison, Wisconsin2; Pathology & Laboratory Medicine, School of Medicine and Public Health,

University of Wisconsin, Madison, Wisconsin3; Research Animal Resources Center and UW Comprehensive Cancer Center,Graduate School and School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin4;

Department of Surgery, School of Medicine and Public Health, University of Wisconsin, Madison,Wisconsin5; and Department of Pathology, University of North Carolina,

Chapel Hill, North Carolina6

Received 20 July 2010/Accepted 14 October 2010

Epstein-Barr virus (EBV) infects cells in latent or lytic forms, but the role of lytic infection in EBV-inducedlymphomas is unclear. Here, we have used a new humanized mouse model, in which both human fetal CD34�

hematopoietic stem cells and thymus/liver tissue are transplanted, to compare EBV pathogenesis and lym-phoma formation following infection with a lytic replication-defective BZLF1-deleted (Z-KO) virus or alytically active BZLF1� control. Both the control and Z-KO viruses established long-term viral latency in allinfected animals. The infection appeared well controlled in some animals, but others eventually developedCD20� diffuse large B cell lymphomas (DLBCL). Animals infected with the control virus developed tumorsmore frequently than Z-KO virus-infected animals. Specific immune responses against EBV-infected B cellswere generated in mice infected with either the control virus or the Z-KO virus. In both cases, forms of virallatency (type I and type IIB) were observed that are less immunogenic than the highly transforming form (typeIII) commonly found in tumors of immunocompromised hosts, suggesting that immune pressure contributedto the outcome of the infection. These results point to an important role for lytic EBV infection in thedevelopment of B cell lymphomas in the context of an active host immune response.

Epstein-Barr virus (EBV) is a human herpesvirus thatcauses infectious mononucleosis and is associated with both Bcell and epithelial-cell malignancies (20, 32). EBV-positive Bcell lymphomas include endemic Burkitt lymphoma (BL),Hodgkin lymphoma (HL), lymphoproliferative disease (LPD)in immunocompromised hosts (32), and diffuse large B celllymphomas (DLBCL), particularly in elderly patients andAIDS patients (29, 30). EBV does not infect rodent cells,making it difficult to study EBV pathogenesis and EBV-in-duced lymphomas by using small-animal models. EBV-positivetransformed B cell lines (lymphoblastoid cell lines [LCLs]) canbe grown in SCID mice, but these animals do not have afunctional immune system and cannot model the different hu-man B cell differentiation states seen in various types of EBV-positive tumors. Recently, mice with partially reconstitutedhuman immune systems from engraftment of human CD34�

hematopoietic stem cells (HSCs) have provided improvedmodels for studying EBV pathogenesis (4, 38, 43, 44).

A critical advantage of these new model systems is the ability

to investigate the role of different forms of viral infection (e.g.,latent versus lytic) in the context of selective pressure exertedby the host immune system. This is important because multipleviral and host factors likely determine whether EBV-infected Bcells eventually proliferate into lymphomas, and the relativecontributions of different factors are not well understood. Forexample, EBV-positive lymphomas primarily contain cellsshowing latent forms of infection, but the role of lytic infectionin their genesis is unclear. Lytic infection kills the host cell;however, it also allows horizontal spread of EBV from cell tocell and may increase the pool of latently infected B cells fromwhich transformed cells arise. Chronic acyclovir therapy inpatients treated for herpes simplex virus reactivation also de-creases EBV viral loads (14), suggesting that horizontal EBVtransmission may be required to replenish the reservoir oflatently infected cells. Consistent with a tumorigenic role forlytic infection, prophylactic treatment of transplant patientswith antiviral drugs that inhibit lytic replication may reduceEBV-associated lymphomas (6, 10). Moreover, LCLs derivedfrom a lytic replication-defective EBV mutant have an im-paired ability to form LPD-like lesions in SCID mice (12).Additionally, lytically infected B cells secrete factors that maypromote B cell tumors through a variety of mechanisms, in-cluding the B cell growth factor interleukin 6 (IL-6) (18), twodifferent angiogenesis factors (vascular endothelial growth fac-

* Corresponding author. Mailing address: University of Wisconsin,Department of Oncology, McArdle Laboratory, University of Wiscon-sin, Madison, WI 53706. Phone: (608) 265-0533. Fax: (608) 262-2824.E-mail: skenney@wisc.edu.

� Published ahead of print on 27 October 2010.

165

tor [VEGF] and IL-8) (13, 15), and immunosuppressive cyto-kines (cellular IL-10 [23, 37], viral IL-10 [25, 39], and trans-forming growth factor � [TGF-�] [2]).

The major counterbalancing element to EBV-driven lym-phomagenesis is the host immune response. Much of the hu-man T cell response to EBV is directed against lytic viralproteins (36), as well as against latency proteins that are asso-ciated with more aggressive B cell proliferation (3), and thushost immune activity probably particularly limits these highlypathogenic forms of viral infection. Several different types ofEBV latency have been described, each characterized by dif-ferent patterns of EBV gene expression and correspondingdifferences in immunogenicity (20). Type III latency, in whichall 9 latent viral proteins are made, is the only type able totransform primary B cells in vitro; however, this type of infec-tion is highly immunogenic, and tumors with type III latencyare usually observed only in immunosuppressed patients (5,11). At the other end of the spectrum are the type 0/type Iforms of latency, in which either no viral protein (type 0) oronly EBNA1 (type I) is expressed. Cells with this type oflatency persist throughout life in the peripheral memory B cellcompartment following recovery from primary EBV infection;type I latency is also found in EBV-positive Burkitt lymphomas(40). Cells with type IIA latency (EBNA1�/LMP1�/EBNA2�)are found in memory B cells and germinal center cells in thetonsils of healthy EBV carriers, as well as in EBV-positiveHodgkin lymphomas (41). Cells with type IIB latency(EBNA1�/EBNA2�/LMP1�) are found in the tonsils of pa-tients with infectious mononucleosis (22, 27) but have not yetbeen reported to be the predominant form of infection in anytype of EBV-positive human tumor. Thus far, animal modelsof EBV infection have demonstrated a predominance of typeIII latency, and it has been difficult to model types 0, I, IIA, andIIB, perhaps because the immune responses in these modelshave not provided sufficient control of cells with type III la-tency to allow for the outgrowth of the other less aggressivelatency forms.

In the current study, we investigate the role of lytic viralinfection in EBV pathogenesis and tumorigenesis by testing alytic replication-defective EBV mutant compared to a lyticallyactive control strain. We use a new humanized NOD/LtSz-scid/IL2R�null (hNSG) mouse model, in which both human CD34�

hematopoietic stem cells and human thymus/liver tissue areengrafted. This model allows for the development of human Tcells that are restricted by the major histocompatibility com-plex (MHC) molecules expressed by their autologous B cells,which is critical for efficient immune control of EBV infection.Thus, the analysis presented here provides the first assessmentof the impact of lytic viral infection on lymphoma developmentwithin the context of a self-educated human immune system.

MATERIALS AND METHODS

Humanized NOD/LtSz-scid/IL2R�null mice. Immunodeficient nonobese dia-betic/severe combined immunodeficient (NOD/LtSz-scid/IL2R�null [NSG]) micewere purchased from Jackson Labs (catalogue no. 005557) and used at 6 to 10weeks of age. Human fetal thymus and liver tissues (gestational age, 17 to 20weeks) were obtained from Advanced Bioscience Resource (Alameda, CA).Mice were humanized by following the procedure described previously (31). Inbrief, the recipient mice were conditioned with sublethal (2 to 3 Gy) whole-bodyirradiation and implanted with fetal thymus and liver fragments under the re-cipient kidney capsule after irradiation. The mice also received an intravenous

injection of purified CD34� cells isolated from the same fetal liver by themagnetically activated cell sorter (MACS) separation system (Miltenyi Biotec,Auburn, CA). The purity of the injected CD34� cells was at least 80 to 90%. At10 weeks after immune reconstitution, the levels of human hematopoietic cells inmice were determined by multicolor flow cytometric (FCM) analysis using var-ious combinations of the following antibodies: pan-anti-CD45 (clone HI30),-CD4 (clone RPA-T4 or OKT4), -CD8� (clone RPA-T8), -CD19 (HIB19), and-CD3 (SP34-2). Antibodies directly conjugated to fluorescent dyes were pur-chased from commercial sources. For flow cytometric analysis, blood was col-lected from mice 10 weeks after implantation of human cells. Samples weresubjected to ACK lysis to remove red blood cells and further purified by densitygradient centrifugation using Histopaque (Sigma). Fluorescence-activated cellsorting (FACS) analysis was performed on a FACSCalibur (Becton Dickinson,Mountain View, CA). Dead cells were excluded from the analysis.

In one experiment, humanized mice were made without cotransplanting fetalthymic tissue. In brief, NSG mice were irradiated with 2 Gy 24 to 48 h after birth,and purified fetal liver CD34� cells (0.5 million) were injected intrahepaticallyinto each mouse. Immune reconstitution was documented as above.

Viruses and plasmids. Wild-type (WT) EBV (B95-8 strain), expressing thegreen fluorescent protein (GFP) and a hygromycin B resistance gene, and theBZLF1-deleted mutant virus (Z-KO) were constructed using bacterial artificialchromosome technology as described previously (7, 8) and were a gift fromHenri-Jacques Delecluse. WT and Z-KO 293 cells were maintained in Dulbec-co’s modified Eagle medium (DMEM) containing 10% fetal bovine serum(FBS), 1% penicillin-streptomycin, and hygromycin B (100 �g/ml; Roche). Raji,an EBV-positive Burkitt lymphoma cell line, was maintained in RPMI 1640medium with 10% FBS and 1% penicillin-streptomycin. Plasmid DNA was pu-rified through columns as described by the manufacturer (Qiagen). The EBVpSG5-BZLF1 and pSG5-gp110 expression vectors were gifts from Diane Hay-ward (34) and Henri-Jacques Delecluse (26), respectively.

EBV infection of mice. 293 cells latently infected with the WT or Z-KO viruseswere converted to the lytic form of EBV infection by transfecting them withBZLF1 and gp110 expression vectors by using the FuGENE 6 transfectionreagent (Roche) as per the manufacturer’s protocol. Supernatants were har-vested at 72 h posttransfection and filtered through a 0.45-um-pore-size filter.The virus was concentrated by centrifuging at 18,000 rpm for 3 h using an SW27rotor, resuspended in phosphate-buffered saline (PBS) overnight at 4°C, andthen stored at �80°C. To determine the titer of the EBV stock, Raji cells wereinfected with serial 10-fold dilutions of virus. After 48 h, cells were treated with50 ng/ml phorbol-12-myristate-3-acetate (PMA; Sigma) and 3 mM sodium bu-tyrate (Sigma), and the GFP-expressing Raji cells were counted 24 h later byfluorescence microscopy. The amount of virus required to form one GFP-posi-tive Raji cell was defined as one green Raji unit (GRU).

Mice were injected intraperitoneally (i.p.) with 2,750 GRUs of WT or Z-KOEBV in 250 �l PBS or mock infected with PBS alone. In most experiments, halfof the mice derived from a particular donor (8 to 12 mice per donor) wereinfected with the WT virus, and the other half were infected with the Z-KO virus.In one experiment, half of the mice reconstituted with a particular donor wereinfected with the Z-KO virus, and the other half were mock infected (and thusmore animals were infected with Z-KO virus than with the control WT virus).One WT EBV-infected animal and one Z-KO virus-infected animal were sacri-ficed at days 3 and 20 postinfection, and the rest of the animals were sacrificedat 60 to 65 days postinfection unless they exhibited clinical signs, including weightloss, disturbed gait, or hair loss. Organs, including spleen, lymph node, lung,liver, pancreas, kidney, heart, implanted human fetal thymus, head, bone mar-row, muscle, stomach, bowel, and salivary gland, were collected and fixed with10% formalin and then paraffin embedded.

Detection of EBV-encoded RNAs (EBERs) by in situ hybridization. EBER insitu hybridization studies were performed using the PNA ISH detection kit(DakoCytomation) according to the manufacturer’s protocol. Briefly, tissueswere deparaffinized and rehydrated and then treated with 1:10 diluted proteinaseK at room temperature for 20 min. After tissues were washed, EBER probe wasadded on top of the tissue and covered with a cover slide. Hybridization wasperformed at 55°C for 90 min. After hybridization, tissues were washed withstringent washing solution at 55°C for 30 min. Anti-fluorescein isothiocyanate(anti-FITC)/AP was added and incubated at room temperature for 30 min, andsubstrate was then added to develop the color. Tissues were counterstained withnuclear fast red. To compare the number of EBV-infected cells in animalsinfected with the WT versus that in animals infected with the Z-KO virus, wecalculated the number of EBER-positive cells in each mouse by counting thepositive cells in slides that included sections containing the kidney, liver, im-planted fetal thymus, spleen, and lung from each animal. Although other tissuesor organs, such as lymph nodes, also contained some EBER-positive cells, since

166 MA ET AL. J. VIROL.

these tissues were not consistently present on the slides of all animals, they wereexcluded in the comparison.

Measurement of EBV viral loads. Plasma samples (approximately 100 �l) werecollected from jugular vein bleeds after euthanasia of animals and frozen. DNAwas extracted using the QIAamp DNA minikit and eluted in 100 �l nuclease-freewater. Five microliters of the DNA was applied in duplicate quantitative PCR(Q-PCR) targeting the EBV BamH1W segment [EBV W1, 5�-GCAGCCGCCCAGTCTCT-3�; EBV W2, 5�-ACAGACAGTGCACAGGAGCCT-3�; andEBVWprobe, 5�-(6-carboxyfluorescein)AAAAGCTGGCGCCCTTGCCTG(6-carboxytetramethylrhodamine)-3�] as previously described (33). Human ApoBDNA was quantitated in a separate well as a marker of the efficacy of extractionand amplification. The EBV PCR assay has a sensitivity of 6 copies per PCR.

Immunohistochemistry. Formalin-fixed, paraffin-embedded tissue sectionsand cells were deparaffinized and hydrated and then treated with 10 mM citratebuffer (0.05% Tween 20, pH 6.0) for 20 min in a water bath at 98°C. Endogenousperoxidase activity was blocked with 0.3% hydrogen peroxidase solution, andnonspecific labeling was blocked in a 5% goat serum blocking solution. Sectionswere incubated with the first antibody for 1 h at room temperature. The SuperSensitive polymer-horseradish peroxidase immunohistochemistry detection sys-tem (BioGenex Inc.) was used by following the manufacturer’s instruction. Col-ors were developed with the diaminobenzidine tetrachloride (DAB) substrate kit(Vector Laboratories Inc.) by following the manufacturer’s instruction. In thecases of double staining, a combination of the DAB substrate kit (nickel solutionwas added to DAB) and the Vector VIP substrate kit (Vector Laboratories Inc.)was used by following the manufacturer’s instruction. Antibodies (and dilutions)used were as follows: anti-CD3 (polyclonal; DakoCytomation; 1:200), anti-CD20(clone H1, BD Pharmingen; 1:600), anti-CD27 (clone M-T271, BD Pharmingen;1:300), anti-LMP1 (CS.1-4, DakoCytomation; 1:800), anti-EBNA1 (EB14, a giftfrom Richard Burgess, University of Wisconsin; 1:2,000), anti-EBNA2 (PE2,Abcam; 1:100), anti-BZLF1 (BZ1, Santa Cruz; 1:200), anti-BMRF1 (G3-E31,Vector Laboratories; 1:200), anti-GP350/220 (OT6, gift from Jaap Middeldorp,Vrije Universiteit University Medical Center, Netherlands; 1:2,000), anti-CD4(BC/1F6, Biocare Medical; 1:25), and anti-CD8 (SP16, Biocare Medical; 1:50).In addition, to confirm type IIB latency status (EBNA2�/LMP1�), we repeatedLMP1 staining using a different anti-LMP1 antibody (OTC21C, a gift from JaapMiddeldorp; 1:200) in some animals. EBV 293 cells were converted to the lyticform of EBV infection by transfecting them with the SG5-BZLF1 vector andthen fixed with 2% agar in 1% formalin and paraffinized for immunohistochem-istry detection to serve as a positive control for lytic viral protein immunohisto-chemical (IHC) assays.

T cell proliferation assays. hNSG mice were infected with 2,750 GRUs of WTor Z-KO EBV suspended in PBS or mock treated with PBS alone. Twenty to 35days postinfection, animals were sacrificed and splenocytes were harvested. Thesplenocytes were depleted of B cells by magnetic sorting using anti-CD19 beads(Miltenyi Biotech) and labeled with 2.5 �M carboxyfluorescein succinimidylester (CFSE; Molecular Probes, Eugene, OR). Autologous uninfected B cells, orlymphoblastoid cell lines (LCLs) made by infecting autologous B cells with WTEBV or Z-KO EBV, were �-irradiated (7,500 rads) and cultured with the CFSE-labeled splenic cells at a ratio of 1:50. After 7 to 8 days of culture, the cells werefluorescently stained for expression of CD3 to identify T cells, treated with DAPI(4�,6-diamidino-2-phenylindole) to allow for the exclusion of dead cells, andanalyzed by flow cytometry. The percentage of T cells that had undergone celldivision was determined by gating on DAPI-negative CD3-positive cells andassessing the fraction that showed diminished CFSE fluorescence intensity.

T cell killing assays. Splenocytes from WT or Z-KO EBV-infected or mock-treated mice were magnetically depleted of B cells and labeled with CFSE asdescribed above. The splenic cells were incubated in culture medium at 37°C ata 3:1 ratio with autologous uninfected B cells or LCLs made by infecting the Bcells with WT EBV or Z-KO EBV. In parallel, uninfected B cells or WT or Z-KOLCLs were incubated alone (without splenic effector cells) to assess the amountof spontaneous cell killing. After 4 h, the cultures were stained with DAPI toidentify dead or dying cells and analyzed by flow cytometry. Percent target cellkilling was assessed for each culture by determining the fraction of the totalCFSE-negative population that showed elevated DAPI staining. Specific killingwas calculated by subtracting the amount of spontaneous cell death of target cellsalone from the amount of cell death observed in the presence of splenic effectorcells.

Delayed-type hypersensitivity (DTH) assay. WT or Z-KO EBV-infected thy-mus-engrafted hNSG(thy) mice (four animals each) and two mock-infected hNS-G(thy) mice were subcutaneously injected into the right hind footpad with 12.5�g of commercially available EBV antigens (Meridian Life Science, Inc., Mem-phis, TN) in a total volume of 25 �l 1 month postinfection. To control forswelling caused by the injection itself, PBS alone was injected into the left hind

footpad. The hind footpad thickness was measured 24 h postinjection by using adial thickness gauge. The preinjection measurement was subtracted from thepostinjection measurement to obtain specific swelling values. DTH reactivity isshown as the change in thickness of the hind footpad expressed in units of 10�4

inches.

RESULTS

Successful multilineage human hematopoietic reconstitu-tion and thymic engraftment of NSG mice. To generate an invivo model that would allow analysis of EBV pathogenesis andlymphoma formation in the context of a self-educated humanimmune system, we injected irradiated NSG mice intrave-nously (i.v.) with human fetal CD34� cells (purified from fetalliver) and cotransplanted a small piece of the fetal thymus(with liver tissue) under the kidney capsule, as previously de-scribed (31). Reconstitution levels of human cells in the NSGmice were determined using multicolor flow cytometric analy-sis at 10 weeks posttransplantation. At the time point of EBVinfection (10 to 12 weeks after engraftment), human-derivedCD45 cells comprised around 35% of the total leukocyte pop-ulation, and both T cells (including CD4- and CD8-positivecells) and B cells were present (Fig. 1). In addition, the en-grafted human thymic tissue remained viable under the mousekidney capsule for the duration of the experiments.

Both the control EBV and a lytic replication-defective mu-tant establish long-term latency in hNSG(thy) mice and trafficto similar sites. To directly test the role of lytic infection in thepathogenesis of EBV infection, we compared a lytically activecontrol virus to a deletion mutant strain (Z-KO) lackingBZLF1, a transcription factor required for lytic viral geneexpression (20, 32). The Z-KO virus contains a kanamycinresistance cassette (inserted between nucleotides 102389 to103388 using B95.8 coordinates) that specifically disruptsBZLF1 function, and the phenotype of the mutant virus can berescued in trans by the expression of the BZLF1 gene product(7, 8). Immune reconstituted mice were infected i.p. with equaltiters of the control or Z-KO virus and sacrificed at various

FIG. 1. Reconstitution of human hematopoietic cells in NSG mice.Peripheral blood was collected from immune reconstituted hNSG(thy)mice 10 weeks after human fetal CD34 cell transplantation and stainedwith antibodies specific for human CD45, CD3, and CD19 by flowcytometry. Dead cells were excluded from the analysis. The reconsti-tution levels in 12 mice that were subsequently infected with controlEBV, versus 16 mice that were subsequently infected with Z-KO EBV,are shown. Results are presented as the percentages of positivelystaining cells with each antibody in comparison to the total leukocytepopulation.

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 167

time points after infection or sooner if they showed signs ofclinical illness. Multiple different organs were examined for thepresence of EBV-infected cells using the EBER in situ hybrid-ization assay. EBERs are expressed in all latently EBV-in-fected cells regardless of their latency type (20). While thenumber of EBER-positive cells was somewhat variable, everyEBV-infected animal, regardless of whether infected by thecontrol or the Z-KO virus, was found to have at least someEBV-infected (EBER-positive) cells at the time of autopsy,except for the animals sacrificed at day 3 postinfection. No

EBER-positive cells were found in the tissues from uninfectedhNSG(thy) animals (data not shown).

To determine if horizontal viral transmission is required forthe ability of EBV to infect certain cell types or to reach certainsites in the body, we compared the sites of EBER-positive cellsin tumor-free animals infected with the control or Z-KO virus.EBER-positive cells were most frequently observed in theliver, spleen, kidney, lung, and transplanted human thymictissue (Table 1). In some animals, EBER-positive cells werefound in the bone marrow, nasal lymphoid tissues, mesentery,and muscle tissue (Fig. 2 and data not shown). In the trans-planted fetal thymus, EBV-positive cells were most commonlyfound in the medulla area, often near Hassall’s corpuscles (Fig.3A). In the spleen, EBV-positive cells were usually locatedwithin CD20-rich lymphoid zones (Fig. 3B). No clear differ-ences in the sites and numbers of EBER-positive cells wereobserved in animals infected with the control virus or theZ-KO virus (Table 1). Although Z-KO-infected animals weresomewhat more likely than control virus-infected animals tohave EBV-infected B cells in the thymus and spleen, the dif-ferences were not statistically significant. Since the Z-KO virusis not horizontally transmitted, these results indicate thatEBV-infected B cells home to many different organs in thismodel and that transmission of EBV from cell to cell via the

FIG. 2. Both control and Z-KO virus-infected animals have EBER-positive cells at multiple different sites. Hematoxylin and eosin (H&E)staining and EBER in situ hybridization were performed on a varietyof different organs in tumor-free animals infected with the control orZ-KO virus. EBER-positive cells were detected at many different sites,including the nasal lymphoid tissue surrounding the vomeronasal or-gan (VNO; outlined with hatches; 20� magnification) (A) and muscle(20� magnification) (B). Examples of positively staining cells are il-lustrated with arrows.

TABLE 1. The numbers of animals with EBER-positive cells invarious organs after infection with control versus Z-KO virus are

shown; results are from the tumor-free animals only

Organ/tissue EBV

No. of animalswith EBER-positive cells/total no. of

animals

Spleen Control 2/5Z-KO 10/12

Lung Control 5/5Z-KO 9/12

Kidney Control 3/5Z-KO 12/12

Thymus Control 1/5Z-KO 6/12

Liver Control 3/5Z-KO 10/12

FIG. 3. EBER-positive cells travel to specific regions of the trans-planted thymic tissue and reconstituted spleen. H&E, EBER, CD20,and CD3 staining was performed on transplanted thymic tissue (20�magnification) (A) and spleen (40� magnification) (B). Ki67 stainingwas also performed on the cells shown in panel B. EBV-positive cellswere primarily located in the medulla region of the thymic tissue andprimarily localized to CD20-rich lymphoid zones (outlined withhatches) of the spleen.

168 MA ET AL. J. VIROL.

lytic form of viral infection is not required for the establish-ment of long-term viral latency.

B cells are the predominant host cell type infected by EBV inthis model. Although EBV is primarily found in B cells andepithelial cells in healthy humans, it can infect T cells andmonocytes/macrophages in vitro (35) and is found in some Tcell and NK cell lymphomas (17). To determine if EBV caninfect cell types other than B cells in the hNSG(thy) mousemodel, we costained multiple tissues using an anti-EBNA1antibody and antibodies specific for human B cells (CD20), Tcells (CD3), monocytes (CD14 and CD68), hematopoieticstem cells (CD34), and epithelial cells (anticytokeratin). AllEBNA1-positive cells costained with CD20 in every organ,including the transplanted human thymus (Fig. 4A). SomeEBNA1-positive cells also costained with anti-CD27, a markerfor memory B cells (Fig. 4B). Thus, few if any T cells ormonocytes are infected with either the control or Z-KO EBVin this model.

EBV establishes type I and type IIB latency in hNSG(thy)mice. To define the type(s) of viral latency established by thecontrol and Z-KO EBV, we performed an IHC assay in

EBER-positive areas from multiple different types of organs intumor-free animals, using anti-EBNA1, anti-EBNA2, and anti-LMP1 antibodies. As summarized in Table 2, type I and typeIIB latency were the most common forms of EBV infection inboth the control virus-infected and Z-KO-infected mice. As wewere unable to perform EBER in situ hybridization assays onthe same cells used to perform EBNA IHC assays (due to lossof EBNA1 IHC staining under the EBER staining conditionsand vice versa), we could not determine how many cells hadtype 0 latency. However, since we generally found moreEBER-positive cells than EBNA1-positive cells in the tumor-free animals (data not shown), cells with type 0 latency mayhave also been present. Some animals had a mixture of variousdifferent types of latency depending upon the location. Anexample of cells with type I latency is shown in Fig. 5A, andexamples of cells with type IIB latency are shown in Fig. 5B toD. Latent membrane protein 1 (LMP1)-positive cells were rareand/or nonexistent in most tumor-free hNSG(thy) animals in-fected with either the control or Z-KO EBV. Rare LMP1�/EBNA2� cells (type IIA latency) were found in the kidney ofa control EBV-infected mouse (Fig. 6A). In tumor-free ani-mals, cells with type III latency were observed most frequentlyat the day 20 time point (Fig. 6B and C). Thus, in animals thatsuccessfully limit their EBV infection, type III latency is tran-sient, presumably because the host immune response elimi-nates cells with type III latency.

Fewer lymphomas develop in Z-KO-infected hNSG(thy)mice. Although many mice had asymptomatic infection, someEBV-infected mice eventually developed B cell lymphomas inthis model. Importantly, the lytic replication-defective Z-KOvirus produced significantly fewer lymphomas than the controlvirus. Six of the 11 mice infected with the control virus devel-oped lymphomas, versus only 2 of the 14 Z-KO virus-infectedanimals (P 0.05) (Fig. 7A). All EBV-positive tumors wereCD20� DLBCL-like lymphomas. A number of the EBV-in-duced lymphomas in this model were small and discoveredonly after review of a series of slide sections obtained frommultiple different organs. Interestingly, in contrast to the EBV-induced lymphomas found in humanized mouse models thatdo not include engrafted human thymic tissue, which have allbeen reported to have type III latency (4, 38, 43), we foundthat some of the lymphomas in this hNSG(thy) model hadtype I (Fig. 7B) and type IIB (Fig. 7C) latency, althoughothers did have type III latency (Fig. 7D) (summarized inTable 3). No uninfected mice developed lymphomas (datanot shown). These results suggest that one or more lytic viralproteins enhance the development of B cell lymphomas inthe hNSG(thy) mouse model and show that this model pro-vides for the development of EBV-positive lymphomas withtype I and type IIB latency.

TABLE 2. Latency types in tumor-free animalsa

EBVNo. of animals with latency type(s):

I I and IIA IIB I and IIB

Control 1 1 3 0Z-KO 5 0 5 2

a Some animals had different latency types in different organs. Animals har-vested at day 3 and day 20 postinfection were excluded from the analysis.

FIG. 4. All EBV-positive cells are CD20 positive and some areCD27 positive. Dual color immunohistochemistry was performed usinganti-EBNA1 (black) and anti-CD20 (pink) antibodies in transplantedthymic tissue (A) or anti-EBNA1 (black) and anti-CD27 (pink) anti-bodies in the spleen (B) (both, 100� magnification). EBNA1-positivecells with costaining for CD20 or CD27 are indicated by black arrows,and EBNA1-negative CD20- and CD27-positive cells are indicated bypink arrows.

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 169

Lytically infected cells in tumors of mice infected with thecontrol virus. To examine the effect of lytic viral infection uponthe level of EBV DNA in the plasma, we performed viral loadassays in a subset of mock- and Z-KO and control virus-in-fected animals. Somewhat surprisingly, the only animal thathad detectable viral DNA in the plasma was a control virus-infected animal with a large type III lymphoma (Table 4).Although it is possible that EBV plasma viral loads might havebeen higher if measured at earlier time points after controlvirus infection, these results suggest that the development ofan effective anti-EBV T cell response in this model acts toeliminate cells with lytic viral infection.

We also performed an IHC assay using antibodies directedagainst three different classes of lytic viral proteins (the imme-diate-early BZLF1 protein, the early BMRF1 protein, and thelate gp350/220 protein) in all animals infected with the controlEBV strain. Whereas no lytically infected cells were observedby IHC assay in any tissues from tumor-free mice (includingthe one animal examined at 3 days postinfection), rare cells

expressing the BZLF1 and BMRF1 proteins, but not thegp350/220 protein, were found within B cell lymphomas (Fig.8A); as expected, no BZLF1- or BMRF1-positive cells werefound in any of the tumors infected with the Z-KO virus (datanot shown). All BZLF1-positive cells costained with CD20(data not shown), indicating that they were B cells. Notably,the absence of late viral antigen staining in the lymphoma cellssuggests that lytic infection is either abortive or that cells ex-pressing late viral proteins are rapidly eliminated. Interest-ingly, EBV infection of mice that were reconstituted with hu-man CD34� stem cells in the absence of cotransplanted thymictissue resulted in the development of tumors showing greaterevidence of lytically infected cells (Fig. 8B). Thus, the self-educated T cells of the thymus-engrafted mice may contributeto the elimination of lytically infected cells.

Development of EBV-specific immune responses in hNSG(thy) mice. EBV-positive cells in tumor-free animals were al-most always surrounded by CD3� T cells (Fig. 9A), suggestingthat T cells actively interact with EBV-infected B cells prior totumor development. To investigate this directly, we assessedthe responses of primary T cells from infected and uninfectedmice to EBV-infected or uninfected B cells. Splenocytes fromthe control or Z-KO virus-infected or mock-infected mice weredepleted of B cells and then labeled with CFSE and culturedwith uninfected B cells or B cells that were infected in vitro withthe control or Z-KO EBV (i.e., LCLs). After 7 to 9 days, thecultures were stained with anti-CD3 and DAPI and analyzedby flow cytometry for the percentage of live T cells that showed

TABLE 3. Latency types in EBV-positive lymphomas

EBV

No. of animals with latency type:

I(LMP1�/EBNA2�)

IIA(LMP1�/EBNA2�)

IIB(LMP1�/EBNA2�)

III(LMP1�/EBNA2�)

Control 1 0 2 3Z-KO 0 0 1 1

FIG. 5. Control and Z-KO viruses establish long-term type I and type IIB latency. EBER in situ hybridization, as well as anti-EBNA1,anti-EBNA2, and anti-LMP1 staining, was performed as indicated to determine the type(s) of viral latency established in tumor-free animals. Someslides were also stained for CD20 or CD3 as indicated. Arrows show examples of positively staining cells. (A) Type I latency in a lymph node ofa Z-KO virus-infected animal (20� magnification in upper panels and 40� magnification in lower panels). (B) Type IIB latency program in thespleen of a control virus-infected animal; cells expressing EBNA2 are shown with arrows (all, 40� magnification, except the upper right panelshows 100� magnification). Hatches indicate CD20� lymphoid aggregates. (C) Type IIB latency in the transplanted thymic tissue of a Z-KOvirus-infected animal; the Hassal’s corpuscles (HC) are outlined with hatches. EBER analysis (20� magnification) and EBNA2 and LMP1 IHCassay (both 100� magnification) are shown. (D) Type IIB latency program in the kidney of a Z-KO virus-infected animal (all, 100� magnification).

170 MA ET AL. J. VIROL.

diminished CFSE fluorescence (indicating that cell divisionhad taken place). As shown in Fig. 9B, T cells obtained frommice infected with either the Z-KO virus or the control virusshowed significantly higher proliferation when exposed toEBV-infected B cells than when exposed to uninfected B cells.Additionally, T cells derived from mock-infected animals hadminimal proliferative responses to the EBV-infected LCLs(Fig. 9B). Thus, primary T cells from EBV-infected mice pro-liferated specifically in response to EBV-infected LCLs,whereas T cells from uninfected mice did not.

To further investigate the potential for immune-mediatedcontrol of EBV-infected cells in this model, cell killing assayswere performed using splenic effector cells (e.g., T cells andNK cells) and uninfected B cells or LCLs infected with thecontrol or Z-KO virus as target cells. As shown in Fig. 9C, ina 4-h assay, splenic effector cells from a mock-infected animalshowed no specific killing of EBV-infected B cells (either con-trol virus or Z-KO virus infected). In contrast, splenocytesfrom each of the two Z-KO virus-infected animals examinedshowed cytotoxic effector activity to both control virus- andZ-KO virus-infected LCLs, as did splenocytes from one of thetwo control virus-infected animals examined. There was nospecific killing of uninfected B cells by effector cells from anyof the animals, confirming that the cytotoxic response wasspecific to infected LCLs (Fig. 9C). These results suggest thatin this model, primary effector cells from EBV-infected ani-mals can produce a functional killing response against EBV-infected B cells.

As another measure of the host immune response againstEBV, we examined in vivo delayed-type hypersensitivity(DTH) responses to an antigen preparation from EBV-in-fected cells. DTH responses are localized inflammatory re-sponses initiated by the activation of antigen-specific lympho-cytes that result in multicellular infiltration and edema. Micethat were either mock infected or infected with the control orZ-KO virus were inoculated in the footpads with a preparationof protein antigens from a lytically induced EBV-positive cellline or PBS, and the amount of swelling in the footpads wasmeasured 24 h later (Fig. 9D). Both control EBV- and Z-KOvirus-infected animals had significantly more swelling in re-sponse to the EBV antigen than the mock-infected animal. TheDTH responses were not significantly different between con-trol EBV- and Z-KO EBV-infected animals. This assay pro-vides powerful confirmation that mice infected with either thecontrol or ZKO virus strain develop in vivo immune responsesagainst proteins from EBV-infected cells.

EBV-positive tumors of different latency types all attract avigorous inflammatory response in hNSG(thy) mice. Finally,we stained tumors with anti-CD3 antibody (which recognizesall T cells), as well as anti-CD8 and anti-CD4 antibodies. All ofthe EBV-positive lymphomas, including those with type IIIlatency (in which the immune response is obviously inade-quate), were infiltrated with both CD4� and CD8� T cells inthis model (Fig. 10A). Interestingly, the CD4 cell infiltration oftumors with type IIB latency appeared particularly robust (Fig.10B). Levels of T cell infiltration of the tumors containing thecontrol EBV versus the Z-KO virus were not obviously differ-ent. These results indicate that host T cells interact with vari-ous different types of EBV-positive tumors in this model, evenwhen the tumors are not effectively eliminated.

DISCUSSION

The potential role of lytic viral infection in the developmentof various types of EBV-positive malignancies has remainedunclear, in part due to the lack of suitable animal models tostudy this question. In addition, it has been difficult to study therole of the more restricted forms of viral latency in EBV-positive cancers using animal models, since EBV cannot infectrodent cells, and in the absence of a highly functional engrafted

FIG. 6. Latency types IIA and III occur rarely in control and Z-KOvirus-infected mice. EBER in situ hybridization, as well as anti-EBNA1, anti-EBNA2, and anti-LMP1 staining, was performed asindicated to determine the type(s) of viral latency established in tumor-free animals. (A) Type IIA latency in the kidney of a control virus-infected animal (H&E staining shown at 40� magnification; EBER,EBNA1, EBNA2, and LMP1 staining all shown at 100� magnifica-tion). (B) Type III and type IIB latency in transplanted thymic tissueof a Z-KO virus-infected mouse (day 20 postinfection) (all, 100�magnification). Costaining with anti-EBNA2 and anti-LMP1 antibod-ies reveals that only a portion of the EBNA2-positive cells also expressLMP1 (indicated by pink arrows). (C) Type III latency in the liver ofa Z-KO virus-infected mouse (day 20) (all, 100� magnification).

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 171

TABLE 4. EBV viral loada

Animal EBV DPI TumorpresenceNo. of ApoB DNA

copies/5 �lViral load

(copies/5 �l)

1 Mock 51 No 133 Undetected2 Mock 56 No 30 Undetected3 WT 39 Yes (type I) 114 Undetected4 WT 51 No 31 Undetected5 WT 60 Yes (type III) 98 256 WT 30 No 16 Undetected7 WT 32 No 51 Undetected

8 Z-KO 57 No 70 Undetected9 Z-KO 24 No 4 Undetected10 Z-KO 3 No 258 Undetected11 Z-KO 60 Yes (type III) 25 Undetected12 Z-KO 51 No 68 Undetected13 Z-KO 32 No 8 Undetected

a Q-PCR was performed (using primers to detect the BamHI repeat region of the EBV sequence) on purified plasma DNA collected from mock-infected, controlvirus-infected, or Z-KO EBV-infected animals at the time points indicated. The human ApoB gene DNA was quantitated in a separate well as a marker of the efficacyof extraction and amplification. The presence (or absence) of tumors in each animal is indicated. DPI indicates the day postinfection that plasma was collected.

FIG. 7. Control EBV infection induces more tumors than Z-KO virus infection in hNSG(thy) mice, and some EBV-positive tumors haverestricted latency types. (A) The numbers of EBV-positive tumors in control and Z-KO virus-infected animals are shown (relative to the numberof animals infected with each virus). The P value was calculated using a one-tailed Fisher exact test. Animals sacrificed at day 3 and day 20postinfection were excluded from this analysis. H&E, EBER, EBNA1, EBNA2, and LMP1 staining was performed, as indicated, on a tumor inthe liver (type I) of a control virus-infected animal (100� magnification) (B), a tumor in the pancreas of a Z-KO virus-infected animal (type IIB)(100� magnification) (C), and a tumor in the liver of a control virus-infected animal (type III) (100� magnification) (D).

172 MA ET AL. J. VIROL.

human immune system, lymphomas with type III latent infec-tion inevitably develop. In this paper, we have used a newhumanized mouse model to directly assess the effects of a lyticreplication-defective BZLF1-deleted mutant compared tothose of an otherwise identical BZLF1� control strain. Weshow that the lytic replication-defective mutant develops fewerlymphomas in this model. In addition, we demonstrate that thisis the first humanized mouse model able to support the devel-opment of EBV-induced B cell lymphomas with restrictedforms of viral latency (type I and type IIB), a feature that maybe related to the presence of self-educated T cells that are ableto efficiently recognize EBV-infected B cells.

We previously showed that early-passage LCLs infected withEBV are more efficient in producing LPD-like lesions in SCIDmice than LCLs infected with the Z-KO virus (12). However,the SCID mouse model does not have human immune effectorpopulations, cannot support horizontal viral transmission, anddoes not contain the various different stages in human B celldifferentiation. Thus, there is a critical need for the develop-ment of new, more sophisticated models to study the patho-genesis of EBV infection in vivo.

The current study is the first to comprehensively examineEBV pathogenesis in just such a system. A previous study usinga similar model system showed the development of EBV-spe-

cific MHC class I- and class II-restricted adoptive immuneresponses (24) following infection with EBV; however, thebehavior of EBV, including its ability to establish latency andform tumors, was not described. Our results showing enhancedlymphoma formation after infection by lytically active EBVsuggest that horizontal transmission of virus is important forlymphoma formation, which is consistent with the findings ofanother group showing that the ability of EBV to induce LPDin NOD/Shi-scid/IL2r�null (NOG) mice following immune re-constitution with human cord blood hematopoietic stem cells(HSCs) (but not thymic tissue) is dependent on the dose ofvirus used to infect the mice (43). Nevertheless, since we foundthat the lytic replication-defective Z-KO virus efficiently estab-lishes viral latency in this model, another possibility is thatlytically infected cells promote EBV-induced lymphomasthrough paracrine mechanisms and/or immunosuppressive fac-tors.

Similar to another recent study examining EBV infection inhNSG mice (38), we detected few if any cells with lytic EBVinfection in tumor-free animals, although we found lyticallyinfected cells within the EBV-induced tumors. In addition, wedid not detect EBV DNA in the plasma of any tumor-freeanimal. Since the previous study found that much of the CD8T cell response in their model is directed against lytic viralantigens (38), lytic EBV infection must clearly occur at a lowlevel in the tumor-free animals, even if it is difficult to detect byusing IHC methods or by measuring viral DNA in the plasmaat the time points tested in our study. The finding that lyticallyinfected cells are preferentially found in lymphomas in boththis study and a previous study (38) is consistent with thehypothesis that lytically infected cells contribute to tumorgrowth through paracrine mechanisms and are thus selectedfor in tumors. The fact that we detected tumor cells withBZLF1 and BMRF1 expression, but not gp350/220 expression,suggests that lytic infection in tumor cells may be abortive.Interestingly, abortively lytic EBV infection which occursshortly after EBV infection in vitro was recently reported tocontribute to EBV-induced cellular proliferation of naïve andmemory B cells (but not germinal center B cells) (19). Never-theless, detectable lytic infection in the humanized mousemodel could also simply be a marker for an inadequate im-mune response to EBV.

Clearly, the hNSG(thy) mouse model does not always fullycontain EBV infection, as evidenced by the finding that someEBV-infected mice went on to eventually develop EBV-in-duced lymphomas. Another technical problem with humanizedmouse models in general is the tendency of the mice to developgraft-versus-host disease (GVHD) as they age. While we didnot find that EBV infection affected the onset or severity ofGVHD (in comparison to the mock-infected mice) in thisstudy, the development of GVHD at late time points afterimmune reconstitution in this model is currently an impedi-ment for studying EBV pathogenesis for periods longer than 2months postinfection.

Nevertheless, several findings in the current study suggestthat this model will be particularly attractive for the study ofEBV-associated diseases. First, the virus is able to establishtype I latency (and likely type 0 latency) in memory B cells,closely mimicking the behavior of the virus in healthy humansfollowing recovery from primary EBV infection. Second, sim-

FIG. 8. Cells with lytic EBV infection are found within EBV-in-duced lymphomas. (A) H&E, EBNA1, BZLF1, BMRF1, and gp350/220 staining was performed on a tumor in the liver (type III) of acontrol virus-infected hNSG(thy) animal (100� magnification).(B) H&E, EBNA1, BZLF1, BMRF1, and gp350/220 staining was per-formed as indicated on a tumor in the spleen (type III) of a controlvirus-infected animal reconstituted with no thymus implantation(100� magnification).

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 173

FIG. 9. Infection with both the control and Z-KO viruses induces a host immune response. (A) EBER in situ hybridization and CD20 and CD3staining were performed on kidney (upper) and muscle (lower) as indicated in tumor-free animals (100� magnification). (B) Proliferation of Tcells in response to EBV-infected B cells. Splenocytes were harvested from 4 animals that were mock infected, 5 animals infected with Z-KO virus,and 6 animals infected with the control virus. The splenocytes were depleted of B cells, labeled with CFSE, and then incubated with autologousuninfected B cells or B cells (LCLs) immortalized with the control or Z-KO virus. The plots show the percentages of the total T cells from eachmouse that had diluted CFSE fluorescence intensity (indicating that they had proliferated) after 7 to 8 days of culture with the three B cell typesshown on the x axes. P values were calculated using a one-tailed paired t test. (C) Cytotoxic responses to EBV-infected B cells. Splenocytes fromthe indicated virus- or mock-infected animals were depleted of B cells and tested in a 4-h cytotoxicity assay for the ability to kill autologousuninfected B cells or B cells immortalized with the control or Z-KO viruses. The bars show the specific killing (means and standard deviations of

174 MA ET AL. J. VIROL.

ilar to healthy humans, tumor-free animals in this model havefew if any cells with the most immunogenic (and transforming)form of latent viral infection (type III latency), suggesting thatthe immune response is successfully recognizing and destroy-ing such cells. We found that primary T cells from infectedhNSG(thy) mice showed specific responses to EBV-infected Bcells in vitro and developed antigen-dependent DTH responsesin vivo. Since the T cell responses were also present in Z-KOvirus-infected animals, at least a portion of the anti-EBV im-mune response in the hNSG(thy) model is directed againstlatent viral antigens. Finally, although some animals do even-tually develop EBV-positive lymphomas, only a portion ofthese lymphomas have type III latency (similar to EBV-in-duced LPD in immunosuppressed humans), whereas otherlymphomas have forms of viral latency (such as type I) thatoccur in EBV-positive tumors of immunocompetent humans.Thus, although a variety of different approaches have beenrecently developed to examine EBV pathogenesis in mice with

a reconstituted functional human immune system (4, 24, 38,43), this is the first human immune-reconstituted mouse modelfound to support the development of EBV-induced tumorsthat have type I and type IIB latency.

Interestingly, similar to the results of several other recentstudies using human immune reconstituted mice (4, 43), we didnot obtain EBV-positive lymphomas with type IIA latency(typical of EBV-positive Hodgkin’s disease). The lack of typeIIA lymphomas may reflect the fact that cells with this form oflatency are very rare in the tumor-free animals (perhaps re-flecting an imperfect reconstitution of the normal B cell dif-ferentiation in the lymphoid structures) and/or the require-ment for multiple different cellular mutations to complementthe role of EBV in such tumors. To date, NOD/SCID micetransplanted with human HSCs (and no thymus) is the onlymouse model reported to support EBV-positive lymphomaswith type IIA latency (16). Since the method used to diagnosetype IIA latency in this study was reverse transcription-PCR(RT-PCR) based, the results need to be confirmed by IHCstaining of the tumors. In any event, since human T cells do notengraft efficiently in the NOD/SCID mouse model, NOD/SCID mice are not useful for studying EBV pathogenesis inthe context of an intact immune system.

Nevertheless, our results here, as well as those of anothergroup (4), suggest that the hNSG(thy) model should proveuseful for studying other types of viral latency, particularly thetype IIB form. Type IIB latency has been found in EBV-infected B cells within the tonsils of infectious mononucleosis(IM) patients (22, 27), but currently very little is known aboutthis type of latency. While type IIB latency has been proposedto be a marker of newly infected cells (since EBNA2 is ex-pressed prior to LMP1 during B cell infection) (27), our find-ing that Z-KO virus-infected hNSG(thy) mice have cells withtype IIB latency long after the initial infection indicates thattype IIB latency is not restricted to newly infected cells. In-stead, since type IIB latency preferentially occurs in the follic-ular region of IM tonsils, whereas type IIA latency primarilyoccurs in the germinal centers (22), interactions occurring be-tween B cells and T cells within normal lymphoid structuresmay regulate the EBV latency state.

The development of tumors with predominantly type IIBlatency in our model was unexpected and suggests that thistype of latency, although nontransforming by itself, must pro-vide some sort of survival advantage for the tumor cells. Al-though cells with this form of EBV gene expression patternhave been described within AIDS-associated lymphomas andposttransplant lymphoproliferative disorders (PTLDs) (9, 28),they represent only a portion of the cells in such tumors, andit has been generally assumed that the cells with type III la-tency are driving the growth of such tumors. Type IIB latencyalso occurs following EBV infection of chronic lymphocyticleukemia (CLL) B cells in vitro (1), but such cells are not stably

results from 3 replicate samples) by effector cells from the indicated animals against the target cell types shown in the legend. The asterisks indicatethe cases where no specific killing was detected. (D) Delayed-type hypersensitivity response of EBV-infected animals in response to EBV antigens.Mock- or virally infected animals were injected in the footpad with an EBV antigen preparation or with PBS. The plot shows the change in thethickness of the footpads after the antigen injection compared to preinjection measurements; the results from two mock-infected animals, fourcontrol virus-infected animals, and four Z-KO virus-infected animals are included.

FIG. 10. EBV-induced lymphomas are infiltrated with T cells.(A) A type III tumor in the kidney of a Z-KO virus-infected animal wasstained for EBER, CD3, CD8, and CD4 as indicated (100� magnifi-cation); (B) EBER/CD20/CD3/CD4/CD8 staining of a type IIB tumorsurrounding the bile duct in a control virus-infected animal (100�magnification).

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 175

transformed by EBV. Interestingly, CD40 ligand stimulationhas been shown to replace the growth stimulation function ofLMP1 in LCLs derived with an LMP1-deleted virus (21). Sincethe CD4 cells are a major source of CD40 ligand (42), the closeproximity of CD4 cells with EBV-positive B cells in the tumorswith type IIB latency (Fig. 10) may induce CD40 signaling inthe EBV-infected cells (reducing the need for LMP1 expres-sion). In the future, it will be important to determine whethertype IIB infection is more common than previously recognizedduring normal EBV infection in humans and what role it plays,if any, in different types of EBV-associated malignancies.

ACKNOWLEDGMENTS

We thank our collaborators, Annette Gendron-Fitzpatrick andDrew Allan Vandenack, for their technical support. Thanks to Henri-Jacques Delecluse for providing valuable reagents. Thanks to JaapMiddeldorp for providing the anti-LMP (OTC21C) and anti-gp350/220(OT6) antibodies and Richard Burgess for kindly providing the anti-EBNA1 antibody (EB14).

This research was supported by grants R21-CA12643, R01-CA58853, and R01-CA66519 from the National Institutes of Healthand University of Wisconsin Cancer Center Support Grant P30CA014520.

REFERENCES

1. Bandobashi, K., A. Liu, N. Nagy, L. L. Kis, J. Nishikawa, M. Bjorkholm, G.Klein, and E. Klein. 2005. EBV infection induces expression of the tran-scription factors ATF-2/c-Jun in B lymphocytes but not in B-CLL cells. VirusGenes 30:323–330.

2. Cayrol, C., and E. K. Flemington. 1995. Identification of cellular target genesof the Epstein-Barr virus transactivator Zta: activation of transforminggrowth factor �igh3 (TGF-�igh3) and TGF-�1. J. Virol. 69:4206–4212.

3. Chapman, A. L., and A. B. Rickinson. 1998. Epstein-Barr virus in Hodgkin’sdisease. Ann. Oncol. 9(Suppl. 5):S5–S16.

4. Cocco, M., C. Bellan, R. Tussiwand, D. Corti, E. Traggiai, S. Lazzi, S.Mannucci, L. Bronz, N. Palummo, C. Ginanneschi, P. Tosi, A. Lanzavecchia,M. G. Manz, and L. Leoncini. 2008. CD34� cord blood cell-transplantedRag2�/� �c�/� mice as a model for Epstein-Barr virus infection. Am. J.Pathol. 173:1369–1378.

5. Cohen, J. I., C. M. Bollard, R. Khanna, and S. Pittaluga. 2008. Currentunderstanding of the role of Epstein-Barr virus in lymphomagenesis andtherapeutic approaches to EBV-associated lymphomas. Leuk. Lymphoma49(Suppl. 1):27–34.

6. Darenkov, I. A., M. A. Marcarelli, G. P. Basadonna, A. L. Friedman, K. M.Lorber, J. G. Howe, J. Crouch, M. J. Bia, A. S. Kliger, and M. I. Lorber.1997. Reduced incidence of Epstein-Barr virus-associated posttransplantlymphoproliferative disorder using preemptive antiviral therapy. Transplan-tation 64:848–852.

7. Delecluse, H. J., T. Hilsendegen, D. Pich, R. Zeidler, and W. Hammer-schmidt. 1998. Propagation and recovery of intact, infectious Epstein-Barrvirus from prokaryotic to human cells. Proc. Natl. Acad. Sci. U. S. A.95:8245–8250.

8. Feederle, R., M. Kost, M. Baumann, A. Janz, E. Drouet, W. Hammer-schmidt, and H. J. Delecluse. 2000. The Epstein-Barr virus lytic program iscontrolled by the co-operative functions of two transactivators. EMBO J.19:3080–3089.

9. Hamilton-Dutoit, S. J., D. Rea, M. Raphael, K. Sandvej, H. J. Delecluse, C.Gisselbrecht, L. Marelle, H. J. van Krieken, and G. Pallesen. 1993. Epstein-Barr virus-latent gene expression and tumor cell phenotype in acquiredimmunodeficiency syndrome-related non-Hodgkin’s lymphoma. Correlationof lymphoma phenotype with three distinct patterns of viral latency. Am. J.Pathol. 143:1072–1085.

10. Hanto, D. W., G. Frizzera, K. J. Gajl-Peczalska, K. Sakamoto, D. T. Purtilo,H. H. Balfour, Jr., R. L. Simmons, and J. S. Najarian. 1982. Epstein-Barrvirus-induced B-cell lymphoma after renal transplantation: acyclovir therapyand transition from polyclonal to monoclonal B-cell proliferation. N. Engl.J. Med. 306:913–918.

11. Heslop, H. E. 2005. Biology and treatment of Epstein-Barr virus-associatednon-Hodgkin lymphomas. Hematology Am. Soc. Hematol. Educ. Program2005:260–266.

12. Hong, G. K., M. L. Gulley, W. H. Feng, H. J. Delecluse, E. Holley-Guthrie,and S. C. Kenney. 2005. Epstein-Barr virus lytic infection contributes tolymphoproliferative disease in a SCID mouse model. J. Virol. 79:13993–14003.

13. Hong, G. K., P. Kumar, L. Wang, B. Damania, M. L. Gulley, H. J. Delecluse,

P. J. Polverini, and S. C. Kenney. 2005. Epstein-Barr virus lytic infection isrequired for efficient production of the angiogenesis factor vascular endo-thelial growth factor in lymphoblastoid cell lines. J. Virol. 79:13984–13992.

14. Hoshino, Y., H. Katano, P. Zou, P. Hohman, A. Marques, S. K. Tyring, D.Follmann, and J. I. Cohen. 2009. Long-term administration of valacyclovirreduces the number of Epstein-Barr virus (EBV)-infected B cells but not thenumber of EBV DNA copies per B cell in healthy volunteers. J. Virol.83:11857–11861.

15. Hsu, M., S. Y. Wu, S. S. Chang, I. J. Su, C. H. Tsai, S. J. Lai, A. L. Shiau,K. Takada, and Y. Chang. 2008. Epstein-Barr virus lytic transactivator Ztaenhances chemotactic activity through induction of interleukin-8 in nasopha-ryngeal carcinoma cells. J. Virol. 82:3679–3688.

16. Islas-Ohlmayer, M., A. Padgett-Thomas, R. Domiati-Saad, M. W. Melkus,P. D. Cravens, P. Martin Mdel, G. Netto, and J. V. Garcia. 2004. Experi-mental infection of NOD/SCID mice reconstituted with human CD34� cellswith Epstein-Barr virus. J. Virol. 78:13891–13900.

17. Jaffe, E. S., L. Krenacs, S. Kumar, D. W. Kingma, and M. Raffeld. 1999.Extranodal peripheral T-cell and NK-cell neoplasms. Am. J. Clin. Pathol.111:S46–S55.

18. Jones, R. J., W. T. Seaman, W. H. Feng, E. Barlow, S. Dickerson, H. J.Delecluse, and S. C. Kenney. 2007. Roles of lytic viral infection and IL-6 inearly versus late passage lymphoblastoid cell lines and EBV-associated lym-phoproliferative disease. Int. J. Cancer 121:1274–1281.

19. Kalla, M., A. Schmeinck, M. Bergbauer, D. Pich, and W. Hammerschmidt.2010. AP-1 homolog BZLF1 of Epstein-Barr virus has two essential func-tions dependent on the epigenetic state of the viral genome. Proc. Natl.Acad. Sci. U. S. A. 107:850–855.

20. Kieff, E. D., and A. B. Rickinson. 2007. Epstein-Barr virus and its replication,p. 2603–2654. In D. M. Knipe and P. M. Howley (ed.), Fields’ virology, 5thed. Lippincott Williams & Wilkins, Philadelphia, PA.

21. Kilger, E., A. Kieser, M. Baumann, and W. Hammerschmidt. 1998. Epstein-Barr virus-mediated B-cell proliferation is dependent upon latent membraneprotein 1, which simulates an activated CD40 receptor. EMBO J. 17:1700–1709.

22. Kurth, J., M. L. Hansmann, K. Rajewsky, and R. Kuppers. 2003. Epstein-Barr virus-infected B cells expanding in germinal centers of infectious mono-nucleosis patients do not participate in the germinal center reaction. Proc.Natl. Acad. Sci. U. S. A. 100:4730–4735.

23. Mahot, S., A. Sergeant, E. Drouet, and H. Gruffat. 2003. A novel function forthe Epstein-Barr virus transcription factor EB1/Zta: induction of transcrip-tion of the hIL-10 gene. J. Gen. Virol. 84:965–974.

24. Melkus, M. W., J. D. Estes, A. Padgett-Thomas, J. Gatlin, P. W. Denton,F. A. Othieno, A. K. Wege, A. T. Haase, and J. V. Garcia. 2006. Humanizedmice mount specific adaptive and innate immune responses to EBV andTSST-1. Nat. Med. 12:1316–1322.

25. Miyazaki, I., R. K. Cheung, and H. M. Dosch. 1993. Viral interleukin 10 iscritical for the induction of B cell growth transformation by Epstein-Barrvirus. J. Exp. Med. 178:439–447.

26. Neuhierl, B., R. Feederle, W. Hammerschmidt, and H. J. Delecluse. 2002.Glycoprotein gp110 of Epstein-Barr virus determines viral tropism and ef-ficiency of infection. Proc. Natl. Acad. Sci. U. S. A. 99:15036–15041.

27. Niedobitek, G., A. Agathanggelou, H. Herbst, L. Whitehead, D. H. Wright,and L. S. Young. 1997. Epstein-Barr virus (EBV) infection in infectiousmononucleosis: virus latency, replication and phenotype of EBV-infectedcells. J. Pathol. 182:151–159.

28. Oudejans, J. J., M. Jiwa, A. J. van den Brule, F. A. Grasser, A. Horstman, W.Vos, P. M. Kluin, P. van der Valk, J. M. Walboomers, and C. J. Meijer. 1995.Detection of heterogeneous Epstein-Barr virus gene expression patternswithin individual post-transplantation lymphoproliferative disorders. Am. J.Pathol. 147:923–933.

29. Oyama, T., K. Ichimura, R. Suzuki, J. Suzumiya, K. Ohshima, Y. Yatabe, T.Yokoi, M. Kojima, Y. Kamiya, H. Taji, Y. Kagami, M. Ogura, H. Saito, Y.Morishima, and S. Nakamura. 2003. Senile EBV� B-cell lymphoprolifera-tive disorders: a clinicopathologic study of 22 patients. Am. J. Surg. Pathol.27:16–26.

30. Oyama, T., K. Yamamoto, N. Asano, A. Oshiro, R. Suzuki, Y. Kagami, Y.Morishima, K. Takeuchi, T. Izumo, S. Mori, K. Ohshima, J. Suzumiya, N.Nakamura, M. Abe, K. Ichimura, Y. Sato, T. Yoshino, T. Naoe, Y. Shi-moyama, Y. Kamiya, T. Kinoshita, and S. Nakamura. 2007. Age-relatedEBV-associated B-cell lymphoproliferative disorders constitute a distinctclinicopathologic group: a study of 96 patients. Clin. Cancer Res. 13:5124–5132.

31. Rajesh, D., Y. Zhou, E. Jankowska-Gan, D. A. Roenneburg, M. L. Dart, J.Torrealba, and W. J. Burlingham. 2010. Th1 and Th17 immunocompetencein humanized NOD/SCID/IL2r�null mice. Hum. Immunol. 71:551–559.

32. Rickinson, A. B., and E. D. Kieff. 2007. Epstein-Barr virus, p. 2655–2700. InD. M. Knipe and P. M. Howley (ed.), Fields’ virology, 5th ed. LippincottWilliams & Wilkins, Philadelphia, PA.

33. Ryan, J. L., H. Fan, S. L. Glaser, S. A. Schichman, N. Raab-Traub, and M. L.Gulley. 2004. Epstein-Barr virus quantitation by real-time PCR targetingmultiple gene segments: a novel approach to screen for the virus in paraffin-embedded tissue and plasma. J. Mol. Diagn. 6:378–385.

176 MA ET AL. J. VIROL.

34. Sarisky, R. T., Z. Gao, P. M. Lieberman, E. D. Fixman, G. S. Hayward, andS. D. Hayward. 1996. A replication function associated with the activationdomain of the Epstein-Barr virus Zta transactivator. J. Virol. 70:8340–8347.

35. Savard, M., C. Belanger, M. Tardif, P. Gourde, L. Flamand, and J. Gosselin.2000. Infection of primary human monocytes by Epstein-Barr virus. J. Virol.74:2612–2619.

36. Steven, N. M., N. E. Annels, A. Kumar, A. M. Leese, M. G. Kurilla, and A. B.Rickinson. 1997. Immediate early and early lytic cycle proteins are frequenttargets of the Epstein-Barr virus-induced cytotoxic T cell response. J. Exp.Med. 185:1605–1617.

37. Stewart, J. P., F. G. Behm, J. R. Arrand, and C. M. Rooney. 1994. Differentialexpression of viral and human interleukin-10 (IL-10) by primary B celltumors and B cell lines. Virology 200:724–732.

38. Strowig, T., C. Gurer, A. Ploss, Y. F. Liu, F. Arrey, J. Sashihara, G. Koo,C. M. Rice, J. W. Young, A. Chadburn, J. I. Cohen, and C. Munz. 2009.Priming of protective T cell responses against virus-induced tumors in micewith human immune system components. J. Exp. Med. 206:1423–1434.

39. Swaminathan, S., R. Hesselton, J. Sullivan, and E. Kieff. 1993. Epstein-Barr

virus recombinants with specifically mutated BCRF1 genes. J. Virol. 67:7406–7413.

40. Thorley-Lawson, D. A. 2005. EBV the prototypical human tumor virus—justhow bad is it? J. Allergy Clin. Immunol. 116:251–261.

41. Thorley-Lawson, D. A., and A. Gross. 2004. Persistence of the Epstein-Barrvirus and the origins of associated lymphomas. N. Engl. J. Med. 350:1328–1337.

42. van Kooten, C., and J. Banchereau. 2000. CD40-CD40 ligand. J. Leukoc.Biol. 67:2–17.

43. Yajima, M., K. Imadome, A. Nakagawa, S. Watanabe, K. Terashima, H.Nakamura, M. Ito, N. Shimizu, M. Honda, N. Yamamoto, and S. Fuji-wara. 2008. A new humanized mouse model of Epstein-Barr virus infec-tion that reproduces persistent infection, lymphoproliferative disorder,and cell-mediated and humoral immune responses. J. Infect. Dis. 198:673–682.

44. Yajima, M., K. Imadome, A. Nakagawa, S. Watanabe, K. Terashima, H.Nakamura, M. Ito, N. Shimizu, N. Yamamoto, and S. Fujiwara. 2009. Tcell-mediated control of Epstein-Barr virus infection in humanized mice.J. Infect. Dis. 200:1611–1615.

VOL. 85, 2011 EARLY LYTIC PROTEIN EXPRESSION IN LYMPHOMA DEVELOPMENT 177