CD40 ligand is a critical effector of Epstein–Barr virus in host cell survival and transformation

Abstract

Epstein–Barr virus (EBV), implicated in numerous human diseases, including lymphoid malignancies, persistently infects peripheral B cells and transforms them into lymphoblastoid cell lines. Here we found that EBV equally infected B cells from patients with X-linked hyper IgM syndrome and those from healthy donors; however, it hardly transformed X-linked hyper IgM syndrome B cells, because of the dysfunctional gene of CD40 ligand (CD40L) of the patients. Unlike CD40, CD40L is not usually expressed on B cells. However, we found that EBV infection of normal B cells induced CD40L expression as a critical effector in host cell transformation and survival. Moreover, chronic active EBV infection of peripheral T cells, implicated in T cell malignancies, was associated with ectopic expression of CD40, and, in Jurkat T cells, EBV infection induced CD40 expression. These results suggest that EBV infection induces CD40L/CD40 signaling in host cells, which appears to play an essential role in its persistent infection and malignancies of lymphocytes.

Epstein–Barr virus (EBV), a ubiquitous human lymphotropic herpesvirus, is a cause of lymphoproliferative diseases in immunosuppressed patients and infectious mononucleosis and is tightly associated with lymphoid malignancies such as Burkitt's lymphoma and T cell/natural killer cell lymphoma (1). EBV infection is also associated with epithelial malignancies such as nasopharyngeal carcinoma and gastric carcinoma. An important biological property of EBV, which rationalizes its tight link to cancer, is an ability to transform peripheral B cells in terms of their continuous growth in vitro and to establish latently infected lymphoblastoid cell lines (LCLs), which eventually become immortalized (1). LCLs express nine viral proteins: six EBV nuclear antigens (EBNA1–EBNA6) and three latent membrane proteins (LMP1, LMP2A, and LMP2B). Among them, an integral membrane protein, LMP1, is believed to be a key regulator of the B cell transformation, mainly because it transforms fibroblasts or epithelial cells and also induces B cell lymphoma in transgenic mice (1, 2). However, LMP1 expression is insufficient to maintain B cell proliferation, which needs, at least, a second signal (3).

CD40 is a membrane-bound protein of the tumor necrosis factor (TNF) receptor family and is expressed on many cell types including B cells. Its ligand, CD40 ligand (CD40L), is a member of the TNF family and expressed mainly on activated T cells. CD40–CD40L interaction is crucial to B cells for their proliferation, survival, Ig istotype switching, and germinal center reaction upon stimulation by activated T cells (4). For instance, mutations in the CD40L gene were identified as the cause of X-linked hyper IgM syndrome (XHIM), a disease associated with drastic, if not complete, inhibition in T cell-dependent humoral immune responses (4, 5). Mice null for CD40 or CD40L had severe defects not only in their Ig isotype switching, but also in germinal center formation and establishment of B cell memory (4, 6). That we had very few LCLs from XHIM B cells upon EBV infection led us to investigate whether CD40L and CD40 play a role in EBV infection and/or subsequent B cell transformation.

Materials and Methods

Reagents. For flow cytometry, mAbs to CD40 (5C3, PharMingen), CD40L (TRAP1, PharMingen), CD3 (Leu-4, PharMingen), and CD19 (HD37, DAKO) and isotype-matched control Ig (PharMingen) were used. For immunoblot analysis, mAb to LMP1 (S12, a gift from E. Kieff, Harvard Medical School, Boston) (7, 8) and a goat polyclonal antibody to β-actin (I-19, Santa Cruz Biotechnology) were used. For CD40 stimulation in LCL analysis, an agonistic mAb to CD40 (mAb89, Immunotech, Luminy, France) was used (9). For CD40L blocking, CD40Ig, a fusion protein of mouse CD40 (amino acids 1–193) and the Fc region of mouse IgG_2a, was used. The CD40Ig was expressed in Sf9 cells by using the baculovirus vector plasmid pFastBac-mCD40/mγ2a (a gift from M. R. Kehry, Boehringer Ingelheim Pharmaceuticals, Ridgefield, CT) and purified to homogeneity (>95%) with protein A-Sepharose (Amersham Pharmacia). The CD40Ig blocked human CD40L, but not IL-4, from stimulating peripheral blood B cell proliferation (data not shown).

Human Peripheral Blood Lymphocytes. Ethical approval was obtained from the ethical boards of the Department of Medicine and Medical Research Institute of Tokyo Medical and Dental University, and informed consent was obtained from all blood donors. B cells were isolated from peripheral blood mononuclear cells (PBMCs) with anti-CD19 Dynabeads M-450 (Dynal, Great Neck, NY) according to the manufacturer's directions. T cells were isolated from PBMCs with mAb to CD3 by using a cell sorter (FACSVantage, Becton Dickinson). Each preparation contained >98% CD19⁺ or CD3⁺ lymphocytes.

EBV Infection. EBV was prepared from culture medium of B95–8 cells as described (10) and concentrated (200-fold) in RPMI medium 1640 supplemented with 10% FCS. The virus suspension was filtered (0.45 μm) and recipient cells (2 × 10⁶ to 1 × 10⁷) were incubated in 1 or 5 ml of the suspension for 1 h, then rinsed twice with the culture medium (RPMI medium 1640 supplemented with 10% FCS). The efficiency of infection was >90% as judged by EBNA staining. For inactivation of the EBV genome, 1 ml of virus suspension in a 100-mm dish was irradiated with UV (254 nm) at 1 J/cm² by using a FUNA-UV-LINKER FS-800 (Funakoshi, Tokyo). For mock infection, culture medium was added instead of the EBV suspension.

LCL Assay. After EBV infection, B cells were cultured in 96-well plates (1 × 10³ cells per well or as indicated) for 4 weeks. Then, of 96 or 32 wells, the wells containing growing cells (LCL) were enumerated. No LCL was generated in the absence of EBV infection in this assay. In LCL-negative wells, no living cell was observed at 4 weeks postinfection, irrespective of B cell type (normal or XHIM), and another 4 weeks of extended culture did not generate any LCL in those LCL-negative wells. Half of the culture medium was replaced with fresh medium every 3 days.

EBNA Staining. Cytospin slides were prepared from EBV-infected or mock-infected B cells and fixed in acetone/methanol (1:1). Then, EBNAs were visualized by using anticomplement immunofluorescence as described (11). Briefly, the fixed B cells were first treated with human serum to EBNA containing complement (a gift from F. Mizuno, Tokyo Medical University, Tokyo). Then, they were washed and stained with FITC-labeled rabbit polyclonal antibody to human C3c complement (F0201, DAKO).

RT-PCR and Southern Blotting. Total RNA of cells was prepared with ISOGEN (Nippon Gene, Toyama, Japan). RT-PCR was performed with 5 μg total RNA by using the SuperScript OneStep RT-PCR system (Invitrogen). Conditions for the RT reaction and PCR were optimized to produce a specific cDNA product for each gene tested. Primers used were: 5′-TGC CAG CCA GGA CAG AAA CT-3′ (CD40 sense) and 5′-GGG ACC ACA GAC AAC ATC AG-3′ (CD40 antisense), 5′-TGC GGC ACA TGT CAT AAG-3′ (CD40L sense) and 5′-CGG AAC TGT GGG TAT TT-3′ (CD40L antisense). PCR products were subjected to Southern blotting with the following oligonucleotide probes: 5′-AAG AAG GCT GGC ACT GTA-3′ (CD40 probe) and 5′-CAT CTG TGT TAC AGT GGG-3′ (CD40L probe). RT-PCR primers and a probe for EBNA1 have been described (12). RT-PCR primers for β-actin (Takara, Osaka) were purchased.

Apoptosis Assay. After EBV infection, cells were plated to a 96-well plate (1 × 10⁴ cells per well) and cultured for 0, 24, or 48 h in the presence or absence of the reagent(s) indicated in each assay. These cells were stained with acridine orange (2 μg/ml) and ethidium bromide (2 μg/ml) and examined under a fluorescent microscope for counting apoptotic cells, which were identified by nuclear morphology according to an established protocol (13). An apoptotic index was calculated from the results of counting 200 cells in total. Apoptotic cells were otherwise detected as indicated with an ApopTag ISOL kit (Intergen, Purchase, NY), which probes double-strand DNA breaks with a hairpin oligo-DNA probe and T4-DNA ligase.

EBV DNA Assay. Copy numbers of EBV DNA per cellular DNA (μg) in CD3⁺/CD40^- or CD3⁺/CD40^+/- chronic active EBV infection (CAEBV) T lymphocytes were determined by the real-time PCR method with a LightCycler DNA Master Hybridization Probes system (Roche Diagnostics) by using a set of primers, 5′-CGC ATA ATG GCG GAC CTA-3′ and 5′-CAA ACA AGC CCA CTC CCC-3′, and a pair of fluorogenic probes, LC-5′-AAC CAT AGA CCC GCT TCC TG-3′, and 5′-AAA GAT AGC AGC AGC GCA GC-3′-FITC for fluorescence resonance energy transfer-based detection (14). These oligonucleotides were selected from the BWRF1 gene of EBV (1).

Results and Discussion

CD40L Is Indispensable for EBV-Induced B Cell Transformation. Trying to establish LCLs from peripheral B cells of XHIM patients carrying the dysfunctional CD40L gene, we found that EBV infection hardly transformed those B cells into LCLs (Fig. 1 A and B). This was not caused by a low efficiency of infection because EBV infected XHIM and normal B cells equally and expressed comparable amounts of LMP1 (Fig. 1 C and D). Furthermore, additional CD40 stimulation restored the transformation of XHIM B cells upon EBV infection (Fig. 1E), indicating a critical role for CD40L and CD40 signaling in EBV-mediated B cell transformation. Indeed, CD40Ig, a specific antagonist to CD40L (15), severely impaired transformation of normal peripheral B cells upon EBV infection (Fig. 1F). It is also reported that XHIM patients appeared at high risk of developing cytomegalovirus infection but not infection of other herpesviruses, including EBV (16). In contrast, EBV activities were frequently observed in patients with other primary or secondary immune deficiencies (17), implying a role for CD40L in EBV infection. Because peripheral B cells from healthy donors don't express CD40L, we hypothesized that EBV infection induces ectopic expression of functional CD40L on B cells as a critical effector in cellular transformation.

Fig. 1.

CD40L is indispensable for EBV-induced B cell transformation. (A) EBV-infected (solid line) or mock-infected (dotted lines) XHIM peripheral B cells were subjected to flow cytometry with the indicated mAb at 24 hpi. (B) XHIM peripheral B cells generated only marginal numbers of LCLs upon EBV infection in contrast to normal B cells. Data represent mean values and standard errors from sextuplicate experiments of LCL assay. (C) EBV equally infected peripheral B cells of XHIM patients and those of healthy donors (Normal). EBNAs were stained with anti-EBNA human serum at 24 hpi, and each staining, in total, showed 95% (122 of 128) or 94% (118 of 126) EBNA-positive (EBNA⁺) cells in EBV-infected cells. (D) LMP1 or β-actin in EBV-infected XHIM or normal peripheral B cells at the indicated hpi was examined by immunoblotting. (E) Stimulation of CD40 with an agonistic mAb, mAb89 (10 μg/ml), restored the EBV-induced transformation of XHIM peripheral B cells to LCLs (CD40 mAb). Data represent mean values and standard errors from triplicate experiments of LCL assay. (F) Blocking of CD40L with CD40Ig (200 μg/ml) inhibited transformation of normal peripheral B cells to LCLs. Data represent mean values and standard errors from sextuplicate experiments of LCL assay.

EBV Infection Induces Ectopic Expression of CD40L on B Cells. To test our hypothesis, we first studied EBV-negative Ramos B cells upon EBV infection and found that CD40L mRNA was induced as early as 24 h postinfection (hpi) (Fig. 2A). Surface expression of CD40L was also confirmed at 48 hpi (Fig. 2B). However, EBV with a UV-inactivated genome induced neither CD40L nor a viral gene, EBNA1, indicating that expression of viral gene(s) is required for CD40L induction (Fig. 2 A). CD40L mRNA expression was also obvious in Ramos(EB) and Bjab(EB) cell lines, which had been rendered EBV-positive by the viral infection of parental Ramos and Bjab cells (Fig. 2C), suggesting long-term expression of CD40L on EBV-positive B cells (18). Consistently, three of four EBV-positive B cell lines (Raji, Akata, AG876, and P3HR1) derived from Burkitt's lymphomas and two LCLs (LCL-MS and LCL-TM) from normal peripheral B cells expressed CD40L (Fig. 2 D and E). In addition, when human peripheral B cells were infected with EBV, surface expression of CD40L was induced as early as 24 hpi (Fig. 2F). These results indicate that EBV infection of B cells induces ectopic expression of CD40L on their surface.

Fig. 2.

EBV infection induces ectopic expression of CD40L on B cells. (A) Expression of CD40, CD40L, or EBNA1 mRNA in EBV-infected Ramos B cells at the indicated hpi was examined by RT-PCR followed by Southern blotting (EBV). EBV with an UV-inactivated genome induced neither CD40L nor EBNA1 mRNA (EBV+UV). Expression of β-actin mRNA was examined by RT-PCR and visualized by ethidium bromide staining. (B) Flow cytometry with mAb to CD40L shows its surface expression on EBV-infected (solid line), but not mock-infected (dotted line), Ramos cells at 48 hpi. (C) Expression of the indicated mRNA in the EBV-positive B cells, Bjab(EB) and Ramos(EB), and their EBV-negative parental cells, Bjab and Ramos, was examined as described in A.(D) Expression of the indicated mRNA in EBV-positive B cells or EBV-negative B or T cells (Ramos or Jurkat) was examined as described in A. (E) Surface expression of CD40L was examined by flow cytometry with mAb to CD40L (solid line) or isotype-matched control (dotted line). (F) Surface expression of CD40L and CD19 on EBV-infected (solid line) or mock-infected peripheral B cells (dotted line) was examined at 24 hpi.

EBV-Induced CD40L Inhibits B Cell Apoptosis. Encountering antigens through its receptor, B cells need additional signals from their CD40 interacting with CD40L on activated T cells to escape from apoptosis. Thus, we examined whether EBV-induced CD40L inhibits apoptosis of LCL-MS cells and found that blocking of CD40L with CD40Ig, but not the control Ig, up-regulated the Ca²⁺-induced apoptosis (Fig. 3A). In contrast, neither reagent influenced the apoptosis of CD40L-negative Ramos cells, indicating that CD40L plays an antiapoptotic role in the EBV-positive LCL-MS cells. Then, we examined whether this is the case in peripheral blood B cells. Although those B cells underwent apoptosis when cultured in vitro, EBV infection inhibited the apoptosis and suppressed the reduction in live cell numbers as early as 24 hpi, followed by a moderate proliferation (Fig. 3 B and C). However, blocking of CD40L with CD40Ig, but not the control Ig, abrogated the antiapoptotic function of EBV (Fig. 3D). Peripheral B cells lacking CD40L from XHIM patients also showed no inhibition of apoptosis, irrespective of CD40Ig, upon EBV infection (Fig. 3 E and F). In addition, CD40Ig treatment increased normal peripheral B cells carrying double-strand DNA breaks, which are characteristic of apoptotic cells, upon EBV infection (Fig. 3G). Together, these results indicate that CD40L induced by EBV infection plays an essential role in the inhibition of host cell apoptosis, thereby facilitating persistent infection of EBV and host cell transformation.

Fig. 3.

EBV-induced CD40L inhibits B cell apoptosis. Apoptotic indices are shown as mean values with standard errors from sextuplicate experiments of apoptosis assay. (A) Blocking of CD40L with CD40Ig (200 μg/ml) up-regulated apoptosis of LCL-MS cells, but not Ramos cells, in the presence of calcium ionophore A23178 (1 μM) for 24 h. (B) EBV infection inhibited apoptosis of peripheral B cells. (C) Live cell numbers in each culture well of EBV-infected (EBV) or mock-infected peripheral B cells in B were counted by standard trypan blue staining. Data represent mean values and standard errors from sextuplicate experiments. (D) Blocking of CD40L with CD40Ig (200 μg/ml) up-regulated apoptosis of EBV-infected (EBV) peripheral B cells to the level for mock-infected B cells (Left). Control Ig did not affect the EBV-mediated inhibition of B cell apoptosis (Right). (E) EBV-infected and mock-infected XHIM peripheral B cells showed the same level of apoptosis. (F) Neither CD40Ig (200 μg/ml) nor control Ig influenced apoptosis of EBV-infected XHIM peripheral B cells. (G) Apoptotic double-strand breaks of cellular DNA (brown spots) upon EBV infection of peripheral B cells in the presence of CD40Ig (200 μg/ml) or control Ig were probed at 24 hpi. (Left) Total of 76 cells. (Right) Total of 71 cells.

EBV Infection of T Cells Is Associated with Ectopic Expression of CD40. Although EBV is a B lymphotropic agent, its infection of T cells or natural killer cells is seen in patients with T cell or natural killer cell type CAEBV, which is often associated with clonal expansion of the infected cells and consequent malignancies (1, 19). Interestingly, ectopic expression of CD40 was found in EBV-positive T cell lines (SKN-P, SIS) derived from T cell type CAEBV patients (Fig. 4 A and B) (20). Furthermore, we found that a fraction of the peripheral T cells of T cell type CAEBV patients expressed CD40 on their surface (Fig. 4C), whereas those of healthy donors were negative for it (data not shown). Because a majority of T cells are not infected with EBV even in T cell type CAEBV patients (1, 19), we further divided those peripheral T cells into CD40-negative or CD40-enriched fractions and found that the latter T cells have a high viral load (Fig. 4 C and D). In contrast, T cells of the CD40-negative fraction showed undetectable levels of EBV DNA, demonstrating a tight link between EBV infection and ectopic expression of CD40 on peripheral T cells. In addition, T cells of the CD40-enriched fraction, but not of the CD40-negative one, showed up-regulation of Ca²⁺-induced apoptosis in the presence of CD40Ig (Fig. 4D). These results suggest that EBV infection induces the expression of functional CD40 on peripheral T cells, which may contribute to the clonal expansion of host cells and consequent malignancies of CAEBV patients. To test directly whether EBV infection induces CD40 expression on T cells, we studied Jurkat T cells upon EBV infection and found that it did in terms of mRNA and surface protein expression. (Fig. 4 E and F). EBV with a UV-inactivated genome induced neither CD40 nor EBNA1, indicating that viral gene expression is required for CD40 expression (Fig. 4E).

Fig. 4.

EBV infection of T cells is associated with ectopic expression of CD40. (A) Expression of the indicated mRNA in the EBV-positive T cells (SKN-P and SIS) and the EBV-negative T (Jurkat) and B (Ramos) cells was examined as described in Fig. 2 A. (B) Surface expression of CD40 was examined by flow cytometry with mAb to CD40 (solid line) or isotype-matched control Ig (dotted line). (C) Flow cytometry with mAb to CD40 (solid line, Left) shows its surface expression on a fraction of peripheral T cells of CAEBV patients. Isotype-matched Ig was used for control (dotted line, Left). These CAEBV T cells were sorted into CD40-negative (-) and CD40-enriched (±) fractions, which were monitored by additional flow cytometry (Right). (D) Peripheral T cells of the CD40-enriched fraction (±) prepared in C contained high copy numbers of EBV DNA and showed up-regulation of apoptosis with CD40Ig (200 μg/ml) in the presence of Ca²⁺ ionophore A23178 (1 μM) for the indicated periods (Right). Peripheral T cells of the CD40-negative fraction (-) contained undetectable levels of EBV DNA, and their apoptosis was insensitive to CD40Ig (Left). Apoptotic indices are shown as mean values with standard errors from sextuplicate experiments of apoptosis assay. (E) EBV infection induces CD40 mRNA expression in Jurkat T cells. Expression of the indicated mRNA in EBV-infected Jurkat T cells at the indicated hpi was examined as described in Fig. 2 A (EBV). EBV with an UV-inactivated genome induced neither CD40 nor EBNA1 mRNA (EBV+UV). (F) Flow cytometry with mAb to CD40 shows its surface expression on EBV-infected (solid line), but not mock-infected (dotted line), Jurkat T cells at 48 hpi.

Implications. It is widely accepted that LMP1 plays an essential role in B cell activation and transformation by triggering an intracellular signaling cascade similar to that downstream of CD40, which recruits the TRAF family of signal mediators (1, 4, 6, 21, 22). For example, both LMP1 and CD40 recruit TRAF2, TRAF3, and TRAF5 to induce activation of c-Jun N-terminal kinase and NF-κB. However, as reported here, CD40L/CD40 signaling is still required for the antiapoptotic function of EBV and B cell transformation even in the presence of LMP1 upon EBV infection. There are reports that LMP1 cannot fully compensate for a deficiency of CD40 in mice and that LMP1 and CD40 do not interact with exactly the same sets of signaling molecules, indicating that their signaling pathways differ in some respects (6, 22, 23). Thus, we propose that EBV-induced CD40L/CD40 signaling cooperates with LMP1 in B cell survival and transformation, as they can cooperate for B cell activation (24). On the other hand, it is noteworthy that several autoimmune diseases, such as systemic lupus erythematosus, are associated with EBV infection and the ectopic expression of CD40L on B cells (25, 26). Because transgenic mice expressing CD40L on their B cells display symptoms of lupus-like disease, EBV-induced CD40L may be one of the etiologic agents for such human autoimmune diseases (27).

We have demonstrated that EBV infection of B cells induces expression of CD40L as a crucial effector in host cell survival and transformation and that EBV infection of T cells is associated with ectopic expression of CD40 in CAEBV patients. It was recently shown that constitutive expression of CD40L on B cells can stimulate CD40-mediated growth signals (4, 27–30), suggesting that EBV-induced CD40L facilitates B cell proliferation as well. The CD40L expression on LCLs and the other EBV-positive B cell lines suggests long-term expression of EBV-induced CD40L. Therefore, CD40L might play a role not only in the establishment but also in the maintenance of persistent infection of EBV, which would last for a lifetime, and it might contribute to B cell malignancies or other EBV-associated diseases during this period. Understanding how CD40L/CD40 signaling is induced and works on EBV infection of lymphocytes may provide a novel therapeutic target.