Functional Analysis of the CD4⁺ T-Cell Response to Epstein-Barr Virus: T-Cell-Mediated Activation of Resting B Cells and Induction of Viral BZLF1 Expression

Abstract

In contrast to the major role played by Epstein-Barr virus (EBV)-specific CD8⁺ cytotoxic T-cell responses in immunosurveillance, recent studies have offered the apparently paradoxical suggestion that development of EBV-driven human B-cell lymphoproliferative disorders and tumors in SCID/hu mice is dependent on the presence of T cells, in particular CD4⁺ T cells. This study presents a functional analysis of the CD4⁺ T-cell response to EBV and shows that while CD4⁺ T cells may be cytotoxic, they also express Th2 cytokines and CD40 ligand (gp39) and possess B-cell helper function. We show that EBV-specific CD4⁺ T cells can provide non-HLA-restricted help for activation of resting B cells via a gp39-CD40-dependent pathway and are able to induce expression of BZLF1, a viral lytic cycle transactivator in latently infected resting B cells, ultimately resulting in rapid outgrowth of transformed B-cell colonies. These results support the proposal that CD4⁺ T cells may play a key role in reactivation of latent EBV infection and may thus contribute to the pathogenesis of EBV-driven lymphoproliferative disorders.

There is a general consensus that HLA class I-restricted CD8⁺ cytotoxic T-cell responses play a major role in the control of latent Epstein-Barr virus (EBV) infection, by virtue of their ability to recognize viral antigens associated with B-cell transformation (9, 19). While EBV-specific CD8⁺ T-cell responses have thus been widely studied, little is known of the CD4⁺ T-cell response to EBV. Virus-specific CD4⁺ T cells that lyse in vitro-transformed lymphoblastoid cell lines (LCL) have been described (12–14, 17), and it has recently been shown that CD4⁺ T cells can induce Fas-mediated apoptosis in LCL (23), suggesting a possible role for CD4⁺ T cells in EBV-specific immunosurveillance.

In contrast to the acknowledged importance of EBV-specific cytotoxic T-cell responses in maintenance of asymptomatic latent infections in healthy individuals, recent studies have presented the apparently paradoxical finding that EBV-driven tumorigenesis in SCID/hu mice is dependent on the presence of T cells (5, 22). The work of Veronese and colleagues showed that both CD4⁺ and CD8⁺ T cells were capable of fulfilling the helper function necessary for the development of EBV-induced tumors in SCID/hu mice, with CD4⁺ T cells being the more efficient (22). These observations support the proposal that helper CD4⁺ T cells can play an important role in the pathogenesis of EBV-associated disease, through non-HLA-restricted bystander activation of resting B cells and resultant reactivation of latent EBV infection. In this report, we show that EBV-specific CD4⁺ T-cell function in a healthy individual covers a broad spectrum, encompassing cytotoxic responses and expression of both Th1 and Th2 cytokines. In particular, we demonstrate that EBV-specific CD4⁺ T cells can provide highly efficient help for activation of resting B cells, and we further show that CD4⁺-T-cell-mediated activation of latently EBV-infected peripheral blood resting B cells through a CD40-dependent pathway induces expression of BZLF1, an immediate-early transactivator of viral lytic cycle replication.

Derivation and characterization of polyclonal EBV-specific CD4⁺ T cells.

CD4⁺ T cells were purified from the peripheral blood lymphocytes (PBL) of a healthy, EBV-seropositive, adult donor (donor 1) by positive selection with anti-CD4-coupled magnetic beads (Dynabeads; Dynal A.S., Lake Success, N.Y.). EBV-specific CD4⁺ T cells were selected by coculture with autologous, irradiated (7,500 cGy) LCL (grown in the presence of 10% human AB serum, rather than fetal calf serum, to avoid stimulation of a CD4⁺ T-cell response to bovine proteins) and maintained as a polyclonal line by periodic restimulation with LCL in RPMI 1640 plus 10% human AB serum, supplemented with 50 U of recombinant human interleukin 2 (IL-2) (provided by the Biological Response Modifiers Program, National Cancer Institute, Bethesda, Md.) per ml. Donor 1 CD4⁺ T cells strongly proliferated in response to stimulation with autologous donor 1 LCL but failed to respond to allogeneic LCL from an unrelated HLA class II-mismatched donor (donor 2). Furthermore, donor 1 CD4⁺ T cells failed to recognize autologous, normal B lymphoblasts activated with anti-CD40 and IL-4 according to the protocol of Banchereau et al. (1). Collectively, these results confirmed that the donor 1 CD4⁺ T-cell response was EBV specific. Donor 1 CD4⁺ T cells were cytotoxic to autologous LCL (21% specific lysis at an effector/target ratio of 20:1) but showed minimal cytotoxicity to autologous anti-CD40-activated normal B lymphoblasts, allogeneic LCL, or NK-sensitive K562 cells (data not shown).

Intracellular cytokine expression by EBV-specific CD4⁺ T cells.

Two-color flow cytometric analysis allows determination of cytokine expression at the single-cell level (15). CD4⁺ T cells (7.5 × 10⁵/ml) were activated by incubation at 37°C for 6 h in RPMI 1640–10% human AB serum plus 50 ng of phorbol myristyl acetate (PMA) per ml and 500 ng of ionomycin per ml. Brefeldin A at 10 μg/ml was added for the final 3 h of incubation. Control, nonactivated cultures were incubated in the presence of brefeldin A only. The cells were harvested, washed, and fixed with 2% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at room temperature. For intracellular staining, the cells were washed and permeabilized by incubation in PBS plus 1% bovine serum albumin (BSA) and 0.5% saponin (S-7900; Sigma, St. Louis, Mo.) for 10 min at room temperature. Monoclonal antibodies (MAb) used were fluorescein isothiocyanate (FITC)-conjugated anti-gamma interferon (anti-IFN-γ), FITC–anti-IL-2, FITC–anti-tumor necrosis factor alpha (FITC–anti-TNF-α), phycoerythrin (PE)-conjugated anti-IL-4, and PE–anti-IL-13 (all from Becton Dickinson, San Jose, Calif.). After staining for 30 min at room temperature, cells were washed twice with PBS plus 1% BSA and 0.5% saponin, washed once with PBS plus 0.5% BSA, and fixed a second time with 2% paraformaldehyde in PBS.

Coordinate analysis of IFN-γ and IL-4 expression revealed three distinct but overlapping subsets (Fig. 1A); 35% of CD4⁺ T cells were IFN-γ^hi IL-4^lo, akin to classically described Th1 cells, and 12% were IFN-γ^lo IL-4^hi, typical of Th2 cells. However, the major population (49%) was of an intermediate, double-positive IFN-γ^hi IL-4^hi phenotype. Broadly similar patterns were revealed by coordinate analysis of IL-4 expression versus TNF-α and IL-2 expression, with a double-positive population being the predominant phenotype in each case (Fig. 1C and E). We also recorded the distribution of CD4⁺ T cells that expressed IL-13, a Th2 cytokine that has many of the functional characteristics of IL-4, including the ability to induce proliferation of B cells activated by CD40 engagement (4, 7). Two-color analysis of IL-13 and IFN-γ expression by CD4⁺ T cells again suggested three distinct functional subsets (Fig. 1B), as seen with the IFN-γ–IL-4 analysis, but in this instance only 22% of T cells possessed an IFN-γ^hi IL-13^lo Th1 phenotype, whereas the double-positive IFN-γ^hi IL-13^hi population constituted 62% of the total. Overall, 75 to 80% of CD4⁺ T cells were IL-13 expressers, and a majority of these cells also expressed TNF-α and IL-2 (Fig. 1D and F).

FIG. 1.

Two-color flow cytometric analysis of intracellular expression of IL-4 versus IFN-γ, IL-2, and TNF-α, and intracellular expression of IL-13 versus IFN-γ, IL-2, and TNF-α by EBV-specific donor 1 CD4⁺ T cells. T cells were stimulated for 6 h with PMA and ionomycin, with brefeldin A added for the final 3 h. In all cases, control, unstimulated T cells failed to stain for intracellular cytokine expression.

CD40 ligand (gp39) expression by EBV-specific CD4⁺ T cells.

T-cell-mediated activation of resting B cells requires T cell expression of gp39 and subsequent engagement of B cell CD40. Coordinate expression of cell surface gp39 and intracellular IL-4 or IFN-γ was measured by flow cytometry, using the activation protocol described above for intracellular cytokine assays. In this case, activated and control (nonactivated) T cells were stained with FITC–anti-gp39 MAb or an isotype-matched FITC-conjugated control MAb (both from Ancell Corp., Bayport, Minn.) following PMA and ionomycin activation but prior to the fixation and permeabilization steps necessary for detection of intracellular cytokine expression. Anticytokine antibodies used were PE–anti-IL-4 and PE–anti-IFN-γ (both from Becton Dickinson). EBV-specific CD4⁺ T cells expressed gp39 in an activation-dependent manner (Fig. 2). The gp39⁺ IL-4⁺ subset had a mean fluorescence intensity for IL-4 expression of 37.8, whereas the gp39⁻ IL-4⁺ subset had an IL-4 mean fluorescence intensity of only 20.9, suggesting that gp39 expression may coincide with high IL-4 expression (Fig. 2A). A subset (30%) of IFN-γ-expressing CD4⁺ T cells also expressed gp39 (Fig. 2B).

FIG. 2.

Two-color flow cytometric analysis of CD40 ligand (gp39) expression versus intracellular expression of IL-4 and IFN-γ. T cells were stimulated for 6 h with PMA and ionomycin, with brefeldin A added for the final 3 h.

Activation of resting B cells by EBV-specific CD4⁺ T cells.

The high frequency of IL-4 and IL-13-expressing EBV-specific CD4⁺ T cells, coupled with activation-dependent gp39 expression, suggested that these T cells were likely to provide efficient help for activation of resting B cells. Purified resting B cells from the donor 1 or 2 were prepared by positive selection from PBL using anti-CD19 Dynabeads (Dynal A.S.). B cells were of 95 to 98% purity, as determined by flow cytometric analysis, and residual T-cell contamination was less than 1%. EBV-specific CD4⁺ T cells (5 × 10⁴/well) were irradiated (3,000 rads) and cocultured with B cells (2 × 10⁴/well) in 96-well flat-bottom plates (Falcon; Becton Dickinson, Lincoln Park, N.J.) precoated with anti-CD3 MAb (50 μg of OKT3 per ml in PBS, overnight at 4°C). Where appropriate, blocking anti-gp39 MAb was used at 50 μg/ml. B-cell activation and proliferation were measured by [³H]thymidine ([³H]TdR) incorporation after 4 days. The results illustrated in Fig. 3 represent one of two experiments conducted with EBV-specific CD4⁺ T cells from donor 1. The same results were also seen for B-cell activation experiments with EBV-specific CD4⁺ T cells from two other donors. Coculture of T cells with resting B cells in the presence of plate-bound anti-CD3 MAb showed that EBV-specific CD4⁺ T cells were capable of providing efficient help for B-cell activation and proliferation (Fig. 3). The presence of anti-gp39 blocking MAb markedly reduced the B-cell stimulation index (significant at a P value of <0.01, Wilcoxon two-sample test), indicating that non-HLA-restricted T-cell-mediated B-cell activation was strongly dependent on gp39-CD40 interaction. B-cell activation was also dependent on helper CD4⁺-T-cell activation, as coculture of T and B cells in the absence of plate-bound anti-CD3 MAb failed to induce B-cell activation (not shown).

FIG. 3.

Bystander activation of resting B cells by EBV-specific donor 1 CD4⁺ T cells. Resting B cells were purified from peripheral blood by positive selection with anti-CD19-coupled magnetic beads and then cultured for 4 days with irradiated CD4⁺ T cells in the presence of plate-bound anti-CD3, in the presence or absence of blocking anti-gp39 MAb (50 μg/ml). Stimulation indices were calculated as the ratio of [³H]TdR uptake by B cells cultured with T cells divided by [³H]TdR uptake by B cells cultured alone. Background counts in control wells of irradiated T cells were subtracted prior to calculation of stimulation indices for B-cell proliferation.

Does CD4⁺-T-cell-mediated activation of latently infected resting B cells result in reactivation of the EBV lytic cycle?

Purified resting B cells (5 × 10⁶/well) prepared following leukapheresis of donor 1 were cultured with EBV-specific donor 1 CD4⁺ T cells (5 × 10⁶/well) for 48 h in 8 ml of RPMI 1640–10% human AB serum per well in six-well plates (Costar Corp., Cambridge, Mass.) precoated with anti-CD3 MAb (50 μg of OKT3 per ml in PBS, overnight at 4°C). The cells were then spun down, and the RNA was extracted with TRIzol (1 ml/well; Gibco BRL, Grand Island, N.Y.) and subsequently prepared as described previously (16). T-B coculturing was conducted in the presence or absence of blocking anti-gp39 MAb. Purified B cells were also incubated with anti-CD40 plus IL-4 as a positive control for B-cell activation (1). As the frequency of resting B cells latently infected with EBV is in the range of only 1 to 100 per 10⁶ B cells, we utilized highly sensitive reverse transcription (RT)-PCR assays capable of detecting very rare viral mRNA species. Nested sets of primers were used for amplification of BZLF1 and EBNA2 sequences from cDNA, as described by Prang et al. (16) and Chen et al. (3), respectively. The outer BZLF1 primer sequences were 5′-ATTGCACCTTGCCGCCACCTTTG-3′ and 5′-CGGCATTTTCTGGAAGCCACCCGA-3′, while the nested BZLF1 primer sequences were 5′-GACCAAGCTACCAGAGTCTAT-3′ and 5′-CAGAATCGCATTCCTCCAGCGA-3′. The outer EBNA2 primer sequences were 5′-ACTTTGAGCCACCCACAGTAACCA-3′ and 5′-TGGAGTGTCTGACAGTTGTTCCTG-3′. The nested EBNA2 primer sequences were 5′-CATAGAAGAAGAAGAGGATGAAGA-3′ and 5′-GTAGGGATTCGAGGGAATTACTGA-3′. Viral mRNA transcripts are spliced, and each primer pair utilizes sequences from different exons. This strategy enables us to distinguish between PCR products amplified from cDNA and PCR products amplified from possible contaminating genomic DNA, which would include intron sequences (3, 16) and thus give rise to significantly larger PCR products (Table 1). For qualitative and quantitative control of RNA and cDNA preparations, we used RT-PCR for histone 3.3 (a housekeeping gene), as described previously (16).

TABLE 1.

PCR conditions^a

Primer	Mg2+ (mmol/liter)	Annealing temp (°C)	No. of cycles	Size (bp) of:
				cDNA product	Genomic DNA product
BZLF1 outer	2	59	40	608	732
BZLF1 inner	2	60	45	442	566
EBNA2 outer	2	61	45	338	702
EBNA2 inner	2	58	45	271b	635

^aDenaturation times and temperatures: 20 s at 94°C for the BZLF1 primers and 15 s at 95°C for the EBNA2 primers. Annealing times: 40 s for the BZLF1 primers and 30 s for the EBNA2 primers. Extension times and temperatures: 40 s at 72°C for the BZLF1 primers and 30 s at 72°C for the EBNA2 primers. 

^bBased on the EBNA2 cDNA sequence for EBV B95-8, published by Sample et al. (21). Using the same primers, Chen et al. have predicted a PCR product of 267 bp (3). 

Nested RT-PCR analysis of RNA harvested after 48 h showed that BZLF1 mRNA expression was readily detectable in B cells cocultured with activated CD4⁺ T cells, whereas BZLF1 expression could not be detected in B cells cultured alone (Fig. 4A). The fidelity of the BZLF1 RT-PCR product from the T-B coculture was confirmed by TA cloning (Invitrogen, San Diego, Calif.) and subsequent DNA sequencing (data not shown). CD4⁺-T-cell-induced expression of BZLF1 was completely blocked in the presence of 50 μg of anti-gp39 MAb per ml. While blockade of gp39-CD40 interaction inhibited T-cell-mediated induction of viral lytic cycle activation in latently infected B cells, we also found that BZLF1 expression could be strongly induced by the combination of an agonist anti-CD40 MAb (Immunotech, Westbrook, Maine) and recombinant human IL-4 (Fig. 4A). EBNA2 mRNA expression was not detected by nested RT-PCR in either resting B cells cultured alone or CD4⁺-T-cell-activated B cells (data not shown). RT-PCR for histone 3.3 mRNA sequences gave homogeneous band intensities for all samples, indicating that variations in the levels of BZLF1 expression could not be attributed to variations in RNA or cDNA sample sizes (Fig. 4B).

FIG. 4.

Induction of BZLF1 mRNA expression in resting B cells following CD4⁺-T-cell-mediated activation. Purified resting B cells from the peripheral blood of donor 1 were cultured alone or with EBV-specific donor 1 CD4⁺ T cells (in the presence of plate-bound anti-CD3) or were activated with anti-CD40 MAb plus IL-4. Total RNA was extracted after 48 h, followed by nested RT-PCR for BZLF1 (A) and RT-PCR for histone 3.3 (B). RNA harvested from B95-8 cells was used as a positive control for BZLF1 expression. Nested RT-PCR for BZLF1 yielded a 442-bp product from the inner primers (Table 1), and RT-PCR for histone 3.3 gave a 215-bp product.

Activation of resting B cells with irradiated, anti-CD3-activated CD4⁺ T cells also resulted in rapid outgrowth of EBV-transformed B cells, i.e., “spontaneous” LCL, within 21 days (data not shown). As B-cell activation initially induces BZLF1 expression rather than transformation-associated transcripts such as EBNA2 mRNA, this observation supports the two-step model of EBV transformation (18, 24), involving viral replication and secondary infection rather than direct outgrowth of activated B cells.

Reactivation of latent EBV and outgrowth of transformed LCL have long been held to be spontaneous events (18, 24). However, more recent studies with the SCID/hu mouse model of spontaneous EBV-associated lymphoproliferative disorders have revealed a T-cell requirement for tumorigenesis. SCID/hu mice reconstituted with T-cell-depleted PBL (5, 22) or in which T-cell activation was inhibited with cyclosporine (2) failed to develop EBV-associated human B-cell tumors. Collectively, these observations strongly suggest that spontaneous reactivation does not take place in vivo and that T-cell interaction with latently infected B cells is a requirement for viral reactivation. As the work of Veronese and colleagues showed that CD4⁺ T cells were highly effective at restoring the generation of EBV-driven human B-cell tumors (22), we conducted a functional analysis of interactions between CD4⁺ T cells and B cells. Although we show that EBV-specific CD4⁺ T cells can be cytotoxic to autologous LCL (and potentially contribute to EBV-specific immunosurveillance), we also found that they can express Th2 cytokines, such as IL-4 and IL-13, and that they are efficient mediators of bystander activation of resting B cells. Furthermore, we found that helper CD4⁺-T-cell-mediated B-cell activation induces expression of viral BZLF1, an immediate-early transactivator of EBV lytic cycle replication. Our subsequent finding that CD4⁺-T-cell activation of latently infected resting B cells leads to rapid establishment of LCL transformed by endogenous virus supports the proposal that T cells play a potentially important role in viral reactivation and tumorigenesis in vivo.

EBV-specific helper CD4⁺ T-cell responses to productively infected or transformed B cells may result in bystander activation of other latently infected resting B cells, which would lead to further viral reactivation and progression of disease. However, T-cell interaction with latently infected resting B cells, and consequent viral reactivation, is non-HLA restricted and noncognate, i.e., there is not a requirement for EBV-specific antigen recognition. In the SCID/hu mouse, activation of CD4⁺ helper T cells of any specificity (e.g., through graft-versus-host responses), and concomitant bystander activation of latently infected B cells, is likely to be sufficient for viral reactivation and subsequent tumorigenesis. Conversely, nonspecific blockade of T-cell activation would cut off the required help for B-cell activation and thus prevent tumorigenesis, as is the case in SCID/hu mice treated with cyclosporine (2).

CD4⁺-T-cell-mediated B-cell activation and consequent expression of BZLF1 could be blocked with a MAb specific for T-cell-expressed gp39 (CD40 ligand). We also show that direct engagement of B-cell CD40 with an agonist anti-CD40 MAb (in conjunction with IL-4 treatment) results in strong induction of BZLF1 expression. CD40 early signaling events involve protein kinase C (8), the activation of which can contribute to viral lytic cycle induction (6). Engagement of CD40 may thus play a direct role in BZLF1 transactivation. However, induction of BZLF1 expression may also be a downstream event following T-cell-mediated B-cell activation. Further investigation will be needed to distinguish between these possible pathways.

In conclusion, we show that CD4⁺-T-cell-mediated activation of latently infected resting B cells leads to expression of BZLF1, a key immediate-early lytic cycle transactivator. This event is followed by rapid in vitro outgrowth of transformed B cells. Coupled with earlier reports demonstrating a requirement for T-cell function in EBV-driven tumorigenesis in the SCID/hu mouse model (2, 5, 22), these results point to an important role for T cells in the pathogenesis of EBV-associated lymphoproliferative disorders. Interestingly, it has recently been shown that transplant patients have elevated levels in serum of IL-4 (10, 11), a cytokine which promotes development of Th2 CD4⁺ T-cell responses with strong helper function for B-cell activation and proliferation, i.e., the type of T-cell response which would favor EBV reactivation from latency. A Th2 bias in CD4⁺ T-cell responses may thus contribute to an increased viral load, which is known to be a risk factor for development of posttransplant lymphoproliferative disorders (20).