|Home | About | Journals | Submit | Contact Us | Français|
Epstein-Barr virus (EBV) EBNA2 and Kaposi's sarcoma-associated herpesvirus (KSHV) replication and transcription activator (RTA) are recruited to their responsive elements through interaction with a Notch-mediated transcription factor, RBP-Jκ. In particular, RTA and EBNA2 interactions with RBP-Jκ are essential for the lytic replication of KSHV and expression of B-cell activation markers CD21 and CD23a, respectively. Here, we demonstrate that like EBV EBNA2, KSHV RTA strongly induces CD21 and CD23a expression through RBP-Jκ binding sites in the first intron of CD21 and in the CD23a core promoter, respectively. However, unlike EBV EBNA2, which alters immunoglobulin μ (Igμ) and c-myc gene expression, RTA did not affect Igμ and c-myc expression, indicating that KSHV RTA targets the Notch signal transduction pathway in a manner similar to but distinct from that of EBV EBNA2. Furthermore, RTA-induced expression of CD21 glycoprotein, which is an EBV receptor, efficiently facilitated EBV infection. In addition, RTA-induced CD23 glycoprotein underwent proteolysis and gave rise to soluble CD23 (sCD23) molecules in B lymphocytes and KSHV-infected primary effusion lymphocytes. sCD23 then stimulated primary human lymphocytes. These results demonstrate that cellular CD21 and CD23a are common targets for B lymphotropic gammaherpesviruses and that KSHV RTA regulates RBP-Jκ-mediated cellular gene expression, which ultimately provides a favorable milieu for viral reproduction in the infected host.
Kaposi's sarcoma-associated herpesvirus (KSHV), also called human herpesvirus 8, is thought to be an etiologic agent of Kaposi's sarcoma (KS) (4). KSHV is also associated with two diseases of B-cell origin: primary effusion lymphoma (PEL) and an immunoblast variant of Castleman disease (2, 3). An important step in the herpesvirus life cycle is the switch from latency to lytic replication. The KSHV replication and transcription activator (RTA) plays a central role in this switch. Ectopic expression of KSHV RTA is sufficient to disrupt viral latency and activate lytic replication to completion (8, 23, 35, 38). RTA activates the expression of numerous viral genes in the lytic cycle of KSHV, including its own promoter, polyadenylated nuclear RNA, K12, ORF57, vOX-2, viral G-protein-coupled receptor, and vIRF1. While the details of RTA-mediated transcriptional activation remain unclear, several pieces of evidence suggest that RTA activates its target promoter through direct binding to the specific sequence (20) and/or interaction with various cellular transcriptional factors. In fact, numerous cellular proteins, such as Stat3, KRBP, RBP-Jκ/CBF1, and CBP, interact with RTA, and these interactions act synergistically with RTA transcriptional activity (10, 11, 18, 19, 32, 44). Furthermore, our recent study (9) demonstrated that RTA recruits cellular SWI/SNF and TRAP/mediator complexes through its carboxy-terminal short acidic sequence. Recruitment of these complexes onto viral lytic promoters is essential for their effects on target promoters and KSHV reactivation (9).
Epstein-Barr virus (EBV) EBNA2 and KSHV RTA have been shown to be recruited to their responsive elements through interaction with the transcription factor RBP-Jκ (13, 18, 21). RBP-Jκ binding sites are present in a number of EBNA2- and RTA-regulated viral promoters. RBP-Jκ, which was originally purified and characterized by Kawaichi et al. (17) and Hamaguchi et al. (12), has been highly conserved in the evolution from nematodes to humans. Biochemical and genetic studies have demonstrated that RBP-Jκ acts downstream of the receptor Notch. Activation of the Notch receptor by binding of its ligands (Delta, Jagged, or Serrate) leads to proteolytic cleavage of the receptor at the inner side of the membrane (30). The Notch intracellular domain (NIC) is then translocated to the nucleus, where it activates genes by interacting with RBP-Jκ. EBNA2 and RTA may thus be regarded as functional homologs or mimics of the activated Notch protein. Indeed, NIC has been shown to be capable of functionally substituting for EBNA2 in the context of EBV for primary B-cell transformation (7). However, the cellular targets of cellular NIC do not completely overlap with those of EBNA2: EBNA2 and RTA both activate CD21 (CR2, EBV receptor) gene expression and repress immunoglobulin μ (Igμ) expression, whereas EBNA2, but not NIC, activates CD23a gene expression (37).
Despite detailed studies of RTA-mediated viral gene expression, the cellular targets of RTA have not been characterized. Here, we demonstrate that, similar to EBV EBNA2 and cellular NIC, KSHV RTA activates cellular CD21 and CD23a gene expression through their RBP-Jκ binding sites, resulting in drastic increases in the expression of CD21 and CD23a on the surfaces of RTA-expressing B cells and KSHV-infected PEL cells. RTA-mediated upregulation of CD21 surface expression consequently results in the enhancement of EBV infection, while upregulation of CD23a results in the activation of primary lymphocytes. Thus, RTA interacts with RBP-Jκ and strongly activates the expression of B-cell-specific surface molecules, which may offer a favorable environment for reproduction of KSHV and other viruses.
293T, BJAB, and BCBL-1 cells were grown in Dulbecco's modified Eagle's medium or RPMI 1640 medium supplemented with 10% fetal calf serum. EBV-infected B95 and AGS cells were induced with phorbol-12-tetradecanoate-13-acetate (TPA) (20 ng/ml) (Sigma, St. Louis, Mo.).
Total RNA (10 μg) was used for the synthesis of first strand cDNA according to the manufacturer's instructions. Aliquots of cDNA samples were then PCR amplified with a combination of forward and reverse gene-specific primers. The amplified PCR products were separated on 2% agarose gels, visualized by ethidium bromide staining, and photographed.
Total RNA was isolated from TRExBJAB-cDNA5 and TRExBJAB-RTA cells (stimulated and unstimulated) with the use of Trizol (Invitrogen). For each microarray, 20 μg of a RNA sample was processed according to the manufacturer's standard protocol (Affymetrix, Santa Clara, Calif.) and hybridized to an Affymetrix HGU95A chip. The data from two independent experiments were analyzed by using the Affymetrix Mass 5.0 software and Genespring software (Silicon Genetics, Redwood City, Calif.).
Real-time PCR was performed by the following thermocycling protocol: (i) 1 cycle of 2 min at 50°C and 10 min at 95°C; and (ii) 40 cycles, with 1 cycle consisting of 15 s at 95°C and 1 min at 60°C. The primer pairs against CD21 (forward primer, GCCGACACGACTACCAACC; backward primer, AGCAAGTAACCAGATTCACAGC), CD23 (forward primer, GGAATTGAACGAGAGGAACGAAG; backward primer, AAAGCCGCTGGACACCTG), and β-actin (forward primer, AGGTGACAGCAGTCGGTTG; backward primer, TGGGGTGGCTTTTAGGATGG) were synthesized by Gene Link. Real-time PCR was completed with SYBR green PCR Master Mix and analyzed with MyQ software according to the manufacturer's instructions (Bio-Rad Instrument). For each primer pair, amplification efficiencies were determined by creating a standard curve; the log of the relative target quantity was plotted versus the cycle threshold value. The standard curves demonstrated various amplification efficiencies between primer pairs with slopes ranging from −3.6 to −3.3 representing amplification efficiencies between 90 and 100%, respectively. A dissociation curve was generated for each primer pair and demonstrated the amplification of a single product. The sizes of the amplified products were confirmed by agarose gel electrophoresis. Reactions were completed in duplicate, and no-template controls were included for each primer pair. The induction level of CD21 and CD23 expression in RTA-expressing cells was normalized to the β-actin transcript level and presented as fold induction compared with that in control cells.
Cells were transfected with pCD21-Luc and pCD21-Luc-Intron (kindly provided by J. H. Weis, University of Utah School of Medicine) (46) or CD23-AC-Luc and CD23-ACP-Luc (kindly provided by H. P. Tony, University of Wurzburg, Wurzburg, Germany) (41), β-galactosidase plasmid, and the indicated amount of empty vector, RTA expression vector, or NIC expression vector (kindly provided by P. Ling, Baylor College of Medicine) (7). At 48 h posttransfection, cells were washed twice with phosphate-buffered saline (PBS), lysed, and analyzed by luciferase assay (Promega, Madison, Wis.). Luciferase levels were normalized to β-galactosidase activity and presented as fold induction compared with the control.
EBVfaV-GFP and EBV-ΔTK-GFP were obtained after 5 days of stimulation of B95-EBVfaV-GFP cells (kindly provided by Rich Longnecker, Northwestern Medical School) (36) with TPA (20 ng/ml) or after stimulation of AGS-EBV-ΔTK-GFP cells (1) with butyric acid (3 mM). Viral suspensions were used to infect doxycycline (DOXY)-treated or untreated TRExBJAB-cDNA5 and TREXBJAB-RTA cells by incubating the virus suspension with 105 cells in a volume of 1 ml of complete RPMI 1640 medium containing TPA and butyric acid for 1 h at 37°C with vigorous agitation. At 48 h postinfection, BJAB cells were fixed and immunostained with anti-CD21, anti-CD23, anti-MHC-I (anti-major histocompatibility complex class I), or anti-MHC-II antibodies.
Cells (5 × 105) were washed with RPMI 1640 medium containing 10% fetal calf serum and incubated for 30 min with antibodies. Cells were then incubated for 20 min at 4°C with secondary anti-mouse antibodies (Pharmingen, San Diego, Calif.). After the cells were washed, each cell sample was fixed with 2% paraformaldehyde, and fluorescence-activated cell sorting analysis was performed with a FACScan (Becton Dickinson, Mountain View, Calif.).
Purified anti-CD23 antibody was diluted to 1 μg/ml in PBS (pH 7.5). Each well of a 96-well enzyme-linked immunosorbent assay (ELISA) plate (Dynex) was coated with 100 μl of the anti-CD23 solution in coating buffer (0.2 M sodium phosphate, pH 6.5) overnight at 4°C. The wells were blocked for 1 h with PBS supplemented with 1% bovine serum albumin. After the wells were blocked, they were washed with PBS containing 0.05% Tween 20 and incubated with the supernatants of DOXY-treated TRExBJAB-cDNA5 or TRExBJAB-RTA cells. The wells were washed with PBS containing 0.05% Tween 20 and incubated with alkaline phosphatase-conjugated goat anti-mouse Ig (diluted 1:4,000; Jackson Laboratory) for 1 h at room temperature. After the wells were washed, they were developed by the addition of p-nitrophenyl phosphate (Sigma) in diethanolamine substrate buffer (Pierce, Rockford, Ill.).
To identify cellular genes activated by KSHV RTA, we compared the cellular transcriptional profiles of parental cells and cells expressing RTA. For this assay, we used KSHV- and EBV-negative BJAB cell lines (TRExBJAB-cDNA5 and TRExBJAB-RTA cells) in which the Myc epitope-tagged wild-type RTA gene was integrated into the chromosomal DNA under the control of a tetracycline-inducible promoter (27). Treating these cells with DOXY rapidly and strongly induced RTA expression. After DOXY treatment of TRExBJAB-cDNA5 cells and TRExBJAB-RTA cells for 0, 24, and 48 h, total RNAs were isolated and analyzed by using an Affymetrix microarray. Comparing the transcriptional profiles of TRExBJAB-cDNA5 cells and TRExBJAB-RTA cells to the profile of parental cells indicated that approximately 100 of 12,000 cellular genes were induced by at least threefold, and approximately 13 genes were repressed by at least twofold (Table (Table1).1). Of these cellular genes, the expression levels of CD21 and CD23a genes were consistently induced in TRExBJAB-RTA cells by 7- and 10-fold, respectively, after 24 h of DOXY treatment, and by 11- and 20-fold, respectively, after 48 h of DOXY treatment.
To confirm the induction of CD21 and CD23a gene expression by RTA, TRExBJAB-cDNA5 and TRExBJAB-RTA cells and KSHV-positive TRExBCBL1-cDNA5 and TRExBCBL1-RTA cells were treated with DOXY. After 0, 24, and 48 h of treatment, total RNAs were isolated and analyzed by RT-PCR to measure the expression of CD21 and CD23a genes. In addition, cellular actin, c-myc, and Igμ genes were included as controls. CD21 and CD23a mRNAs in TRExBJAB-RTA and TRExBCBL1-RTA cells were induced and remained at elevated levels after 48 h of treatment with DOXY (Fig. (Fig.1).1). In contrast, CD21 and CD23a mRNAs in TRExBJAB-cDNA5 and TREXBCBL1-cDNA5 cells stayed at low levels throughout the DOXY treatment (Fig. (Fig.1).1). Densitometry analysis indicated a twofold induction of CD21 and a three- to fivefold induction of CD23a mRNA in TRExBJAB-RTA and TRExBCBL1-RTA cells after 48 h of DOXY treatment.
Because of the weak sensitivity of RT-PCR analysis in quantitation, real-time RT-PCR was performed on RNAs from DOXY-treated or untreated TRExBJAB-cDNA5 and TRExBJAB-RTA cells with the primers for cellular CD21, CD23a, and β-actin transcripts. It showed that CD21 and CD23a mRNA levels remained low in control TRExBJAB-cDNA5 cells for both DOXY-treated cells and untreated cells (Fig. (Fig.1B).1B). In striking contrast, the levels of CD21 and CD23a mRNA were greatly increased in TRExBJAB-RTA cells upon DOXY treatment: the mRNA levels of CD21 and CD23a genes were induced by 160- and 23-fold, respectively, after 24 h of DOXY treatment, and by 40- and 20-fold, respectively, after 48 h of DOXY treatment (Fig. (Fig.1B).1B). Interestingly, the level of CD21 mRNA peaked at 24 h of DOXY treatment and declined to the basal level at 48 h of DOXY treatment, whereas CD23a mRNA remained at elevated levels after 48 h of DOXY treatment, showing the different kinetics between RTA-mediated CD21 and CD23a mRNA induction. It should be noted that CD21 and CD23a expression was detected at a high level in untreated TRExBJAB-RTA cells because of the leakiness of DOXY-induced RTA expression (Fig. (Fig.1B).1B). Finally, cellular c-myc and Igμ expression that is altered by EBV EBNA2 and cellular NIC (37) was examined by RT-PCR and immunoblot analyses. These analyses showed that unlike EBV EBNA2 and cellular NIC, KSHV RTA did not induce a detectable level of alteration in c-myc or Igμ gene expression (Fig. (Fig.1A1A and data not shown). Cellular actin and tubulin controls were also unchanged by DOXY treatment (Fig. (Fig.1A1A).
To further demonstrate the RTA-mediated induction of CD21 and CD23a, their level of surface expression was measured by flow cytometry. KSHV-negative TRExBJAB-cDNA5 and TRExBJAB-RTA cells and KSHV-positive TRExBCBL1-cDNA5 and TRExBCBL1-RTA cells were treated with DOXY for 0, 24, and 48 h, fixed with paraformaldehyde, and allowed to react with anti-CD21 and anti-CD23a antibodies. Anti-MHC-I, anti-MHC-II, anti-Igμ, and anti-ICAM1 antibodies were included as controls. The levels of CD21 and CD23a expressed on the cell surfaces of TRExBJAB-RTA cells increased by 10- and 1,000-fold after 48 h of DOXY treatment, respectively (Fig. (Fig.2).2). In contrast, their surface expression was not altered on TRExBJAB-cDNA5 control cells (Fig. (Fig.2A).2A). In addition, the surface expression of MHC-I, MHC-II, Igμ, and ICAM1 were not altered on either cell type with or without DOXY treatment, indicating that RTA specifically induced the surface expression of CD21 and CD23a (Fig. (Fig.2A2A and data not shown). When KSHV-positive TRExBCBL1-cDNA5 and TRExBCBL1-RTA cells were tested in the same assay, the surface expression of CD21 and CD23a increased only on TRExBCBL1-RTA cells, but not on TRExBCBL1-cDNA5 cells (Fig. (Fig.2B).2B). However, we noticed that the levels of CD21 and CD23a expressed on the cell surface induced by RTA were lower on TRExBCBL1-RTA cells than on TRExBJAB-RTA cells (Fig. (Fig.2).2). Furthermore, unlike TRExBJAB-RTA cells on which MHC-I and ICAM1 were not affected by RTA expression, TRExBCBL1-RTA cells displayed a considerable reduction of MHC-I and ICAM1 surface expression upon DOXY treatment (Fig. (Fig.2B).2B). The reduction of MHC-I and ICAM1 surface expression was likely mediated by KSHV K3 and K5 proteins, and the levels of expression of K3 and K5 proteins were induced by RTA-initiated lytic replication (5, 16).
To compare RTA with cellular NIC in the regulation of CD21 and CD23 surface expression, cellular NIC was also placed under the control of the DOXY-inducible promoter in TRExBJAB cells. Immunoblotting analysis showed that NIC expression was strongly induced upon DOXY treatment of the TRExBJAB-NIC cells (data not shown). These cells were treated with DOXY for 0, 24, and 48 h, fixed with paraformaldehyde, and allowed to react with ant-CD21, anti-CD23a, anti-MHC-I, and anti-MHC-II antibodies. Flow cytometry showed that NIC expression strongly induced CD21 surface expression in a tetracycline-inducible manner, whereas it did not induce CD23a surface expression under the same conditions (Fig. (Fig.2C).2C). Furthermore, TRExBJAB-NIC cells showed no alteration of MHC-I surface expression and a slight reduction of MHC-II surface expression after 48 h of DOXY treatment (Fig. (Fig.2C).2C). In summary, KSHV RTA and EBV EBNA2 induce cellular CD21 and CD23a gene expression, whereas cellular NIC induces only CD21 gene expression. Both EBV EBNA2 and cellular NIC alter c-myc and Igμ expression (37), whereas RTA does not affect Igμ and c-myc gene expression. These results collectively indicate that KSHV RTA targets the Notch-mediated signal transduction pathway in a manner similar to but distinct from those of EBV EBNA2 and cellular NIC.
While EBV EBNA2 activates cellular CD21 gene expression (47), the detailed mechanism has not been elucidated. Recent studies have indicated that cell- and stage-specific expression of CD21 is regulated by an intronic transcriptional silencer sequence (24) and that this intronic transcriptional silencer sequence contains a binding site for the transcriptional repressor RBP-Jκ (14). To identify the RTA-mediated elements responsible for the activation of human CD21 expression, reporter constructs pCD21-Luc and pCD21-Luc-Intron were used to measure the transcription activity of CD21. pCD21-Luc contained the 5′-proximal promoter region of the CD21 gene (1,272 bp 5′ of the initiating ATG codon), and pCD21-Luc-Intron contained the additional 1.6 kb of intron 1 (14).
To determine the effect of RTA activity on the CD21 promoter, 293T and BJAB cells were transfected with pCD21-Luc or pCD21-Luc-Intron reporter together with vector alone or RTA expression vector. The pGK-β-Gal reporter was also included as a control for transfection efficiency. At 48 h posttransfection, cell lysates were used for luciferase and β-galactosidase assays. RTA was not able to activate the promoter activity of pCD21-Luc in either 293T or BJAB cells, whereas it efficiently activated the promoter activity of pCD21-Luc-Intron under the same conditions (Fig. (Fig.3).3). Note that a low level of CD21-Luc-Intron promoter activation was detected in BJAB cells. This was probably because of the presence of a B-cell-type-specific repressor element, the E-box motif, in the CD21 proximal promoter region; cellular E2A repressor binds to this motif and inhibits CD21 gene expression in a B-cell-specific manner (40). Nevertheless, these results indicate that RTA is recruited to the RBP-Jκ binding site in the first intron of CD21 to activate its transcription.
Most cell lines derived from PEL are coinfected with EBV and KSHV, suggesting a potential cooperation between EBV and KSHV in their infection and/or pathogenesis (26). The initial event required for EBV entry into B cells is the interaction of the major viral envelope gp350/220 glycoprotein with its primary receptor, CD21, while this interaction is not necessary for epithelial infection (6, 29). Since RTA efficiently activated CD21 gene expression, we examined whether the activation of CD21 gene expression by RTA facilitates EBV infection. To test this hypothesis, we used three pairs of EBV-producing cells: B95 B lymphoma cells in which the LMP2A gene had been replaced with the green fluorescent protein (GFP) gene (36), AGS epithelial cells in which the thymidine kinase (TK) gene had been replaced with the GFP gene (1), and Akata B cells in which the thymidine kinase (TK) gene had been replaced with the GFP gene (1). EBVfaV-GFP from B95 B cells (B-EBV-GFP), EBV-ΔTK-GFP from AGS epithelial cells (E-EBV-GFP), and EBV-ΔTK-GFP from Akata B cells (Akata-EBV-GFP) were obtained after 5 days of TPA or Ig stimulation and used to infect DOXY-treated or untreated TRExBJAB-cDNA5 and TRExBJAB-RTA cells. At 48 h postinfection, EBV infectivity was assessed by measuring GFP fluorescence by flow cytometry. No significant difference in the infectivity of B95-EBV-GFP, AGS-EBV-GFP, and Akata-EBV-GFP viruses in TRExBJAB-cDNA5 and TREXBJAB-RTA cells was detected in cells that were not treated with DOXY (Fig. (Fig.4).4). In contrast, DOXY-treated TRExBJAB-RTA cells displayed considerably higher levels of EBV infection than DOXY-treated TRxBJAB-cDNA5 cells did; the levels of infection of B95-EBV-GFP and AGS-EBV-GFP viruses in DOXY-treated TRExBJAB-RTA cells were approximately threefold higher than that in DOXY-treated TRxBJAB-cDNA5 cells, and the level of infection of Akata-EBV-GFP virus in DOXY-treated TRExBJAB-RTA cells was twofold higher than that in DOXY-treated TRExBJAB-cDNA5 cells (Fig. (Fig.4).4). These results demonstrate that RTAmediated upregulation of CD21 surface expression effectively facilitates EBV infection.
Human CD23a is a B-cell differentiation marker involved in inflammatory responses. Analysis of the CD23a core promoter sequence reveals several binding sites for various cellular transcription factors, such as RBP-Jκ, Pax-5, and NF-κB (15). To identify the RTA-responsive sites in the CD23a promoter, we first used two CD23a promoter-luciferase reporter constructs CD23-ACP-Luc and CD23-AP-Luc. CD23-AP-Luc contains 1.2 kb of the CD23a promoter sequence, and CD23-ACP-Luc contains 283 bp of the CD23a core promoter sequence (41). TREx293-cDNA5, TREx293-RTA, TRExBJAB-cDNA5, and TRExBJAB-RTA cells were transfected with CD23-AP-Luc or CD23-ACP-Luc together with pGK-β-Gal and then treated with DOXY (1 μg/ml) or left alone. At 48 h posttransfection, cell lysates were used for reporter assays. RTA expression strongly activated promoter activities of both CD23-AP and CD23-ACP, indicating that the CD23a core promoter sequence is sufficient for RTA-induced activation (Fig. (Fig.5A).5A). Precise inspection revealed five potential RBP-Jκ binding sites in the CD23a core promoter sequence (Fig. (Fig.5B5B).
To identify the RBP-Jκ elements in the CD23a core promoter region that are responsible for RTA activity, we generated luciferase reporter constructs containing serial deletion mutations at the CD23a core promoter sequence (Fig. (Fig.5B).5B). These mutant reporters were Δ1, Δ2, Δ3, and Δ4, containing the sequences of the CD23a core promoter region from positions −161 to +83, −129 to +83, −101 to +83, and −68 to +83, respectively (positions relative to the transcription start site [+1]) (Fig. (Fig.5B).5B). Wild-type or mutant CD23a promoter reporter constructs together with the pGK-β-Gal construct were transfected into TREx293-cDNA5, TREx293-RTA, TRExBJAB-cDNA5, and TRExBJAB-RTA cells, and these cells were then treated with DOXY (1 μg/ml) or left alone. At 48 h posttransfection, cell lysates were used for reporter assays. Deletion of the first RBP-Jκ binding site abolished RTA-mediated activation of CD23a promoter activity in both 293 and BJAB cells, indicating that the first RBP-Jκ binding site is principally responsible for the ability of RTA to induce CD23a gene expression (Fig. (Fig.5C5C).
One of the hallmarks of B-cell chronic lymphocytic leukemia (B-CLL) cells is the overexpression of the transmembrane glycoprotein CD23, which undergoes proteolysis, giving rise to sCD23 molecules (15). In fact, the concentration of sCD23 in sera of B-CLL patients can be several hundredfold higher than in healthy individuals and is closely correlated with the clinical stage of the disease (15, 31, 33, 34). To determine whether RTA induced the production of sCD23, TRExBJAB-cDNA5 and TRExBJAB-RTA cells and KSHV-infected TRExBCBL1-cDNA5 and TRExBCBL1-RTA cells were treated with DOXY (1 μg/ml) for 24 and 48 h or left alone, and their supernatants were collected and analyzed by ELISAs to measure the amount of sCD23. Large amounts of sCD23 were detected in the supernatants of DOXY-treated TRExBJAB-RTA and TRExBCBL1-RTA cells, whereas almost no sCD23 was detected in the supernatants of TRExBJAB-cDNA5 and TRExBCBL1-cDNA5 cells with or without DOXY treatment (Fig. (Fig.6A).6A). These results demonstrate that RTA expression in B cells and KSHV-infected PEL cells markedly induces sCD23 production.
Although CD23a was initially characterized as the low-affinity IgE receptor, it is involved in a variety of biological processes, such as IgE-dependent inflammatory processes and antigen presentation on B cells, and it is particularly important in homotypic or heterotypic interactions of T and B cells (41, 45). Furthermore, a recent study has demonstrated that sCD23 induced from macrophages by human immunodeficiency virus (HIV) Nef contributes to the induction of CD80 and CD86 surface expression on B cells and the HIV permissivity of resting T lymphocytes (39).
To investigate the potential biological effect of sCD23 on the activation of primary B and T lymphocytes, human peripheral blood mononuclear cells (PBMCs) were treated with the supernatants from DOXY-stimulated TRExBJAB-cDNA5 or TRExBJAB-RTA cells. At 48 h, the PBMCs were immunostained with an anti-CD21 antibody to identify B lymphocytes and an anti-CD2 antibody to identify T lymphocytes. CD21-positive B lymphocytes were then further tested for the expression of CD80 and CD86 costimulating molecules on the cell surface, and CD2-positive T lymphocytes were tested for the expression of the early T-lymphocyte activation marker CD38 on the cell surface. Flow cytometry showed that the levels of CD80 and CD86 expressed on the surfaces of primary B lymphocytes increased after treatment with the supernatants of TRExBJAB-RTA cells, whereas they were not affected by treatment with the supernatants of TRExBJAB-cDNA5 cells (Fig. (Fig.6B).6B). Furthermore, expression of CD38 on the surfaces of primary T lymphocytes markedly increased after treatment with supernatants from TRExBJAB-RTA cells, whereas it was not altered after treatment with supernatants from TRExBJAB-cDNA5 cells (Fig. (Fig.6B).6B). These results indicated that sCD23 produced by RTA expression induces the stimulation of primary B and T lymphocytes through the potential paracrine and autocrine pathways, which may facilitate homotypic or heterotypic interactions of T and B lymphocytes as seen in HIV-infected macrophages (39).
The KSHV RTA protein is responsible for the switch from latency to lytic replication, and this RTA activity requires its interaction with the RBP-Jκ transcription factor (18). Here, we demonstrate that RTA strongly induced CD21 and CD23a gene expression through the RBP-Jκ binding sites at the first intron region of CD21 and CD23a core promoter region, respectively. Furthermore, RTA-induced expression of CD21 surface glycoprotein effectively facilitated EBV infection, and RTA-induced expression of sCD23 glycoprotein activated primary human B and T lymphocytes. These results demonstrate that KSHV RTA regulates RBP-Jκ-mediated cellular gene expression, which ultimately provides a favorable milieu for viral reproduction in the infected host.
EBV EBNA2 contributes to B-cell immortalization, most likely by its ability to act as a transcriptional modulator of cellular and viral gene expression. It activates the transcription of B-cell activation markers CD21, CD23, and c-myc (42) and tyrosine kinase Fgr (43) and downregulates the expression of Igμ (37). EBNA2 does not bind to DNA directly but is recruited to EBNA2-responsive elements by interaction with RBP-Jκ transcription factor, indicating that EBNA2 is a functional homolog of the activated Notch protein (47). Indeed, cellular NIC has been shown to functionally substitute for EBNA2 in the context of EBV for primary B-cell transformation (7). However, the cellular targets of cellular NIC do not completely overlap with those of EBNA2: NIC and EBNA2 both activate CD21 gene expression and repress Igμ expression, whereas EBNA2, but not NIC, activates CD23a gene expression (37). We demonstrate that, like EBNA2, RTA activates CD21 and CD23a transcription. Interestingly, the first intronic sequence of CD21 has been shown to control appropriate B-cell-specific expression and also contain binding sites for RBP-Jκ and other transcription factors (25). In concordance with these findings, RTA-induced activation of CD21 promoter activity required the first intronic sequence. This was further confirmed by the requirement of the first intronic sequence of CD21 for NIC-mediated activation (unpublished results). Besides the previously characterized RBP-Jκ binding site of the CD23a core promoter region that EBV EBNA2 primarily targets (22), four additional sequences have also been shown to bind to RBP-Jκ (15). Our serial deletion mutational analysis showed that, as seen with EBNA2, the first RBP-Jκ binding sequence of the CD23 core promoter region was primarily responsible for RTA-induced activation. Despite this similarity between KSHV RTA and EBV EBNA2 for the activation of CD21 and CD23a gene expression, RTA does not affect Igμ and c-myc expression, which are strongly regulated by EBNA2. These results collectively indicate that KSHV RTA targets Notch signal transduction in a manner that is similar to but distinct from that of EBV EBNA2.
CD21 (also known as CR2, C3d, and EBV receptor) is a member of the regulators of complement activation gene family. CD21 is a B-cell receptor for CD23 and possibly for gamma interferon and also serves as the receptor for EBV gp350/220 (28). We demonstrate that the upregulation of CD21 surface expression induced by RTA facilitates EBV infection. Since most PEL cells are coinfected with KSHV and EBV, our results suggest the possibility of hierarchically ordered infection: KSHV infection, followed by EBV infection. This hypothesis is under active investigation.
CD23 (also known as low-affinity IgE receptor and FcRII) is expressed as a type II extracellular protein on a variety of cells, such as B cells, monocytes, and macrophages, and is cleaved from the cell surface to generate several distinct fragments. This cleavage is mediated by a metalloprotease. This sCD23 then acts as a B-cell growth factor and is associated with EBV infection (15). In fact, since the level of sCD23 in serum is higher in patients with autoimmune diseases and B-CLL, it has been used as an indicator of the progression of disease (15, 31, 34). This suggests the intriguing possibility that KSHV-infected PEL patients may also have high levels of sCD23 in serum. Unfortunately, the limited availability of sera from PEL patients did not allow us to determine the levels of sCD23 in sera from PEL patients. Nevertheless, it will be interesting to pursue whether KSHV infection is associated with a high level of sCD23 in serum, which may provide a marker for KSHV infection and progression.
A recent study has demonstrated that sCD23 and soluble ICAM1, which are induced from macrophages by HIV Nef, contribute to the induction of the expression of CD80 and CD86 on the surfaces of B cells and the HIV permissivity of resting T lymphocytes (39). We also showed that the expression of CD80 and CD86 costimulating molecules on the surfaces of primary B lymphocytes and the CD38 activation marker on primary T lymphocytes increased after treatment with the supernatants of TRExBJAB-RTA cells, whereas it was not affected after treatment with the supernatants of TRExBJAB-cDNA5 cells. These results suggest that sCD23 produced by RTA expression induces the stimulation of primary B and T cells through the potential paracrine and autocrine pathways, which may ultimately facilitate homotypic or heterotypic interactions of T and B cells. This also suggests that KSHV RTA potentially influences the activation state of infected B cells, thereby enhancing the ability of neighboring T cells to support the replication of HIV type 1. This activity may require the upregulation of CD80 and CD86 B-cell costimulating receptors involved in the alternative pathway of T-lymphocyte stimulation. Thus, although the full significance of RTA-mediated effects on the other coexisting viruses in KS and PEL patients needs to be studied further, our study may provide insight into the comprehensive mechanism by which KSHV RTA not only activates lytic reactivation by targeting KSHV gene expression but also creates a favorable milieu for viral infection and replication by activating cellular gene expression.
We thank R. Longnecker, P. Ling, D. Hayward, H. P. Tony, and J. H. Weis for providing cells and reagents and J. Macke for reading and editing the manuscript.
This work was partly supported by U.S. Public Health Service grants CA115284, CA106156, CA91819, CA082057, and RR00168. J. U. Jung is a Leukemia & Lymphoma Society Scholar.