Identification of Functional Differences between Prototype Epstein–Barr Virus-Encoded LMP1 and a Nasopharyngeal Carcinoma-Derived LMP1 in Human Epithelial Cells

Virology 272, 204–217 (2000) doi:10.1006/viro.2000.0344, available online at http://www.idealibrary.com on

Identification of Functional Differences between Prototype Epstein–Barr Virus-Encoded LMP1 and a Nasopharyngeal Carcinoma-Derived LMP1 in Human Epithelial Cells Christopher W. Dawson, Aristides G. Eliopoulos, Sarah M. Blake, Rachel Barker, and Lawrence S. Young 1 CRC Institute for Cancer Studies, University of Birmingham Medical School, Birmingham B15 2TJ, United Kingdom Received February 3, 2000; returned to author for revision March 29, 2000; accepted April 4, 2000 The contribution of Epstein–Barr virus (EBV) strain variation to the pathogenesis of virus-associated tumours remains unknown. Given the central role of LMP1 in EBV-induced transformation, much interest has focused on the influence of LMP1 sequence variation on the signaling pathways and multiple downstream phenotypic consequences of LMP1 expression. The identification of LMP1 variants with a common 10-amino-acid deletion and additional point mutations (typified by the CAO-LMP1 isolate) in EBV strains associated with nasopharyngeal carcinoma prompted us to examine the effect of stable prototype B95.8-LMP1 and CAO-LMP1 expression on the phenotype and differentiation of SCC12F human epithelial cells. Both forms of LMP1 were able to induce expression of the antiapoptotic A20 protein and provide protection from tumour necrosis factor-␣-induced cytotoxicity. Although B95.8-LMP1 induced growth inhibition, expression of certain cell surface molecules (CD40, CD44, and CD54), and secretion of interleukin-6 and -8 in SCC12F cells, stable CAO-LMP1 expression failed to elicit these effects. Furthermore, B95.8-LMP1, but not CAO-LMP1, induced alterations in cell morphology and blocked epithelial cell differentiation. Both B95.8-LMP1 and CAO-LMP1 induced similar levels of nuclear factor-␬B activation, but the ability of CAO-LMP1 to activate the AP-1 pathway was relatively impaired. These data highlight significant functional differences between the prototype B95.8-LMP1 and the CAO-LMP1 variant when stably expressed in human epithelial cells and suggest that continued analysis of LMP1 variants will help to further dissect the signaling pathways activated by LMP1 as well as provide insights into the contribution of LMP1 sequence variation to the pathogenesis of EBV-associated tumours. © 2000 Academic Press

INTRODUCTION

expression in various EBV-associated tumours such as Hodgkin’s disease and NPC (Rickinson and Kieff, 1996). In both B cells and epithelial cells, LMP1 has been shown to induce the expression of certain cell surface molecules (CD40, CD54) (Wang et al., 1988,1990) and of cell survival genes such as A20, Bcl-2, and Mcl-1 (Henderson et al., 1991; Laherty et al., 1992; Rowe et al., 1994; Wang et al., 1996; Miller et al., 1997). Additional effects in epithelial cells include up-regulation of interleukin (IL)-6 and IL-8 secretion (Eliopoulos et al., 1997; Eliopoulos et al., 1999) and induction of the EGF receptor (EGFR) (Miller et al., 1995, 1997, 1998a). Of particular relevance to NPC is the ability of LMP1 to transform epithelial cells and to block terminal differentiation (Dawson et al., 1990; Fahraeus et al., 1990; Wilson et al., 1990; Hu et al., 1993). Functional studies have revealed that LMP1 can activate the nuclear factor (NF)-␬B and JNK/p38 mitogen-activated protein (MAP) kinase pathways and that these are responsible for the diverse phenotypic effects of LMP1 (Mosialos et al., 1995; Kieff, 1996; Eliopoulos et al., 1997, 1998, 1999; Kieser et al., 1997; Izumi and Kieff, 1997b; Eliopoulos and Young, 1998). Two major effector regions have been identified in LMP1 (Huen et al., 1995): a membrane proximal CTAR1 domain (residues 194–232), which weakly activates NF-␬B and p38 through direct interaction with members of the TRAF (TNF receptorassociated factor) family, and a membrane distal CTAR2

Epstein–Barr virus (EBV), a ubiquitous human herpesvirus, is associated with the development of tumours of B cell origin (Burkitt’s lymphoma, immunoblastic lymphoma, Hodgkin’s disease) and of epithelial cell origin [nasopharyngeal carcinoma (NPC), gastric adenocarcinomas] (Osato and Imai, 1996; Raab-Traub, 1996; Rickinson and Kieff, 1996). An in vitro corollary of the oncogenic capacity of EBV is its ability to transform resting B cells into permanently growing lymphoblastoid cell lines in which every cell carries multiple episomal copies of the virus genome and expresses a limited set of virus-encoded nuclear (EBNA) and membrane (LMP) proteins (Kieff, 1996). Of these viral proteins, the latent membrane protein 1 (LMP1) is particularly interesting because it displays classic oncogene activity in rodent fibroblast transformation assays and is capable of inducing a wide range of phenotypic effects in both B cells and epithelial cells (Kieff, 1996). Recent work demonstrates that LMP1 is essential for EBV-induced B cell transformation (Kaye et al., 1993; Izumi and Kieff, 1997a) consistent with its

1 To whom reprint requests should be addressed. Fax: 21-414-5376. E-mail: [email protected].

0042-6822/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

204

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

domain (residues 351–386), which is the dominant NF-␬B and JNK/p38 activating region of LMP1 and directly interacts with TRADD (TNF receptor-associated death domain) (Huen et al., 1995; Izumi and Kieff, 1997b; Kieser et al., 1997; Eliopoulos and Young, 1998; Eliopoulos et al., 1998, 1999). Recent work demonstrates that the LMP1 repeat region between CTAR1 and CTAR2 interacts with JAK3, thereby activating the STAT pathway in B cells, but the functional significance of this interaction in epithelial cells is unknown (Gires et al., 1999). Genetic analysis has demonstrated that although CTAR2 is the major effector domain of LMP1, its activity is not absolutely essential for B cell transformation (Kaye et al., 1993, 1995, 1999; Izum and Kieff, 1997a). Although the LMP1 gene is generally well conserved between different virus isolates (Miller et al., 1994; Sample et al., 1994), the existence of LMP1 sequence variation between geographically distinct EBV isolates has led to speculation that these LMP1 variants might contribute to the development of EBV-associated tumours (Chen et al., 1992, 1996; Knecht et al., 1995; Cheung et al., 1998; Sung et al., 1998; Walling et al., 1999). The cloning and sequencing of the LMP1 gene from EBV isolates derived from either a Chinese or a Taiwanese NPC have identified several mutations compared with the prototype B95.8 strain, including a point mutation leading to loss of an XhoI restriction site in the first exon, a 30-bp deletion in the carboxyl-terminus immediately upstream of CTAR2, and multiple point mutations (Hu et al., 1991; Chen et al., 1992; Miller et al., 1994; Cheung et al., 1998; Sung et al., 1998). These so-called delLMP1 variants (typified by CAO-LMP1) display increased tumourigenicity in rodent fibroblasts (Li et al., 1996) and the RHEK-1 epithelial cell line (Hu et al., 1991, 1993). Although initial studies suggested that EBV isolates carrying delLMP1 are more frequently detected in lymphoproliferative disorders (Knecht et al., 1995; Chen et al., 1996), a more thorough analysis has shown that the incidence of delLMP1 in EBV-associated tumours reflects the frequency with which this variant is detected in healthy virus carriers from the same geographical region (Khanim et al., 1996). However, given that delLMP1 is the predominant form of LMP1 in Chinese populations, it remains possible that this variant contributes to the development of NPC. A recently published functional analysis in lymphoid cells has revealed that CAO-LMP1 is impaired in its ability to up-regulate CD40 and CD54 relative to B95.8-LMP1 even though CAO-LMP1 can induce greater activation of NF-␬B than B95.8-LMP1 (Johnson et al., 1998). These studies concluded that the 30-bp deletion was not responsible for these differences and that sequences outside CTAR2 were involved. Similar studies using a delLMP1 isolated from a different NPC (C15) have shown that this LMP1 isolate is also more efficient at activating NF-␬B than B95.8-LMP1 with resultant enhanced induction of the EGF receptor in the C33A

205

carcinoma cell line; these effects of C15-LMP1 were not due to the 30-bp deletion (Miller et al., 1998b). Many of the previous studies analysing the functional differences between distinct LMP1 isolates have used transient transfection approaches and not examined the phenotypic consequences of stable LMP1 expression. Thus we have analysed the effect of stable expression of B95.8-LMP1 versus CAO-LMP1 on the phenotype of SCC12F cells, an immortalised but nontumorigenic cell line derived from stratified squamous epithelium (Dawson et al., 1990, 1995, 1998). We now show that although B95.8-LMP1 is able to induce a range of phenotypic effects in SCC12F cells, CAO-LMP1 has only limited effects. The cell signaling basis for this differential behaviour is also explored. RESULTS Stable expression and cellular localization of B95.8LMP1 and CAO-LMP1 in SCC12F cells Expression of either B95.8-LMP1 or CAO-LMP1 in clones of SCC12F cells was confirmed by immunoblotting (Fig. 1A). Clones of SCC12F transfected with B95.8-LMP1 (clones 1, 4, and 9) show clear expression of the 63-kDa form of LMP1 at levels that are comparable to those observed in the EBV-transformed LCL, X50–7. Similar levels of expression were also observed in three CAO-LMP1-expressing clones (clones, 1, 5, and 6), with the CAO-derived LMP1 protein migrating at the expected higher molecular weight of 66 kDa due to an increase in the number of internal repeats. No expression was observed in four vector control clones (neo-1, 3, 4, and 13). Immunostaining for LMP1 in stable clones revealed no major differences in their patterns of subcellular location. Both B95.8-LMP1 (Fig. 1C) and CAOLMP1 (Fig. 1D) proteins localised to the plasma membrane of SCC12F cells and, to a lesser extent, within internal membranes. Immunostaining of vector control clones gave only background fluorescence with the CS1–4 monoclonal antibody mix confirming the specificity of the staining (Fig. 1B). Both B95.8-LMP1 and CAO-LMP1 induce expression of the NF-␬B regulatable gene A20 and partially protect SCC12F from TNF-␣-induced cytolysis LMP1 activation of NF-␬B results in the induction of A20, a protein that plays a role in protecting cells from TNF-␣induced cytotoxicity (Laherty et al., 1992). To investigate whether B95.8-LMP1 and CAO-LMP1 induced expression of this protein to similar degrees, immunoblotting analysis was performed on cell extracts using a monoclonal antibody specific for A20. As shown in Fig. 2A, both B95.8LMP1 and CAO-LMP1 induced significant expression of A20 with all LMP1-positive clones expressing similar amounts of A20. In contrast, negligible A20 expression was observed in all four vector control clones. To investigate

206

DAWSON ET AL.

FIG. 1. Stable expression of B95.8-LMP1 and CAO-LMP1 in SCC12F clones. (A) Immunoblot showing expression of B95.8-LMP1 and CAO-LMP1 in stable clones of SCC12F. X50/7, an EBV transformed B cell line, is included as a positive control, whereas three vector control clones of SCC12F serve as negative controls. LMP1 expression was detected by probing the blots with CS1–4, a pool of monoclonal antibodies specific for LMP1. Immunostaining of stable clones of SCC12F demonstrates the distribution of B95.8-LMP1 and CAO-LMP1 in situ (B–D). Cells grown for 48 h on Teflon-coated microslides were fixed in methanol and immunostained with the anti-LMP1 monoclonal antibodies CS1–4. Representative B95.8-LMP1, clone 4 (C), and CAO-LMP1, clone 1 (D), showed strong staining with the LMP1-specific monoclonal antibodies, whereas only background levels of staining were observed on vector control clones (B).

whether LMP1 induction of A20 translated into increased protection from TNF-␣-induced cytolysis, short-term viability assays were performed. Both B95.8-LMP1- and CAOLMP1-expressing clones were significantly more protected from TNF-␣-induced cytolysis compared with vector control clones (Fig. 2B). Although all LMP1-expressing clones were similarly resistant at very low concentrations of TNF-␣, significant differences became apparent at concentrations of ⬎1 ng/ml. At a concentration of 100 ng/ml, both B95.8LMP1- and CAO-LMP1-expressing clones were significantly more resistant to TNF-␣-induced cytolysis than vector control clones. Unlike B95.8-LMP1, CAO-LMP1 does not influence cell growth As previously reported (Dawson et al., 1990), the expression of B95.8-LMP1 in SCC12F cells is associated

with growth inhibition. One striking feature of SCC12F cells expressing CAO-LMP1 was the apparent lack of any growth inhibitory effects. To examine this in more detail, the growth kinetics of SCC12F cells expressing B95.8-LMP1 or CAO-LMP1 was examined over a 10-day period. The results of a representative experiment are shown in Fig. 3. Although no obvious differences were observed at early time points, clear differences in the growth rates of B95.8-LMP1- and CAO-LMP1-expressing clones were observed at day 6. Compared with SCC12F vector control and CAO-LMP1 clones, which had achieved five to seven population doublings by day 6, the two B95.8-LMP1 clones (LMP1/1 and LMP1/9) were severely growth-impaired with only marginal increases of one or two population doublings over the same period. At day 10, these effects were even more pronounced, with both SCC12F vector control and CAO-LMP1-expressing

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

FIG. 2. Both B95.8-LMP1 and CAO-LMP1 induce A20 induction in SCC12F cells and protect these cells from TNF-␣-induced cytotoxicity. (A) Immunoblot showing expression of A20 in SCC12F clones stably expressing B95.8-LMP1 or CAO-LMP1. X50/7, an EBV transformed B cell line, is included as a positive control, whereas three vector control clones of SCC12F serve as negative controls. A20 expression was detected by probing with a monoclonal antibody specific for the A20. (B) The effect of B95.8-LMP1 and CAO-LMP1 on TNF-␣ induced cytotoxicity was assessed in MTT assays in response to increasing concentrations of TNF-␣. Data are shown relative to that of cells treated with 10 ␮g/ml cycloheximide alone and are representative of three similar experiments.

clones achieving 10 population doublings compared with 2.5–5 population doublings for the two B95.8-LMP1 clones. Unlike B95.8-LMP1, CAO-LMP1 fails to induce expression of cell surface markers and interleukin production The induction of CD40 and CD54 (ICAM-1) by LMP1 are useful indicators of its biological activity (Wang et al., 1988, 1990; Dawson et al., 1990; Rowe et al., 1994; Huen et al., 1995; Johnson et al., 1998). As a previous study had shown that CAO-LMP1 was defective in its ability to induce cell surface CD40 and CD54 expression in lym-

207

phoid cells (Johnson et al., 1998), we were interested in investigating whether the same phenomenon was observed in epithelial cells stably expressing CAO-LMP1. The cell surface profiles of SCC12F cells expressing B95.8-LMP1 and CAO-LMP1 were investigated by flow cytometry using monoclonal antibodies specific for CD40, CD54, and CD44. A monoclonal antibody specific for CD23 served as a negative control in these studies. The results from a representative analysis are shown in Fig. 4A. Expression of the lymphoid specific marker CD23 gave only low levels of background staining on both vector control and LMP1-expressing clones [mean fluorescence intensity (MFI) values of 10–15]. In contrast, the expression of CD40, CD54, and CD44 was clearly up-regulated on cells expressing B95.8-LMP1. Although the magnitude of the induction was variable for each marker, the expression of CD40 was almost twice as high on B95.8-LMP1-expressing clones compared with vector control clones (MFI of 30–40 versus 18–19). The levels of CD54 expression were more variable but were significantly higher than vector control clones (MFI values of 80–150 versus 50–60), whereas the expression of CD44 was clearly higher on all three B95.8-LMP1 clones examined (MFI values of 95–110 versus 40–65). In marked contrast, all three CAO-LMP1 clones expressed levels of CD40, CD54, and CD44 that were equivalent to those observed on vector control clones. An additional phenotypic consequence of LMP1 expression is induction of the secretion of IL-6 (Eliopoulos et al., 1997, 1999) and IL-8 (Eliopoulos et al., 1999). To evaluate the effects of B95.8-LMP1 and CAO-LMP1 on the levels of IL-6 and IL-8 secretion from SCC12F cells, supernatants were collected from subconfluent SCC12F cultures seeded at the same cell density and the levels of IL-6 and IL-8 secreted into the growth medium were determined by ELISA. As demonstrated in Fig. 4B, B95.8LMP1-expressing SCC12F clones secreted far greater quantities of both IL-6 and IL-8 than vector control clones, whereas the levels of IL-6 and IL-8 secretion from the two CAO-LMP1 expressing clones were only marginally greater than those of the controls. B95.8-LMP1 but not CAO-LMP1 induces alterations in cell morphology and blocks terminal differentiation Our previous work has demonstrated that expression of B95.8-LMP1 induced a marked alteration in the morphology of SCC12F cells in monolayer culture (Dawson et al., 1990). This is confirmed in the present study as stable clones expressing B95.8-LMP1 displayed poorer intercellular contact and failed to stratify to the same extent as representative vector control clones (data not shown). This was in marked contrast to clones expressing CAO-LMP1, which maintained a cuboidal morphology and were almost indistinguishable from vector control clones (data not shown). In organotypic raft culture, vec-

208

DAWSON ET AL.

FIG. 3. CAO-LMP1 does not induce the growth-inhibitory effects associated with B95.8-LMP1 expression in SCC12F cells. The growth kinetics of representative SCC12F vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones was analysed over a 10-day period. Cell number was determined at 48-h intervals by trypan blue exclusion to estimate viable cell number. Data are the means of triplicate counts and are representative of two experiments.

tor control SCC12F clones routinely gave rise to simple differentiating epithelial structures that generated a clearly defined basal cell layer and two or three layers of flatter differentiating cells as demonstrated by immunostaining with the AE1 pan-keratin monoclonal antibody (Fig. 5A). In agreement with our previous findings, B95.8LMP1-expressing clones produced much thicker and less well-organised epithelial structures that were routinely three or four times the thickness of vector control rafts (Fig. 5C). However, unlike B95.8-LMP1, CAO-LMP1expressing clones gave rise to raft structures that were identical to those formed by SCC12F vector control cells and retained their capacity for terminal differentiation (Fig. 5E). Thus, although the differentiation-associated antigen involucrin was clearly not expressed in suprabasal cell layers of B95.8-LMP1-expressing rafts (Fig. 5D), expression of this protein was clearly evident in differentiating suprabasal cell layers of both the SCC12F vector control (Fig. 5B) and CAO-LMP1 raft structures (Fig. 5F). Similar results were obtained in at least three separate raft experiments with representative B95.8LMP1- and CAO-LMP1-expressing clones.

Both B95.8-LMP1 and CAO-LMP1 induce NF-␬B, but CAO-LMP1 weakly activates AP-1 One possible explanation for the inability of CAOLMP1 to induce the various phenotypic effects observed with B95.8-LMP1 in SCC12F cells is differences in the relative activation of signaling pathways such as the NF-␬B and JNK/AP-1. To address this possibility, we used electrophoretic mobility shift assays (EMSAs) to examine the constitutive levels of NF-␬B and AP-1 activity in the stable SCC12F clones expressing either B95.8-LMP1 or CAO-LMP1. NF-␬B complexes were detected using a probe containing concensus ␬B binding sites from the HIV LTR, and AP-1 complexes were detected using a probe containing concensus AP-1 binding sites from the collagenase promoter (Gires et al., 1999). Significant NF-␬B activity was observed in nuclear extracts derived from B95.8-LMP1- and CAO-LMP1-expressing clones, although somewhat lower NF-␬B activity was observed in CAO-LMP1 clones 5 and 6 compared with the other LMP1-positive clones (Fig. 6A). Although significant AP-1 activity was observed in SCC12F clones expressing

FIG. 4. Unlike B95.8-LMP1, CAO-LMP1 fails to induce cell surface antigen expression and cytokine secretion. (A) The ability of B95.8-LMP1 and CAO-LMP1 to induce cell surface antigen expression in SCC12F cells was determined by FACS analysis. Data are presented as MFI on an arbitrary log scale. Monoclonal antibodies used include antibodies to CD23, CD40, CD44, and CD54. (B) The ability of B95.8-LMP1 and CAO-LMP1 to induce the secretion of IL-6 (i) and IL-8 (ii) was determined by ELISA. Conditioned medium from representative vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones was analysed. The results are presented as the mean of duplicate determinations (shown in pg/ml) and are representative of two experiments.

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

209

210

DAWSON ET AL.

whether this was due to an inability of CAO-LMP1 to activate c-jun N-terminal kinase activity to the same extent as B95.8-LMP1. In vitro kinase assays were performed on extracts derived from SCC12F vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones, and the ability of endogenous JNK to phosphorylate GST-jun substrate was examined. Interestingly, all LMP1-expressing clones were found to have elevated JNK activity at levels that were clearly higher than those observed with vector control clones (Fig. 6D). The degree of JNK activation was similar between B95.8-LMP1 and CAO-LMP1 clones (Fig. 6D). DISCUSSION

FIG. 5. Expression of B95.8-LMP1 but not CAO-LMP1 is associated with alterations in the differentiation capacity of SCC12F cells in organotypic rafts. Immunostaining of sections of control clone neo-1 (A and B), B95.8-LMP1 clone 9 (C and D), and CAO-LMP1 clone 1 (E and F) using monoclonal antibody AE1 (Woodcock-Mitchell et al., 1982) against type 1 keratins (A, C, and E) and a polyclonal rabbit antiserum against involucrin (Biogenesis) (B, D, and F). Magnification, 250⫻.

B95.8-LMP1, much weaker activity was found in CAOLMP1-expressing cells (Fig. 6B); however, AP-1 activity in the CAO-LMP1 clones was still greater than that observed in the vector control clones. To investigate whether B95.8-LMP1 and CAO-LMP1 activated qualitatively different species of NF-␬B, supershift experiments were performed with antibodies specific for p50, p65, and c-rel subunits. As can be seen in Fig. 6C, both B95.8and CAO-LMP1 activated NF-␬B complexes that were composed of both p50 homodimers and p50/p65 heterodimers. The apparent inability of CAO-LMP1 to induce a robust AP-1 activation prompted us to examine

In this study, we have demonstrated that the CAOLMP1 isolate fails to induce many of the morphological and phenotypic changes associated with B95.8-LMP1 expression in a human epithelial cell background. This is perhaps surprising given previous work suggesting that CAO-LMP1 is more oncogenic than B95.8-LMP1 in both rodent fibroblasts and in the RHEK-1 epithelial cell line (Hu et al., 1993). However, other studies with RHEK-1 cells failed to identify significant differences between B95.8-LMP1 and CAO-LMP1 with regard to induction of proliferation and blockade of differentiation (Zheng et al., 1994). It is nevertheless clear that RHEK-1 cells respond very differently than SCC12F cells with regard to the effects of LMP1 expression, and this may reflect the distinct nature of the epithelial cell environment in the two cell lines. RHEK-1 cells were generated by immortalisation with an adenovirus 12–SV40 hybrid virus and are exceptionally susceptible to transformation by a variety of agents, including oncogenes, chemical carcinogens, and ionizing radiation (Yang et al., 1992; AllenHoffmann et al., 1993; Lee et al., 1993; Razzzaque et al., 1993). SCC12F cells are a subclone derived from a squamous cell carcinoma, and they behave like primary keratinocytes in vitro and are relatively resistant to transformation (Rheinwold et al., 1981; Parkinson et al., 1983). Our previous work suggests that SCC12F cells are ideal for studying the effects of stably expressed transfected genes on the epithelial cell phenotype, particularly with regard to terminal differentiation (Dawson et al., 1990, 1995, 1998). Thus it is surprising that despite equivalent levels of stable expression, CAO-LMP1 is unable to induce most of the phenotypic effects observed with B95.8LMP1 in the same cell background. These data, however, are consistent with recent work using transient transfection in human B and T cell lines that demonstrates that although CAO-LMP1 is able to induce robust activation of NF-␬B, it is relatively impaired in its ability to induce various phenotypic effects such as up-regulation of CD40 and CD54 (Johnson et al., 1998). Our data extend this observation by demonstrating that CAO-LMP1 not only is unable to induce expression of CD40 and CD54

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

211

but also fails to stimulate secretion of IL-6 and IL-8, to inhibit cell growth, and to block epithelial cell differentiation. The inability of CAO-LMP1 to induce phenotypic effects does not appear to be due to defective activation of NF-␬B as the B95.8-LMP1 and CAO-LMP1 transfectants displayed comparable levels of activity and the NF-␬B species present in the activated complexes appear to be identical on supershift analysis. In support of this observation is the finding that both LMP1 proteins were able to induce similar levels of A20, a protein known to be under the dominant control of NF-␬B (Laherty et al., 1992; Miller et al., 1997), and that this A20 expression correlated with relative protection from TNF␣-induced cytotoxicity. However, CAO-LMP1 was unable to induce expression of other genes that also contain NF-␬B sites in their promoters, such as CD40, CD54, IL-6, and IL-8 (Grimaldi et al., 1991; Voraberger et al., 1991). This emphasises the complexity of transcriptional regulation and highlights the involvement of multiple signals in controlling promoter function. The inability of CAO-LMP1 to induce many of the phenotypic effects associated with B95.8-LMP1 expression despite activating similar levels of NF-␬B suggests that CAO-LMP1 is defective in engaging other LMP1activated signaling pathways. Recent work demonstrates that LMP1 can stimulate both the JNK and p38 MAP kinase pathways (Kieser et al., 1997; Eliopoulos and Young, 1998; Eliopoulos et al., 1999) and that these contribute to the transcriptional activity of certain genes such as IL-6 and IL-8. It is therefore interesting that CAO-LMP1 appears to be less efficient at inducing AP-1 activity than B95.8-LMP1, and this might explain the failure of CAO-LMP1 to induce up-regulation of cell surface markers and cytokine secretion, as these genes may be coregulated by NF-␬B and AP-1. As JNK activation by LMP1 is mediated via CTAR-2 and requires direct interaction with TRADD (Izumi and Kieff 1997a; Eliopoulos et al., 1998), it is possible that the CTAR-2 region of CAOLMP1 is less efficient at engaging JNK due to a weaker interaction with TRADD. However, the amino acid sequence over this region is well conserved between B95.8 and CAO, and consistent with this, we found that JNK activation by CAO-LMP1 is similar to that induced by B95.8-LMP1. Thus the weaker AP-1 activity induced by CAO-LMP1 may be due to additional factors that impinge

FIG. 6. Both B95.8-LMP1 and CAO-LMP1 activate NF-␬B and AP-1 transcription factors. Nuclear extracts isolated from SCC12F cells expressing B95.8-LMP1 or CAO-LMP1 were subjected to EMSA to detect the basal levels of NF-␬B (A) and AP-1 activity (B). (A) A 32P-labeled NF-␬B probe from the HIV LTR was used to detect NF-␬B activity. The predominant complex, indicated by an arrow, contains p65/p50 het-

erodimers. (B) The same nuclear extracts were then subjected to EMSA to detect the basal levels of AP-1 activity using a 32P-labeled collagenase TRE oligonucleotide as a probe. (C) Supershift experiment with antibodies to p50, p65, and c-rel demonstrating that B95.8-LMP1 and CAO-LMP1 activate NF-␬B transcription factor complexes that are composed of both p50 and p65 subunits. Arrows denote supershifted complexes containing either p50 or p65. (D) In vitro kinase assay showing that both B95.8-LMP1 and CAO-LMP1 are able to activate c-Jun N-terminal kinase activity to similar extents. Arrow indicates the phosphorylated substrate, GST-c-jun.

212

DAWSON ET AL.

on the efficient activation of AP-1, such as the influence of other signaling pathways. In this regard, recent work indicates that both the p38 and ERK MAP kinases can affect AP-1 activation in keratinocytes by regulating c-fos expression (Chen and Bowden, 1999), and thus the ability of CAO-LMP1 to activate these pathways when stably expressed in epithelial cells is worthy of examination. Our preliminary data indicate that both types of LMP1 activate the p38 kinase to the same extent in the stable SCC12F transfectants. Because many of the amino acid differences between CAO-LMP1 and B95.8-LMP1 lie within the N-terminus and transmembrane loops, these regions of the protein could have a direct effect on cell signaling or influence the activity of CTAR1 and CTAR2 by affecting the conformation of the LMP1 cytoplasmic tail. The recent demonstration of a cooperative effect between CTAR1 and CTAR2 on the up-regulation of CD54 may be of relevance because it suggests that a physical interaction between these domains is necessary for efficient LMP1 signaling (Floettmann et al., 1998). This cooperation between CTAR1 and CTAR2 may be defective in CAO-LMP1, and further experiments addressing this using chimeric combinations of CAO-LMP1 and B95.8-LMP1 are under way. Alternatively, it is possible that additional yet-to-be-identified signaling pathways are differentially activated by B95.8-LMP1 versus CAOLMP1. The recent demonstration that LMP1 activates members of the STAT family of transcription factors through the binding and activation of JAK proteins is a relevant case in point (Gires et al., 1999). Interestingly, JAK3 binding does not map to the known C-terminal activating regions CTAR1 or CTAR2 but maps to prolinerich domains within the repeat regions. Although JAK3 is not expressed in epithelial cells, these observations demonstrate that LMP1 is able to engage additional signaling molecules through regions of its cytoplasmic C-terminus that are distinct from those that engage NF-␬B and JNK/p38. However, the possibility that differential activation of other members of the JAK/STAT pathway or of other yet-to-be-identified LMP1-activated signals accounts for the biological differences between B95.8-LMP1 and CAO-LMP1 cannot be excluded and is worthy of further investigation. The ability of B95.8-LMP1 to block the terminal differentiation of human stratified squamous epithelial cells is confirmed in this study. Although the relative contribution of different signaling pathways to this effect remains unknown, both AP-1 and NF-␬B have been implicated in the regulation of epithelial differentiation. Within stratified squamous epithelium, differentiation-dependent expression of AP-1 family members has been observed, suggesting a role for these proteins in the temporal and spatial regulation of gene expression during epithelial differentiation (Welter and Eckert, 1995). The precise relevance of these observations in terms of the possible role of different AP-1 family members in controlling the

commitment of epithelial cells to the differentiation process remains unknown. By contrast, increasing evidence implicates the NF-␬B family as key regulators of epithelial cell growth and differentiation. In normal epidermis, NF-␬B is activated as cells migrate from the basal layer, suggesting a role for NF-␬B in epithelial cell differentiation (Seitz et al., 1998). Functional inhibition of NF-␬B by expression of dominant negative NF-␬B inhibitory proteins (I␬Bs) has been shown to block differentiation and promote epithelial hyperplasia (Seitz et al., 1998). More recent studies on the function of the I␬B kinase, a large multiprotein complex consisting of I␬B kinase kinase (IKK) subunits, support a role for NF-␬B in the terminal differentiation of epithelial cells. Thus IKK␣ (I␬B kinase kinase ␣) null mice are born with limb and skeletal abnormalities and develop a hyperplastic epidermis characterised by increased cell proliferation and severely diminished differentiation (Hu et al., 1999; Takeda et al., 1999). This implicates IKK␣-mediated NF-␬B activation in the regulation of normal epithelial cell differentiation. Interestingly, cells from IKK␣ null mice are still able to activate NF-␬B in response to TNF-␣ and IL-1 as this appears to be predominantly regulated by IKK␤ (Li et al., 1999). Thus it appears that IKK␣ plays a fundamental role in regulating cell fate (developmental patterning) and epithelial cell differentiation and is activated by morphogenic signals. The degree to which NF-␬B activation through IKK␣ or IKK␤ influences the subsequent transcriptional regulation of different downstream genes remains unknown. A recent study demonstrated that both IKK␣ and IKK␤ are required for LMP1-induced NF-␬B activation from CTAR1 and CTAR2 (Sylla et al., 1998); a more profound inhibitory effect of dominant negative IKK␣ was observed. Reconciling these data with the ability of B95.8-LMP1 to block epithelial differentiation is more complex. It could be that the constitutive activation of NF-␬B by B95.8-LMP1 blocks the morphogenic signals responsible for IKK␣-induced NF-␬B activation while promoting the IKK␤ pathway. This would result in the inhibition of epithelial differentiation but the stimulation of the NF-␬B responsive genes involved in cytokine-mediated responses such as CD54 and the interleukins. The ability of the I␬B kinase complex to transmit and regulate different signals through differential activation of the IKKs has been demonstrated and may further explain the effects of LMP1 (May and Gosh, 1999). Extending this argument to account for the inability of CAO-LMP1 to block epithelial differentiation while still retaining the ability to activate NF-␬B is more difficult. Apart from the possible role of other signaling pathways as discussed earlier, defects in both the differential activation of the IKKs and the downstream pathways responsible for regulating NF-␬B responsive gene transcription may account for the severely impaired activity of CAO-LMP1. Such differential effects may account for the ability of CAO-LMP1 to induce expression of the NF-␬B-regulated

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

A20 gene but not of other NF-␬B-regulatable genes. Further studies examining the relative contribution of IKK␣ and IKK␤ to NF-␬B activation induced by B95.8LMP1 and CAO-LMP1 are clearly warranted, and mutants of B95.8-LMP1 should help to map the signaling pathways responsible for the inhibition of epithelial differentiation. Overall, this study has identified functional differences between the prototype B95.8-LMP1 and the NPC-derived CAO-LMP1 isolate when stably expressed in human epithelial cells. Our findings, along with those of others, clearly demonstrate that sequence differences between prototype and variant delLMP1 genes translate into functional differences in vivo. With regard to CAO-LMP1, recent work indicates that sequences outside the CTAR2 domain are responsible for the distinct effects of transiently expressed CAO-LMP1 in lymphoid cells (Johnson et al., 1998), and these may also be responsible for the functional differences observed in this study. To address this issue, studies are currently under way to determine the effects of B95.8-LMP1 mutants and of chimerae of CAO- and B95.8-LMP1 on the epithelial cell phenotype. Whether the CAO-LMP1 isolate is representative of other delLMP1 variants or whether it is unique among variants carrying the 30-bp deletion requires further assessment. Thus it is important to isolate and sequence other LMP1 genes from NPC biopsies and study their behaviour in vitro to help determine the role of variant LMP1 genes in the pathogenesis of NPC. Although the predominantly inert phenotypic behaviour of CAO-LMP1 when stably expressed in human epithelial cells is puzzling, continued study of this and other variant LMP1 isolates should help to further dissect the signaling pathways activated by LMP1 as well as provide insights into the contribution of LMP1 sequence variation to the pathogenesis of EBVassociated tumours. MATERIALS AND METHODS

213

taining 400 ␮g/ml G418 (GIBCO). Single G418-resistant clones were cloned and subsequently expanded for further analysis. Analysis of B95.8- and CAO-LMP1 expression in stable SCC12F cells LMP1 expression was confirmed using standard immunoblotting procedures. Briefly, 100 ␮g of a total protein extract from each clone (equivalent to ⬃10 6 cells) was separated on a 7.5% SDS-polyacrylamide gel. Separated proteins were electroblotted onto nitrocellulose membranes and, after blocking for 2 h with PBS-Tween (0.1%) containing 5% fat-free skimmed milk, incubated overnight with CS1–4, a pool of murine monoclonal antibodies specific for the cytosolic C-terminus of LMP1 (Rowe et al., 1987). After two 60-min washes in PBSTween (0.1%), the nitrocellulose membrane was incubated for 90 min with horseradish peroxidase-conjugated goat anti-mouse immunoglobulin (1:1000) (Sigma). After an additional two 60-min washes in PBS-Tween (0.1%), antibody–protein complexes were visualised by enhanced chemiluminescence (ECL; Amersham International, Buckinghamshire, UK). Analysis of cell growth kinetics Exponentially growing cultures of SCC12F vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones were recovered as single-cell suspensions by trypsinisation. Cells were plated out at a concentration of 10 4 cells/well in a 6-well plate in duplicate (Nunc). Twenty-four hours after seeding, the medium was replaced to removed nonadherent cells; thereafter, cells were cultured for a period of 10 days with a medium change at days 3 and 6. Cell growth was assessed at 48-h intervals, and viable cell number was determined by trypan blue exclusion. Cell number is expressed as cells ⫻ 10 4/ml.

Cell culture and isolation of SCC12F cells stably expressing B95.8- and CAO-LMP1

Sensitivity of SCC12F cells expressing B95.8-LMP1 and CAO-LMP1 to TNF-␣

The SCC12F cell line was grown in DMEM/F-12 (3:1 mixture) supplemented with 5% foetal calf serum (GIBCO, Paisley, Scotland), 2 mM glutamine, hydrocortisone (0.4 ␮g/ml) (Up-John Ltd., UK), and the antibiotics penicillin (1000 U/ml) and streptomycin (1 mg/ml) (Sigma Chemical, Poole, Dorset, UK). For routine culture, SCC12F cells were grown at clonal density (10 5 cells/9-cm petri dish) on an irradiated 3T3 fibroblast feeder cell layer (Parkinson et al., 1983). Stable clones of the SCC12F cell line expressing either B95.8 or CAO-LMP1 were generated by electroporation with either pCM1B95.8 or pCM1CAO, respectively. The empty vector pCM1 was also included as a negative control. Stable, drug-resistant clones were isolated after seeding at clonal density on a G418-resistant 3T3 feeder cell layer and selection in medium con-

Cultures of vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones were trypsinised and plated out at 10 4 cells/well onto a 96-well plate in triplicate. The next day, cells were treated with various concentrations of TNF-␣ (0.1–100 ng/ml) (R and D Systems Europe, Oxford, UK) in the presence of 10 ␮g/ml cycloheximide (Sigma). At 24–48 h after treatment, cell viability was assessed using the MTT (3-[4,4-dimethylthiazol-2yl]-2,5-diphenyl tetrazolium bromide; Sigma) assay (Mossmann, 1983). Briefly, 20 ␮l of a 5 mg/ml stock of MTT in PBS was added to each well, and the cells were incubated with the substrate for 5 h. After the incubation, spent medium was aspirated from the wells, and the formazan crystals solubilised in 200 ␮l of DMSO (Fisons). The absorbance was then determined at 550

214

DAWSON ET AL.

nm on a Becton Dickinson (Le Pont de Claix, France) Multiscan. In each case, viability is expressed as a percentage relative to cells treated with cycloheximide (10 ␮g/ml) alone. Flow cytometric analysis for cell surface antigen expression Subconfluent cultures of vector control, B95.8-LMP1expressing, and CAO-LMP1-expressing clones were recovered by mild trypsination (0.0125% trypsin–0.02% EDTA). Single-cell suspensions were collected by centrifugation, counted, and distributed into the wells of a microtitre plate at 2.5 ⫻ 10 5 cells/well. Then, 50 ␮l of prediluted antibody was added to each well, and the plate was incubated on ice for 45 min. Cell surface markers examined in this study included monoclonal antibodies specific for CD23 (DAKO, Bucks, UK), CD40 (G28.5; a gift from E. Clark), CD44H (Novocastra, Newcastle, UK), and CD54 (Serotec, Oxford, UK). After three washes in cold PBS (4°C), the cells were resuspended in 50 ␮l of FITC-labeled goat anti-mouse antibodies (Sigma) and incubated on ice for an additional 45 min. After an additional three washes in precooled PBS (4°C), the cell suspensions were fixed in 1% paraformaldehyde and processed for FACS analysis. Fluorescent staining of cell surface markers was analysed by flow cytometry with a Coulter EPICS XL FACSCAN. Values are presented as MFI values on an arbitrary log scale. Analysis of cytokine secretion Cultures of actively growing vector control, B95.8LMP1-expressing, and CAO-LMP1-expressing clones were trypsinised and plated out at 5 ⫻ 10 5 cells/5-cm plate. The next day, growth medium was aspirated and replaced with growth medium supplemented with 1% serum. 48 hrs later, the medium was collected for analysis. Cell debris was removed by centrifugation, and conditioned medium was frozen as aliquots at ⫺70°C before analysis. The amount of IL-6 and IL-8 present in the conditioned medium was determined by ELISA using commercially available kits from Pelikine (Amsterdam). Cytokine levels are expressed as pg/ml. Collagen raft culture and immunostaining Vector control, B95.8-LMP1-expressing, and CAOLMP1-expressing clones were analysed for their ability to terminally differentiate in the collagen raft system as previously described (Dawson et al., 1990, 1995, 1998). Briefly, 2 ⫻ 10 5 epithelial cells were seeded onto a collagen lattice (Collaborative Research, UK) containing 3T3 fibroblast feeder cells (10 5 cells/ml) and grown until confluent. Thereafter, the lattice was carefully transferred to a stainless steel grid, and the epithelial culture was exposed at the air/liquid interface for an additional 3 weeks. Under such conditions, SCC12F cells stratify and

terminally differentiate. After the appropriate time, individual rafts were snap-frozen in liquid nitrogen. Frozen sections (6 ␮m) were cut from representative rafts, fixed in cold acetone and stored at -70°C prior to use. Analysis of terminal differentiation was performed using a monoclonal antibody specific for acidic keratin family members, AE1 (ICN), and a polyclonal antisera to involucrin (Biogenesis). Bound antibody was visualised using the appropriate FITC conjugated secondary antibodies (Sigma). Raft sections were examined using an Olympus UV-fluorescence microscope. Analysis of NF-␬B and AP-1 binding activity by EMSA Cultures of vector control, B95.8-LMP1-expressing, and CAO-LMP1-expressing clones were trypsinised and plated out at 2 ⫻ 10 6 cells/9-cm plate. The next day, cell nuclei were isolated by resuspending cells in a solution containing 10 mM HEPES (pH 7.9), 1.5 mM magnesium chloride, 10 mM potassium chloride, and 0.5 mM DTT, and protease inhibitors and were subjected to lysis in a buffer of 20 mM HEPES (pH 7.9), 25% glycerol, 1.5 mM magnesium chloride, 0.42 M sodium chloride, 0.2 mM EDTA, 0.5 mM DTT, and protease inhibitors. Protein concentration of isolated nuclear extracts was determined by the Bio-Rad protein assay according to manufacturer’s instructions (Bio-Rad, Hercules, CA). For EMSAs, a 29-bp HIV-␬B probe (5⬘-GATCAGGGACTT-TCCGCTGGGGACTTTCC-3⬘) or a 22-bp collagenase TRE probe (5⬘-AGCTTGATGAGTC-AGCCGGATC-3⬘) were made by annealing complementary synthetic oligonucleotides and end-labeling using [␥- 32P]ATP (Amersham) and T4 polynucleotide kinase (Boehringer-Mannheim). Binding reactions containing 10 ␮g of nuclear protein extract, 1 ␮g of poly(dI/dC) (Pharmacia), 0.1 ng of HIV-kB or TRE probe labeled to a specific activity of 2 ⫻ 10 8 cpm/␮g, and binding buffer (4% glycerol, 1 mM EDTA, 5 mM DTT, 0.01 M Tris–HCl, pH 7.5, and 5 mM KCl) in a volume of 20 ␮l were incubated for 30 min at room temperature and then resolved by electrophoresis through a 5% polyacrylamide gel in a 0.55⫻ TBE buffer. Gels were then dried down and exposed to X-ray film for autoradography (Kodak). For supershift experiments, nuclear extracts were incubated with 1 ␮l of polyclonal antisera to the subunit of interest for 30 min on ice before electrophoresis. Polyclonal antisera to p65, p50, and c-rel were purchased from Santa Cruz. Analysis of Jun-N-terminal kinase activity by in vitro kinase assay JNK assays were performed as previously described (Eliopoulos and Young, 1998). Briefly, 80–90% confluent cultures of SCC12F vector control, B95.8-LMP1-expressing, or CAO-LMP1-expressing clones were lysed in 500 ␮l of kinase lysis buffer (20 mM Tris, pH 7.6, 0.5% Triton X-100, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 2 mM

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1

sodium orthovanadate, 10 ␮g/ml aprotinin, 10 ␮g/ml leupeptin, and 1 mM DTT) on ice for 15 min. Nonsolubilised material was pelleted by centrifugation, and the protein concentration was determined by Lowry assay using a commercially available assay (Bio-Rad). c-Jun N-terminal kinases were immunoprecipitated from 250 ␮g of total protein extracts using 1 ␮g of JNK1 (C17; Santa Cruz) by incubating end-over-end at 4°C for 1 h. Protein–antibody complexes were collected using Protein G–Sepharose (Pharmacia) for 2–3 h. After immunoprecipitation, the beads were washed once with kinase lysis buffer and twice with kinase assay buffer (20 mM HEPES, pH 7.5, 20 mM ␤-glycerophosphate, 10 mM MgCl, 1 mM DTT, 50 ␮M sodium vanadate, and 1 ␮g/ml leupeptin). After the last wash, the beads were drained with a fine needle and resuspended in 40 ␮l of kinase assay buffer containing 2 ␮g of GST-jun substrate (Stratagene, La Jolla, CA), 20 ␮M cold ATP, and 3 ␮Ci of [␥- 32P]ATP. Kinase reactions were carried out at 30°C for 30 min and stopped by the addition of 40 ␮l of 6⫻ Laemmli’s buffer and boiling for 5 min. Samples were then analysed on a 10% SDS–polyacrylamide gel. After staining with Coomassie blue to confirm that equal amounts of substrate were used, the gel was dried and exposed to X-ray film (Kodak). The degree of JNK kinase activation was then determined by densitometric scanning of the X-ray film and phosphorimaging. ACKNOWLEDGMENTS This work was supported by the Cancer Research Campaign (London, UK). A.G.E. is supported by a fellowship from the Medical Research Council. We would like to thank Sue Williams for her assistance with the photographic work and Kim Ward for assistance with FACS analysis.

REFERENCES Allen-Hoffmann, B. L., Schlosser, S. J., and Rhim, J. S. (1993). Use of RHEK-1 immortalized human keratinocytes for detection of induced mutation at the locus. Int. J. Oncol. 3, 619–625. Chen, M. L, Tsai, C. N., Liang, C. L., Shu, C. H., Huang, C. R., Sulitzeanu, D., Liu, S. T., and Chang, Y. S. (1992). Cloning and characterization of the latent membrane protein (LMP) of a specific Epstein-Barr virus variant derived from the nasopharyngeal carcinoma in the Taiwanese population. Oncogene 7, 2131–2140. Chen, W. G., Chen, Y. Y., Bacchi, M. M., Bacchi, C. E., Alvarenga, M., and Weiss, L. M. (1996). Genotyping of Epstein-Barr virus in Brazilian Burkitt’s lymphoma and reactive lymphoid tissue-type A with a high prevalence of deletions within the latent membrane protein gene. Am. J. Pathol. 148, 17–23. Chen, W. X., and Bowden, G. T. (1999). Activation of p38 MAP kinase and ERK are required for ultraviolet-B induced c-fos expression in human keratinocytes. Oncogene 18, 7469–7476. Cheung, S. T., Leung, S. F., Lo, K. W., Chiu, K. W., Tam, J. S. L., Fok, T. F., Johnson, P. J., Lee, J. C. K., and Huang, D. P. (1998). Specific latent membrane protein 1 gene sequences in type 1 and type 2 EpsteinBarr virus from nasopharyngeal carcinoma in Hong Kong. Int. J. Cancer 76, 399–406. Dawson, C. W., Dawson, J., Jones, R., Ward, K., and Young, L. S. (1998). Functional differences between BHRF1, the Epstein-Barr virus-en-

215

coded Bcl-2 homologue, and Bcl-2 in human epithelial cells. J. Virol. 72, 9016–9024. Dawson, C. W., Eliopoulos, A. G., Dawson, J., and Young, L. S. (1995). BHRF1, a viral homologue of the Bcl-2 oncogene, disturbs epithelial cell differentiation. Oncogene 9, 69–77. Dawson, C. W., Rickinson A. B., and Young, L. S. (1990). Epstein-Barr virus latent membrane protein inhibits human epithelial cell differentiation. Nature 344, 777–780. Eliopoulos, A. G., Blake, S. M. S., Floettmann, J. E., Rowe, M., and Young, L. S. (1998). Epstein-Barr virus-encoded latent membrane protein 1 activates the JNK pathway through its extreme C-terminus via a mechanism involving TRADD and TRAF2. J. Virol. 73, 1023–1035. Eliopoulos, A. G., Stack, M., Dawson, C. W., Kaye, K. M., Hodgkin, L., Sihota, S., Rowe, M., and Young, L. S. (1997). Epstein-Barr virusencoded LMP1 and CD40 mediate IL-6 production in epithelial cells via an NF-␬B pathway involving TNF-receptor-associated factors. Oncogene 14, 2899–2916. Eliopoulos, A. G, and Young, L. S. (1998). Activation of the c-jun Nterminal kinase (JNK) pathway by the Epstein-Barr virus-encoded latent membrane protein 1 (LMP1). Oncogene 16, 1731–1742. Eliopoulos, A. G., Gallagher, N. J., Blake, S. M. S., Dawson, C. W., and Young, L. S. (1999). Activation of the p38 mitogen-activated protein kinase pathway by Epstein-Barr virus-encoded latent membrane protein 1 coregulates interleukin-6 and interleukin-8 production. J. Biol. Chem. 274, 16085–16096. Fahraeus, R., Rymo, L., Rhim, J. S., and Klein, G. (1990). Morphological transformation of human keratinocytes expressing the LMP gene of Epstein-Barr virus. Nature 345, 447–449. Floettmann, J. E., Eliopoulos, A. G., Jones, M., Young, L. S., and Rowe, M. (1998). Epstein-Barr virus latent membrane protein-1 (LMP1) signalling is distinct from CD40 and involves physical cooperation of its two C- terminus functional regions. Oncogene 17, 2383–2392. Gires, O., Kohlhuber, F., Kilger, E., Baumann, M., Kieser, A., Kaiser, C., Zeidler, R., Scheffer, B., Ueffing, M., and Hammerschmidt, W. (1999). Latent membrane protein 1 of Epstein-Barr virus interacts with JAK3 and activates STAT proteins. EMBO J. 18, 3064–3073. Grimaldi, J. C., Torres, R., Kozak, C. A., Chang, R., Clark, E. A., Howard, M., and Cockayne, D. A. (1991). Genomic structure and chromosomal mapping of the murine CD40 gene. J. Immunol. 149, 3921–3926. Henderson, S., Rowe, M., Gregory, C., Croom-Carter, D., Wang, F., Longnecker, R., Kieff, E., and Rickinson, A. B. (1991). Induction of bcl-2 expression by Epstein-Barr virus latent membrane protein 1 protects infected B cells from programmed cell death. Cell 65, 1107–1115. Hu, L. F., Chen, F., Zheng, X., Ernberg, I., CAO, S. L., Christensson, B., Klein, G., and Winberg, G. (1993). Clonability and tumorigenicity of human epithelial cells expressing the EBV-encoded membrane-protein LMP1. Oncogene 8, 1575–1583. Hu, L. F., Zabarovsky, E. R., Chen, F., CAO, S. L., Ernberg, I., Klein, G., and Winberg, G. (1991). Isolation and sequencing of the Epstein-Barr virus BNLF-1 gene (LMP1) from a Chinese nasopharyngeal carcinoma. J. Gen. Virol. 72, 2399–2409. Hu, Y., Baud, V., Delhase, M., Zhang, P., Deerinck, T., Ellisman, M., Johnson, R., and Karin M. (1999). Abnormal morphogenesis but intact IKK activation in mice lacking the IKK␣ subunit of the I␬B kinase. Science 284, 318–320. Huen, D. S., Henderson, S. A., Croom-Carter, D., and Rowe, M. (1995). The Epstein-Barr virus latent membrane protein 1 (LMP1) mediates activation of NF-␬B and cell-surface phenotype via 2 effector regions in its carboxy terminal cytoplasmic domain. Oncogene 10, 549–560. Izumi, K. M., and Kieff, E. D. (1997a). The Epstein-Barr virus oncogene product latent membrane protein 1 engages the tumour necrosis factor receptor-associated death domain protein to mediate B lymphocyte growth transformation and activate NF-␬B. Proc. Natl. Acad. Sci. USA 94, 12592–12597. Izumi, K. M., and Kieff, E. D. (1997b). The Epstein-Barr virus LMP1 amino acid sequence that engages tumour necrosis factors is critical for

216

DAWSON ET AL.

primary B lymphocyte growth transformation. Proc. Natl. Acad Sci. USA 94, 1447–1452. Johnson, R. J., Stack, M., Hazlewood, S. A., Jones, M., Blackmore, C. G., Hu, L. F., and Rowe, M. (1998). The 30-base-pair deletion in Chinese variants of the Epstein-Barr virus LMP1 gene is not the major effector of functional differences between variant LMP1 genes in human lymphocytes. J. Virol. 72, 4038–4048. Kaye, K. M., Izumi, K. M., and Kieff, E. (1993). Epstein-Barr virus latent membrane protein 1 is essential for B-lymphocyte growth transformation. Proc. Natl. Acad. Sci. USA 90, 9150–9154. Kaye, K. M., Izumi, K. M., Li, H., Johannsen, E., Davidson, D., Longnecker, R., and Kieff, E. (1999). An Epstein-Barr virus that expresses only the first 231 LMP1 amino acids efficiently initiates primary B-lymphocyte growth transformation. J. Virol. 73, 10525–10530. Kaye, K. M., Izumi, K. M., Mosialos, G., and Kieff, E. (1995). The EpsteinBarr virus LMP-1 cytoplasmic carboxy terminus is essential for Blymphocyte transformation: Fibroblast cocultivation complements a critical function within the last 155 residues. J. Virol. 69, 675–683. Khanim, F., Yao, Q-Y., Niedobitek, G., Sihota, S., Rickinson, A. B., and Young, L. S. (1996). Analysis of Epstein-Barr virus gene polymorphisms in normal donors and in virus-associated tumors from different geographical locations. Blood 88, 3491–3501. Kieff, E. (1996). Epstein-Barr virus and its replication. In “Fields’ Virology,” 3rd ed., Vol. 2 (B. N. Fields, D. N. Knipe, and P. M. Howley, Eds.), pp. 2343–2396. Lippincott-Raven Publishers, Philadelphia. Kieser, A., Kilger, E., Gires, O., Ueffing, M., Kolch, W., and Hammerschmidt, W. (1997). Epstein-Barr virus latent membrane protein-1 triggers AP-1 activity via the c-Jun N-terminal kinase cascade. EMBO J. 16, 6478–6485. Knecht, H., Bachmann, E., Brousset, P., Rothenberger, S., Einsele, H., Lestou, V. S., Delsol, G., Bachmann, F., Ambros, P. F., and Odermatt, B. F. (1995). Mutational hot-spots within the carboxy-terminal region of the LMP1 oncogene of Epstein-Barr virus are frequent in lymphoproliferative disorders. Oncogene 10, 523–528. Laherty, C. D., Hu, H. M., Opipari, A. W., Wang, F., and Dixit, V. M. (1992). The Epstein-Barr virus LMP1-gene product induces A20-zinc finger protein expression by activating nuclear factor-kappa-B. J. Biol. Chem. 267, 24157- 24160. Lee, M. S., Yang J. H., Salehi, Z., Arnstein, P., Chen, L. S., Jay, G., and Rhim, J. S. (1993). Neoplastic transformation of a human keratinocyte cell-line by the v-Fos oncogene. Oncogene 8, 387–393. Li, S.-N., Chang, Y.-S., and Liu, S.-H. (1996). Effect of a 10-amino-acid deletion on the oncogenic activity of latent membrane protein-1 of Epstein-Barr virus. Oncogene 12, 2129–2135. Li, Q., D. Van Antwerp, D., Mercurio, F., Lee, K-F., and Verma, I. M. (1999). Severe liver degeneration in mice lacking the I␬B kinase 2 gene. Science 284, 321–325. May, M. J., and Ghosh, S. (1999). I␬B kinases: Kinsmen with different crafts. Science 284, 271–272. Miller, W. E., Cheshire, J. L., and Raab-Traub, N. (1998a). Interaction of tumor necrosis factor receptor-associated factor signaling proteins with the latent membrane protein 1 PXQXT motif is essential for induction of epidermal growth factor receptor expression. Mol. Cell. Biol. 18, 2835–2844. Miller, W. E, Cheshire, J. L., Baldwin, A. S., and Raab-Traub, N. (1998b). The NPC derived C15 LMP1 protein confers enhanced activation of NF-␬B and induction of the EGFR in epithelial cells. Oncogene 16, 1869–1877. Miller, W. E., Earp, H. S., and Raab-Traub, N. (1995). The Epstein-Barr virus latent membrane protein 1 induces expression of the epidermal growth factor receptor. J. Virol. 69, 4390–4398. Miller, W. E., Edwards, R. H., Walling, D. M., and Raab-Traub, N. (1994). Sequence variation in the Epstein-Barr virus latent membrane protein-1. J. Gen. Virol. 75, 2729–2740. Miller, W. E., Mosialos, G., Kieff, E., and Raab-Traub, N. (1997). EpsteinBarr virus LMP1 induction of the epidermal growth factor receptor is

mediated through a TRAF signaling pathway distinct from NF-␬B activation. J. Virol. 71, 586–594. Mosialos, G., Birkenbach, M., Yalamanchili, R., VanArsdale, T., Ware, C., and Kieff, E. (1995). The Epstein-Barr virus transforming protein LMP 1 engages signalling proteins for the tumour necrosis factor receptor family. Cell 80, 389–399. Mosmann, T. (1983). Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays. J. Immunol. Methods 65, 55–63. Osato, T., and Imai, S. (1996). Epstein-Barr virus and gastric carcinoma. Semin. Cancer. Biol. 7, 175–182. Parkinson, E. K., Grabham, P., and Emmerson, A. (1983). A sub-population of cultured human keratinocytes which is resistant to the induction of terminal differentiation-related changes by phorbol,12myristate,13-acetate: Evidence for an increase in the resistant population following transformation. Carcinogenesis 4, 857–861. Raab-Traub, N. (1996). Pathogenesis of Epstein-Barr virus and its associated malignancies. Semin. Virol. 7, 315–323. Razzzaque, A., Williams, O., Wang, J. H., and Rhim, J. S. (1993). Neoplastic transformation of immortalized human epidermal keratinocytes by 2 HHV-6 DNA clones. Virology 195, 113–120. Rheinwald, J. G., and Beckett, M. A. (1981). Tumourigenic keratinocyte lines requiring anchorage and fibroblast support cultured from human squamous cell carcinomas. Cancer Res. 41, 1657–1663. Rickinson, A. B., and Kieff, E. (1996). Epstein-Barr virus. In “Fields’ Virology,” 3rd ed., Vol. 2 (B. N. Fields, D. N. Knipe, and P. M. Howley, Eds.), pp. 2397–2446. Lippincott-Raven Publishers, Philadelphia,. Rowe, M., Evans, H. S., Young, L. S., Hennessy, K., Kieff, E., and Rickinson, A. B. (1987). Monoclonal antibodies to the latent membrane protein of Epstein-Barr virus reveal heterogeneity of the protein and inducible expression in virus-transformed cells. J. Gen. Virol. 68, 1575–1586. Rowe, M., Pengpilon, M., Huen, D. S., Hardy, R., Croom-Carter, D., Lundgren, E., and Rickinson, A. B. (1994). Up-regulation of Bcl-2 by the Epstein-Barr virus latent membrane protein LMP1: A B-cellspecific response that is delayed relative to NF-␬B activation and to induction of cell-surface markers. J. Virol. 68, 5602–5612. Sample, J., Kieff, E. F., and Kieff, E. D. (1994). Epstein-Barr virus type-1 and type-2 have nearly identical LMP1 transforming genes. J. Gen. Virol. 75, 2741–2746. Seitz, C. S, Lin, Q., Deng, H., and Khavari, P. A. (1998). Alterations in NF-␬B function in transgenic epithelial tissue demonstrate a growth inhibitory role for NF-␬B. Proc. Natl. Acad. Sci. USA 95, 2307–2312. Sung, N. S., Edwards, R. H., Seillier-Moiseiwitsch, F., Perkins, A. G., Zeng, Y., and Raab-Traub, N. (1998). Epstein-Barr virus strain variation in nasopharyngeal carcinoma from the endemic and non-endemic regions of China. Int. J. Cancer. 76, 207–215. Sylla, B. S., Hung, S. C., Davidson, D. M., Hatzivassiliou, E., Malinin, N. L., Wallach, D., Gilmore, T. D., Kieff, E., and Mosialos, G. (1998). Epstein-Barr virus-transforming protein latent infection membrane protein 1 activates transcription factor NF-␬B through a pathway that includes the NF-␬B-inducing kinase and the I␬B kinases IKK␣ and IKK␤. Proc. Natl. Acad. Sci. USA 95, 10106–10111. Takeda, K. Takeuchi, O., Tsujimura, T., Itami, S., Adachi, O., Kawai, T., Sanjo, H., Yoshikawa, K., Terada, N., and Akira, S. (1999). Limb and skeletal abnormalities in mice lacking IKK␣. Science 284, 313–316. Voraberger, G., Schafe R., and Stratowa C. (1991). Cloning of the human gene for intercellular-adhesion molecule-1 and analysis of its 5⬘regulatory region: Induction by cytokines and phorbol ester. J. Immunol. 147, 2777–2786. Walling, D. M., Shebib, N, Weaver, S. C., Nichols, C. M., Flaitz, C. M., and Webster-Cyriaque, J. (1999). The molecular epidemiology and evolution of Epstein-Barr virus: Sequence variation and genetic recombination in the latent membrane protein-1 gene. J. Infect. Dis. 179, 763–774. Wang, D, Liebowitz, D., Wang, F., Gregory, C., Rickinson, A. B., Larson, R., Springer, T., and Kieff, E. (1988). Epstein-Barr virus latent infection

FUNCTIONAL DIFFERENCES BETWEEN B95.8-LMP1 AND CAO-LMP1 membrane protein alters the human B lymphocyte phenotype: Deletion of the amino terminus abolishes activity. J. Virol. 62, 4173–4184. Wang, F., Gregory, C. Sample, C., Rowe, M., Liebowitz, D., Murray, R., Rickinson, A. B., and Kieff, E. (1990). Epstein-Barr virus latent membrane protein (LMP1) and nuclear protein 2 and protein 3C are effectors of phenotypic changes in B lymphocytes: EBNA-2 and LMP1 cooperatively induced CD23. J. Virol. 64, 2309–2318. Wang, S., Rowe, M., and Lundgren, E. (1996). Expression of the Epstein Barr virus transforming protein LMP1 causes a rapid and transient stimulation of the Bcl-2 homologue Mcl-1 levels in B-cell lines. Cancer Res. 56, 4610–4613. Welter, J. F., and Eckert, R. L. (1995). Differential expression of the Fos and Jun family members c-Fos, FosB, Fra-1, Fra-2, c-Jun, JunB, and

217

JunD during human epidermal keratinocyte differentiation. Oncogene 11, 2681–2687. Wilson, J. B., Weinberg, W., .Johnson, R., Yuspa, S., and Levine, A. J. (1990). Expression of the BNLF1–1 oncogene of Epstein-Barr virus in the skin of transgenic mice induces hyperplasia and aberrant expression of keratin-6. Cell 61, 1315–1327. Yang, J. H., Thraves, P., Dritschilo, A., and Rhim, J. S. (1992). Neoplastic transformation of immortalized human keratinocytes by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Cancer Res. 52, 3478–3482. Zheng, X., Yuan, F., Hu, L. F., Klein, G., and Christensson, B. (1994). Effect of B-lymphocyte- and NPC-derived EBV-LMP1 gene expression on in vitro growth and differentiation of human epithelial cells. Int. J. Cancer 57, 747–753.