INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Through evolution, multicellular organisms have developed an intricate cellular signaling network to cope with changes in the cellular microenvironment. External stimuli trigger signaling through various cellular molecules and so influence the fate of the cell. Tyrosine kinases and the molecules upon which they act constitute components of an important and tightly regulated network of pathways. Under normal physiological conditions, these tyrosine kinases are key cellular elements determining cell proliferation, differentiation, and apoptosis (48). On the other hand, tumor formation that escapes the regular control loop usually involves a breakdown of restricted tyrosine kinase regulation, leading to rapid or irregular growth or even metastasis (26). Furthermore, it has been reported that the transforming ability of several oncogenic viruses may be attributable to activation or structural mimicking of tyrosine kinase receptors by viral oncogenes. For example, v-sis of the simian sarcoma retrovirus is a homolog of platelet-derived growth factor (PDGF) and can activate PDGF receptor, and the E5 gene product of bovine papillomavirus can activate PDGF receptor  (10, 22, 39, 40). Besides their direct effects in promoting cell growth, the oncogenic mutated ras and v-src may induce the activation of vascular endothelial growth factor, which is involved in tumor angiogenesis, and this activation may play a role in enhancing tumor formation (44).

Epstein-Barr virus (EBV), a human herpesvirus, is capable of immortalizing primate B cells and human primary epithelial cells in vitro (25, 37). The oncogenic potential of EBV is evident in vivo, since it causes B-cell lymphomas and promotes metastasis in animal models (49, 50). Seroepidemiologic and pathologic data reveal a strong association between EBV and various malignancies, such as nasopharyngeal carcinoma (NPC), B-cell proliferative disease in immunodeficient patients, and lymphoepithelioma-like gastric carcinoma (33, 36, 58). An unusual characteristic that differentiates EBV-associated tumors from other human malignancies is that they rarely have mutations in well-known tumor suppressor genes, such as the Rb and p53 genes, or in well-defined proto-oncogenes (55, 56). Therefore, we hypothesized that EBV may encode proteins that influence the normal cellular signaling cascade, especially the versatile tyrosine kinase family, in causing tumor formation.

A likely candidate, among EBV gene products, to influence the cellular signaling cascade is the Zta protein. Structurally, Zta is a leucine zipper DNA-binding protein; functionally, it acts as the key viral transactivator and initiates the EBV lytic cascade (25), as well as influencing many cellular genes (53). Based on data from previous reports, there are three mechanisms through which Zta may exert potent effects on cell signaling. First, Zta can activate several promoters, such as c-Fos and transforming growth factor  (TGF-), or suppress c-Jun (5, 14, 47). The ability of Zta to regulate AP1 protein expression and to compete with Fos-Jun heterodimers for AP1 sites suggests that Zta has the potential to interfere with AP1-mediated cell proliferation and differentiation (53). Second, Zta may inhibit or cooperate with cellular proteins by interacting with them, for example, NF-B, p53, c-Myb, and the retinoic acid receptor (11, 17, 24, 52, 66). Third, Zta may modulate cellular transcription indirectly, through a competitive interaction with the important cellular regulator CREB-binding protein, since the interaction between Zta and CREB-binding protein may influence the transcriptional efficiency of other cellular proteins (1, 65). However, whether Zta directly or indirectly influences any cellular tyrosine kinase remains unknown. Utilizing a newly developed kinase display assay (27), we found that a cellular receptor tyrosine kinase, TKT (TRK-related tyrosine kinase, also called DDR2, Tyro 10, and CCK-2), could be upregulated by Zta at both the RNA and protein levels in a cell culture system (2, 23, 28). Importantly, the upregulation of phosphorylated TKT and the TKT downstream effector matrix metalloproteinase 1 (MMP-1) hinted that TKT might initiate a signaling cascade in Zta-expressing cells. Moreover, correlation of expression levels of Zta and TKT transcripts in NPC biopsy specimens was clearly demonstrated by quantitative PCR (Q-PCR), providing the first evidence for an effect of Zta on cellular gene expression in vivo. These findings offer insight into the virus-cell interactions and may help us elucidate the role of EBV in tumorigenesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. The RHEK-1 cell line was established by immortalizing primary human foreskin keratinocytes with adenovirus type 12-simian virus 40 (45). NPC-TW01 is an EBV-negative NPC cell line (29). Both lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum plus 100 U of penicillin per ml and 100 µg of streptomycin per ml. The other NPC-derived cell line, HONE-1, was grown in RPMI 1640 medium with the same supplements (15). Stable, Zta expression transfectants or vector controls were cultured in the same medium as that used for the parental cells, with 300 µg of neomycin (Gibco BRL, Gaithersburg, Md.) per ml. For detection of MMP-1 transcripts, cells were starved in Dulbecco's modified Eagle's medium without fetal bovine serum 24 h prior to 3 days of 50 nM collagen I (Gibco BRL) treatment.

Plasmid construction. pRc/CMV-Zta is a Zta expression plasmid (B95-8 sequences; nucleotides 103742 to 101947) driven by the cytomegalovirus promoter and was generously provided by M.-T. Liu (National Taiwan University, Taipei, Taiwan).

Transfection and establishment of Zta stable expression clones. Cells were transfected using a modified calcium phosphate method (7). Briefly, 20 µg of plasmid DNA in 360 µl of H₂O was mixed with 40 µl of 2.5 M CaCl₂ and 0.4 ml of 2× BBS [50 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (pH 6.95), 280 mM NaCl, 1.5 mM Na₂HPO₄] and incubated for 20 min at room temperature. The mixture was added dropwise to 50% confluent cells, which were then incubated for 20 h at 35°C under 3% CO₂. After the medium was refreshed, transfectants were moved to a 37°C incubator containing 5% CO₂. For the transient-transfection assay, cells were harvested 24 h after transfection. For selection of stable Zta-expressing clones, the transfectants were cultured in the presence of 400 µg (NPC-TW01 and HONE-1 cells) or 800 µg (RHEK-1) of neomycin per ml until sufficient cells were available for assay. To establish the inducible system, pUHD10.3-Zta and pUHD172-1neo were cotransfected into HONE-1 cells at a ratio of 10:1 and the clones were selected using neomycin (400 µg/ml).

Immunofluorescence assay. Transfected cells were fixed to the slides with acetone-methanol (1:1) at 20°C for 30 min. After being incubated with anti-Zta antibody 4F10 or control antibody for 1 h at 37°C (59), the cell smears were washed and incubated with 100-fold-diluted fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin G antiserum (Jackson, West Grove, Pa.). After being washed with phosphate-buffered saline (PBS), the cells were counterstained with Evans blue, mounted in 90% glycerol-PBS solution, and examined under a UV fluorescence microscope.

Western blot analysis. Cells were lysed in KBO buffer (20 mM octyglucoside, 0.5% Triton X-100, 0.3 M NaCl, 0.025 M NaPO₄ [pH 7.4], 10 µg of leupeptin per ml, 10 µg of aprotinin per ml, 1 mM phenylmethylsulfonyl fluoride) for analysis of Zta and TKT protein expression. A 20-µg portion of extracted protein was denatured at 100°C for 3 min and resolved on sodium dodecyl sulfate (SDS)-10% polyacrylamide gels. The proteins were electrotransferred onto a polyvinylidene difluoride nitrocellulose membrane (Millipore, Bedford, Mass.), and the blot was blocked and then incubated with anti-Zta antibody 4F10 (59), anti-TKT antibody (Santa Cruz, Santa Cruz, Calif.), anti-tubulin antibody (Amersham Pharmacia, Piscataway, N.J.), or anti-actin antibody (Sigma, St. Louis, Mo.) for 1 h. After being washed with Tris-buffered saline containing Tween 20 (TBS-T), the blot was incubated with peroxidase-labeled goat anti-mouse or rabbit anti-goat antibody (Jackson, Santa Cruz, Calif.). Finally, bands were visualized using the Renaissance kit (NEN, Boston, Mass.).

Kinase display. A tyrosine kinase display strategy was used to visualize the tyrosine kinase profile in Zta and vector transfectants (27). Total cellular RNA was extracted as specified by the manufacturer (REzol C&T; Protech, Taiwan, Republic of China). First, the cDNAs of tyrosine kinase were generated using 25 µg of total RNA as template and 2.4 µM degenerate primers in the reverse transcription (RT) reaction. The sequences of the degenerate primers for the RT reaction are as follows: TK-RTA, 5'-CCRHANGMCCA-3'; and TK-RTB, 5'-CCRHAVMTCCA-3'. The mixtures were then denatured at 65°C for 10 min, and 1 mM deoxynucleoside triphosphate (dNTP), 10 U of RNasin and 0.5 U of reverse transcriptase were added to give a 25-µl reaction mixture. RT was carried out at 42°C for 30 min. The mixtures were then purified by phenol-chloroform extraction and CHROMA SPIN-200 column chromatography (Clontech, Palo Alto, Calif.).

PCR and Q-PCR. cDNA was synthesized using 5 µg of total RNA as the template and random hexamers as primers, as described in the manufacturer's protocol (Gibco BRL).

Northern blot analysis. mRNA was purified from total cellular RNA using magnetic conjugated oligo(dT) beads (MACS, Auburn, Calif.). Denatured mRNA was separated on a 0.8% agarose gel and transferred onto membranes (Hybond-N; Amersham Pharmacia). To detect the TKT transcripts in Zta-expressing cells, a ³²P-labeled RNA probe was prepared by in vitro transcription (TKT nucleotide sequences 425 to 1550 and 1616 to 2958). The blot was prehybridized for 2 h in prehybridization buffer (5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.2% SDS, 20 mM sodium phosphate buffer, 0.6% polyvinylpyrrolidone, 0.1% Ficoll, 50% foramide, 0.1% bovine serum albumin, 250 µg of denatured salmon sperm DNA per ml, 10 µg of tRNA per ml) and hybridized overnight in hybridization buffer (5× SSC, 10% sodium dextran sulfate, 0.2% SDS, 0.1% bovine serum albumin, 0.6% polyvinylpyrrolidone, 20 mM sodium phosphate buffer, 250 µg of denatured salmon sperm DNA per ml, 10 µg of tRNA per ml, 50% foramide) at 68°C. After stringent washing in 0.1% SDS-0.1× SSC for 5 h at 68°C with five or six changes of buffer, the signals were visualized using a PhosphorImager (Storm 840; Molecular Dynamics, Sunnyvale, Calif.).

Immunoprecipitation. Cells treated with 1 mM sodium vanadate for 90 min were solubilized with KBO lysis buffer as detailed above for the Western blot analysis. A 500-µg sample of protein was incubated overnight with 4 µg of antiphosphotyrosine antibody (4G10; Upstate Biotechnology, Lake Placid, N.Y.) or irrelevant mouse anti-actin antibody (Sigma) at 4°C. Then the immunocomplexes were precipitated with protein A-Sepharose at 4°C for 2 h and washed with PBS five times. Finally the immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) (10% polyacrylamide) and the Western blot analysis was carried out as described above.

Biopsy samples. Ten primary NPC biopsy samples, pathologically validated, were obtained from the Department of Otolaryngology, National Taiwan University Hospital. Ten samples of nasopharyngeal tissues from patients with lymphohyperplasia (LH), containing nonmalignant epithelium and lymphocytes without pathological evidence of cancer cells, were also tested by Q-PCR.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Establishment of Zta expression cell clones. It has been known for a long time that the EBV lytic cycle occurs in vivo in epithelial cells, and expression of the Zta protein has been documented in NPC and oral hairy leukoplakia (OHL) (8, 30, 64). However, it is difficult to obtain sufficient material to investigate the effect of Zta in epithelial cells. To overcome this obstacle, three human epithelial cell lines which express Zta have been established in RHEK-1, NPC-TW01, and HONE-1. Zta protein expression may be detected by Western blot analysis in transient transfectants and in stable clones of RHEK-Z2, RHEK-Z3, and NPC-TW01-Z cells (Fig. 1A). The amount of Zta protein expressed in all transfectants was equivalent to or below that in P3HR-1 cells which were activated in the lytic cycle (Fig. 1A). An immunofluorescence assay was carried out to determine the percentage of cells expressing the Zta protein and its location. As expected, most of the Zta protein was located in the nuclei of the cells. Large quantities of the protein were detected in transient transfectants but only in a small percentage of the cells (Fig. 1B). Conversely, more than 90% of stable transfectants express the Zta protein but at a lower level (data not shown).

View larger version (93K): [in this window] [in a new window]   FIG. 1.   Detection of Zta protein in transfectants by Western blotting and immunofluorescence assay. (A) Western blot analysis of Zta protein expression in transient transfectants or stable clones by anti-Zta antibody 4F10. pRc/CMV-Zta, the Zta-expressing plasmids, were transfected into RHEK-1 and NPC-TW01, yielding transient transfectants or stable clones RHEK-Z2, RHEK-Z3, and NPC-TW01-Z. P3HR1 cells stimulated with 40 ng of TPA per ml and 3 mM sodium butyrate served as the positive control. The same blot was probed with an anti-tubulin monoclonal antibody as an internal control. (B) An immunofluorescence assay was carried out to determine the location of the Zta protein. TW01 cells were transiently transfected with pRc/CMV-Zta or pRc/CMV. Cell smears were reacted with anti-Zta antibody 4F10 or anti-EA-D antibody 88A9 as an unrelated antibody control and then incubated with fluorescein isothiocyanate-conjugated goat anti-mouse antiserum. The bright nuclear fluorescence indicates Zta protein expression. (C) Establishment of a tetracycline-regulating Zta expression system in HONE-1 cells. A Western blot analysis of Zta protein expression in tetracycline-regulating HONE-1 cells is shown. Zta expression in HONE-1-Zta (lane Z) and in control cells (lane V) was determined in the absence or present of 1 µg of doxycycline per ml for 48 h. The same blot was probed with an anti-actin antibody as an internal control.

Identification of a unique tyrosine kinase in Zta-expressing clones using the kinase display assay. Because of the importance of tyrosine kinases in cellular control, we focused our investigation on the identification of Zta-regulated tyrosine kinases. Taking advantage of a two-step kinase display system, the tyrosine kinase profiles of Zta-expressing and control cells were displayed following amplification of tyrosine kinase transcripts using degenerate primers, which were based on the conserved DVW and DFG motifs found in most tyrosine kinases (27). The advantage of this newly developed assay is that each tyrosine kinase may be recognized by its unique restriction enzyme digestion pattern in the variable regions between the DVW and DFG motifs (27). Analysis of the restriction profiles revealed that the most prominent, differentially regulated tyrosine kinase in the Zta-expressing clones is TKT (also known as DDR2) (23). The unique TKT profile, which was apparent in four different restriction enzyme profiles (Bsp1286I, BstNI, HpaII, and AciI), was seen only in Zta-expressing stable clones and not in vector controls (Fig. 2). Similar findings were observed following transient transfection (data not shown). The differences in the HpaII and AciI digestion profiles for NPC-TW01-Z are not as clear as those for RHEK-Z2 and RHEK-Z3. This could be due to the comigration of bands derived from other kinases, which may be present in the NPC-TW01 cell line but not in the RHEK-1 cell line.

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Exhibition of expression profiles of TKT using the kinase display assay. Total cellular RNA extracted from Zta-expressing stable clones NPC-TW01-Z, RHEK-Z2, and RHEK-Z3 and vector controls was subjected to RT, amplified by PCR, and digested with restriction enzymes as described in Materials and Methods. Four differentially displayed bands are evident, using the restriction enzymes Bsp1286I, BstNI, HpaII, and AciI, in Zta transfectants but not in the vector controls. From the digestion pattern and fragment sizes, this expression profile was recognized as TKT (27).

Verification of the induction of TKT expression by Zta. To confirm the induction of TKT transcripts in Zta-expressing cells, TKT-specific primers located in the extracellular region were used for RT-PCR. The upregulation of TKT transcripts was demonstrated clearly in Zta-expressing transient transfectants and in three independent stable clones, NPC-TW01-Z, RHEK-Z2, and RHEK-Z3, but was not apparent in control cells (Fig. 3A). Similar results were obtained with HONE-1 cells with an inducible Zta expression system (Fig. 3B). In the presence of doxycycline, the expression of TKT increased significantly followed by an increase in Zta expression; however, trace levels of TKT transcripts could be detected in the absence of doxycycline due to the low level of Zta leakage. Certainly, there were no detectable TKT transcripts in the pUHD10.3 vector control, even in the presence of doxycycline (Fig. 3B). These results indicated that Zta could upregulate TKT at the transcriptional level, independent of cell type (NPC-TW01, RHEK-1, or HONE-1 cells) and transfection system (transient transfection, constitutive expression stable clones, or tet-on system).

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Induction of TKT transcription by Zta is revealed by RT-PCR and Northern blotting. (A) Detection of TKT transcripts in Zta-expressing transient transfectants or in stable clones of NPC-TW01 and RHEK-1 cells using RT-PCR. The reaction without reverse transcriptase served as a control to rule out the possibility of contamination with genomic DNA. The reaction detecting hypoxanthine phosphoribosyltransferase (HP) transcripts was an internal control for RNA quality and amounts. (B) Detection of TKT transcripts by RT-PCR in HONE-1-Zta cells, with or without induction by 1 µg of doxycycline per ml, using the same procedure. (C) Northern blot analysis of TKT expression in cells expressing Zta. mRNA (2 µg) from Zta-expressing clones NPC-TW01-Z, RHEK-Z2, and RHEK-Z3 and vector controls was subjected to agarose gel electrophoresis (0.8% agarose), blotted, and hybridized with a 32P-labeled TKT probe as described in Materials and Methods. Human placental mRNA (lane P) served as a positive control. The lane marked with an asterisk was exposed for a shorter period. The same blot was probed with GAPDH as an internal control. The numbers to the left indicate the molecular size markers in kilobases.

Detection of TKT protein expression and status in Zta transfectants. We investigated TKT expression at a translational level by Western blot analysis. As shown in Fig. 4A, a protein of approximately 125 kDa was detected by a TKT-specific antibody in Zta transfectants but not in control cells; the molecular mass is consistent with a previous report (61). In this experiment, the pRc/CMV-TKT plasmid, carrying the entire TKT open reading frame, was used as a positive control.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Expression and status of TKT protein. (A) Detection of TKT protein in Zta-expressing cells by Western blot analysis. Total protein (20 µg) was extracted from Zta-expressing transfectants RHEK-Z2, RHEK-Z3, and NPC-TW01-Z and control cells and was blotted and reacted with a TKT-specific antibody. After being incubated with peroxidase-labeled goat anti-mouse immunoglobulin G, the membrane was visualized with a chemiluminescence kit. The same blot was reacted with Zta-specific antibody 4F10 to measure Zta expression and with the tubulin-specific antibody as an internal control. The two lanes marked with asterisks contained NPC-TW01 cells transfected with vector control (lane V*) or transfected with pRc/CMV-TKT plasmid (lane TKT*) and were the controls for TKT expression. (B) Phosphorylation status of TKT in Zta-expressing stable clones. RHEK vector control (lane V) and RHEK-Z2 (lane Z2) cells were immunoprecipitated with antiphosphotyrosine antibody (4G10) or irrelevant antibody control (con), subjected to SDS-PAGE (10% polyacrylamide), and blotted with anti-TKT antibody as described in Materials and Methods. The numbers to the left indicate the molecular mass markers in kilodaltons. (C) Expression of the TKT downstream effector MMP-1 transcripts detected by RT-PCR. RNA extracted from RHEK parental (lane P), vector control (lane V), and RHEK-Z2 (lane Z2) cells was subjected to RT and amplified by PCR using Zta-, TKT-, MMP-1-, and hypoxanthine phosphoribosyltransferase (HP)-specific primers. The reaction without reverse transcriptase served as a control to rule out the possibility of contamination with genomic DNA. The reaction detecting HP transcripts was an internal control for RNA quality and amounts.

Investigation of the Zta domain required for TKT induction. Structurally, Zta is a leucine zipper, DNA-binding protein containing a DNA-binding domain, a transactivation domain, and a dimerization domain (53). We wished to determine which domains of the Zta protein are required for TKT induction. Various plasmids containing Zta sequences with deletions in the transactivation or dimerization domains were transfected into NPC-TW01 cells (13) and the levels of TKT expression were estimated using Q-PCR. It may be seen that amino acids 52 to 78 in the transactivation domain (Fig. 5A) and the dimerization domain of the Zta protein (Fig. 5B) were both crucial for TKT induction. In this experiment, the input mass of RNA was standardized by GAPDH RNA levels and the data are presented after standardizing the Zta expression levels for transfection efficiency.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   The transactivation and dimerization domains of the Zta protein are required for TKT activation. To investigate the requirement of Zta protein domains for upregulation of TKT, Q-PCR was carried out to quantitate the TKT transcripts in transfectants with wild-type Zta or various deletion mutants of Zta. The Zta plasmids used in this assay included SV (control plasmid), SV/Z WT (wild-type Zta, full length), SV/Z d27/53 (mutant Zta, codons 27 to 53 in the transactivation domain deleted), SV/Z d52/78 (mutant Zta, codons 52 to 78 in the transactivation domain deleted), SV/Z d77/103 (mutant Zta, codons 77 to 103 in the transactivation domain deleted), SV/Z d102/128 (mutant Zta, codons 102 to 128 in the transactivation domain deleted), and SV/Z d127/153 (mutant Zta, codons 127 to 153 in the transactivation domain deleted) (A) and pRc/CMV (control plasmid), pRc/CMV-Z (wild-type Zta, full length), and pRc/CMV-Z DIM () (mutant Zta, codons 194 to 245; the entire dimerization domain was deleted) (B) transfected into NPC-TW01. The y axis shows the relative amounts of TKT transcripts standardized by GAPDH and Zta transcripts as described in Materials and Methods. The error bars indicate the standard deviation of triplicate tests.

Correlation of Zta and TKT expression in NPC biopsy specimens. The possibility that trace amounts of Zta protein are sufficient to induce TKT expression (Fig. 3 and 4) led us to question whether Zta induces TKT expression in vivo. Q-PCR was carried out to determine the levels of expression of Zta and TKT in NPC biopsy specimens (Fig. 6A). There was a significant correlation between the levels of expression of Zta and TKT. As controls, samples from 10 patients with LH (normal nasopharyngeal tissues containing nonmalignant squamous epithelium and lymphocytes) were tested for Zta and TKT expression. TKT transcripts were detectable only with high levels of Zta transcription, with one exception (Fig. 6B). In addition, normal epithelial cells isolated from the nasal cavity were chosen to examine the expression of Zta and TKT. In the three cases tested, Zta transcripts were barely detected in the RT-PCR assay and no or only trace amounts of TKT transcripts were demonstrated in the control cells (data not shown). Moreover, through combination of NPC and LH data, the correlation coefficient is 0.8 and the P value calculated by the F test is <0.001. Simultaneously, the correlation between latent membrane protein 1 (LMP-1), another EBV protein, and TKT was also investigated in the NPC biopsy specimens. However, we did not find any correlation between LMP-1 and TKT (unpublished data). Thus, it seemed that Zta may upregulate TKT expression in vivo, as well as in cell culture.

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Correlation of Zta and TKT transcription in NPC and LH biopsy specimens. Q-PCR was performed to detect the expression of Zta and TKT in the specimens, and GAPDH primers were used as an internal control to standardize the mass of input RNA. The y axis shows the relative amounts of TKT or Zta transcripts indicated by the fluorescence intensity, as described in Materials and Methods. The x axis shows the case numbers of NPC biopsy specimens (A) and of LH biopsy specimens (B).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Several reports suggest that EBV lytic transcripts, proteins, or antibodies to them may be detected in EBV-related diseases (25), but their contribution to carcinogenesis remains obscure. For example, Zta transcripts may be detected in NPC and OHL biopsy specimens and high titers of antibodies to Zta may be detected in individuals with NPC or other EBV-related diseases but not in healthy EBV carriers (8, 21, 25, 30, 57, 64). To date, many reports have documented that Zta is the control switch in the EBV lytic cycle and is responsible for the initiation of the lytic cascade (25, 53). Moreover, Zta affects or cooperates with not only EBV-encoded proteins but also many important cellular targets such as c-Fos, TGF-, p21, c-Jun, NF-B, p53, c-Myb, the retinoic acid receptor, and p300/CBP through a variety of mechanisms of action (1, 5, 6, 11, 14, 17, 24, 47, 52, 53, 65, 66).

In this study, we adapted the newly developed tyrosine kinase display system, which specifically displays transcripts of most of the known tyrosine kinase genes (27), to explore the effects of Zta on cellular signaling. Using kinase display, we found that TKT-specific transcripts were induced in Zta-expressing cells but not in vector control cells (Fig. 2). This finding was confirmed by RT-PCR and Northern and Western blot analysis (Fig. 3 and 4A). Furthermore, the importance of TKT in Zta-expressing cells was suggested by the tyrosine-phosphorylated status of the TKT protein and by the upregulation of TKT downstream effector MMP-1 (Fig. 4B and C). These data imply that TKT may initiate a signaling cascade. Interestingly, and consistent with the data from cultured cells, the levels of Zta and TKT transcripts in NPC and LH biopsy specimens correlated closely with each other (correlation coefficient is 0.8; P < 0.001) (Fig. 6). To our knowledge, this is the first evidence that Zta may influence the expression of cellular genes both in vitro and in vivo. Moreover, the requirement for the transactivation and dimerization domains of the Zta protein in upregulating TKT in transfected cells (Fig. 5) favors the direct involvement of Zta.

TKT is a receptor tyrosine kinase that comprises an extracellular domain with discoidin I homology, an extended single-stretch transmembrane domain, and an intracellular catalytic tyrosine kinase domain (23). Typically, proteins containing discoidin I-homologous motifs play an important physiological role in cell-cell contact or adhesion (54). For example, neurexin, which has a discoidin I domain, is believed to participate in neuron-glial cell adhesion (38). Del-1, which is expressed in endothelial cells, is the ligand for intergrin ₅₃ receptor. This ligand-receptor interaction promotes the adhesion of endothelial cells (19). However, the physiological function of TKT is not known. Recently, several studies of the expression pattern, ligands, and downstream signals of TKT have provided clues to its biological functions.

Several collagens (I, II, III, and V) have been identified as ligands which activate TKT, leading to the upregulation of MMP-1 (51, 61). Biologically, these collagens are major components of the extracellular matrix (ECM), which separates epithelial and stromal cells (32). Under certain physiological conditions, such as wound healing or tissue development, the epithelial or surrounding stromal cells secrete proteases which degrade the ECM, leading to cell migration and tissue reconstruction (32). MMPs play a major role in this process, and MMP-1 is essential for wound healing (41).

As well as facilitating tissue reconstruction and cell migration through the collagen matrix during wound healing (41, 42), MMP-1 is overexpressed in head and neck tumors, Ras-transformed papilloma-derived cells, and lung cancer and is associated with a poor prognosis for colon tumors (4, 31, 34, 35, 43). Moreover, MMP-1 expression leads to epidermal hyperplasia and increased susceptibility to tumor formation in a transgenic mouse model (9). MMP-1 may play an important role in carcinogenesis by destabilizing the ECM and basal lamina. This could allow tumor cells to grow or to metastasize by gaining access to the vascular and lymphatic systems (32).

Interestingly, TKT transcripts were suggested to be expressed in a collagen I-dependent hepatic wound-healing model in the rat (3). In this model, collagen is needed for cell cycle progression and TKT may be involved in the wound-healing process (3). Like its downstream effector, MMP-1, TKT is believed to be expressed in tumors such as prostate carcinoma, pediatric brain tumors, primary colon adenocarcinoma, and SK-Mel-2 (simian virus 40-transformed cell lines) (2, 46, 62). All these studies suggest that TKT may be involved in tissue-remodeling processes, such as wound healing or carcinogenesis (2, 3, 46, 60, 62). Collagen, TKT, and MMP-1 may be involved in regulating the microenvironment of the collagen barrier and may affect the migration of epithelial cells during wound healing or malignancy (60).

A model for NPC development may be proposed on the basis of Zta expression leading to TKT activation and MMP-1 upregulation. This newly synthesized MMP-1 can break down the ECM and promote cancer cell proliferation or migration. Furthermore, MMP-1 may cooperate with MMP9, which can be upregulated by another EBV protein, LMP-1, to facilitate tumor progression (63). These factors may explain the extraordinarily high incidence of metastases in NPC and are consistent with the finding of an in vivo increase in the level of more than one type of MMP in carcinogenesis (12, 20).