Consistent Sequence Variation of Epstein-Barr Virus Nuclear Antigen 1 in Primary Tumor and Peripheral Blood Cells of Patients with Nasopharyngeal Carcinoma¹

EBV³ is a ubiquitous human γ-herpes-virus that infects >90% of the population worldwide. It has been demonstrated that EBV infection is associated with a variety of epithelial and lymphoid neoplasms (1, 2).

In vitro, EBV infects and transforms B lymphocytes. Lymphoblastoid cell lines derived from EBV-infected B cells express six EBV nuclear antigens (EBNA-1, -2, -3A, -3B, -3C, and EBV-LP), two EBV latent membrane proteins (LMP-1 and -2), as well as two small untranslated RNAs (EBER1 and EBER2). Infection in EBV-associated tumors is often latent, with only a limited range of viral genes being expressed by cells (3). EBNA-1 is essential for the persistence and replication of the EBV genome in latently infected cells. It is the only viral protein required for the maintenance of latent EBV infection (4) and is expressed in all EBV-associated malignant tissues (5). EBNA-1 may, therefore, play a critical role in the onset, progression, or maintenance of these tumors (6). This hypothesis was supported further by the fact that two independent lines of EBNA-1 transgenic mice succumbed to monoclonal B-cell lymphoma (7).

The predicted aa sequence of EBNA-1, which consists of 641 aa, can be separated into unique N- and COOH-terminal domains joined by internal, glycine/alanine-rich short repeat sequences (8). Several studies have been carried out to map functional domains within EBNA-1, including dimerization, transactivation, nuclear localization, DNA looping, and RNA binding (9, 10, 11, 12). Most of the domains associated with these activities were localized to the COOH-terminus of EBNA-1, except for RNA binding, which involved the RGG motifs at aa 34–55, 330–349, and 355–376 (12).

Recently, sequence variations within the functionally important unique domains of EBNA-1 were reported by various groups (13, 14, 15, 16, 17, 18, 19, 20, 21, 22). It was proposed that this polymorphism could be critical for the apparent evasion of host cytotoxic T-lymphocyte responses by EBNA-1, and for disease progression, or it might be important for EBNA-1 transactivation within different cell types. However, no obvious functional differences between these variants have been documented.

In this study, we examined sequence variations of EBNA-1 gene from the freshly frozen tumor tissue of NPC and their corresponding peripheral-blood cells by PCR and direct sequencing. We observed that all tumor cells of NPC patients from Taiwan and their corresponding peripheral-blood cells harbored the same subtype, V-val.

A small piece of primary tumor from 13 patients with previously untreated, biopsy-confirmed NPC was collected under nasopharyngoscopy after obtaining informed consent. The specimens were freshly frozen in liquid nitrogen until DNA extraction. Venous blood (10 ml) was collected before treatment from each patient. After erythrocytes were lysed, the pellet of peripheral-blood cells was washed and subjected to DNA extraction with TRIzole (Life Technologies, Bethesda, MD) according to the manufacturer’s instructions.

EBNA-1 gene was amplified as four subfragments (Fig. 1): a 407-bp fragment including the N-terminal coding region upstream of the Gly-Ala repeat region; a 226-bp fragment including nucleotides 57 to 82 upstream of the Gly-Ala repeat region and nucleotides 235 downstream of the Gly-Ala repeat region, excluding the whole Gly-Ala repeat; and two overlapping fragments (349 and 719 bp) that cover the entire carboxyl terminus downstream of the Gly-Ala repeat region.

The primer sequences and PCR conditions are summarized in Table 1. In brief, PCR for the N-terminal coding region and two overlapping COOH-terminal regions of EBNA-1 gene were carried out using 100 ng of extracted DNA in a 50-μl mixture containing 10 mm Tris-HCl (pH 9.0), 50 mm KCl, 0.04% (w/v) gelatin, approximately 1.5–6.5 mm MgCl_2, 0.1%Triton X-100, 50 μm deoxynucleotide triphosphate, 2 units of Taq polymerase (Protech), and 10 pmol of each primer. The DNA solution was preheated to 94°C for 5 min with a final extension for 5 min at 72°C.

However, PCR for the Gly-Ala repeat region of EBNA-1 gene was amplified under the following conditions: 43 cycles of 94°C for 30 s then 60°C for 60 s. The DNA solution was preheated to 95°C for 9 min with a final extension for 10 min at 60°C. The reaction was carried out using 100 ng of extracted DNA in a 50 μl mixture containing 10 mm Tris-HCl (pH 8.3), 50 mm KCl, 0.001% (w/v) gelatin, 3 mm MgCl₂, 200 μm deoxynucleotide triphosphate, 1.25 units of Taq polymerase (AmpliTaq Gold), and 10 pmol of each primer.

The PCR products were separated on a 1.5% agarose gel and purified with Amicon Microcon-PCR centrifugal filter devices (Millipore Corporation). The sequencing reactions were performed with a dye-labeled terminators kit (Perkin-Elmer). Sequence analysis was performed in an ABI PRISM Genetic Analyzer, Model 377 (Perkin-Elmer, Foster City, CA).

Because the relative proportion of EBV DNA extracted from the peripheral-blood cells was extremely low, nested PCR was used to amplify a 308-bp fragment of COOH-terminal part of the EBNA-1 gene. The primer sequences and PCR conditions were the same as described previously (23) and are summarized in Table 1. Positive samples were purified and subjected to DNA sequencing.

The negative control (consisting of reaction mixture with no added DNA) and positive control (DNA extracted from an EBV-positive cell line, B95-8) were processed in parallel with patient samples in every PCR run. In addition, several measurements were taken to prevent contamination during each experiment, such as frequently changing gloves and cleaning the equipment, using aerosol-resistant pipette tips for PCR, and performing different procedures in separate areas.

All tumor tissues from 13 NPC patients (one keratinizing squamous cell carcinoma, nine nonkeratinizing squamous cell carcinoma, and three undifferentiated carcinoma) harbored EBV DNA. All tumor samples had the same clustered point mutations. They differed in the sequence pattern when compared with the B95-8 strain. The sequence variation included 21 nucleotides exchange, leading to 14 aa mutations and 6 silent exchanges (Table 2). Two of 13 patients exhibited an additional nucleotide exchange at position 109703 (C→T), resulting in an aa exchange at codon 585 (Thr→Ile).

The positive rate of EBNA-1 DNA from peripheral-blood cells of NPC patients by our nested PCR assay was 69.2% (9 of 13). Sequencing data of this COOH-terminal part (codons 440–541) of the EBNA-1 gene revealed that all positive blood samples have the same point mutations as their primary tumor. These included one silent nucleotide exchange at position 109509 (A→C) and six clustered point mutations at position 109409 (C→T), 109446 (C→G), 109454 (C→A), 109520 (C→T), 109531 (A→G), and 109546 (C→A). The resulting aa exchanges were at codons 487 (Ala→Val), 499 (Asp→Glu), 502 (Thr→Asn), 524 (Thr→Ile), 528 (Ile→Val), and 533 (Leu→Ile). One of nine EBNA-1-positive peripheral-blood samples exhibited an additional nucleotide exchange at position 109472 (T→C), resulting in an aa exchange at codon 508 (Phe→Ser).

All patients received combined radiotherapy and chemotherapy. After a median follow-up period of 29 months (range, 14–40 months), there were four failures: two distant metastases alone, one primary recurrence, and one distant metastasis plus neck recurrence. Four of nine patients with EBV-carrying cells and none of the four patients without EBV-carrying cells in the peripheral blood developed tumor relapse (Table 3).

EBNA-1 is the only EBV antigen consistently expressed in all EBV-infected cells studied to date and is essential for the persistence and replication of the virus in latently infected cells (4). Recent reports have described several variants of EBNA-1 relative to the prototype virus B95-8 (13, 14, 15, 16, 17, 18, 19, 20, 21, 22). Each of these EBNA-1 variants has been identified as the exclusive virus subtypes in tumor biopsies and/or tumor or lymphoblastoid cell lines known to carry monoclonal EBV, strongly suggesting that these variants represent distinct virus subtypes. Five subtypes, based on the EBNA-1 sequence in the COOH-terminal region were identified. There were two prototypes (P-ala and P-thr) and three variant subtypes (V-pro, V-leu, and V-val) according to the aa at position 487 in EBNA-1, because this aa residue is highly predictive of the pattern of mutations observed.

NPC has been proven as an EBV-associated cancer for a long time. It has been demonstrated that EBV is harbored in almost every NPC tumor irrespective of the degree of differentiation and geographic distribution. Thus far, EBNA-1 sequence data in NPC biopsy is seldom reported (13, 14, 15). Snudden et al. (13) determined the sequence of the EBNA-1 gene except for the internal glycine/alanine repeat region in seven NPC patients from Hong Kong Chinese and identified several mutations (V-val) relative to B95–8 (P-ala). Many of the nucleotide changes would lead to substantial aa alterations in apparently functionally significant regions of the protein. In our study, identical sequence variations were demonstrated in 13 biopsy samples from NPC patients of Taiwanese. We further found a new aa mutation at codon 364 (Gly→Arg) attributable to nucleotide exchange at positions 109039–109041 (GGA→AGG).

Gutierrez et al. (14) reported that EBV with an EBNA-1 subtype (V-val) frequently associated with NPC in 9 of 13 tumor tissue samples from patients from Hong Kong and France. The remaining four samples belonged to a P-thr subtype. However, the authors did not indicate whether or not a V-val subtype had a predilection for Hong Kong Chinese. Chen et al. (16) from Taiwan determined the COOH-terminal part of EBNA-1 sequence (codons 363–641) from several EBV-containing cell lines and two NPC biopsies. These two NPC samples also expressed a V-val subtype. Sandvej et al. (15) sequenced the COOH-terminal part of the EBNA-1 gene (codons 438–547) in Danish and Chinese NPC patients. They found that two of three Chinese NPC samples were of a V-val subtype but none of the Danish NPC samples (0 of 11) harbored a V-val subtype. Thus, we deduce that EBV containing a V-val subtype of EBNA-1 is frequently associated with Chinese NPC.

Sandvej et al. (15) also showed that two of three Hodgkin’s disease, both the nasal T-cell lymphoma, and four of five throat wash samples from Chinese patients contained a V-val EBNA-1 subtype, which was never seen in Danish patients. In Danish patients, P-thr was the most common EBNA-1 subtype irrespective of source (Hodgkin’s disease, NPC, and throat wash samples). Chen et al. (19) from the United States recently reported V-val EBNA-1 subtypes in tested cases of EBV-positive gastric carcinoma (25 of 25) and EBV-positive reactive lymphoid follicular hyperplasia (8 of 8) in Japanese patients. Only 1 in 28 cases of known EBV-positive American gastric carcinoma or reactive lymphoid tissue was of a V-val subtype. MacKenzie et al. (20) investigated tumor samples of EBV-associated lymphoma from Brazil and the United Kingdom, and no V-val subtype could be found. Chang et al. also reported that no V-val subtype occurred in Brazilian and American patients with Hodgkin’s disease and reactive lymphoid tissue. These findings may suggest that V-val is the dominant subtype of EBV latent infection in the Asian region.

NPC is distinguished from other cancers of the head and neck by its epidemiology, histopathology, clinical characteristics, and therapy. Radiotherapy is the primary treatment for NPC. Recent trials support the use of combined chemoradiotherapy for patients with advanced NPC. Because of recent advances in radiation oncology, the failure pattern has predominantly changed to distant metastasis (23, 24, 25). Distant metastasis is the most important and urgent issue to be solved in the management of NPC. Using real-time quantitative PCR, Lo et al. (26, 27) have suggested that EBV DNA load in cell-free plasma can be used as a prognostic indicator for tumor relapse. Instead of using plasma samples, we demonstrated that the presence of EBV DNA in the peripheral-blood cells was a poor prognostic factor for NPC patients with a significantly higher risk of developing distant metastasis, in addition to a lower survival rate (23). EBV has been proven in metastatic cancer cells of NPC patients with neck nodes or distant organ metastases by in situ hybridization (28, 29, 30). Thus, the circulating EBV-carrying cells we detected may arise from disseminated cancer cells. Sequencing data in the current study showing the match mutation pattern in codons 487, 499, 502, 524, 528, 533 from both primary tumor and peripheral-blood cells give further support to this hypothesis. These finding and clinical observations reflect a higher relapse rate for patients with the presence of EBV DNA in the peripheral blood and should be considered in treatment planning of future patients.