Common cytological and cytogenetic features of Epstein-Barr virus (EBV)-positive natural killer (NK) cells and cell lines derived from patients with nasal T/NK-cell lymphomas, chronic active EBV infection and hydroa vacciniforme-like eruptions

Authors


Dr Norio Shimizu, Department of Virology, Division of Virology and Immunology, Medical Research Institute, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113–8510, Japan. E-mail: nshivir@mri.tmd.ac.jp

Abstract

Summary. In this study, we describe the cytological and cytogenetic features of six Epstein-Barr virus (EBV)-infected natural killer (NK) cell clones. Three cell clones, SNK-1, -3 and -6, were derived from patients with nasal T/NK-cell lymphomas; two cell clones, SNK-5 and -10, were isolated from patients with chronic active EBV infection (CAEBV); and the other cell clone, SNK-11, was from a patient with hydroa vacciniforme (HV)-like eruptions. An analysis of the number of EBV-terminal repeats showed that the SNK cell clones had monoclonal EBV genomes identical to the original EBV-infected cells of the respective patients, and SNK cells had the type II latency of EBV infection, suggesting that not only the cell clones isolated from nasal T/NK-cell lymphomas but also those isolated from CAEBV and HV-like eruptions had been transformed by EBV to a certain degree. Cytogenetic analysis detected deletions in chromosome 6q in five out of the six SNK cell clones, while 6q was not deleted in four control cell lines of T-cell lineage. This suggested that a 6q deletion is a characteristic feature of EBV-positive NK cells, which proliferated in the diseased individuals. The results showed that EBV-positive NK cells in malignant and non-malignant lymphoproliferative diseases shared common cytological and cytogenetic features.

The Epstein-Barr virus (EBV) was originally described in cultured lymphoblasts from African Burkitt's lymphoma in 1964 (Epstein et al, 1964), and subsequently identified as the causative agent for infectious mononucleosis (Henle et al, 1968), which is usually a benign and self-limiting lymphoproliferative disease of B lymphocytes. The virus has subsequently been identified in various neoplasms, including nasopharyngeal carcinoma (zur Hausen et al, 1970), acquired immunodeficiency-related lymphoma (Doll & List, 1982; Ziegler et al, 1982) and post-transplantation B-cell lymphoma (Klein & Purtilo, 1981; Hanto et al, 1983).

It was not recognized until late the 1980s, however, that EBV infects T/natural killer (NK) cells to cause life-threatening lymphoproliferative diseases. In 1988, EBV was identified in T lymphocytes in a boy with chronic active EBV infection (CAEBV) (Kikuta et al, 1988) and in T-cell lymphomas arising in patients with CAEBV (Jones et al, 1988). In 1990, two groups identified EBV in nasal T-cell lymphomas (Harabuchi et al, 1990; Ho et al, 1990). Subsequent studies revealed that many cases with CAEBV and hydroa vacciniforme (HV)-like eruptions are lymphoproliferative diseases of NK or T cells (Jones et al, 1988; Kikuta et al, 1988; Ishihara et al, 1989, 1997; Kawa-Ha et al, 1989; Fukunaga et al, 1994; Tanaka et al, 1994; Kanegane et al, 1996, 1998; Imai et al, 1996; Ohshima et al, 1998; Ohga et al, 1999; Tsuge et al, 1999; Okamura et al, 2000; Quintanilla-Martinez et al, 2000; Kawa et al, 2001; Kimura et al, 2001; Nagata et al, 2001a). Moreover, it has been shown that cases of nasal T/NK-cell lymphomas consist of more NK-cell than T-cell lymphoma cases, and tumours of both lineage were consistently associated with EBV (Suzumiya et al, 1994; Emile et al, 1996; Jaffe et al, 1996, 1999; Petrella et al, 1996; Chiang et al, 1996a, 1997; Chan, 1999). In addition, other malignant diseases of T/NK-cell lineage such as nasal type (extra nasal) T/NK-cell lymphoma and NK-cell leukaemia proved to be associated with EBV (Chan, 1999; Jaffe et al, 1999). It is, thus, now generally accepted that EBV plays a role in the development of various types of lymphoproliferative diseases of T/NK-cell lineage, although the precise role of the virus in such diseases is poorly understood.

Among the rare lymphoproliferative diseases of NK-cell lineage, nasal T/NK-cell lymphoma and CAEBV are relatively frequent, especially in Asia. These are regarded as representative lymphoproliferative diseases that are positive for EBV. Indeed, not only every case of CAEBV, but also almost all cases of nasal T/NK-cell lymphomas, are associated with EBV (Harabuchi et al, 1990; Ho et al, 1990; Weiss et al, 1992; Arber et al, 1993; Borisch et al, 1993; Kanavaros et al, 1993; Chan et al, 1994; van Gorp et al, 1994). In addition, many CAEBV cases, as well as cases of nasal T/NK-cell lymphomas, show clonal expansion of EBV-positive NK cells (Ishihara et al, 1989, 1997; Kawa-Ha et al, 1989; Ohshima et al, 1998; Ohga et al, 1999; Kimura et al, 2001; Nagata et al, 2001a). However, in spite of methodological development for the investigation of lymphoproliferative diseases, studies of these NK-cell diseases have been limited because they are relatively rare. Moreover, nasal T/NK-cell lymphoma is associated with a tendency for necrosis to occur, making it difficult to analyse, and studies on CAEBV have been hampered by the scarcity of EBV-positive cells in the peripheral blood and the lack of a method to enrich these cells. In order to study these diseases in detail, EBV-positive NK-cell lines have recently been isolated from patients with nasal T/NK-cell lymphomas (Tsuchiyama et al, 1998; Nagata et al, 2001a,b) and CAEBV (Tsuge et al, 1999). In previous studies, we observed that high-dose recombinant interleukin 2 (IL-2) was useful for the long-term culture of cells from the primary lesions and peripheral blood of patients with nasal T/NK-cell lymphomas and CAEBV (Nagata et al, 2001a,b).

Thus, we have been trying to establish cell lines from more patients with nasal T/NK-cell lymphomas and patients with CAEBV. We have succeeded in obtaining five EBV-positive NK-cell clones to date. In addition, we have cultured an EBV-positive NK-cell clone from the skin lesion of a patient with hydroa vacciniforme (HV)-like eruptions (Tabata et al, 1995); this has supported the recent finding that some skin lesions such as atypical HV are associated with EBV (Iwatsuki et al, 1999).

With this in mind, in the present study we endeavoured to characterize the EBV-positive NK cells which accumulate in lymphoproliferative diseases by obtaining cell clones. We sought to analyse their phenotype, genotype, type of latency of EBV and chromosomal abnormalities. We found common cytological and cytogenetic features in these NK cell clones. The mechanism for the development of nasal T/NK-cell lymphoma, CAEBV and HV-like eruptions is discussed.

Patients and methods

Patients, NK-cell clones.  Six cell clones, designated SNK-1, -3, -5, -6, -10 and -11, were isolated and used in this study. The diagnosis, clinical remarks, sources of the cell clones and the titres of antibodies to EBV in patients are summarized in Table I. SNK-1, -6 and -10 were established cell lines, while SNK-3, -5 and -11 were maintained in culture media for over 3–5 months, but could not be established as a cell line. SNK-1, -3 and -6 were isolated from patients with nasal NK-cell lymphomas; the patient for SNK-1 had suffered from CAEBV and subsequently developed nasal NK-cell lymphoma. SNK-5 and -10 were isolated from the peripheral blood of patients with CAEBV. SNK-11 was derived from the skin lesion of a patient with HV-like eruptions, which was histologically diagnosed as lymphomatoid papulosis (LP) (Table I). Some characteristics of SNK-1 and SNK-6 and case reports have been described previously (Nagata et al, 2001a,b). The clinical features of the patient with HV-like eruptions have also been previously described (Tabata et al, 1995).

Table I.  Origin of the cell clones.
Cell cloneCulture period (months)Patient Age/SexDiagnosisCell sourceSerum anti-EBV titresPBMC phenotype (%)Prognosis/clinical remarks
VCA –IgGEA-DR IgGEBNA IgGCD3CD56
  1. Case reports for SNK-1, -6 and -11 were reported previously (Tabata et al, 1995; Nagata et al, 2001a,b). CAEBV, chronic active Epstein-Barr virus infection; HV-like LP, hydroa vacciniforme-like lymphomatoid papulosis; VCA, viral capsid antigen; EA-DR, early antigen-diffuse and restricted; EBNA, Epstein-Barr nuclear antigen; LPD, lymphoproliferative disease. ND, not done; w/o, without.

SNK-1≥3624/FNasal T/NK cell lymphoma with CAEBVPeripheral blood1280 32040132Died of disease
SNK-3344/MNasal T/NK cell lymphomaPrimary lesion80< 1040726Died of disease
SNK-5514/FCAEBVPeripheral blood25602560203645Died of disease
SNK-6≥3662/MNasal T/NK cell lymphomaPrimary lesion320  10205918Alive w/o disease
SNK-10≥3617/MCAEBVPeripheral blood80  80402242Died of disease
SNK-11416/FHV-like LPPrimary lesion640 160403143NK-LPD (peripheral blood)
SNT-8≥3648/FNasal T/NK cell lymphomaPrimary lesion640  1020474Alive w/o disease
SNT-13≥3613/FCAEBVPeripheral blood1280 320∼64020∼4093NDDied of disease
SNT-15≥3615/FCAEBVPeripheral blood1280 1601072NDDied of disease
SNT-16≥3613/FCAEBVPeripheral blood640 1604093NDDied of disease

Cell clones.  The methods used for establishing SNK-1 and SNK-6 have been described in our previous reports (Nagata et al, 2001a,b), and essentially the same methods were used for establishing other cell clones. Briefly, for SNK-3 and SNK-11, tissues excised from a nasal T/NK-cell lymphoma and an HV-like eruption were minced and strained through a 70-µm cell strainer (Becton Dickinson, San Jose, CA, USA). The cells obtained through the strainer were then processed in the same manner as for SNK-6 (Nagata et al, 2001b); they were then suspended in Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with antibiotics, 10% heat-inactivated human serum and 700 U/ml of recombinant IL-2. In terms of SNK-5 and SNK-10, peripheral blood mononuclear cells were isolated from 10 ml of the patients' blood by the Ficoll-Hypaque method, and suspended in 10 ml RPMI-1640 medium supplemented with 10% heat-inactivated human serum and 700 U/ml of IL-2. T lymphocytes were removed by using anti-CD3 monoclonal antibody-conjugated magnetic beads (Dynal, Oslo, Norway). The T lymphocyte-depleted cell fraction was cultured by the same method as that used for SNK-1 (Nagata et al, 2001a). SNK-1, -6 and -10 were maintained in the presence of IL-2 for over 3 years; SNK-3, -5 and -11 were maintained for 3–5 months, but could not be established as a cell line.

For controls, we used the EBV-positive γδ T-cell line SNT-8, which was established from a nasal T-cell lymphoma (Nagata et al, 2001b), and the novel γδ T-cell clones SNT-13 and -15, which were recently established from the peripheral blood of patients with CAEBV. In addition, we used the novel αβ T-cell clone SNT-16, which was established from the peripheral blood of a patient with CAEBV, as a further control (Table I). The methods used to establish EBV-infected T-cell lines were the same as those used for NK-cell clones.

Cytological analysis for cell clones. Morphological evaluation: cell smears of SNK-cell clones were prepared and stained with Wright's Giemsa for observation under a light microscope.

Flow cytometric analysis: the cell clones were analysed by two-colour immunofluorescence with a flow cytometer (EPICS XL, Beckman Coulter, Hialeah, FL, USA) for the expression of surface markers. The following antibodies, conjugated with fluorescein isothiocyanate (FITC) or phycoerythrin (PE) (Becton Dickinson), were used: anti-CD3, CD4, CD8, CD16, CD19, CD25, CD56, HLA-DR, TCR α/β and TCR γ/δ.

Assessment of the number of EBV-terminal repeats.  To determine the clonality of the EBV-positive cell clones, original lesions and patient's peripheral blood mononuclear cells (PBMC), the number of EBV-terminal repeats (EBV-TR) was assessed by Southern blot analysis. DNA extracted from the cell clones was digested with BamHI, separated on a 0·7% agarose gel, and transferred to a nylon membrane. A 1·9-kb XhoI subfragment of BamHI-Dhet derived from EBV termini was used as a probe for EBV-TR DNA (Raab-Traub & Flynn, 1986). Hybridizations were visualized using a Fluorescein Gene Images System (Amersham, Buckinghamshire, UK) according to the manufacturer's instructions. Raji and B95-8 cells were used as controls for monoclonal cell expansion with a single EBV-TR and polyclonal cell proliferation for EBV replication.

Southern blot analysis of T cell receptor (TCR) and immunoglobulin genes.  The cell clones were analysed for rearrangements of the TCR β-, γ- and δ-chain, and the immunoglobulin heavy-chain genes by Southern blotting. DNA (3 µg) extracted from each of the cell clones was digested with a restriction endonuclease, BamHI, EcoRI, HindIII or KpnI (TAKARA, Kyoto, Japan), electrophoresed through a 0·7% agarose gel and transferred onto a nylon membrane. Human placenta DNA was used as a germ-line control. The membrane was exposed to a fluorescein-labelled Cβ1, Jγ1, Jδ1 or Jδ3 probe for the TCR genes, or a JH probe for the immunoglobulin heavy-chain gene (Korsmeyer et al, 1983; Sims et al, 1984; Quertermous et al, 1986; Tkachuk et al, 1988; Takihara et al, 1988). The Jδ3 probe was a 1·5-kb XbaI genomic fragment of the Jδ3 region. Hybridizations were visualized by the same methods used for Southern blot analysis of the EBV-TR.

Western blot analysis for expression of EBV genes.  Cells (5 × 106) were washed twice in phosphate-buffered saline (PBS), and cell pellets were lysed in electrophoresis sample buffer and boiled for 10 min. The equivalent of 2·5 × 105 cells was loaded onto each lane of an 8% Tris-glycine gel. The separated proteins were blotted onto a nitrocellulose filter and blocked in 5% skimmed milk in PBS overnight, and then incubated with a human serum containing high-titre antibodies against Epstein-Barr virus nuclear antigen (EBNA)-1, with a mouse monoclonal antibody against latent membrane protein-1 (LMP-1) (S12, IgG2a, Boxtel, the Netherlands), or with a monoclonal antibody to EBNA-2 (PE-2, mouse IgG1, DAKO, Denmark) in 5% skimmed milk in PBS for 90 min at room temperature. After rinsing twice in 5% skimmed milk for 15 min, the blots were incubated for 30 min with horseradish peroxidase (HRP)-linked protein-A (Amersham Life Science) for the human serum or HRP-linked anti-mouse IgG (Amersham Life Science) for the monoclonal antibodies. The filter was rinsed twice for 10 min in 5% skimmed milk and then rinsed in PBS. Signals were visualized using an enhanced chemiluminescence Western blotting detection system (Amersham Life Science).

Cytogenetic studies.  Cell cultures were treated with Colcemid (0·05 µg/ml) for 2 h, followed by treatment with 0·075 mol/l KCl for 20 min. The cells were then fixed in an acetic acid and methanol solution (1 : 3), and chromosome slides were prepared by the standard method. Karyotype analyses were made according to ISCN, 1995) after either Q-staining or G-staining.

Results

Cytological analysis of cell clones

Morphologically, the cells of six SNK-cell clones appeared as large mononuclear cells. The cells had pleomorphic nuclei, which were round, indented or lobulated, with dense chromatin, and some appeared atypical. They had a low nuclear cytoplasmic ratio and intense basophilic cytoplasma that contained azurophilic granules (data not shown).

Phenotype of cell clones: flow cytometry showed that all of the six SNK-cell clones shared a common phenotype of NK-cell lineage: CD3 CD4 CD8 CD16+ CD19 CD56+ HLA-DR+TCR α/β TCR γ/δ. On the other hand, the γδ T-cell lines SNT-8, -13 and -15 were CD3+ CD4 CD8 CD16 CD19 CD56+ HLA-DR+TCR α/β TCR γ/δ+. The αβ T-cell line SNT-16 was CD3+ CD4+ CD8 CD16 CD19 CD56 HLA-DR+TCR α/β+ TCR γ/δ(Table II).

Table II.  Phenotype of cell clones.
Cell clonesCD3CD4CD8CD19CD16CD56CD25HLA-DRTCRα/βTCRγ/δ
  1. –, negative; +, < 10%; + +, 10–75%; + + +, > 75%.

SNK-1++ + ++ + ++ + +
SNK-3+ ++ + ++ ++ + +
SNK-6++ ++ + ++ + +
SNK-5+ ++ + ++ ++ + +
SNK-10++ + ++ + ++ + +
SNK-11+ ++ ++ ++ + +
SNT-8+ + ++ + ++ + ++ + ++ + +
SNT-13+ + ++ + ++ ++ + ++ + +
SNT-15+ + ++ + ++ + ++ + ++ + +
SNT-16+ ++ + ++ + ++ + ++ +

Clonality of SNK-cell clones

Southern blot analysis for the number of EBV-TR showed a monoclonal band in each of the six SNK and four SNT-cell clones (Fig 1). We have already reported that the clonality of SNK-1 and SNK-6 was identical to that of their original tumours (Nagata et al, 2001a, b). In addition, the clonality of SNK-3 and SNK-11 was also consistent with that of their original lesions. Likewise, the clonality of SNK-5 and SNK-10 was the same as that of the EBV-positive cells in the patient's blood (data not shown).

Figure 1.

Southern blot analysis for the number of EBV-TR as evidence for monoclonal expansion of EBV-positive cells in SNK-1 (lane 1), SNK-3 (lane 2), SNK-6 (lane 3), SNT-8 (lane 4), SNK-5 (lane 5), SNK-10 (lane 6), SNK-11 (lane 7), SNT-13 (lane 8), SNT15 (lane 9) and SNT-16 cells (lane10). Controls are B95-8 cells (B) and Raji cells (R) for cells with polyclonal and monoclonal cellular expansion respectively.

Genotype of cell clones

None of the SNK-cell clones showed rearrangements of the immunoglobulin heavy-chain or TCR genes, while the γδ T-cell line SNT-15 showed rearrangements of the Cβ1, Jγ1, Jδ1 and Jδ3 genes (Fig 2). These results supported the conclusion that all of the six SNK-cell clones were of NK-cell lineage.

Figure 2.

Southern blot analysis for TCR genes. Southern blot analysis shows no rearrangement of the TCR genes in SNK-3 and rearrangements of the β-, γ- and δ-chain genes in SNT-15. Panels show hybridization with the following probes: (A) the Cβ1 probe, (B) the Jγ1 probe, (C) the Jδ1 probe, (D) the Jδ3 probe. Human placenta DNA was used as a germ-line control. The letters above the blots indicate BamHI (B), EcoRI (E), HindIII (H) and KpnI (K) digests.

Expression of EBV genes

Western blot analysis for the expression of EBNA-1, EBNA-2 and LMP-1 showed that all the SNK-cell clones were positive for the serum against EBNA-1 (data not shown), but negative for the anti-EBNA-2 antibody (Fig 3). With regard to LMP-1, the SNK-1, -3, -5, -6 and -10 cells were positive for the monoclonal antibody, but the SNK-11 cells did not express a detectable amount of LMP-1 protein (Fig 3). The SNK-11 cells, however, expressed mRNA for LMP-1, which was detected by reverse transcription polymerase chain reaction (RT-PCR, data not shown).

Figure 3.

Western blot analysis for the expression of EBNA-2 and LMP-1 in SNK-1 (lane 3), SNK-3 (lane 4), SNK-6 (lane 5), SNK-5 (lane 6), SNK-10 (lane 7) and SNK-11 cells (lane 8). The SNK-1, -3, -5, -6 and -10 cells clearly express LMP-1, but not EBNA-2. EBNA-2 and LMP-1 are not detected in the SNK-11 cells. WaY cells were used as the negative control (lane 1) and a lymphoblastoid cell line (LCL) was used as the positive control (lane 2).

Cytogenetic findings

The results of chromosomal analysis in the six NK-cell clones are summarized in Table III. All the cell cultures revealed abnormal karyotypes including clonal chromosomal abnormalities, and chromosome 6 was found to be commonly involved. In particular, five of the SNK-cell clones (SNK-1, -3, -6, -10 and -11) consistently showed a partial loss of the long arm of chromosome 6, either by simple deletion (SNK-1 and -11) or by unbalanced translocation between 6q and an unidentified chromosome (SNK-3, -6 and -10). These 6q abnormalities were found in almost all the cells analysed in each cell clone. Of special significance was the finding that a simple partial deletion of 6q was the sole chromosomal abnormality in SNK-11 (Table III, Fig 4). Deletions in chromosomes were confirmed by fluorescence in situ hybridization studies using the painting probe for 6q and probes for the telomeric region specific for the long arm of chromosome6 (data not shown, unpublished observations).

Table III.  Cytogenetic analysis of cell clones.
Cell clonesChromosomal findings
SNK-145,X,add(3)(p25)[8],del(6)(q11 or q12)[cp22]
SNK-346,XY,del(3)(q23q27)[17],add(5)(q35)[16], add(6)(q21)[16],add(12)(p13)[16],del(15) (q22q26)[16][cp17]
SNK-546,XX,t(6;12)(p12;q24)[14],del(9)(q22q34) [2][cp14]
SNK-686–89,XXY,i(1)(q10)[10],der(2)t(2;4)(p21;q12) [10],der(4)t(2;4)(p21;q12) × 2[10],add(6) (q1?3) × 2[10],−13,−14, add(14)(p13)[10],−15, −16,−17,i(17)(q10)[10],−18, + 20,−21, +mar1, +mar2, +mar3[cp10]
SNK-1047,XY,der(6)t(3;6)(p23;p25),add(6)(q15),  + 20[10]
SNK-1146,XX,del(6)(q21q25)[20]
SNT-846–47,del(X)(q11q13),der(X)add(X)(p?21)dic (X;9)(q?26;p?13), + 1,del(1)(p13p22),der(1;9) (q10;p10), add(4)(p16), + 8,ins(8;?)(p?21;?),  +?add(9)(p22)[cp20]
SNT-1344–46,X, + X,−Y,del(15)(q22q24),der(15)t(3;15) (q?22;q?26)[cp14]
SNT-1547–48,XX, + 1,add(1)(p13),i(5)(q10),add(10) (q26),add (20)(p13)[cp7]
SNT-1646,XX,der(11)t(1;11)(q25;q13),add(22) (p11.2)[17]
Figure 4.

Chromosome 6 in the EBV-positive NK cells and T cells. Abnormalities in the cells are as follows: SNK-1, del(6)(q11 or q12); SNK-3, add(6)(q21); SNK-5, t(6;12)(p12;q24); SNK-6, add(6)(q1?3) × 2; SNK-10, der(6)t(3;6)(p23;p25),add(6)(q15); SNK-11, del(6)(q21q25). No abnormality of chromosome 6 was present in the T-cell lines. Karyotype analysis of SNK-1, -3, -5, -6 and -10 was carried out by the Q-banding method. Karyotype analysis of SNK-8, -11, -13, -15 and -16 was carried out using the G-banding method.

On the other hand, the four T-cell clones (SNT-8, -13, -15 and -16) which were used as controls showed complex clonal chromosomal abnormalities, but none exhibited a 6q abnormality.

Discussion

The present report described the cytological and cytogenetic features of six cell clones of NK-cell lineage that were positive for EBV. Three of these were obtained from patients with nasal T/NK-cell lymphomas, two were established from the peripheral blood of patients with CAEBV, and the other cell clone was derived from the skin lesion of a patient with HV-like LP. We described the NK cells as clones because the cells were isolated from patients with T/NK lymphoma, chronic EBV infection and hydroa vacciniforme-like eruptions, and all these showed clonal expansion of EBV-infected cells; the isolated cells and the cells of primary lesions showed monoclonal bands of EBV-TR respectively. All these cell clones possessed a phenotype and genotype typical of NK cells. Analysis of the number of EBV-TR showed that the cell clones consisted of the respective monoclonal cells (Fig 1). Moreover, the clonality of SNK-1, -3, -6 and -11 was identical to that of the original tumours or skin lesion. In two CAEBV patients, CD56+ (putative NK) cells were increased in patient's peripheral blood. We speculated that these increased cells were EBV positive. Similarly, the clonality of the SNK-5 and -10 cells was consistent with that of EBV-positive cells in the patient's peripheral blood. Thus, we concluded that the SNK-cells clones were derived from the NK cells, which were responsible for the development of the original lymphoproliferative diseases in our patients.

To the best of our knowledge, SNK-11 is the first EBV-positive NK-cell clone to be cultured long-term from a skin lesion. The diagnosis of the skin lesion was HV-like LP, which clinically resembled HV, but was histologically diagnosed as LP (Tabata et al, 1995). Long-term culture of SNK-11 provided evidence that proliferative NK cells, which were affected by EBV, might cause some cases of HV-like eruptions (Tokura et al, 1998). In addition, the present results appear to be consistent with previous reports of atypical HV, which frequently progressed to lymphoid malignancies; latent EBV infection of lymphoid cells was probably involved in the disease processes (Oono et al, 1986; Asada et al, 1994; Ruiz-Maldonado et al, 1995; Cho et al, 1996; Magana et al, 1998; Iwatsuki et al, 1999). Moreover, some of these atypical HV cases were diagnosed as LP (Iwatsuki et al, 1999). We therefore conclude that some cases of HV-like skin lesions should be characterized as lymphoproliferative diseases of NK cells.

NK cell clones have, to date, been established from patients with NK-cell leukaemia (Yodoi et al, 1985; Fernandez et al, 1986; Gong et al, 1994; Yoneda et al, 1992; Robertson et al, 1996), leukaemic-state nasal angiocentric NK-cell lymphoma (Tsuchiyama et al, 1998) or CAEBV (Tsuge et al, 1999). Each of these studies showed that IL-2 or T-cell growth factor putative for IL-2 was necessary for the culture. Moreover, we recently reported that the technique using high-dose IL-2 was useful for the isolation and maintenance of EBV-positive NK-cell lines from patients with nasal T/NK-cell lymphomas (Nagata et al, 2001a,b). Furthermore, the present study showed that high-dose IL-2 is also useful for the establishment of NK-cell lines from patients with CAEBV and HV-like eruptions that had not yet developed into malignant lymphoma. The usefulness of IL-2 in the culture of EBV-positive NK cells may derive from the elevated expression of the IL-2 receptor α chain (IL-2Rα), as it has been reported that expression of IL-2Rα by exogenous IL-2 was elevated in EBV-infected NK cells compared with EBV-negative NK cells (Tsuge et al, 1999). However, we found that high-dose IL-2 did not always establish NK-cell lines even from nasal T/NK-cell lymphomas. Therefore, the factors that determine the proliferative capacity of NK cells in vitro remain to be clarified, although IL-2 appears to be necessary for culturing NK cells derived from various lymphoproliferative diseases.

The monoclonal expansion of the SNK-cells indicated that the cells were transformed to a certain degree. In terms of the expression of the EBV genes, all the SNK-cells expressed EBNA-1. Furthermore, SNK-cells other than SNK-11 were clearly positive for LMP-1 protein (Fig 3), and SNK-11 expressed mRNA for LMP-1, as detected by RT-PCR. The expression of EBNA-1 and LMP-1 is characteristic for the type II latency of EBV infection (Rowe et al, 1987; Klein, 1994), and this type of latency has been reported in EBV-associated malignancies such as nasal T/NK-cell lymphoma (Chiang et al, 1996b; van Gorp et al, 1996), nasopharyngeal carcinoma (Brooks et al, 1992; Busson et al, 1992) and Hodgkin's disease (Pallesen et al, 1993). LMP-1 is known to possess transforming activity for both rodent and human cells (Wang et al, 1985; Fahraeus et al, 1990) and is thus believed to play a role in these malignancies. Our results suggest that EBV transformed the proliferative NK cells in our patients with nasal T/NK-cell lymphomas, CAEBV and HV-like eruptions. However, it is not known why some NK cells developed malignant lymphomas and the others formed non-malignant lymphoproliferative diseases.

In the present study, deletions in chromosome 6q were commonly found in SNK-cells (Fig 4), an observation apparently consistent with previous reports (Ohshima et al, 1997; Tien et al, 1997; Wong et al, 1997). According to a recent review (Wong et al, 1999), it was demonstrated that deletions of chromosome 6 around the q21-q23 region were the most common recurrent chromosomal abnormality in NK-cell lymphomas and leukaemia. Wong et al (1997) observed three cases of del(6)(q21q25) and one case of del(6)(q21q23) in a series of 11 cases of NK-cell lymphomas and leukaemia. Tien et al (1997) independently reported cytogenetic studies on four cases of nasal T/NK-cell lymphomas, and all were associated with a deletion of 6q in the q21-q23 region. In terms of CAEBV, Ohshima et al (1998) reported interesting observations on chromosomal abnormalities in the 6q region. They analysed four patients with CAEBV and detected oligoclonal del(6q) including the q21-q23 region in one case with oligoclonal EBV-TR bands and a clonal del(6)(q15q23) in another case with a monoclonal EBV-TR band. Thus, the present results together with these previous reports strongly suggest that deletions at chromosome 6q, especially in the q21-q23 region, are involved in lymphoproliferative diseases of NK-cell lineage, irrespective of whether the disease is malignant or non-malignant. Therefore, it appears to be necessary to analyse in future studies whether cases without major deletions of 6q have genetic abnormalities in the genes located in this region. Because there is a possibility that SNK-5 has a small deletion in the 6q region that could not be identified by karyotype analyses, SNK-5 may provide an important tool in the analysis.

Ohshima et al (1998) reported the subsequent clinical courses of those CAEBV cases with del(6q) described above; the patient with the oligoclonal del (6q) (case #12 in the report) died of a virus-associated haemophagocytic syndrome, and the case with the monoclonal del(6)(q15q23) (case #13 in the report) died of EBV-positive NK-cell leukaemia. Among our patients with del(6q), those from which SNK-10 and SNK-11 were established have not developed malignant lymphomas. It is thus suggested that chromosomal deletions in the 6q region do not always cause malignancies, while the deletions appear to be a risk factor for malignant lymphoma/leukaemia of NK-cell lineage. Indeed, the development of NK-cell lymphoma/leukaemia in patients with CAEBV has been reported (Ohshima et al, 1998; Nagata et al, 2001a). In order to clarify the mechanisms for the malignant transformation of NK cells in these non-malignant diseases, more cell clones derived from the blood of the individual with the original disease or primary lesion would be useful.

In summary, proliferating NK cells in nasal T/NK-cell lymphomas, CAEBV and HV-like eruptions shared common features, including the type II latency of EBV infection and chromosomal abnormalities in the 6q region. The present study thus raises an important question: why did NK cells with similar pathological changes develop into malignant lymphomas or into non-malignant lymphoproliferative diseases? We recently proposed that the type and/or degree of transformation of EBV-positive NK cells could be variable even in a single patient with CAEBV (Nagata et al, 2001a); the patient was the one from whom SNK-1 was established. This patient showed the proliferation of three NK-cell clones, and analyses suggested that one of the three clones was identical to the SNK-1 cells and formed the nasal NK-cell lymphoma. Similar conditions showing variability in the transformation of EBV-positive cells have been reported in the post-transplantation lymphoproliferative disorders (PT-LPD) of B lymphocytes (Klein & Purtilo, 1981; Hanto et al, 1983; Shearer et al, 1985; Cleary et al, 1988; Knowles et al, 1995); these reports hypothesized that the development of PT-LPD is a multistep process, initiated as a polyclonal expansion of EBV-carrying B cells on the basis of immunosuppression, and may progress to form lymphomas. The present study, together with these reports, suggests that transformation of NK cells by EBV infection in itself is not sufficient for the development of a malignant lymphoma, and an additional abnormality such as a genetic mutation is likely to be involved in the malignant transformation of NK cells. The necessity of clarifying such genetic abnormalities in the aetiology of nasal T/NK cell lymphomas warrants the utilization of the cultured cell clones for future studies, as well as further efforts to establish new cell lines of NK-cell lineage.

Acknowledgments

This study was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan (no.11670207 and 09253103), Grants-in-Aid for Education and Research Promotion Program 2001 from Tokyo Medical and Dental University, and Grants-in-Aid for Medical Research from the Atsuko Ohuchi Memorial Research Fund.

Ancillary