Establishment and characterization of three transplantable EBV-containing nasopharyngeal carcinomas

lnt. J. Cancer: 42, 599-606 (1988) Publication of the International Union Against Cancer Publication de I‘Union Internationale Conire le Cancer 0 1988 Alan R. Liss, Inc.

ESTABLISHMENT AND CHARACTERIZATION OF THREE TRANSPLANTABLE EBV-CONTAINING NASOPHARYNGEAL CARCINOMAS P. BuSSoN”3, G. CANEM’, P. FLOmS’, F. MUGNERET~, B. CLAUSSE’, B. CAILLOU~, K. BRAHAM’, H. WAKASUGI’, M. LIPINSKI’ and T. TURSZ’ ‘Luboratoire d’lmmunobiologie des Tumeurs, U. A. I156 C. N . R. S. , lnstitut Gustave Roussy, 94805 Villejuif Cedex; and ’Luboratoire de Cytogtnttique, Facultk de Mkdecine, Universit.4 de Dijon, 21033 Dijon Cedex, France.

Three transplantable nasopharyngeal carcinoma (NPC) tumors, designated CIS, C17 and C18, have been obtained and characterized. C15, derived from a primary NPC tumor, has been propagated in nude mice for 30 passages. C17 and C18, derived from metastatic NPC tissue, have been passaged 10 times. Desmosomes, present in every case, provided confir- mation of the epithelial origin of all 3 tumors. The Epstein- Barr virus (EBV) genome is contained in C15, C18 and C17 tumor cells with 30, I2 and 3 copies, respectively. The Epstein- Barr virus nuclear antigen (EBNA) was stained by the classical anti-complement immunofluorescence (ACIF) technique. Flu- orescence intensity was strong in CIS, moderate in C18, and hardly detectable in C17 cells. No expression of the EA and VCA antigens was detected. Flow cytometry analysis performed on monocellular suspensions showed the absence of detectable CR2 molecules (the EBV receptor on B lymphocytes) in all 3 tumors, and the constitutive expression of HLA class-ll antigens in CIS and C17 cells. IL-l activity was demon- strated in the supernatant of C15 and C17 cells cultivated in vitro for 3 days. These data confirm that the constitutive synthesis of MHC class-ll molecules and the release of IL-I- like activities are frequent features of NPC cells. These characteristics could be of importance in relation with the T-cell infiltrate found in NPC primary tumors.

Nasopharyngeal carcinoma (NPC) is unique among epithelial malignancies because of its epidemiological and biological characteristics (Klein, 1979; Henle and Henle, 1985). It occurs at a very high frequency in Southern China and to a lesser extent in Northern and Eastern Africa. It is consistently associated with Epstein-Barr virus (EBV) regardless of the geo- graphic origin. In addition, it presents with an exceptional histological pattern characterized by a heavy admixture of infiltrating lymphocytes within the malignant epithelial cells (Shanmugaratnam et al.. 1979; Herait et al., 1987). All these characteristics provide attractive clues for biological investigation. But, in contrast with Burkitt’s lymphoma (BL), another EBV-associated malignancy, all studies have been hampered so far by the rarity of tumor material and the absence of any NPC cell line propagated in vitro. Little is known about NPC cells, only scarce data being available on NPC cytogenetics (Mitelman et al., 1983), presence of EBV-encoded proteins, characterization of cell membrane antigens and growth factor production.

So far transplantation in nude mice appears the only way of obtaining abundant, homogeneous NPC material (Klein et al. , 1974). However, the difficulty of growing NPC material in nude mice has already been underlined (Henle and Henle, 1985). A few instances of NPC transplantation in nude mice have been reported (Klein et al., 1974; Trumper et al., 1977; Mitelman et al., 1983). However, most, if not all of these transplanted tumors are no longer available for further investigation.

With this in mind, a program of NPC transplantation in nude mice has been conducted at the Institut Gustave Roussy from 1983 to 1987. From each non-treated NPC patient undergoing retro-nasal biopsy, a tumor fragment was obtained for heterotransplantation. In addition, fresh tumor material was occasionally obtained by surgery on lymph-nodes or cu-

taneous metastases. Three transplantable NPC tumors termed C15, C17 and C18 are now routinely carried in animals; C15 was derived from a primary tumor, and C17 and C18 from metastatic tissue. These tumors have been extensively characterized with respect to cell morphology, EBV status, chromosome abnormalities and immunological characteristics.

METHODS

Patients Clinical and serological data are summarized in Table I. C15

tumor was obtained from a 13-year-old girl with a locally advanced tumor and early metastatic spread. She had not been treated previously. The biopsy was taken from the primary nasopharyngeal tumor.

The C17 biopsy was derived from a cutaneous metastasis in a 38-year-old male patient, after 4 years of NPC disease progression. This patient had initially undergone head and neck radiotherapy for the primary nasopharyngeal tumor and cervical lymph nodes. After bone and skin metastatic relapse, he received systemic polychemotherapy (bleomycin. cis-dichloro di-hydro platinum, 5-fluorouracil). Metastases of bone marrow, liver and skin were apparent at the time of biopsy.

C18 was derived from a 47-year-old man with locally advanced tumor and cervical lymph-node involvement. He had previously undergone a 2-month course of polychemotherapy (bleomycin, cis-dichloro-di-hydro platinum, 5 fluorouracil). The biopsy was taken from a cervical lymph node.

All 3 patients had elevated anti-VCA (viral capsid antigen) and anti-EA (early antigen) IgG. In addition, patient C18 had high titers of anti-VCA and anti-EA IgA. Transplantation

The initial fragments obtained from fresh biopsies were trimmed into 2-mm3 pieces, which were then inoculated by skin incision S.C. into the backs of 5 previously irradiated (5 Gy in one fraction) Swiss nude mice. The animals’ drinking water was supplemented with 0.5 n M estrone (Sigma, St. Louis, MO) from the day of tumor inoculation.

Tumors (1-2 cm3) were recovered 5 to 7 weeks later, frag- mented into pieces of 2 mm3 and used for subsequent passages. When viable monocellular suspensions were prepared from C 15 tumors (see below), successful grafting was also obtained by S.C. injection of 5 X lo6 viable cells in 0.5 ml of RPMI 1640 medium (Flow, Puteaux, France).

Light microscopy Tumor fragments were fixed in 10% buffered formalin and

embedded in paraffin, then 4-pm-thick sections were stained with hematoxylin and eosin.

’To whom reprint requests should be addressed.

Received: January 14, 1988 and in revised form March 22, 1988.

600 BUSSON ET AL.

TABLE I - CLINICAL DATA ON DONOR PATIENTS'

Length of Source of previous uansplanted Clinical

status Sex,

Patient age,, oriein chemotheraov tissue

EBV serological markers Histological diagnosis IgGVCA IgAVCA IgGEA IgAEA

C15 F Early advanced None 13 disease Morocco T4 N3

C17 M Met as t a t i c 2 years 39 relapse; France Initially

T3 N3

41 disease Algeria T4 N3

C18 M Early advanced 2 months

Nasopharyngeal Poorly 1/2,560 1/20 1/640 1/10 primary tumor differentiated

nasophary ngeal carcinoma

Cutaneous Poorly 111,280 1/40 1/160 1/10 metastasis differentiated

nasopharyngeal carcinoma

Ly mph-node Poorly 1/640 1/160 1/320 11160 metastasis differentiated

nasophary ngeal carcinoma

'Histological diagnosis is indicated according to the WHO classification (Shanmugaratnam ef a/. , 1979). Clinical staging was assessed according to the TNM classification of the International Union Against Cancer (Spiessl era / . . 1982). Titles of anti-VCA and anti-EA antibodies are indicated as measured at first testing.

Electron microscopy Tumor fragments (1 mm3) were fixed with 4% glutaralde-

hyde at 4"C, followed by osmium, then dehydrated and embedded in epoxy resin. Ultra-thin sections were cut on an LKB-111 ultramicrotome, stained for contrast with uranyl ace- tate and examined with a Philips EM-400 electron microscope. Preparation of monocellular suspensions

Tumor fragments (1 cm3) were minced into 1-mm3 pieces which were washed twice in RPMI medium. Tissue fragments were treated with 0.8% [w:v in RPMI 1640 supplemented with 20% fetal calf serum (FCS)] collagenase (type I, Sigma), and 0.02% DNase (DNase I, Sigma), for 15 rnin at 37°C. Viable cells were dispersed by pipetting collagenase-treated fragments in a solution containing 0.14 M NaC1, 6 mM KCI, 5.5 mM NaOHC03 6 m ~ EDTA, 5.5 mM glucose and 2 X M trypsin (Choay, Paris, France). Cells were trans- ferred into 5 ml FCS at 4°C to stop trypsin hydrolysis. After filtration through a sterile wire gauze, cells were washed, centrifuged and resuspended in RPMI medium. Approxi- mately 5 X lo7 viable cells were obtained from C15 tumors, and lo7 viable cells from C17 and C18 tumors.

Immunological reagents Three polyclonal sera from EBV-positive donors were used

as EBNA-positive reagents. Their antibody titers were 1 :640 for anti-VCA IgG, 1: 160 for anti-EA IgG, and 1:640 for anti- EBNA. They were checked for the absence of anti-nuclear antibodies.

The following monoclonal antibodies (MAbs) directed against EBV lytic-cycle antigens were kindly provided by Prof. G.R. Pearson, Washington, DC (Pearson and Luka, 1986): R3 directed against the 52-kDa polypeptide associated with the diffuse component of the early antigens (EA-D); R63 directed against the 85-kDa polypeptide associated with the restricted component of the early antigens (EA-R); V3 and L2 against the 160- and 125-kDa polypeptides associated with the viral capsid antigens (VCA) respectively; 2L10 directed against the gp 300/350 component of the membrane antigen (MA) complex.

We used the following MAbs directed against various cell- membrane antigens: W6-32 against a monomorphic determinant of HLA class-I molecules (Barnstable et al., 1978), D1- 12 against a monomorphic determinant of HLA class-I1 DR molecules (Carrel et al . , 1981), B7-21 against a monomorphic determinant of HLA class-II DP molecules (Royston et al . , 1981), L2 against a monomorphic determinant of HLA class- I1 DQ molecules (own data not shown); all 3 anti-class-I1 were kindly provided by Dr. D. Charron, Paris, France. The anti- CDw40 (328-5 antibody (Ledbetter et al., 1987) was kindly

provided by Dr. J.A. Ledbetter, Seattle, WA. HB5 (Becton- Dickinson, Grenoble, France) and OKE37 (Ortho, Aubervil- liers, France) detect 2 different epitopes on the CR2 (C3d/ EBV receptor) or CD21 molecule (Siaw et al., 1986).

Immunofluorescence assays on fixed cells Monocellular suspensions were derived from all 3 tumors

and smears were prepared with a cytocentrifuge and fixed in acetone (10 min at -20°C). Slides were incubated with appropriate dilutions of ascites fluid (45 min, 37"C), washed in phosphate-buffered saline (PBS, 137 mM NaC1, 2.6 mbi KCI, 1.14 mM Na2HP04, 3.21 mM KH2P04) PH 7.6 (10 rnin), then incubated with a dilution of fluoresceinated goat anti-mouse I g G ~ ~ ~ 9 2 (Tago, Burlingame, CA; 1:25, 45 min, 37°C). After washing in PBS, slides were mounted in PBS:glycerol: 1/1 and examined for fluorescence under a Leitz Ortholux microscope. Controls were treated with mouse ascites fluid.

EBNA staining was performed by the anti-complement immunofluorescence (ACIF) technique on acetone-methanol- fixed imprints (Reedman and Klein, 1973). Fresh tumor fragments were applied on glass slides which were air-dried, and fixed in acetone-methanol (v:v) for 10 min at -20°C. Serum from an EBV-negative donor was used as a negative control. Flow cytometry analysis of membrane immunojluorescence

Cells were incubated with appropriate dilutions of ascites fluid or purified MAb (5 X 105 cells in 50 pl; 4"C, 45 min), washed 3 times, incubated with fluoresceinated goat antisera to mouse (goat anti-mouse IgGFabo2; 1 5 0 in 50 p1 PBS; 4"C, 45 rnin), washed 3 times and resuspended in 200 pI PBS. Controls were treated with mouse ascitic fluid. Fluorescence of live cells (3 X 104/sample) was quantitatively analyzed by flow cytometry (Epics C, Coultronics, Margency, France). Percentage positive cells and mean fluorescence intensity was determined by computer analysis using the Irnrnuno Easy 88 program (Coultronics).

Detection of interleukin-l in culture supernatants C 15, C 17 and C 18 conditioned media were prepared accord-

ing to the following procedure: viable cells (106/ml) were incubated for 72 hr in RPMI plus 5% FCS. Interleukin-1 (IL- 1) activity was measured in the murine thymocyte co-mito- genic assay. Briefly, lo6 thymocytes from C3HIHej mice were incubated in 96-well plates, in RPMI medium supplemented with 5% FCS and 1 pglml phytohemagglutinin (DIFCO, De- troit, MI) in the presence of serial dilutions of tumor-cell- conditioned medium. After 48 hr of culture, thyrnocytes were pulse-labelled for 12 hr with 3H thymidine at 37 X lo3 Bq per well. IL-1 activity units were defined by reference to standard human IL-1 preparations tested in the same assay: recombinant

EBV-CONTAINING NPC TUMORS 601

IL-1 alpha (Dainippon, Osaka, Japan; lng/ml, 4 unitslml), and recombinant IL-1 beta (Otsuka, Tokoshima, Japan; lng/ ml, 4 units/ml). The IL-1 activity derived from NPC tumor cells was tested in the presence of rabbit sera directed to IL-1 alpha and IL-1 beta species. The anti-IL-1 alpha rabbit serum was a gift from Dr. T. Kasahara (Jichi, Japan; Kasahara et al., 1987) (dilution 1/1,O00). The anti-IL-1 beta rabbit serum was obtained from Otsuka (dilution 1: 100). A non-immune rabbit serum was used as a control at a dilution of 1 : 100. DNA extraction and Southern blot analysis

Frozen tumors (1 cm") were mixed with dry ice and pulver- ized in an IKA A10 grinder (Staufen, FRG). The resulting powder was suspended in 10 ml of lysis buffer containing 50 mM Tris HCI PH 7.6, 10 mM EDTA PH 8, 1 % sodium dodecyl sulfate (SDS) and proteinase K 200 pglml, then incubated for 2 hr at 60°C for digestion. Tumor DNA was obtained by phenol-chloroform extraction, ethanol precipitation, RNase digestion, a new round of phenol extraction and final ethanol precipitation. DNA samples (10 pg) were digested with restriction enzymes and subjected to electrophoresis through 0.8% agarose gel. After migration, gels were soaked in 0.15 M HCl for 10 min, then in 0.4 N NaOH for 30 min. Transfer onto Biotrace membrane (Gelman, Ann Arbor, MI) was done in 0.4 N NaOH. The filters were hybridized in 5 X SSPE [l X = 0.15 M NaCl, 30 m~ NaH2P04 and EDTA (PH 7.4) 1 m ~ , 1 % SDS, Denhardt's solution 0.1 % (1 % = 1 % each of bovine serum albumin, polyvinylpyrrolidone and Ficoll 400)], soni- cated salmon sperm DNA (100 pg/ml) and radioactive probe (5 X lo6 cpm/ml). Hybridization was performed at 42°C for 24 hr. Filters were then washed to a final stringency of 0.2 X SSPE and 0.1 % SDS at 65 "C and autoradiographed for 1 to 7 days at -70°C with an intensifying screen.

For EBV DNA quantitation, NPC DNA and DNA extracted from the BL line Raji were adjusted at the same optical den- sity; Raji DNA was serially diluted. Samples of equal volume were digested with Barn H1, electrophoresed, blotted and hybridized with the same probes in order to compare the intensity of hybridization.

RESULTS

Growth in nude mice Out of 60 attempts to heterotransplant NPC tissue derived

from primary tumor of untreated patients, one resulted in successful grafting of a tumor designated C15. In the other cases, grafted fragments failed to grow or stopped growing as early as passage 2. The C15 tumor has now been grown successfully in Swiss nude mice for 30 passages. In addition to direct transplantation of tumor fragments, successful grafting has also been obtained by s . ~ . injection of viable monocellular suspensions derived from C 15 tumors and cryopreserved in liquid nitrogen. Five million viable C15 cells are sufficient to generate a 1-cm3 tumor after 5 weeks of growth. This suggests a doubling time of approximately 3.5 days.

After 4 attempts to heterotransplant NPC tissue derived from metastases, we obtained 2 permanently transplantable tumors, designated C17 and C18. Both have been grown for 10 passages by direct transplantation of tumor fragments. Histology and cell mophology

The initial C15 biopsy fragment was derived from a primary tumor identified as a poorly differentiated carcinoma, with heavy admixture of nonmalignant lymphocytes (Fig. la). By passage 3 in nude mice, all infiltrating lymphocytes had dis- appeared, thus yielding a highly homogeneous tissue. On histological section, it consisted entirely of large cells with a syncytial appearance and large, clear nuclei with prominent nucleoli (Fig. lb). This pattern remained unchanged at various passages (3,4,20). Under no circumstances was an infiltration of mouse stromal cells morphologically detectable.

The C17 tumor was derived from a cutaneous metastasis. The primary tumor was a poorly differentiated carcinoma with few infiltrating lymphocytes. Within the metastatic tissue, carcinoma cells were trapped in a network of fibroblasts and collagen fibrosis (data not shown). At passages 2 and 6 in the nude mouse (Fig. lc), fibroblasts were completely eliminated and there was no evidence of infiltration by mouse stromal cells. NPC cells were organized into pseudo-stratified masses

FIGURE 1 - Histological sections of NPC material. (a) Primary tumor corresponding to the initial C15 transplant; presence of abundant tumor- infiltrating lymphocytes. (b) C15 tumor at passage 20; homogeneous layer of malignant epithelial cells with syncytial appearance. (c) C17 tumor at passage 6 . Scale bar, 50 pm.

602 BUSSON ET AL.

separated by wide areas of tumor necrosis. They exhibited clear cell margins with some rare intercellular bridges. Mini- mal keratinization was seen.

C18 was derived from a lymph-node metastasis of a poorly differentiated nasopharyngeal carcinoma with extra-capsular spreading (data not shown). The primary tumor was not available for histology. There were no infiltrating mouse stromal cells in the transplanted tumor; NPC cells were clearly limited without intercellular bridges or keratin deposits. Electron microscopy

In electron micrographs of each tumor, cells had a homogeneous appearance; they were electron-lucent with clear nuclei and prominent nucleoli (as exemplified by C18 cells in Fig. 2a). Desmosomes were observed in all 3 tumors (Fig. 2b and c ) . In addition, many bundles of intermediate filaments were

FIGURE 2 - Electron micrographs of transplanted tumors. (a) CIS tumor; cells are electron-lucent with clear nuclei and prominent nucleoli. (b) Desmosomes in C18 tumor (thin arrows). (c) C17 cells at high magnification; one desmosome (thin arrow), bundles of intermediate filaments (thick arrow). Scale bar, 1 pm.

observed in C17 cells (Fig. 2c). There was no structure reminiscent of EBV virions in any of the 3 tumors. Numerous C- type particles were visible in C15 tumors either in the intercellular spaces, or budding at the cell membrane.

Detection of EBV-encoded antigens Cells derived from C15 tumors were regularly checked for

EBNA antigen expression by the ACIF technique. In each test, at least 98% of the cells gave a positive EBNA reaction with a granular pattern of nuclear fluorescence (Fig. 3). EBNA antigens were also detected on C18 imprints with a moderate intensity of fluorescence and a similar granular pattern. In 5 attempts with 5 different anti-EBNA sera, fluorescence was extremely weak, at best, on C17 cells and the positivity of the test remained questionable.

MAbs directed against proteins related to MA, VCA and EA antigens did not stain any cells on fixed smears prepared from the 3 tumors.

Detection and quantitation of EBV DNA Total genomic DNA from C15, C17 and C18 tumors was

digested with the Bam HI endonuclease and probed with the Bam HI W and the Sal 1 F restriction fragments of the B95 EBV strain DNA (Baer et al., 1984). The restriction pattern (Fig. 4) was identical in all 3 tumors, consistent with the pattern predicted from the B95-8 EBV strain sequence and restriction map.

EBV-CONTAINING NPC TUMORS 603

For EBV DNA quantitation, we assumed that the 3 tumors contained homogeneous malignant cells of human origin and that all malignant cells within any one tumor contained the same amount of viral DNA.

The intensity of the hybridization obtained was compared on the major bands of 4.9 kb and 3.2 kb revealed by the Sal 1F probe (Fig. 4a) and the Bam H1 W probe (Fig. 4b), respectively. The bands in C15, C18, and C17 lanes compared in intensity with those obtained with Raji material diluted 1:2, 1:5 and 1:25, respectively. The same result was obtained with the Bam H1 W probe, which detects a repetitive fragment as with the Sal 1F probe, which hybridizes to a non-repetitive fragment of the EBV genome. With approximately 60 EBV DNA copies in every Raji cell (Sternas et al., 1986), the mean number of viral DNA copies can be approximated to 30, 15 and 3 in C15, C18 and C17 cells, respectively. Cell membrane antigen analysis

Viable monocellular suspensions could be obtained from all 3 tumors, thus allowing analysis of membrane antigens by flow cytometry (Table 11, Fig. 5). A strong and stable expression of both HLA class-I and class-I1 antigens was demon- strated at the surface of C15 cells. As shown in Figure 5 and Table 11, 89 % of the cells obtained after tumor disaggregation were HLA class-I-positive; most of them were also HLA DR-, DP- and DQ-positive. Similar data were obtained at various

F I C U R E ~ - Southern blots of C15, C17 and C18 tumor DNA or dilutions of Raji cell DNA. NPC DNA (10 pg) and decreasing amounts of Raji DNA (1/1 = 10 pg) were digested with the restriction enzyme Bam H1. (a) Filter hybridized with the Sal 1 F fragment of EBV DNA. The following Bam HI fragments are visible: B (= 8.5 kb), K (= 4.9 kb) and R (= 3.8 kb). (b) Filter hybridized with the Bam H1 W fragment of EBV DNA.

passages. The CD40 antigen was significantly and permanently expressed on C15 cells.

C17 cells exhibited a strong expression of HLA class-I, HLA DR, DP and DQ antigens. CD40 antigen was detected at a lower intensity, and on a smaller proportion of the cells. In contrast, HLA class-I antigens were weakly expressed, and HLA class-I1 and CD 40 were consistently negative on C18 cells.

The CR2 molecule (C3d/EBV receptor) was undetectable on all 3 tumors by 2 MAbs directed against 2 different epitopes of the protein. Interleukin-I production

A highly significant IL-1 activity was regularly observed in C15 and C17 short-term culture supernatant (Table 111). We could not detect any IL-1 activity in C18 conditioned media.

Immunological analysis of IL-1 activity yielded different results for C15 and C17 cells. IL-I activity recovered from CIS cells was completely inhibited by rabbit serum directed against recombinant 1L-1 alpha; no significant inhibition was

FIGURE 3 - Immunofluorescence staining of EBNA antigens on C15 observed with serum directed against recombinant IL-1 beta. cells. (a) Tumor imprint treated with EBNA-positive serum. (b) Con- In contrast, IL-1 activity from C17 cells was not inhibited trol treated with EBNA-negative human serum. Scale bar, 10 pm. either by anti-IL-1 alpha or anti-IL-1 beta serum.

604 BUSSON ET AL.

A , epithelial cells retain the ability to proliferate in nude mice, ,n. thus allowing the tumor cells to be separated from non-malignant, infiltrating human elements. In our experience, and in contrast with findings of others (Chen et al., 1978), human infiltrating cells were not replaced by abundant mouse stromal cells. Infiltration of the nude-mouse-grown tumors by murine

. k ! 4. [ti.

“L,+

TABLE 11 - SURFACE PHENOTYPE OF TRANSPLANTED NPC CELLS ANALYZED BY MEMBRANE IMMUNOFLUORESCENCE AND FLOW CYTOMETRY. MOUSE ASCITES FLUID WAS USED AS A NEGATIVE CONTROL. THE PERCENTAGE OF POSITIVE CELLS AND THE VALUE OF MEAN FLUORESCENCE

ARE MENTIONED FOR EACH ANTIGEN

Complement Complement receptor 2 receptor 2 CD 40

DQ (G28-5) HLA class I HLA class I1 HLA class I1 HLA class I1

(om71 A - B - C DR DP Antigens Negative

(W6-32) (DI-12) (B7-21) (L2) (antibodies) control

C15 14% 20 89% 180 85% 168 73% 96 81% 130 81% 67 15% 29 15% 27 C17 27% 23 89% 122 92% 137 83% 87 86% 100 64% 65 21% 37 26% 34 C18 10% 31 62% 69 10% 15 5% 7 17% 33 4% 28 19% 28 8% 35

TABLE 111 - DETECTION OF IL-1 ACTIVITY IN NPC-CELL-CONDITIONED M,EDIA; ANALYSIS WITH ANTI-IL-1 ALPHA AND ANTI-IL-1 BETA ANTIBODIES

Tumor-cell-conditioned medium

C15 c 1 7 C18

IL-I unitslml Recombinant Recombinant human IL-1 alpha human IL-1 beta Rabbit serum

None 5.3 3.1 0.1 4.9 5.2 Non-immune 5.1 2.9 ND 3.9 4.6 Anti-IL-1 alpha 0.1 2.8 ND 0.1 4.4 Anti-IL- 1 beta 4.9 3.1 ND 5 0.1 ‘Results represent means of triplicate tests. Controls: Recombinant IL-1 alpha (Inglml. 4 unitslml) and recombinant IL-1

beta (lng/ml, 4 unitdml).

605 EBV-CONTAINIh'G NPC TUMORS

EBV genome was probed in all 3 tumors, with 2 probes corresponding to a repetitive (Bam H1 W) and a unique (Sal 1 F) sequence of the viral genome. With these probes-which account for approximately 6% of EBV DNA-we observe the same Bam H1 restriction pattern in C15, C17 and C18 material; it was consistent with the pattern deduced from the pro- totype B95-8 genome sequence (Baer et al., 1984). In contrast, the quantitation of EBV genomes varied widely from approximately 30 copies/cell for C15 to 12 and 3 copies for C18 and C17 cells, respectively. These data are reminiscent of the situation observed in E,BV-positive BL: while the structure of EBV DNA is grossly conserved in most BL-derived cell lines (Dambaugh et al., 1986), the number of viral genomes per cell varies from 100 in the Daudi cell line to only one or 2 in the Namalwa cell line (Sternas et al., 1986).

EA, VCA and MA complexes, which correspond to the lytic cycle of EBV, were not detected in any of the cells derived from the 3 tumors. Accordingly, no EBV particles were visible on electron micrographs. While EBV particles are not observed in fresh NPC biopsies (Klein, 1979; Henle and Henle, 1985), transplanted NPC cells may be permissive of EBV replication. Some tumors spontaneously produce EBV particles which are detectable in degenerated cells or in cystic fluid; in many other cases, cells derived from non-producing tumors may support complete replication of the virus when treated in vitro by bromodeoxyuridine (Trumper et al., 1977; Crawford et al., 1979b). Further investigation will be required to assess the inducibility in vitro of EBV replication in the C15, C17 and C18 cells.

C-type particles were observed in the C15 tumor. Similar structures, reported in transplanted NPC tumors (Crawford et al., 1979a) were ascribed to infection by murine xenotropic retroviruses.

The analysis of membrane antigens on NPC transplanted cells yielded 2 main results: the absence of CR2 molecules in all 3 tumors, and the constitutive expression of HLA class-I1 antigens on C15 and C17 cells.

The CR2 molecule, a 140-kDa glycoprotein, which is the receptor of the C3d fraction of the complement, has been identified as the EBV membrane receptor on B lymphocytes (Frade et al., 1985; Siaw et al., 1986). More recently, using MAb HB5 directed to the CR2 molecule, Young et al. (1986) were able to stain the basal and parabasal layers of malpighian epithelia in various anatomic sites including normal nasopharyngeal mucosa. Sixbey et al. (1987) have observed a corre- lation of HB5 staining with EBV binding on normal epithelial cells in primary culture. Referring to these data, they have both suggested that the CR2 molecule was also the EBV receptor on epithelial cells. Nevertheless, we could find no evidence of the presence of the CR2 molecule on malignant NPC cells carried on nude mice. This negative result was obtained with 2 different MAbs, HB5 and OKB7, the latter being directed to a CR2 epitope specifically involved in EBV binding (Siaw et al., 1986). Similar negative results were obtained by immunofluorescence staining performed on tumor imprints (data not shown), thus confirming that CR2 negativity was not due to trypsin or collagenase treatment. The CR2 molecule could have been present initially on nasopharyngeal epithelial cells, and switched off at some later stage of tumor progression. An alternative hypothesis implies that EBV epithelial target cells are indeed CR2-negative and that the virus uses another membrane receptor to infect nasopharyngeal cells.

We (Herait et al., 1987) and others (Thomas et al., 1984) have previously noted that HLA class-I1 antigens are consistently expressed by malignant epithelial cells from NPC primary tumors. Expression of HLA class-I1 antigens rarely occurs in non-malignant epithelial cells, except when closely

associated with T lymphocytes, as for example in grafi-versus- host disease (Volc-Platzer et al., 1984). In contrast, HLA class-I1 antigens remain strongly expressed on C15 cells up to passage 30, despite the disappearance of the T-cell infiltrate present in the original biopsy. Similarly, HLA class-I1 antigens remained permanently expressed on C17 cells which were derived from metastatic tissue. This provides substantial evidence that class-I1 antigens are constitutively expressed by malignant NPC cells and do not depend upon a T-cell microen- vironment. Moreover, all 3 DR, DP and DQ subregions of HLA class I1 are expressed on C15 and C17 cells. This pattern of expression contrasts with that found in other malignant epithelial cells. For instance, colon carcinoma cells express mainly HLA DR, with minimal amounts of DP and DQ antigens (Ghosh et al., 1986). NPC tumors could thus provide a new model for exploring the regulation of HLA class-II expression in epithelia.

Another point of interest is the expression of the CD40 antigen on C15 and C17 cells. CD40 designates a 50-kDa antigen shared by B lymphocytes and some species of carcinoma such as lung and breast carcinomas (Ledbetter et al., 1987). It has been suggested that CD40 and MHC class-I1 antigens could be co-regulated in B-cells (Ledbetter et al., 1987). Their simultaneous expression in NPC epithelial cells might support this hypothesis.

The presence of a massive lymphoid infiltrate is a striking characteristic of NPC tumors. Most of the tumor-infiltrating lymphocytes are T cells, 10 to 20% of which display an activated phenotype (co-expression of the DR antigen and of the IL-2 receptor; Herait et al., 1987). It is attractive to consider that malignant epithelial cells may contribute to the development of the T-cell infiltrate through the release of lymphotactic agents. Among these agents, IL-1, which is a key factor in T-cell migration and activation, appears a good candidate. Two molecular species of IL-1 have been characterized so far, IL-1 alpha and IL-1 beta (March et al., 1985). These 2 species have different amino-acid sequences and are encoded by 2 different genes. The production of IL-1 alpha is a frequent feature of NPC cells (Busson et al., 1987). Cells derived from the C15 tumor or from NPC primary tumors produce an IL-1 activity with strong biochemical and immunological homologies to IL-1 alpha. In addition, C15 total RNA specifically hybridizes with a cDNA probe to IL-1 alpha. Data reported here show that C18 cells did not produce any IL-1 activity, as detected using the murine thymocyte assay. In contrast, a significant IL-1 activity was recovered from C17 cells but was inhibited neither by anti-IL-1 alpha nor by anti- IL-1 beta antibodies. These results suggest that the IL-1 activity derived from C17 cells might be supported by another molecule distinct from both IL-1 alpha and IL-1 beta species. Similar observations have previously been made in EBV-positive lymphoblastoid cell lines (Wakasugi et al., 1987). In the 3B6 cell line investigated in this laboratory, an IL-1-like molecule was shown to act as an autocrine growth factor (Waka- sugi et al., 1987). The production of IL-1 activity by EBV- infected cells of another kind, i . e . , NPC, thus appears of great interest, and would obviously deserve further study.

ACKNOWLEDGEMENTS

We thank Dr. C. Gosse for technical assistance in the man- agement of nude mice and Drs. G. Schwaab, E. Cvitkovic and C. Micheau for helpful discussions. P. Busson was the recipient of a grant from the Fondation pour la Recherche MCdicale, Paris. G. Ganem was the recipient of a grant from the Institut Gustave Roussy, Villejuif. This work was supported by a grant from the Association pour la Recherche sur le Cancer, Paris, and from the Ligue Nationale Franqaise contre le Cancer, Paris.

606 BUSSON ET AL.

REFERENCES

BAER, R., BANKIER, A.T., BIGGIN, M.D., DEININGER, P.L., FARELL, P.J., GIBSON. T.J., HATFULL, G., HUDSON, G.S., SATCHWELL, S.C., SEGUIN, C., TUFFNELL, P.S. , and BARRELL, B.G., DNA sequence and expression of the B95-8 Epstein-Barr virus genome. Nature (Lond.), 310, 207-211 (1984). BARNSTABLE, C.J.. BODMER, W.F., BROWN, G., GALFRE, G. , MILSTEIN, C., WILLIAMS. A.F., and ZIEGLER, A., Production of monoclonal antibodies to group A erythrocytes, HLA and other human cell surface antigens: new tools for genetic analysis. Cell, 14, 9-20 (1978). BUSSON, P., BRAHAM, K., GANEM, G., LIPINSKI, M., GRAUSZ, D., WAK- ASUGI, H., and TURSZ, T., EBV containing epithelial cells from nasopharyngeal carcinoma produce interleukin-1 alpha. Proc. nat. Acad. Sci.

CARREL, S. , Tosi, R . , GROSS, N., TANIGAKI, N., CARMAGNOLA, A.L., and ACCOLA, R.S., Subsets of human Ia-like molecules defined by monoclonal antibodies. Mol. Immunol., 18,403-411 (1981). CHEN, H.C., KAWAMURA, A., MURATA, J.M., HAMAJIMA, K., OSONO, M., SUZUKI, K., and SUDO, K., A histopathological study of lymphoid tissue reaction to metastatic nasopharyngeal carcinoma in nude mice. In: G. de-The and Y. Ito (eds.), Nasopharyngeal carcinoma: eriology and conrrol, pp. 65-84, IARC, Lyon (1978). CRAWFORD, D.H., ACHONG, B.G., TEICH, N.M., FINERTY, S. , THOMP- SON, J.L., EPSTEIN, M.A., and GIOVANELLA, B.C., Identification of murine endogenous xenotropic retrovirus in cultured multicellular tumour spheroids from nude-mouse-passaged nasopharyngeal carcinoma. Itzt. J . Cancer, 23, 1-7 (1979~). CRAWFORD, D.H., EPSTEIN, M.A., BORNKAMM, G.W., ACHONG, B.G., FINERTY. S. , and THOMPSON, J.L., Biological and biochemical observations on isolates of EB virus from the malignant epithelial cells of two nasopharyngeal carcinomas. fnt. J . Cancer, 24, 294-302 (19796). DAMBAUGH, T., HENNESSY, K., FENNEWALD, S. , and KIEFF, E. , The virus genome and its expression in latent infection. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus. Recenr advances, pp. 13-45, Heineman, London (1986). FRADE, R., BAREL, M., EHLIN-HENRIKSSON, B., and KLEIN, G., gp140, the C3d receptor of human B lymphocytes, is also the Epstein-Barr virus receptor. Proc. nut. Acad. Sci. (Wash.), 82, 1490-1493 (1985). GHOSH, A.K., MOORE, M., STREET, A.J., HOWAT, J.M.T., and SCHO- FIELD, P.F., Expression of HLA-D sub-region products on human colorec- tal carcinoma. Int. J . Cancer, 38,459-464 (1986). GIOVANELLA, B.C., and FOGH, J., The nude mouse in cancer research. Advanc. Cancer Res., 44,69-120 (1985). HENLE, W., and HENLE, G. , Epstein-Barr virus and human malignancies. In: G . Klein (ed.), Advances in viral oncology, Vol. 5 , pp. 201-238, Raven, New York (1985). HERAIT, P., GANEM, G., LIPINSKI, M., CARLU, C. , MICHEAU, C., SCHWAAB. G.. DE-THE. G.. and TURSZ. T.. LvmDhocvte subsets in tumours

(Wash.), 84,6262-6266 (1987).

of patients with undifferentiated nasopharyhgeal cakinoma: presence of lymphocytes with the phenotype of activated T cells. Brit. J . Cancer, 55, 135-139 (1987). KASAHARA, T., MUKAIDA, N., SHINOMIYA, H., IMAI, M., MATSUSHIMA, K., WAKASUGI, H., and NAKANO, K., Preparation and characterization of polyclonal and monoclonal antibodies against human interleukin-1 alpha. J . Immunol., 138, 1804-1812 (1987). KLEIN, G., The relationship of the virus to nasopharyngeal carcinoma. In: M.A. Epstein and B.G. Achong (eds.), The Epsrein-Burr virus, pp. 339- 350, Springer, New York (1979). KLEIN, G . , GIOVANELLA, B.C., LINDAHL, T.. FLALKOW, P.J., SINGH, S. , and STEHLIN, J.S., Direct evidence for the presence of Epstein-Barr virus

DNA and nuclear antigen in malignant epithelial cells from patients with poorly differentiated carcinoma of the nasopharynx. Proc. nat. Acad. Sci. (wash.), 71,4737-4741 (1974). LEDBETTER, J.A.. SHU, G., GALLACHER, M., and CLARK, E.A., Augmen- tation of normal and malignant B cell proliferation by monoclonal antibody to the B cell-specific antigen BP 50 (CDw40). J. Immunol., 138,

MARCH, C.J., MOSLEY. B., LARSEN, A., CERETTI, D.P., BRAEDT, G., PRICE, V., GILLIS, S., HENNEY, C.S., KRONHEIM, S.R., GRABSTEIN, K., CONLON, P.J., HOW, T.P., and COSMAN, D., Cloning, sequence and expression of two distinct human interleukin-1 complementary DNAs. Nature (Lond.), 315, 641-647 (1985). MITELMAN, F.. MARK-VENDEL, E., MINEUR, A, , GIOVANELLA, B. . and KLEIN. G . , A 3q' marker chromosome in EBV-carrying nasopharyngeal carcinomas. fnt. J. Cancer, 32, 651-655 (1983). PEARSON, G.R.. and LUKA, J., Characterisation of the virus-determined antigen. Iti: M.A. Epstein and B.G. Achong (eds.), The Epstein-Barr virus. Recenr advances, pp. 47-73, Heinemann, London (1986). REEDMAN, B.M., and KLEIN, G., Cellular localization of an Epstein-Barr virus (EBV)-associated complement-fixing antigen in producer and non- producer lymphoblastoid cell lines. fnt. J. Cancer, 11,499-520 (1973). ROYSTON, I., OMARY, M.B., and TROWBRIDGE, I.S., Monoclonal antibodies to a human T-cell antigen and Ia-like antigen in the characterization of lymphoid leukemia. Transplant. Proc., 13, 761-766 (1981). SHANMUGARATNAM, K., CHAN, S.H., DE-THE, G., GOH, J.E.H., KHOR, T.H., SIMONS, M.J., and TYE, C.Y., Histopathology of nasopharyngeal carcinoma. Correlations with epidemiology, survival rates and other biological characteristics. Cancer, 44. 1029- 1044 (1979). SIAW, M.F.E., NEMEROW, G.R., and COOPER, N.R., Biochemical and antigenic analysis of the Epstein-Barr virusK3d receptor (CR2). J. fm- munol., 136,4146-4151 (1986). SIXBEY, J. W., DAVIS, D.S., YOUNG, L.S., HUTT-FLETCHER, L., TEDDER, T.F., and RICKINSON, A.B., Human epithelial cell expression of an Ep- stein-Barr virus receptor. J. gen. Virol., 68, 805-811 (1987). SPIESSL, B., SCHEIBE, O., and WAGNER, G. (eds.), International Union Against Cancer, TNMAtlas, p. 22, Springer, New York (1982). STERNAS, L., ELIASSON, L., LERNER, R., and KLEIN, G., Quantitation of Epstein-Barr virus (EBV)-determined nuclear antigen (EBNA) by a two- site enzyme immunoassay. in parallel with EBV-DNA. J. imniunol. Meth.,

THOMAS, J.A., ILIESCU. V., CRAWFORD, D.H., ELLOUZ, R., CAMMOUN, M. I and DE-THE, G., Expression of HLA-DR antigens in nasopharyngeal carcinoma: an immunohistological analysis of the tumour cells and infiltrating lymphocytes. Inr. J . Cancer, 33, 813-819 (1984). TRUMPER, P.A., EPSTEIN, M.A., GIOVANELLA, B.C., and FINERTY, S. , Isolation of infectious EB virus from the epithelial tumour cells of nasopharyngeal carcinoma. Int. J. Cancer, 20, 655-662 (1977).

W., LECHNER, K., and STINGL, G., Evidence of HLA-DR antigen biosyn- thesis by human keratinocytes in disease. J . exp. Med., 159, 1784-1789 (1984). WAKASUGI, H., RIMSKI, L., MAHE, Y., MAHMOUD KAMEL, A,, FRADELIZI, D. , TURSZ, T., and BERTOGLIO, J., Epstein-Barr virus-containing B-cell line produces an interleukin 1 that it uses as a growth factor. Proc. nut. Acad. Sci. (Wash.), 84, 804-808 (1987). YOUNG, L.S., CLARK, D., SIXBEY. J.W., and RICKINSON, A.B., Epstein- Barr virus receptors on human pharyngeal epithelia. Lancet, I, 240-242 ( 1986).

788-794 (1987).

89, 151-158 (1986).

VOLC-PLATZER, B., MAIDIC, O., KNAPP, W., WOLFF, K . . HINTERBERGER,

Documents

Establishment and characterization of three transplantable EBV-containing nasopharyngeal carcinomas