Epstein-Barr Virus Is Detected in Undifferentiated Nasopharyngeal Carcinoma but not in LymphoepitheliomaLike Carcinoma of the Urinary Bladder MARGARET L. GULLEY, MD, MAHUL B. AMIN, MD, JOHN M. NICHOLLS, MBBS, FRCPA, PETER M. BANKS, MD, ALBERTO G. AYALA, MD, J. R. SRIGLEY, MD, PHYLLIS A. EAGAN, BS, AND JAE Y. RO, MD The Epstein-Barr virus (EBV) is associated with nasophmyngeal carcinoma (NPC) and with lymphoepithelioma-like carcinomas devel-
oping in certain anatomic sites. In this study, an in situ hybridization was used to identify EBV-encoded ribonucleic acid (RNA) (EBER1) transcripts in 32 of 45 cases of NPC but not in any of the 11 lymphoepithefioma-like carcinomas developing in the urinary bladder. EBER1 was most commonly detected in those NPCs having undifferentiated or nonkeratinizing squamous histology rather than the kerafinizing squamous cell subtype of NPC. The EBV-encoded latent membrane protein 1 (LMPI) was expressed focally in only seven of 21 EBERl-positive NPCs by an immunohistochemical technique. These findings imply that EBER1 hybridization is more sensitive than LMP1 immunohistochemistry on paraffin sections in detecting carcinoma-associated virus. Previous in vitro studies have suggested that LMP1 expression might be a function of differentiation, but this study of naturally infected NPCs s h o w e d n o strong correlation be-
Lymphoepithelial carcinomas develop in diverse anatomic sites, most commonly in the nasopharynx. They are characterized by the presence of large undifferentiated malignant epithelial cells associated with a prominent infiltrate of benign reactive lymphocytes. Epstein-Barr virus (EBV) is frequently present in the malignant epithelial cells of such carcinomas arising in the nasopharynx, stomach, lung, thymus, and salivary gland, but not in those of the skin, breast, and uterine cervix, a-~°The restricted anatomic distribution of EBVassociated carcinomas has led to the hypothesis that only foregut-derived organs are susceptible to EBV-associated carcinogenesis, possibly because these organs are in close proximity to sites of natural viral replication.~ The portal of entry of the virus is thought to be From the Departments of Pathology, University of Texas Health Science Center at San Antonio and Audie L Murphy Memorial Veteran's Hospital, San Antonio, TX; Henry Ford Hospital, Detroit MI; University of Hong Kong, Hong Kong; University of Texas MD Anderson Cancer Center, Houston TX; and Sunnybrook Medical Center, Toronto, Ontario, Canada. Accepted for publication April 19, 1995. Presented in part at the 1994 U.S. and Canadian Academy of Pathology Meeting in San Francisco and the 1994 Symposium on Epstein-Barr Virus and Associated Diseases in Cold Spring Harbor. Supported by grants from the National Institutes of Health (K08CA01615) and the Veteran's Administration. Address correspondence and reprint requests to Margaret L. Gulley, MD, Department of Pathology, University of Texas Health Science Center, 7703 Floyd Curl Dr, San Antonio, TX 78284-7750. This is a US government work. There are no restrictions on its use,
0046-8177/95/2611~000750.00/0
tween LMP1 positivity and degree of tumor differentiation, albeit a limited spectrum of differentiation that could be examined. In two cases in which f r o z e n tissue was available, the NPCs were monoclonal with respect to viral D N A structure, implying that the virus was present before malignant transformation. Unlike NPCs, the lymphoepithelioma-like carcinomas of the bladder were uniformly EBV negafive, lending further evidence to the growing body o f literature linking EBV with lymphoepithelial carcinomas of foregut-derived tissues but not with similar-appearing tumors developing in other anatomic sites. HUM PATHOL 26:1207--1214. This is a US government work. There are no restrictions on its use. Key words: Epstein-Barr virus, nasopharyngeal carcinoma, bladder carcinoma, lymphoepithelioma-like carcinoma. Abbreviations: EBV, Epstein-Barr virus; NPC, nasopharyngeal carcinoma; RNA, ribonucleic acid; EBER1, Epstein-Barr v i r u s - e n c o d e d ribonucleic acid; LMP1, latent membrane protein 1.
oropharyngeal mucosa that also serves as the site of production of infectious virus shed periodically in saliva. ~2Infection of lymphocytes permits systemic dissemination of the virus, and EBV lies latent in a few lymphocytes for the duration of life following primary infection. EBV-containing lymphomas are thought to develop from this subset of latently infected lymphocytes, especially when immunodeficiency results in failure to recognize and destroy infected cells. In contrast, immunodeficiency does not seem to predispose to EBVassociated carcinomas, and the precise role of the virus in the development of either carcinoma or lymphoma is not understood. One viral factor that may play a role in tumorigenesis is the latent membrane protein 1 (LMP1). In vitro studies have shown that LMP1 transfection confers tumorigenicity is and inhibits keratinocyte differentiaLion. 14 However, the role of LMP1 in viral carcinogenesis in vivo is less certain. Expression of LMP1 in the skin of transgenic mice induces hyperplastic dermatitis and aberrant keratin production but not carcinoma, a5 LMP1 is not always expressed in nasopharyngeal carcinomas (NPC),I~2° but it is almost always expressed in infectious mononucleosis 21'22 and in the Reed-Sternberg cells of EBV-associated Hodgkin's disease. 23 Another EBV gene product, the nonpolyadenylated EBV-encoded-RNA (EBER), is abundantly expressed in latently infected cells and in all types of EBVinfected tumors examined t o d a t e . ~4-~6 The function of EBER is still uncertain, but the sheer abundance of
1207
HUMAN PATHOLOGY Volume26, No. 11 (November 1995) EBER transcripts provides a c o n v e n i e n t m a r k e r o f tumor-associated viral infection. I n this study, 45 cases o f N P C a n d 11 cases o f lymphoepithelioma-like c a r c i n o m a o f the b l a d d e r were evaluated for the p r e s e n c e o f EBER1 transcripts by in situ hybridization to investigate associations between a n a t o m i c site, stage o f t u m o r differentiation, a n d the p r e s e n c e o f tumor-associated EBV. I n cases in which f r o z e n tissue was available, t u m o r clonality was analyzed with respect to EBV g e n o m i c structure. I m m u n o h i s t o chemistry was u s e d to m e a s u r e a n d localize LMP1 expression a n d to c o m p a r e the diagnostic utility o f LMP1 staining with that o f EBER1 staining f o r the d e t e c t i o n o f carcinoma-associated EBV.
MATERIALS AND METHODS Tissue Aquisition a n d Histopathological Diagnosis Paraffin-embedded tissues were collected retrospectively from 34 cases of NPC and 11 cases of lymphoepithelioma-like carcinoma of the bladder in the United States and Canada. The clinical and histopathologic details of nine of the bladder carcinomas have been previously published, 27 and two additional cases from the University of Toronto were included because they had similar clinical and histological features. Every bladder carcinoma included in this study was histologically indistinguishable from NPC, at least in focal areas of the tumor, and these areas were the target of this investigation. An additional 11 cases of NPC were obtained from the University of Hong Kong. In each of the cases from Hong Kong, a biopsy was performed on the patient's nasopharyngeal tumor twice, and one was sent to the Pathology Department, whereas the other was fixed immediately in 95% ethanol and stored at 4°C in the Department of Radiation Oncology for up to 2 months before histological processing, embedding, and sectioning. Diagnosis of NPC was confirmed, and the predominant histological pattern was assigned by one author (PMB) according to standard histopathologic criteria, 2s and without knowledge of clinical or experimental data. NPCs were categorized as either undifferentiated (Regaud or Schmincke variants), keratinizing, or nonkeratinizing squamous cell carcinomas. In cases of undifferentiated NPC in which Regaud and Schmincke patterns coexisted, the predominant pattern was assigned. The Fisher's exact test (two tailed) was used to assess statistical significance among groups.
In Situ Hybridization Tissue specimens that had been routinely formalin fixed, processed, and paraffin embedded were sectioned onto silane-coated slides. Hybridization was performed as previously described 29 using digoxigenin-labeled (Boehringer-Mannheim, Indianapolis, IN) riboprobes complementary to EBER1 and U6 transcripts. EBER1 expression is characteristic of latent EBV infection. 24'26 U6 is a ubiquitous cellular RNA whose expression serves as an indicator of RNA preservation. U6 is similar in copy number, size, and nuclear localization 3° to EBER1, but is encoded by a cellular gene whose transcription is thought to be essential for RNA splicing. U6 RNA was detected in cytologically malignant epithelial ceils as well as in most reactive lymphocytes and stromal cells except in one case of bladder carcinoma that was excluded from the study. Focal or diminished U6 RNA preservation was observed in
some tissues, particularly in zones of necrosis or in small biopsy specimens. Because RNA may be easily degraded, EBER1 results were interpreted in direct comparison with visualization of U6 RNA preservation on an adjacent section. Probe templates for EBER1 (RA386) and U6 (RA390) were kindly donated by Richard Ambinder, MD, PhD, of Johns Hopkins University.
Immunohistochemistry LMP1 immunostains were performed on paraffin sections using a cocktail of antibodies (CS1-431; DAKO, Carpinteria, CA) and a peroxidase-labeled streptavidin-biotin detection system (DAKO).32 Antigen availability was enhanced, and false nuclear reactivity was minimized by pretreatment with either (1) microwaving for 10 minutes in 10-mmol/L citrate buffer pH 6.0 or (2) a 5-minute incubation at 37°C in I rag/ mL pronase E in phosphate-buffered saline. Both techniques were equally effective. The sensitivity and specificity of the stains was evaluated using EBV-positive tissues from nine cases of infectious mononucleosis, 22 19 cases of Hodgkin's disease, and the C15 NPC cell line propagated in nude mice. 33Specific LMP1 staining was localized to the cytoplasm and membrane of affected cells.
Southern Blot Analysis Tissues were snap frozen at the time of biopsy and stored at -70°C until analysis. High molecular weight DNA was isolated by SDS-proteinase K lysis and phenol-chloroform extraction, and analyzed by the Southern blot method using 32p_ labeled riboprobes complementary to the EBV X h o l a and EcoRlI fragments 34 (provided by Nancy Raab-Traub, PhD, of the University of North Carolina at Chapel Hill). DNA from the Raji Burkitt lymphoma cell line was used as a standard by which to estimate the amount of EBV DNA in each tumor specimen.
RESULTS
EBERI In Situ Hybridization EBER1 transcripts were f o u n d in 32 o f 45 cases o f NPC. In the 32 positive cases, the EBER1 signal was p r e s e n t in virtually all o f the m a l i g n a n t epithelial cells b u t n o t in the infiltrating lymphocytes o r o t h e r n o n m a lignant stromal cells. N o r m a l c o l u m n a r epithelium o f the n a s o p h a r y n x was u n i f o r m l y negative. In two instances in which s q u a m o u s metaplasia coexisted with EBERl-positive c a r c i n o m a , the metaplastic epithelium did n o t express EBER1. W h e n EBER1 was observed, it was restricted to the n u c l e u s o f the m a l i g n a n t cells with sparing o f the n u c l e o l u s (Fig 1). Table 1 shows the correlation between EBER1 expression a n d the histopathologic subtype o f NPC. T h e r e was a strong association between EBER1 expression a n d u n d i f f e r e n t i a t e d histology. EBV was c o m m o n in the R e g a u d p a t t e r n o f u n d i f f e r e n t i a t e d N P C (15 o f 19; 79% positive) characterized by islands o f cohesive, u n d i f f e r e n t i a t e d epithelial cells s u r r o u n d e d by small, well-differentiated lymphocytes. T h e Schrnincke pattern o f u n d i f f e r e n t i a t e d NPC, characterized by individual m a l i g n a n t - a p p e a r i n g ceils dispersed in a s t r o m a o f small lymphocytes, was also EBV-associated at a similar f r e q u e n c y (10 o f 13; 77%). Two additional cases o f u n -
1208
EPSTEIN-BARR VIRUS IN _YMPHOEPITHELIOMAS (Gulley et al)
FIGURE 1. (A) In situ hybridization to EBER1transcripts in paraffin sections of an undifferentiated NPC shows that EBER1 is localized to the nucleus of all malignant epithelial cells. (B) Hematoxylin-eosin stain of an adjacent section shows the cytological detail of the tumor ceils and the surrounding reactive lymphocytes. (C) Hybridization of an EBERl-negative NPC to a control probe (U6) shows that RNA was well preserved in the tumor cells and in most of the reactive lymphocytes, thus validating the EBER1 result as true negative. (A & B, case no. 601; C, case no. 985; original magnification ×600.)
differentiated NPC were EBERl-positive but were not histologically subtyped because of the low n u m b e r of tumor cells. EBER1 transcripts were found in five of eight (63%) cases of nonkeratinizing squamous cell NPC
TABLE I . C o r r e l a t i o n of EBV G e n e Expression w i t h Histological S u b t y p e of N a s o p h a r y n g e a l C a r c i n o m a and Geographic Location EBER1
Undifferentiated Regaud Schmincke Unclassified Nonkeratinizing squamous Keratinizing squamous Total
United States
Hong Kong
Total
LMP1 positive/ E B E R 1positive
17/23 10/13 6/9 1/1
10/11 5/6 4/4 1/1
27/34 15/19 10/13 2/2
5/16 3/9 2/6 0/1
5/8
0
5/8
2/5
0/3 21/34
0 10/11
0/3 32/45
NA 7/21
NOTE. LMP1 stain was not done in one case from the United States and is not reported for 11 cases from Hong Kong where alternative tissue fixation may have influenced the outcome. Abbreviations: EBER1, Epstein-Barr virus-encoded ribonucleic acid 1; LMP1, EBV latent membrane protein 1.
characterized by sheets of anaplastic epithelial cells with prominent nucleoli and distinct cytoplasmic borders accompanied by a mild lymphoid infiltration. The remaining three NPC cases were keratinizing squamous cell carcinomas of the nasopharynx characterized by sheets of anaplastic epithelial cells with dense pink cytoplasm but lacking keratin pearl formation. All three of these keratinizing carcinomas were EBER1 negative. The trend toward absence of EBV in the more differentiated cases is borne out by a statistically significant difference in the rate of EBER1 positivity in the undifferentiated NPCs compared with the nonkeratinizing and keratinizing squamous carcinomas (P = .015). Regarding geographic location, there was a significant difference between the American cases and those from Hong Kong with respect to histological spectrum and EBER1 association. Although the American cases were 77% undifferentiated and 62% EBV-associated, the H o n g Kong cases were 100% undifferentiated and 91% EBV-associated. Because some investigators have suggested that Asian NPC is virtually always EBV associated, further examination of the single EBER1negative carcinoma from H o n g Kong was undertaken. The patient was of Chinese descent, and morphological examination of the nasopharyngeal biopsy tissue showed a typical Regaud's pattern of undifferentiated NPC. The EBER1 stain was repeatedly negative, and
1209
HUMAN PATHOLOGY
Volume 26, No. 11 (November 1995)
the control hybridization indicated that RNA was well preserved (Fig 1C). The lack of EBV in this tumor was supported by serological studies that revealed immunoglobulin A anti-viral capsid antigen titer of less than 1:5. Clinically, the t u m o r involved the sphenoid and ethmoid sinuses as well as the nasal cavity, and, therefore, is likely to have developed either in the sinuses or in the nasopharynx. Neither clinical n o r morphological examination could distinguish the site of origin, but if it developed in the sinuses it should be classified as a sinonasal undifferentiated carcinoma rather than NPC. EBV association alone cannot be used to distinguish these entities because NPC-like sinonasal carcinomas were reportedly EBERl-positive in four of five cases, according to Lopategui et al. 35 EBER1 was not detected in any of 11 cases of lymphoepithelioma-like carcinoma o f the bladder (95% confidence interval = 0% to 27%). T h e clinical and morphological features o f nine of these bladder carcinomas were previously reported. 27 All cases exhibited syncytial clusters of undifferentiated epithelial cells surr o u n d e d by small lymphocytes reminiscent o f the Regaud pattern o f undifferentiated NPC. Unlike NPC, however, the lymphoepithelioma-like carcinomas of the bladder exhibited no association with latent EBV. I m m u n o h i s t o c h e m i c a l D e t e c t i o n of EBV LMP1 in NPC
Paraffin immunohistochemistry was used to detect and localize EBV LMP1 oncoprotein in 44 cases o f NPC. As expected, n o n e o f the 13 EBERl-negative NPCs expressed LMP1. O f the 32 EBERl-positive NPCs, only seven expressed LMP1 protein. LMP1 was noticeably undetectable in the NPCs from H o n g Kong. T h e absence of LMP1 in these tissues may represent a technical artifact resulting from the m a n n e r in which these tissues were fixed (in 95% ethanol) before histological processing. Ethanol fixation yielded excellent preservation o f tissue RNA as j u d g e d by U6 hybridizations, but the effect on antigen preservation is not known. Therefore, the LMP1 results of the H o n g Kong cases are excluded from the summary data displayed in Table 1. In the seven cases o f NPC from the United States that expressed LMP1, the immunohistochemical signal was weak to moderate in intensity and was localized to the cytoplasm and m e m b r a n e o f the malignant epithelial cells. In two instances (one Schminke and one nonkeratinizing squamous carcinoma) the LMP1 was diffusely positive a m o n g all tumor cells, but in the remaining five instances the LMP1 was restricted to rare scattered malignant cells (Fig 2). The sparsity of the LMP1 signal was unlike that of 28 lymphoid control tissues (nine cases of infectious mononucleosis 22 and 19 cases o f Hodgkin's disease with EBERl-positive ReedSternberg cells) where LMP1 was strongly expressed in the cytoplasm and m e m b r a n e of most infected cells. With the exception of 11 cases from H o n g Kong, the NPC samples were handled similarly to the lymphoid control tissues in terms o f m e t h o d of fixation, tissue
FIGURE 2. Immunohistochemical staining for LMP1 in paraffin sections shows focal weak expression in rare scattered malignant epithelial cells of NPC. (Case no. 373, original magnification x600.)
processing, and staining. Furthermore, any bias toward p o o r fixation in the NPCs c o m p a r e d with the lymphoid samples seems unlikely because RNA was well preserved in both sample types. Therefore, it seems that the observed differences in staining pattern and intensity are likely to result from biological differences in LMP1 expression. These findings suggest that LMP1 is more strongly expressed in EBV-infected lymphocytes and in ReedSternberg cells than in the malignant epithelial cells of NPC. The C15 NPC specimen (a human-derived carcinoma propagated in T-cell-deficient mice) also expressed LMP1 in a weak and focal pattern, suggesting that host T-cell surveillance is not the only factor limiting high LMP1 expression in the carcinoma cells. Nor was the degree of tumor differentiation a strong predictor of LMP1 expression; excluding the H o n g Kong cases, LMP1 was detected in five of 16 (31%) of undifferentiated NPCs, and two of five (40%) of the slightly more differentiated nonkeratinizing squamous cell NPCs. The proportional numbers of t u m o r cells staining for LMP1 did not seem to depend on tumor morphology, and the cytological features of the LMPl-positive rumor cells were generally indistinguishable from their LMPl-negative cohorts. EBV G e n o m i c Clonality b y Southern Blot Analysis
In two cases of NPC diagnosed in Texas, frozen tissue was available for Southern blot analysis of tumor clonality with respect to the structure of EBV DNA. O n e case (no. 601) was an undifferentiated NPC of the Schmincke pattern, and the other (no. 519) was a nonkeratinizing squamous cell carcinoma. Both tumors expressed EBER1 transcripts and focal LMP1 protein. On Southern blots hybridized with X h o l a probe to the right EBV terminus, each t u m o r exhibited a single band (Fig 3). T h e presence of a single band signifies that the EBV DNA is monoclonal, and, therefore, that the
1210
EPSTEIN-BARR VIRUS IN LYMPHOEPITHELIOMAS (Gulley et al)
Raji BL 10
1
NPCs 519
601
--- 2 3 k b
--
9.4
FIGURE 3. Southern blot analysis of BamHl-digested DNA hybridized to the EBV X h o l a probe shows a single high molecular weight b a n d in e a c h NPC specimen (nos. 519 and 601 ) indicating the monoclonal structure of the associated EBV DNA. Each NPC specimen contains approximately 10 copies of EBV DNA per ceil as estimated by comparing b a n d intensity with dilutions of Raji Burkitt lymphoma (BL) cell line DNA (shown at 10 and one c o p y per cell).
infected cells in the sample are also monoclonal. 34 The same single bands were present when the samples were hybridized to a p r o b e (EcoRlI) from the left terminus of the viral genome, suggesting that each t u m o r contained monoclonal, episomal EBV DNA as has been rep o r t e d 34'36previously in o t h e r cases of NPC. T h e r e were an average of 10 copies of the EBV g e n o m e per cell as j u d g e d by relative band intensity in comparison with dilutions o f DNA from the Raji Burkitt lymphoma cell line. No low molecular weight bands were seen, consistent with the absence o f a substantial a m o u n t of infectious viral replication in these tumors. DISCUSSION This is the first r e p o r t to show that EBV is absent in lymphoepithelioma-like carcinomas of the bladder (95% confidence interval -- 0% to 27%). In contrast, EBV EBER1 transcripts were detected in most NPCs. These data confirm that EBV is not a prerequisite for the development of lymphoepithelioma-like carcinomas. In concert with previously published data on lymphoepithelioma-like carcinomas developing in other anatomic sites, these findings suggest that EBV is most likely to be associated with carcinomas arising from foregut-derived organs. Perhaps the proximity of
these organs to sites of natural viral replication is not coincidental. T h e undifferentiated subtypes of NPC were more commonly (P = .015) EBV positive than were the keratinizing carcinomas of the nasopharynx, supporting previous work by others 4'~7-s9 using different molecular techniques. This study also addressed whether there was any difference in EBV positivity between the Regaud and Schmincke variants of undifferentiated NPC. EBER1 was f o u n d in 15 of 19 (79%) of cases with Regaud histology and in a nearly identical proportion 10 of 13 (77%) with Schmincke histology. These data suggest that there is n o virological basis for these alternative histological appearances. Furthermore, according to Batsakis, 28 there is no prognostic significance to the distinction of Regaud from Schmincke histologies in patients with NPC. Also pertinent to this topic was our observation that lymphoepithelioma-like carcinomas of the bladder exhibited Regaud's histology even though they were not EBV associated. These findings imply that there is no obvious histological marker for EBV-associated carcinoma. Regarding geographic location, NPCs from H o n g Kong were m o r e likely to be undifferentiated and EBV associated than were NPCs from the United States. These data support studies from Europe 4'4°42 and other parts of the United States 1 in showing that undifferentiated NPCs frequently but not always h a r b o r EBV. In contrast, undifferentiated NPCs from Taiwan, H o n g Kong, Southern China, and Malaysia virtually always harbor E B V . 12'43'44 Regardless of geographic location, those variants of NPC that are differentiated e n o u g h to produce visible keratin in histological sections are less likely to harbor EBV. 4'4°48 NPCs are 50 times more c o m m o n in H o n g Kong than they are in the United States or Europe, and this has been attributed to both genetic and environmental factors. 48 T h e variation in incidence and EBV positivity by geographic location is analogous to what has been described in Burkitt's lymphomas where endemic African cases are 99% EBV positive, whereas n o n e n d e m i c American cases are 50-fold less c o m m o n and are only 15% EBV associated. 49 Likewise, it is feasible that NPCs can develop from either EBV-dependent or EBV-indep e n d e n t pathways and that the EBV-negative tumors comprise a greater proportion o f cases in n o n e n d e m i c parts of the world. Alternatively, it is feasible that all o f the EBER1negative carcinomas were in the nasopharynx only because they e x t e n d e d to that location from other sites; therefore, they may not really represent NPC. The single EBERl-negative carcinoma from H o n g Kong is one such example in which it was impossible to distinguish by either clinical or histological criteria whether the tumor developed in the nasopharynx or in an adjacent sinus. Even sinus undifferentiated carcinomas are occasionally EBV associated, particularly if they develop in Asian patients or exhibit the distinct lymphoepithelioma-like histological features of NPC. s~ It is not known whether EBV relatedness is a m o r e important biological predictor o f clinical behavior than are anatomic site or
1211
HUMAN PATHOLOGY
Volume 26, No. 11 (November 1995)
histological appearance, but further study of this issue is warranted. Another novel aspect o f this study was the appraisal o f LMP1 expression in NPC c o m p a r e d with lymphoid lesions. These data showed that LMP1 is detectable by immunohistochemical staining of paraffinized tissue in a small fraction of t u m o r cells in only seven of 21 (33%) of infected NPC cases in the United States. It could be argued that this technique was insensitive and identified only those tumor cells producing the greatest a m o u n t of LMP1. But using the same technique, LMP1 was strongly and uniformly expressed in the infected cells o f lymphoid processes including nine cases of infectious mononucleosis 22 and 19 cases of EBV-associated Hodgkin's disease. No technical problem was identified that might explain the observed differences in LMP1 staining, so it is concluded that LMP1 is less strongly and less uniformly expressed in NPC cells than in EBV-infected lymphoid lesions. Because LMP1 is an immunogenic target for T-cell cytotoxicity, the downregulation of LMP1 in carcinoma cells might permit escape from i m m u n e destruction. O u r finding o f weak and focal LMP1 expression in the C15 NPC tumor, a human-derived carcinoma propagated in T-cell-deficient mice, suggests that host T-cell surveillance is not the only factor limiting high LMP1 expression in carcinoma cells. Instead, these data imply that malignant epithelial cells inherently downregulate LMP1 c o m p a r e d with Reed-Sternberg cells and infected lymphocytes of infectious mononucleosis. O t h e r investigators have reported LMP1 positivity in 0% to 66% of NPC cases d e p e n d i n g on the m e t h o d u s e d . 16-20'39'42'50-53 When methods were directly compared, paraffin immunohistochemistry was less sensitive than frozen-section immunohistochemistry or Western blotting of the same sample, s° T h e most sensitive m e t h o d of LMP1 detection was probably polymerase chain reaction (rtPCR), which showed LMP1 expression i n only three o f 18 cases of EBV-positive NPC. 5° After a second r o u n d of amplification using nested primers, detection improved to 15 of 18 cases, suggesting that low levels of LMP1 are p r o d u c e d in most infected cases. These data combined with immunohistochemical localization studies by ourselves and o t h e r s 4'19 suggest that LMP1 expression often varies from cell to cell with only rare tumor cells expressing high e n o u g h levels to detect on immunostains of routinely p r e p a r e d paraffin e m b e d d e d tissues. In vitro studies have shown that LMP1 expression is influenced by epithelial cell differentiation, but this study is the first in vivo assessment of this issue. In the in vitro model, LMP1 was not expressed in epithelial cell lines infected with EBV except following induction of virus replication by differentiating agents. TM Any effect o f differentiation on viral gene expression has therapeutic implications in that differentiating agents might permit immune-mediated recognition and destruction of cells expressing the highly immunogenic viral replicative antigens. This study could not measure whether the individual NPC cells expressing LMP1 were more differentiated than their nonexpressing cohorts. How-
ever, when whole tumors were c o m p a r e d with respect to overall degree o f tumor differentiation, there was no strong correlation between LMP1 expression and degree of differentiation, albeit a small sample size and a limited spectrum of differentiation. Specifically, LMP1 expression was in five o f 16 (31%) undifferentiated EBERl-positive NPCs and two of five (40%) of five slightly more differentiated nonkeratinizing squamous cell NPCs. In terms of diagnostic pathology, an important finding in this study was that routine paraffin immunostains for LMP1 are an insensitive means of detecting carcinoma-associated virus. Neither is any other immunohistochemical marker sensitive for detection of latent EBV in paraffin e m b e d d e d tissues. 26 Instead, EBER in situ hybridization seems to be the more sensitive test for detecting tumor-associated virus. Compared with Southern analysis, the chief advantage of EBER in situ hybridization is the rapidity o f obtaining results and the suitability of paraffin-embedded tissue. Most other techniques, such as culture or PCR, are unable to confirm the tumor-associated nature of a positive signal. Therefore, it seems that the current preferred m e t h o d for detection of EBV in NPC tissues is EBER in situ hybridization. In two EBV-positive NPC specimens examined by the Southern blot technique, the virus existed in a latent episomal genomic configuration without detectable viral replication. Both of the tumors were monoclonal with respect to the structure of the viral DNA in keeping with their malignant nature and consistent with the findings of previous investigators. 34'36T h e presence of monoclonal EBV DNA implies that the virus was present before malignant transformation and, therefore, that the virus may have played a role in tumor development. T h e strong link between EBV and NPC has led to the suggestion that EBV testing can be used to confirm a diagnosis of NPC in metastatic lesions. 54 Indeed, an enlarged cervical lymph n o d e is the most frequent clinical presentation o f NPC. 55 However, the diagnostic use of EBV testing seems hazardous in light of potential for EBV-positive carcinomas of the stomach, lung, or salivary gland to metastasize to cervical nodes. Moreover, a significant fraction of NPCs do not contain EBV. Therefore, EBV positivity may be helpful in pinpointing possible primary sites, but a negative result is not helpful in excluding any sites.
Acknowledgment. The authors thank Dr John W. Williams, Jr, for statistical analysis, Drs Paula Rogers and John N. Eble for case retrieval, and Dr J. S. T. Sham for biopsies of the patients in Hong Kong. REFERENCES 1. Weiss LM, Mohaved LA, Butler AE, et at: Analysisof lymphoepithelioma and lymphoepithelioma-like carcinomas for EBV genomes by in situ hybridization.Am J Surg Pathol 13:625-631, 1989 2. Shibata D, Weiss LM: Epstein-Barrvirus-associatedgastric carcinoma. AmJ Pathol 140:769-774, 1992
1212
EPSTEIN-BARR VIRUS IN LYMPHOEPITHELIOMAS (Gulley et al) 3. Pittaluga S, Wong MP, Chung LP, et al: Clonal Epstein-Barr virus in lymphoepithelioma-like carcinoma of the lung. Am J Surg Pathol 17:678-682, 1993 4. Niedobitek G, Hansmann ML, Herbst H, et al: Epstein-Barr virus and carcinomas: Undifferentiated carcinomas but not squamous cells carcinomas of the nasopharynx are regularly associated with the virus. J Pathol 165:17-24, 1991 5. Hamilton-Dutoit SJ, Hamilton Therkildsen M, Nielsen NH, et al: Undifferentiated carcinoma of the salivary gland in Greenlandic Eskimo: Demonstration of Epstein-Barr virus DNA by in situ nucleic acid hybridization. HUM PATHOL 22:811-815, 1991 6. Leyvraz S, Henle W, Chahinian AP, et al: Association of Epstein-Barr virus with thymic carcinoma. N EnglJ Med 312:1296-1299, 1985 7. Wu TC, Kuo TT: Study of Epstein-Barr virus early RNA-1 (EBER1) expression by in situ hybridization in thymic epithelial tumors of Chinese patients in Taiwan. HuM PATHOL 24:235-238, 1993 8. Carr KA, Bulengo S, Weiss LM, et al: Lymphoepithelioma-like carcinoma of the skin: A case report with immunophenotypic analysis and in situ hybridization for Epstein-Barr viral genome. Am J Surg Pathol 16:909-913, 1992 9. Kumar S, Kumar D: Lymphoepithelioma-like carcinoma of the breast. Mod Pathol 7:129-131, 1994 10. Weinberg E, Hoisington S, Eastman AY, et al: Uterine cervical lymphoepithelial-like carcinoma: Absence of Epstein-Barr virus genomes. A m J Clin Pathol 99:195-199, 1993 11. Miller G: Epstein-Barr virus, biology, pathogenesis and medical aspects, in Fields BN, Knipe DM, et al (eds): Virology (ed 2). New York, NY, Raven, 1990, pp 1928-1933 12. Pearson GR: Epstein-Barr virus and nasopharyngeal carcinoma. J Cell Biochem S17F:150-154, 1993 (suppl) 13. Wang D, Liebowitz D, Kieff E: An EBV membrane protein expressed in immortalized lymphocytes transforms established rodent cells. Cell 43:831-840, 1985 14. Li QX, Young LS, Niedobitek G, et al: Epstein-Barr virus infection and replication in a human epithelial cell system. Nature 356:347-350, 1992 15. Wilson JB, Weinberg W, Johnson R, et al: Expression of the BNLF1 oncogene of Epstein-Barr virus in the skin of transgenic mice induces hyperplasia and aberrant expression of keratin 6. Cell 61:1315-1327, 1990 16. Fahreus R, Fu HL, Ernberg I, et al: Expression of EpsteinBarr virus-encoded proteins in nasopharyngeal carcinoma. Int J Cancer 42:329-338, 1988 17. Klein G: Viral latency and transformation: The strategy of Epstein-Barr virus. Cell 58:5-8, 1989 18. Young LS, Dawson CW, Clark D, et al: Epstein-Barr virus gene expression in nasopharyngeal carcinoma. J Gen Virol 69:10511065, 1988 19. Niedobitek G, Young LS, Sam CK, et al: Expression of Epstein-Barr virus genes and of lymphocyte activation molecules in undifferentiated nasopharyngeal carcinoma. Am J Pathol 140:879-887, 1992 20. Brousset P, Butet V, Chittal S, et al: Comparison of in situ hybridization using different nonisotopic probes for detection of Epstein-Barr virus in nasopharyngeal carcinoma and immunohistochemical correlation with anti-latent membrane protein. Lab Invest 67:45% 464, 1992 21. Isaacson PG, Schmid C, Pan L, et al: Epstein-Barr virus latent membrane protein expression by Hodgkin and Reed-Sternberg-like cells in acute infectious mononucleosis. J Pathol 167:26%271, 1992 22. Reynolds DJ, Banks PM, Gulley ML: New characterization of infectious mononucleosis and a phenotypic comparison with Hodgkin's disease. Am J Pathol 146:379-388, 1995 23. Deacon EM, PaUesen G, Niedobitek G, et al: Epstein-Barr virus and Hodgkin's disease: Transcriptional analysis of viral latency in the malignant cells. J Exp Med 177:339-349, 1993 24. GuUey ML, Raab-Traub N: Detection of Epstein-Barr virus in h u m a n tissues by molecular genetic techniques. Arch Pathol Lab Med 117:1115-1120, 1993 25. Wu TC, Mann RB, Epstein J, et al: Abundant expression of EBER1 small nuclear RNA in nasopharyngeal carcinoma: A morphologically distinctive target for detection of Epstein-Barr virus in forma-
lin-fixed paraffin-embedded carcinoma specimens. Am J Pathol 138:1461-1469, 1991 26. Ambinder RF, Mann RB: Epstein-Barr encoded RNA in situ hybridization. HUN PATHOL 25:602-605, 1994 27. Amin MB, RoJY, Lee K, et al: Lymphoepithelioma-like carcinoma of the bladder. A m J Surg Pathol 18:466-473, 1994 28. Batsakis JG: Carcinomas of the nasopharynx, in Tumors of the Head and Neck: Clinical and Pathological Considerations (2nd ed). Baltimore, MD, Williams & Wilkins, 1979, pp 190-193 29. Gulley ML, Eagan PA, Quintanilla-Martinez L, et al: EpsteinBarr virus DNA is abundant and monoclonal in the Reed-Sternberg cells of Hodgkin's disease: Association with mixed cellularity subtype and Hispanic American ethnicity. Blood 83:1595-1602, 1994 30. Andersen J, Zieve GW: Assembly and intracellular transport of snRNP particles. Bioessays 13:5%64, 1991 31. Rowe M, Evans H, Young L, et al: Monoclonal antibodies to the latent membrane protein of Epstein-Barr virus reveal heterogeneity of the protein and inducible expression on virus-transformed cells. J Gen Virol 68:1575-1586, 1987 32. EliasJM, Margiotta M, Gaborc O: Sensitivity and detection efficiency of the peroxidase antiperoxidase (PAP), avidin-biotin peroxidase complex (ABC) and peroxidase4abeled avidin-biotin (LAB) methods. Am J Clin Pathol 92:62-67, 1989 33. Busson P, Ganem G, Flores P, et al: Establishment and characterization of three transplantable EBV-containing nasopharyngeal carcinomas. I n t J Cancer 42:599-606, 1988 34. Raab-Traub N, Flynn K: The structure of the termini of the Epstein Barr virus as a marker of clonal cellular proliferation. Cell 47:883-889, 1986 35. Lopategui JR, Galley MJ, Frierson HF, et al: Detection of Epstein-Barr viral RNA in sinonasal undifferentiated carcinoma from Western and Asian patients. A m J Surg Pathol 18:391-398, 1994 36. Abdel-Hamid M, ChenJ-J, Constantine N, et al: EBV strain variation: Geographical distribution and relation to disease state. Virology 190:168-175, 1992 37. Chang YS, Tyan YS, Liu ST, et al: Detection of Epstein-Barr virus DNA sequences in nasopharyngeal carcinoma cells by enzymatic DNA amplification. J Clin Microbiol 28:2398-2402, 1990 38. Hording U, Nielsen HW, Daugaard S, et al: Human papillomavirus types 11 and 16 detected in nasopharyngeal carcinomas by the polymerase chain reaction. Laryngoscope 104:99-102, 1994 39. Lu Q-L, Ella G, Lucas S, et al: Bcl2 protooncogene expression in Epstein-Barr virus-associated nasopharyngeal carcinoma. IntJ Cancer 53:29-35, 1993 40. Della Torte G, Pilotti S, Donghi R, et al: Epstein-Barr virus genomes in undifferentiated squamous nasopharyngeal carcinomas in Italian patients. Diagn Mod Pathol 3:32-37, 1994 41. Hording U, Nielsen HW, Albeck H, et al: Nasopharyngeal carcinoma: Histopathological types and association with Epstein-Barr virus. E u r J Cancer B Oral Oncol 29B:137-139, 1993 42. Kouvidou CH, Kanavaros P, Papaioannou D, et al: Expression of bel-2 and p53 proteins in nasopharyngeal carcinoma: Absence of correlation with the presence of EBV encoded EBER-2 transcripts and latent membrane protein-1. J Clin Pathol: Molec Pathol 48:M1% M22, 1995 43. Kositanont U, Kondo K, Chongkolwatana C, et al: Detection of Epstein-Barr virus DNA and HHV-6 DNA in tissues biopsies from patients with nasopharyngeal carcinoma. Southeast Asian J Trop Med Public Health 24:455-460, 1993 44. Dickens P, Srivastava G, Loke SL, et al: Epstein-Barr virus DNA in nasopharyngeal carcinoma from Chinese patients in Hong Kong. J Clin Pathol 45:396-397, 1992 45. Yip TfC, Ngan RKC, Lau WH, et al: A possible prognostic role of immunoglobulin-G antibody against recombinant EpsteinBarr virus BZLF-1 transactivator protein ZEBRA in patients with nasopharyngeal carcinoma. Cancer 74:2414-2424, 1994 46. Niedobitek G, Agathanggelou A, Barber P, et al: p53 overexpression and Epstein-Barr virus infection in undifferentiated and squamous cell nasopharyngeal carcinomas.J Patho1170:457-461, 1993 47. Yeung WM, Zong YS, Chiu KH, et al: Epstein-Barr virus carNage by nasopharyngeal carcinoma in situ. I n t J Cancer 53:746-750, 1993 48. Hidesheim A, Levine PH: Etiology of nasopharyngeal carcinoma. Epidemiol Rev 15:466-485, 1993
1213
HUMAN PATHOLOGY
Volume 26, No. 11 (November 1995)
49. Shiramizu B, Barriga F, Neequaye J, et al: Patterns of chromosomal breakpoint locations in Burkitt's lymphoma: Relevance to geography and Epstein-Barr virus. Blood 77:1516-1526, 1991 50. Brooks L, Yao QY, Rickinson AB, et al: Epstein-Barr virus latent gene transcription in nasopharyngeal carcinoma cells: Coexpression of EBNA1, LMP1, and LMP2 transcripts. J Virol 66:26892697, 1992 51. Stewart JP, Arrand JR: Expression of the Epstein-Barr virus latent membrane protein in nasopharyngeal carcinoma biopsy specimens. HUM PA'rHOL24:239-242, 1993 52. Li-Fu H, Minarovits J, Shi-Long C, et al: Variable expression of latent membrane protein in nasopharyngeal carcinoma can be
related to methylation status of the Epstein-Barr virus BNLF-1 5'flanking region. J Virol 65:1558-1567, 1991 53. Zheng X, Hu L, Chen F, et al: Expression of Ki67 antigen, epidermal growth factor receptor, and Epstein-Barr virus-encoded latent membrane protein (LMP1) in nasopharyngeal carcinoma. Eur J Cancer 30B:290-295, 1994 54. Feinmesser R, Miyazaki I, Cheung R, et al: Diagnosis ofnasopharyngeal carcinoma by DNA amplification of tissue obtained by fine-needle aspiration. N EngJ Med 326:17-21, 1992 55. Chan MK, McGuire LJ, LeeJC: Fine needle aspiration cytodiagnosis of nasopharyngeal carcinoma in cervical lymph nodes: A study •of 40 cases. Acta Cytol 3:344-350, 1989
1214