The Epstein-Barr virus (EBV) is consistently associated with undifferentiated nasopharyngeal carcinoma (NPC). There is, however, conflicting evidence as to whether squamous cell NPCs are also EBV-associated. Moreover, it has been proposed that other epithelial tumours, particularly thymomas and thymic carcinomas, should be included in the group of EBV-associated neoplasias. However, since the viral DNA in these studies was demonstrated only in extracted DNA, the cellular origin of the viral DNA is uncertain. We have therefore investigated 152 epithelial tumours from various sites for the presence of EBV-DNA by in situ hybridization with 35S-labelled probes. Sixty-eight of 77 undifferentiated NPCs showed an EBV-specific autoradiographic signal, thus confirming the strong association of this tumour type with EBV even in geographical areas where undifferentiated NPC is not endemic. None of eight squamous cell NPCs showed an EBV-specific signal. All of 15 carcinomas with a similar morphology to undifferentiated NPC but from different anatomic sites (thymus, tonsil, breast) were EBV-negative as were 9 thymomas, 26 squamous cell carcinomas of the palatine tonsil, and 14 cervical carcinomas. Our results therefore suggest a unique association of EBV with undifferentiated NPC and support concepts assigning different biological properties to undifferentiated NPC as compared with squamous cell NPC.

Twenty eight tonsillar carcinomas of various histological types were investigated for the presence of Epstein-Barr virus (EBV), cytomegalovirus (CMV), and human papillomavirus (HPV) types 6, 11, and 16 by in situ hybridisation using highly stringent procedures. In six cases an autoradiographic signal was obtained in the tumour cell nuclei with the HPV type 16 specific probe. As all of these cases were negative with EBVspecific probes (G Niedobitek, unpublished observations) we extended our study to other DNA tumour viruses and report here the results obtained with HPV specific DNA probes. MethodsFormalin fixed, paraffin wax embedded biopsy specimens of 28 tonsillar carcinomas were studied. Sixteen (57%) patients were male and 12 (43%) were female. The age of the patients ranged from 43 to 85 years, with a median of 66 years (table). The women were older than the men, the median ages being 70 years and 59 years, respectively. Tobacco smoking or alcohol abuse, or both, were risk factors for most of the patients (table). The tumours were graded according to WHO criteria.'4 Two carcinomas were well differentiated, 12 moderately differentiated, and 11 poorly differentiated squamous cell carcinomas. One case was an in situ carcinoma and two were lymphoepithelioma-like carcinomas with morphological similarities with undif-

Gut associated lymphoid tissue in 15 normal appendices has been characterised in tissue sections using both morphological criteria and immunocytochemical techniques. A panel of monoclonal and polyclonal antibodies was used including antibodies to B-cells, T-cells, macrophages, HLA DR and immunoglobulins. The lymphoid tissue in the appendix was shown to bear a strong resemblance to that in lymph nodes with the exception of the region where the appendix follicles associate with the dome epithelium, which has no lymph node equivalent. This zone of cells between the lymphoid follicles and the dome epithelium termed the 'mixed cell zone' has been shown to contain an abundance of HLA DR-bearing cells, some of which have irregular nuclear morphology and resemble follicle centre cells. These cells were seen to extend into the epithelium of the dome but not the crypts. Using a monoclonal anti-B-cell antibody a population of B-cells was detected in the equivalent areas of mixed cell zone and epithelium and quantitative studies showed that these intraepithelial B-cells comprised approximately 4-5% of the cells in the epithelium. The mixed cell zone was also seen to contain T-cells, S-100 protein-containing macrophages and occasional lysozyme-containing macrophages. Plasma cells were rarely seen in this area.

surround the lymphoid follicles. They were most concentrated on the serosal aspect around the high endothelial venules. Cells with macrophage-like morphology were present in both the lamina propria and the dome region of the follicles; those in the lamina propria containing lysozyme and those in the dome region S100 protein. The results are discussed in relation to the generation and dissemination of antibody producing cells in human gut.

In this study, 32 cases of T-cell lymphoma of angioimmunoblastic lymphadenopathy type (AILD-TCL) were investigated for their association with Epstein-Barr virus (EBV). For this purpose, three different approaches were applied: polymerase chain reaction (PCR) for the presence of EBV-DNA, in situ hybridization (ISH) for EBV-encoded small nuclear RNAs (EBER), and immunohistology for EBV-encoded latent membrane protein (LMP). PCR and EBER-ISH produced almost identical results, showing that all but one case of AILD-TCL contained EBV genomes. Three distinctive patterns of EBV infection were observed after immunophenotypical characterization of EBER-positive cells: (1) in 26% of the cases, B and T cells were infected, the majority of which were B cells of immunoblastic morphology located in the remnants of lymphoid follicles; (2) in 42% of the cases, the vast majority of infected cells were neoplastic T cells diffusely distributed in the lymph nodes, but infected B cells were also present; and (3) in 32% of the cases, there were only a few infected small lymphoid cells. Detectable LMP was frequent in cases exhibiting patterns 1 and 2. These findings suggest that in AILD-TCL patients, B cells and especially T cells are highly susceptible to a persistent EBV infection, which often leads to a growth advantage of the infected cells. Thus EBV, in conjunction with genetic abnormalities and selective defects of the immune system, might be involved in the pathogenesis of AILD-TCL.