The DNA double strand break repair gene XRCC4, an important caretaker of genome stability, is suggested to play a role in the development of human carcinogenesis. However, no evidence has been provided showing that XRCC4 was associated with oral oncology. In this hospital-based case-control study, the association of XRCC4 G-1394T (rs6869366), intron 3 (rs28360071), intron 7 (rs28360317), and intron 7 (rs1805377) polymorphisms with oral cancer risk in a Taiwanese population was first investigated. In total, 318 patients with oral cancer and 318 age- and gender-matched healthy controls were genotyped. We found a significant different distribution in the frequency of the XRCC4 intron 3 genotype, but not the XRCC4 G-1394T or intron 7 genotypes, between the oral cancer and control groups. Those who had heterozygous del/ins at XRCC4 intron 3 showed a 1.57-fold (95% confidence interval=1.12-2.21) increased risk of oral cancer compared to those with ins/ins. As for XRCC4 G-1394T or intron 7 polymorphisms, there was no difference in the distribution between the oral cancer and control groups. There were significant gene-environment interactions between XRCC4 intron 3 genotype with smoking and with betel quid chewing, but not with alcoholism. In smoker and betel quid chewer groups, the XRCC4 intron 3 del variants exhibited 2.57- and 3.03-fold higher risks than the ins genotype, respectively. Our results firstly suggest that the XRCC4 intron 3 del genotype may be associated with oral oncology and may be a novel useful marker for primary prevention and anticancer intervention.

We studied the distribution of the EBV genome in tumour biopsies obtained from 42 patients with poorly differentiated or undifferentiated nasopharyngeal carcinoma (NPC) and 3 patients with well-differentiated NPC. Six carcinoma in situ (CIS) foci were seen in 5 tumour specimens. By in-situ hybridization, multiple copies of the EBV genome were detected in some of the tumour cells in 3 CIS lesions involving the full thickness of the mucosal epithelium, but without microinvasion, while the viral genome was present in the majority of the tumour cells contained in another 3 CIS lesions with microinvasion. In agreement with previous findings, poorly differentiated and undifferentiated carcinomas regularly carried the viral genome, the number of copies of which was similar to that seen in CIS, while some, but not all, of the tumour cells of the well-differentiated histological type carried the virus. The viral genome was otherwise rarely detected in other areas of the mucosal epithelium and, where present, the viral carriage was confined to a few epithelial cells, in which the viral genome contents were markedly lower than in tumour cells. These results suggest that EBV may first become associated with NPC at an early stage of the disease shortly after the tumour has been initiated.

The relationship of Epstein-Barr virus (EBV), type I human T-cell lymphotropic virus (HTLV-I), and parvovirus B19 to histiocytic necrotizing lymphadenitis was studied prospectively in 10 Taiwanese patients using materials obtained by fine-needle aspiration and lymph node biopsy. The presence of EBV was detected by in situ hybridization for EBV-encoded RNA expression. Immunocytochemistry was used to detect virus-encoded protein for EBV and parvovirus B19. DNA in situ hybridization and polymerase chain reaction were performed to determine the existence of HTLV-I provirus. Expressions of EBV-encoded RNA and Fas ligand were detected in all cases. Expression of EBV-encoded protein was identified in only 1 case. Neither HTLV-I nor parvovirus B19 was detected in any case.