Two facts need to be pointed out to help explain why the history of Epstein-Barr virus (EBV) research has been inseparable from that of the studies with human hematopoietic cell lines of neoplastic and non-neoplastic origin. One is that Burkitt lymphoma (BL) cell lines, EBV-positive or-negative, can be established in culture quite easily. Thus, the BL cell lines which Epstein established were indeed some of the first hematopoietic as well as virus-carrying cell lines of human neoplastic origin. The other is that EBV-positive B-cell lymphoblastoid cell lines (B-LCL) of normal origin can be grown from samples of sero-positive individuals. B-LCL were often mistakenly regarded as being of neoplastic origin, but are almost always of normal cell origin. Very rarely, however, B-LCL with the same clonal markers as those of neoplastic cells have also been obtained. While the development of B-LCL has been referred to as the in vitro viral immortalization of human B cells and as a phenomenon representing the potential oncogenicity of EBV, the phenotypic and genotypic differences between B-LCL and EBV-carrying BL cells are obvious, indicating that the development of B-LCL per se does not prove the oncogenic activity of EBV. Two EBV-derived antigens, EBNA2 and latent-infection membrane protein (LMP), which are strongly expressed by B-LCL but not by BL cells, have recently been detected in EBV-positive proliferative B cells in patients with organ transplants, suggesting that the proliferating of B-LCL-like cells may take place as an initial step of the multi-step in vivo oncogenesis of EBV.(ABSTRACT TRUNCATED AT 250 WORDS)