Laboratory of Tumor Cell Biology National Cancer Institute Bethesda, Maryland 20014

I. Introduction

Type-C RNA viruses have been isolated from many species. In several, they have been often associated with leukemia and shown to reproduce the disease on inoculation into recipient animals. In a few species the data appear now to be conclusive that they are the major cause of the natural disease. Two major difficulties in verifying results that the virus causes the disease in some animal systems have been: (1) the long latent period for evident disease, and (2) the fact that many type-C viruses are apparently not oncogenic. Regarding the latter, we have argued for a major subdivision of these viruses based on a molecular hybridization assay (see below).

Conclusions

We believe human myelogenous leukemia blood cells do not frequently permit complete expression of type-C viral information, but this information is at least partially present in many and perhaps all AML patients. This is in contrast to the case of some animals like cats where most animals with leukemia actively produce virus. On the other hand, even with cats there is variation. The occasional (or rare) infected animal does not completely express virus (M. Essex, personal communication). Conversely, patient A. S. (HL-23) may be the unusual or rare human, who after appropriate growth stimulation of her leukemic cells, expresses completely and releases whole virus. One difficulty with our interpretation is our inability to detect the complete provirus.' This results in a paradox revolving around the question how do human leukemic cells become transformed and how can they release virus if they lack the complete genetic information apparently essential in animal model systems for transformation and virus production? We think that generally the integrated complete provirus may be in only a small number of cells, perhaps not even the leukemic cell precursors. Release of fragments of the provirus by the infected cells may be sufficient, in some instances, to transform leukocyte precursors. This model is compatible with the existing data on human leukemia, including the detection of extra sequences in human leukemic DNA by Spiegelman and associates (41). At least one tissue or cell population should contain cells with complete provirus. Portions of this provirus may integrate into leukocyte progenitors, a necessary prelude to leukemic transformation. On occasion complete provirus may integrate into some leukocyte precursors, the necessary event for the rare complete virus production. We suggest that the site of integration for fragments or whole provirus is the "hot spot" region discussed before at this workshop ( 4) and that this may alter gene expression by a mechanism called "paraprocessing" (1) which in turn leads to transformation. If this speculation is correct, detection of the complete provirus as the proof for the involvement of these viruses in man will be extremely difficult.'* Other approaches will be necessary such as additional virus isolates from other laboratories and/or a clear seroepidemiological studies.