Light microscope studies of leukemic tissues do not indicate any pathological alterations of megakaryocytes. These giant cells appear normal in structure and in frequency and apparently do not participate in the leukemic proliferation. Nevertheless, electron microscopy demonstrates the viruses in a suprisingly large number of megakaryocytes (fig. 5), the cells being frequently overloaded with viruses morphologically identical to those at the surface of the leukemic cells. The only particularity of the viruses observed is megakaryocytes is the frequent occurrence of dumbbell or chainlet profiles corresponding to the continuity of the outer virus membrane of several viruses.

This observation probably means that the virus may multiply in a cell that is not necessarily leukemic. In the megakaryocyte, the viruses are not found at the periphery of the cell, but within cytoplasmic vacuoles. It is believed, however, that these vacuoles are derived from the endoplasmic reticulum, and that consequently their membranes could be related to the cell membranes. These vacuoles are therefore equivalent to extracellular spaces, and the particular case of megakaryocytes is not in contradiction to the statement that Swiss mouse leukemia virus is always found outside the cell or at the cell surface.

The hypothesis that the localization of the viruses within megakaryocytes might be the result of the phagocytic activity of these cells does not hold for two reasons: (1) Classical light microscopy studies have demonstrated the absence of phagocytic activity in megakaryocytes (8), and (2) incomplete viruses were observed budding from the membranes limiting the cytoplasmic vacuoles of the megakaryocyte, which showed once more how closely the cell membrane and the endoplasmic reticulum membranes are related.

Problem Of Intracytoplasmic Precursors

Despite the fact that the viruses are budding at the surface of leukemic cells, it has not thus far been possible to identify positively any earlier step in the virus cycle. Before the tiny bulge at the cell membranes is demonstrable, there must be another part of the virus cycle not observable with our present morphologic techniques. In leukemic cells, the Golgi zone is frequently large and presents numerous vacuoles, but no relationship between these Golgi structures and the earlier steps of virus development has been established. In two cases, however, a large vesicle beneath a budding virus was seen. Such a rare observation is, of course, of questionable value, but it is noteworthy that a similar phenomenon was seen by Benedetti and Bernhard (9) in their study of chicken leukosis.

Preliminary And Negative Observation On Ribonu-Clease (Rnase) Digestion

Lead hydroxide staining does not seem to have a cytochemical specificity but, as already mentioned, it has a maximum affinity for the inner virus shell comparable in intensity to that for RNP granules. In an attempt to demonstrate whether or not ribonucleic acid is present in the virus, and with the observations of Epstein (10) in mind,

RNase digestion was used on material fixed with permanganate and embedded in Araldite. So far the results have been entirely negative. Although leukemic specimens embedded in Araldite are useful for low-magnification studies, it is known that the recognition of RNP granules in such permanganate-fixed material is difficult. Unfortunately, this was also true for the virus. It was therefore necessary to devise another technique: After the specimen had been fixed in buffered formalin, it was subjected to RNase digestion. It was then fixed a second time in osmium tetroxide before being embedded in methacrylate.

After this procedure the RNP granules seemed to disintegrate, even though the fixation was poor. The viruses, however, appeared resistant to RNase digestion under these experimental conditions and retained a normal structure. This may be related to the nature of the outer virus shell and is being investigated further.