The focus of our research is devoted to understanding the molecular basis for the pathogenesis of retrovirus-induced diseases. We have been studying retroviruses that cause leukemia or neurological disease in rodents to obtain basic information on how molecular changes in normal cells can result in pathological consequences. We hope to use the information gained from these studies to design rational strategies to counteract the molecular events that are responsible for the diseases induced and apply this to similar diseases in man. Erythroleukemia induced in mice by the Friend spleen focus-forming virus (SFFV) provides an important model for understanding how deregulation of hematopoietic pathways can lead to leukemia. Although normal erythroid cells require erythropoietin (Epo) for proliferation and differentiation, those expressing the SFFV envelope protein, gp55, proliferate and differentiate in the absence of Epo, resulting in acute erythroid hyperplasia and polycythemia. In addition, SFFV integration at the sfpi-1 locus results in expression of non-physiological levels of the myeloid transcription factor PU.1 in erythroid cells, causing a block in differentiation and the outgrowth of transformed leukemic cells. To understand how expression of the SFFV envelope protein in erythroid cells alters their growth and differentiation and how expression of PU.1 leads to their transformation, we have been studying signal transduction pathways known to be activated by Epo to determine if any of these are deregulated after SFFV infection. Our studies indicate that in non-transformed erythroid cells derived from the first stage of SFFV-induced erythroleukemia, most Epo-induced signal transducers, including Stat proteins, components of the Raf-1/MAP kinase pathway, PI 3-kinase, protein kinase C and Akt kinase, are constitutively activated. Our recent studies indicate that the Epo-independent activation of these pathways by the virus involves a truncated form of the receptor tyrosine kinase Stk, which becomes activated after forming a covalent complex with SFFV gp55 and associates with multiple tyrosine-phosphorylated signal transducing molecules. When immortal erythroleukemia cell lines derived from SFFV-infected mice were examined for activation of the same signal transduction pathways, we observed a block in the activation by Epo or SFFV of the DNA binding activity of the transcription factors Stats 1 and 3. The block is at the level of DNA binding, not tyrosine phosphorylation or nuclear transport, and is specific for Stats activated by Epo, not other inducers such as interferon. There is a direct correlation in SFFV-infected erythroid cells between expression of PU.1 and inhibition of Stat DNA binding. PU.1 does not interfere with Stat DNA binding by forming a stable complex with the Stat proteins but rather competes with Stat proteins for binding to the promoters of Stat-responsive genes. Our studies suggest that initial infection of erythroid cells with SFFV results in the constitutive activation of signals needed for both proliferation and differentiation, and that these cells become transformed only when differentiation signals activated by Stat proteins are blocked due to inappropriate expression of the PU.1 protein. As a second retroviral model system, we have been studying PVC-211 murine leukemia virus (MuLV), a variant of the leukemia-inducing Friend MuLV that causes a rapid neurodegenerative disease in rodents. PVC-211 MuLV provides an important model for understanding how retroviruses can undergo genetic changes that alter their interaction with cells in the host to cause novel biological effects. We previously demonstrated that PVC-211 MuLV has undergone two amino acid changes in the receptor binding domain of its envelope protein which allow it to efficiently infect brain capillary endothelial cells (BCEC), which are generally resistant to infection by MuLVs. This expanded host range allows PVC-211 MuLV to be expressed at high levels in the neonatal rodent brain, resulting in a rapid neurological disease. Our recent studies have concentrated on understanding how changes in the viral envelope glycoprotein contribute to BCEC tropism and how expression of the virus in BCEC, which are the only cells in the central nervous system significantly infected, results in neurological damage. We observed that one of the two changes in the receptor binding domain of PVC-211 MuLV associated with acquisition of BCEC tropism created a unique heparin-binding domain, and we were able to show that this change enables PVC-211 MuLV, unlike other MuLVs, to strongly bind to heparin. Thus, PVC-211 MuLV may be able to efficiently bind to the negatively charged surface of BCEC via heparin-like molecules, increasing the probability that the virus will interact with its cell surface receptor and infect these cells. We have also been examining the molecular events that occur in PVC-211 MuLV-infected BCEC. Previous results suggested that nitric oxide (NO), a potential neurotoxin, may be involved in PVC-211 MuLV-induced neurodegeneration and our recent studies show that one of the enzymes associated with NO generation, inducible nitric oxide synthase (iNOS), is activated in BCEC derived from PVC-211 MuLV-infected rats. Furthermore, elevated levels of a 32 kDa cellular protein modified by 3-nitrotyrosine, a hallmark of NO production, were observed in PVC-211 MuLV-infected BCEC. These changes could not be detected in BCEC infected with a non-neuropathogenic version of the virus, suggesting that neurological damage induced by PVC-211 MuLV infection may result from NO produced by infected BCEC.
