Tumorigenic poxviruses: growth factors in a viral context?

Shope fibroma virus (SFV) is one of the few poxviruses that induce cutaneous tumours, whereas myxoma virus, a closely related leporipoxvirus, does not. However, both have a virally encoded homologue of the epidermal growth factor (namely SFGF and MGF, respectively) that is considered to be crucial for poxvirus tumorigenesis. In this study, the role of viral growth factors in the context of infection with SFV, a tumorigenic leporipoxvirus, was investigated. An SFV mutant was engineered with the sfgf gene deleted and replaced with mgf. Macroscopic, histological and cytological examinations led to the conclusion that growth factors are indeed important for the development and maintenance of fibromas, provided that they are expressed in the proper viral context. However, they are not exchangeable and MGF cannot substitute for SFGF in the genesis of fibromas. It is likely that factors other than viral epidermal growth factor homologues influence the development of tumours.

Comparative effects of virulent and avirulent poxviruses on cell cycle progression.

We studied the impact of tumorigenic poxviral infection on key regulators of cell cycle progression. Malignant fibroma virus (MV) is a virulent poxvirus that causes severe immunological impairment in vivo and in vitro. It also directs expression of important cellular regulatory proteins, such as p53. Its avirulent relative, Shope fibroma virus (SFV), has little effect on the immune system or p53. Accordingly we examined the effects of MV and SFV on the cell cycle in RK-13 rabbit kidney fibroblasts. MV caused an accumulation of cells in G2/M phase and decreased the percentage of cells in G0/G1. Prolongation of G2/M phase was associated with increased levels of cyclin B protein, decreases in cyclin A and cdc2 proteins, and diminished cdc2 activity. In contrast SFV did not affect cellular cycling detectably. SFV infection was accompanied by large increases in cyclin A and cdc2 proteins and increased cdc2 activity. Thus alterations in cell cycle transit during virus infection may reflect active direction in which virus induces changes in cell cycle regulators. Such changes may be important in the differences in virulence between MV and SFV.

Shope fibroma virus RING finger protein N1R binds DNA and inhibits apoptosis.

Shope fibroma virus (SFV) N1R gene encodes a RING finger protein that localizes to virus factories within the cytoplasm of infected cells. Altered proteins, with deletions and site-specific mutations, were transiently expressed in vaccinia virus-infected cells to discern regions of the protein that are required for localization. We have determined that at least part of the RING finger region is necessary for localization but that the RING motif alone is not sufficient. A chimeric protein, however, in which the RING finger region of the herpes simplex virus-1 ICP0 protein replaces the SFV N1R RING motif does localize to virus factories. A region of five highly conserved amino acids at the amino terminus of SFV N1R is also critical for localization. We report that the SFV N1R protein binds double- and single-stranded DNA, suggesting a mechanism for localization, and that overexpression of this protein in vaccinia virus-infected cells reduces apoptosis-associated fragmentation of nuclear DNA.

Isolation and characterization of a Chinese hamster ovary mutant cell line with altered sensitivity to vaccinia virus killing.

The Chinese hamster ovary (CHO) cell line is nonpermissive for vaccinia virus, and translation of viral intermediate genes was reported to be blocked (A. Ramsey-Ewing and B. Moss, Virology 206:984-993, 1995). However, cells are readily killed by vaccinia virus. A vaccinia virus-resistant CHO mutant, VV5-4, was isolated by retroviral insertional mutagenesis. Parental CHO cells, upon infection with vaccinia virus, die within 2 to 3 days, whereas VV5-4 cells preferentially survive this cytotoxic effect. The survival phenotype of VV5-4 is partial and in inverse correlation with the multiplicity of infection used. In addition, viral infection fails to shut off host protein synthesis in VV5-4. VV5-4 was used to study the relationship of progression of the virus life cycle and cell fate. We found that in parental CHO cells, vaccinia virus proceeds through expression of viral early genes, uncoating, viral DNA replication, and expression of intermediate and late promoters. In contrast, we detect only expression of early genes and uncoating in VV5-4 cells, whereas viral DNA replication appears to be blocked. Consistent with the cascade regulation model of viral gene expression, we detect little intermediate- and late-gene expression in VV5-4 cells. Since vaccinia virus is known to be cytolytic, isolation of this mutant therefore demonstrates a new mode of the cellular microenvironment that affects progression of the virus life cycle, resulting in a different cell fate. This process appears to be mediated by a general mechanism, since VV5-4 is also resistant to Shope fibroma virus and myxoma virus killing. On the other hand, VV5-4 remains sensitive to cowpox virus killing. To examine the mechanism of VV5-4 survival, we investigated whether apoptosis is involved. DNA laddering and staining of apoptotic nuclei with Hoechst 33258 were observed in both CHO and VV5-4 cells infected with vaccinia virus. We concluded that the cellular pathway, which blocks viral DNA replication and allows VV5-4 to survive, is independent of apoptosis. This mutant also provides evidence that an inductive signal for apoptosis upon vaccinia virus infection occurs prior to viral DNA replication.

Expression of the myxoma virus tumor necrosis factor receptor homologue and M11L genes is required to prevent virus-induced apoptosis in infected rabbit T lymphocytes.

Myxoma virus is a leporipoxvirus that causes a highly lethal virulent disease known as myxomatosis in the European rabbit. An important aspect of myxoma virus pathogenesis is the ability of the virus to productively infect lymphocytes and spread to secondary sites via lymphatic channels. We investigated the infection of the CD4+ T lymphoma cell line RL-5 with myxoma virus and Shope fibroma virus, a related but benign leporipoxvirus, and observed that myxoma virus, but not Shope fibroma virus, was able to productively infect RL-5 cells. We also discovered that infection of RL-5 cells with Shope fibroma virus or attenuated myxoma virus mutants containing disruptions in either the T2 or the M11L gene resulted in the rapid induction of DNA fragmentation, followed by morphological changes and loss in cell integrity characteristic of cell death by apoptosis. Purified exogenous T2 protein was unable to prevent apoptosis, suggesting that T2 functions intracellularly. Thus, myxoma virus T2, originally described as a secreted homologue of the tumor necrosis factor receptor, and M11L, a novel transmembrane species with no known cellular homologue, function to extend virus host range for replication in rabbit T lymphocytes through the inhibition of apoptosis in infected T lymphocytes.

Syphilis superinfection activates expression of human immunodeficiency virus I in latently infected rabbits.

Superinfection of latently human immunodeficiency virus (HIV)-infected rabbits with either Treponema pallidum or Shope fibroma virus (SFV) activates HIV expression. In addition, HIV-infected rabbits demonstrate prolonged cutaneous lesions (chancres) after intracutaneous challenge with T. pallidum, the causative agent of syphilis. Rabbits were infected by intravenous inoculation of 3 x 10(7) human T-cell lymphotrophic virus type III (HTLV-III)/B10 (HIV-1)-infected H9 (human) cells. Five weeks after initial infection, integrated HIV-1-specific DNA sequences were detected in the DNA of the peripheral blood lymphocytes of only one of eight rabbits using polymerase chain reactions (PCR); human DNA could not be detected at this time. Furthermore HIV infection could not be demonstrated by either seroconversion or PCR during the next 6 months. All HIV-infected rabbits remained clinically healthy and had normal white blood cell counts. Six months after HIV infection, four HIV-infected and two noninfected controls were superinfected with 10(6) T. pallidum in eight skin sites in the shaved skin of the back, and four infected and two control animals were challenged with an intradermal injection with SFV. After infection with either syphilis or SFV, the DNA from the white blood cells of all eight HIV-infected rabbits contained HIV sequences, and HIV sequences were demonstrated in dermal mononuclear cells of the syphilitic lesions by in situ hybridization. The SFV-induced tumors were rejected normally in the HIV-infected rabbits, but four of the four rabbits challenged with T. pallidum had delayed development of cutaneous lesions and three of four demonstrated larger and more prolonged lesions. White blood counts, mitogen responses, and interleukin-2 production remained within normal limits, and seroconversion for HIV was not detected. Three of four rabbits in a second group, challenged with T. pallidum 4 months after HIV-inoculation, also had delayed healing of syphilitic lesions. These results indicate that latent HIV-infection of rabbits may be activated by immunostimulation and that latently HIV-infected rabbits have impaired delayed hypersensitivity reactions. It is hypothesized that true latent HIV-infection in the rabbits is in monocytes and postulated that further immunostimulation may produce infection of lymphocytes and activation of disease.

Immunosuppression during viral oncogenesis. IV. Generation of soluble virus-induced immunologic suppressor molecules.

We describe herein functional attributes and generation of immunologic suppressor activity elaborated in response to oncogenic virus infection. Malignant rabbit fibroma virus-induced immunologic suppressor factor (VISF) is a T cell product produced in peak quantities by spleen cells taken from infected rabbits 7 days after infection in vivo. Its production does not appear to require macrophage participation. VISF is highly labile, 3.5 to 12 kDa, and capable of suppressing both B and T lymphocytic responses. Indomethacin and the cyclic nucleotides cAMP and cGMP inhibit its generation. VISF activity is neither antigen nor species specific. It suppresses murine and leporine immune responses to antigens unrelated to the inducing virus. Comparable suppressor activity may be induced by infecting an apparently non-functional rabbit T lymphoma line, RL-5, with malignant rabbit fibroma virus. VISF is principally a suppressor-inducer factor: in vitro, lymphocytes exposed to VISF do not show decreased immunologic responsiveness until 4 days of culture. VISF induces T suppressor cell activity when normal spleen cells are exposed briefly to VISF. Thus, immunosuppressive consequences of malignant fibroma virus infection are partially mediated by a small, non-specific T cell-derived suppressor lymphokine with unique functional characteristics. Non-specific immunologic dysfunction that often attends virus infections may reflect the activity of such factors in humans as well.

Lymphocyte-virus interactions. Identification of a restriction fragment permitting virus replication in lymphocytes.

We are interested in understanding the effects of virus infection on lymphocyte function. To approach this question, we used a unique system of two genetically related leporipoxviruses to generate recombinants. One of these viruses, malignant fibroma virus (MV), replicates in many different cell types, including lymphocytes. The other, Shope fibroma virus (SFV), replicates principally in fibroblasts, but cannot replicate in lymphocytes. Fibroblasts infected with SFV received restriction fragments from MV by transfection. Recombinant viruses were selected in vitro for their ability to replicate in lymphocytes. By these means we have identified one restriction fragment, the 10.8-kb BamHI "C" fragment, capable of transferring from MV to SFV the ability to replicate in lymphocytes. A family of recombinants bearing different sized inserts of this restriction fragment has been isolated and is being characterized. Lymphocytotropic recombinants bearing portions of this restriction fragment produce colony morphology in vitro intermediate between MV's plaques and SFV's foci. On the basis of their ability to grow in and suppress mitogen responsiveness of lymphocytes, these recombinants may be classified into four different groups. Group 1 viruses are the most immunosuppressive, whereas those of group 4 are least. These traits correlate with ability to replicate in lymphocytes. Genetic analysis of recombinants indicates that the most immunosuppressive recombinants do not necessarily contain the most fragment C DNA. Therefore, we have identified a restriction fragment, one or more of the genes of which are sufficient to allow an otherwise nonlymphocytotropic virus to replicate in lymphocytes. Additional genetic and immunologic analysis should permit us to determine the structure and function of the protein responsible for this effect.

High levels of genetic recombination among cotransfected plasmid DNAs in poxvirus-infected mammalian cells.

The frequency of recombination between transfected plasmid DNAs was measured by using cultured cells infected with a variety of poxviruses. Plasmid derivatives of pBR322 containing XhoI linker insertion mutations in the tetracycline gene were used to assess recombination frequencies in rabbit cells infected with the leporipoxviruses Shope fibroma virus and myxoma virus and the orthopoxvirus vaccinia virus. Recombination frequencies were calculated by Southern blotting, which detects novel plasmid restriction fragments generated by genetic recombination, and by a plasmid rescue procedure in which the reconstruction of an intact tetracycline gene in the transfected rabbit cell was monitored by transformation back into Escherichia coli. The highest recombination frequencies were measured in cells infected with Shope fibroma virus and myxoma virus, and a minimum recombination frequency of at least one recombination event per 7 kilobases was calculated within 24 h posttransfection under these conditions. The deduced recombination frequency in vaccinia virus-infected cells was at least fivefold lower and was not detectable in mock-infected cells, suggesting that the induced recombination activity detected by these methods was under viral control. The results of kinetic studies, analysis with methylation-sensitive restriction enzymes, and the use of phosphonoacetic acid, a specific inhibitor of poxvirus DNA polymerase, indicated that recombination between transfecting DNAs occurred concomitantly with DNA replication but that the two processes could be partially uncoupled. We conclude that the dramatic expansion of recombination activities in the cytoplasm of poxvirus-infected cells is virus specific and offers a good model system with which to analyze the mechanism of recombination in a eucaryotic environment.

Virus-lymphocyte interactions during the course of immunosuppressive virus infection.

Malignant rabbit fibroma virus (MV) is a lymphocytotropic leporipoxvirus which produces profound immunological dysfunction and lethal fibromyxosarcoma. We examined virus recovery from splenic lymphocytes as a function of time after inoculation in vivo, and correlated this with both immunological function and expression of virus-induced host suppressor activity. MV was most abundant in lymphocytes obtained 4 days following inoculation. At that time, immune function was relatively normal and host suppressor activity was not observed. By 7 days after infection, when active host immunosuppressor functions were observed, virus recovery was decreased. Eleven days post-inoculation host immune function began to recover despite increasing virus-induced tumours and developing opportunistic infection. Simultaneously, MV was no longer recoverable from spleen cells. Spleen cells from day 11 tumour-bearing rabbits did not support MV replication as efficiently as did normal or day 4 or 7 splenic lymphocytes, but they did not alter the ability of MV to grow in the latter cells. By fluorescence examination and cytofluorography, splenic lymphocytes bearing MV antigens were abundant 7 days after infection but disappeared by 11 days. This was temporally related to production of neutralizing antibody to MV, and development of virus-specific lymphocyte proliferative activity. The composition of splenic lymphocytes changed as well: the normal ratio of about 1:1 for B and T cells changed to 1:2 by day 7, and then inverted to almost 2:1 by day 11. Rabbits infected with MV thus appear to recover their immune function, concurrently eliminate virus-infected lymphocytes, and elaborate high titres of neutralizing serum antibodies despite progressive infections and tumour development.

Inhibition of virus replication does not alter malignant rabbit fibroma virus-induced immunosuppression.

Malignant rabbit fibroma virus (MV) directly suppresses generation of antibody responses and mitogen induced T and B lymphocyte proliferation. We investigated whether this phenomenon required expression of the complete viral genome. Phosphonoacetic acid (PAA) inhibits poxvirus specific DNA polymerases. Adding PAA to cultures reduces both MV replication and mitogen-driven rabbit lymphocyte proliferation in a dose-dependent fashion. A dose of PAA adequate to inhibit MV replication by about 97%, but insufficient to reduce lymphocyte proliferation appreciably, does not affect the ability of MV to suppress lymphocyte proliferation or initiation of antibody production. Spleen cells from MV tumour-bearing rabbits contain very little virus, but inhibit the proliferative and antibody forming responses of normal spleen cells. This activity is shown here to reflect the production by T lymphocytes of a soluble mediator of greater than 25 kD molecular weight. Adding PAA to these mixed spleen cell cultures does not alter the ability of MV to induce T suppressor activity in host lymphocytes. Thus, these immunosuppressive capabilities of MV appear to reflect early MV gene functions.

Replication of vesicular stomatitis virus facilitated by Shope fibroma virus in vivo.

Previous studies show that Shope fibroma virus facilitates replication of vesicular stomatitis virus (VSV) in some rabbit cells grown in vitro. In the present investigation, the possibility that these two viruses can also interact in vivo was determined. Rabbits inoculated intradermally with both viruses together, or each separately, were examined for the formation of lesions or tumors and for the production of infectious virus. The presence of VSV interfered with tumorigenesis by Shope fibroma virus. In tumors already formed, production of infectious VSV was greater than in normal skin. Hence, each virus affected the other. Sera and tissues of normal rabbits were found to contain a substance which inhibits VSV; this may act to limit replication of VSV in rabbit skin. In addition, cultured rabbit skin cells appeared to adsorb VSV inefficiently. When persistently infected by Shope fibroma virus, however, adsorption of VSV was markedly improved. Our results suggest that in vivo Shope fibroma virus may facilitate adsorption of VSV to reduce the effect of a natural inhibitor and consequently enhance production of infectious virus.