Polyomavirus tumor induction in mice: effects of polymorphisms of VP1 and large T antigen.

By testing recombinants between "high tumor" (inducing a high incidence of tumors) and "low tumor" (inducing a low incidence of tumors) strains of polyomavirus, we have previously shown that the key determinant(s) for induction of a high tumor profile resides in coding regions of the high tumor strain (R. Freund, G. Mandel, G. G. Carmichael, J. P. Barncastle, C. J. Dawe, and T. L. Benjamin, J. Virol. 61:2232-2239, 1987). Three single-amino-acid differences between the PTA (high tumor) and RA (low tumor) virus strains have now been identified by DNA sequencing, one each in the large T antigen, in the region common to the middle and small T antigens, and in the major capsid protein VP1. Further tests of appropriate recombinants and oligonucleotide-induced mutants show that VP1 of PTA is the major determinant for induction of a high tumor profile, including all tumors of epithelial origin. The differential effect of the VP1s of PTA and RA on the tumor profile is discussed in terms of a likely contribution of the polymorphic region of VP1 to binding of receptors and infection of different cell types in the animal. The polymorphism in the large T antigen has a more restricted action, which is seen only when tested in virus carrying the VP1 type of PTA; the PTA large T antigen then promotes more rapid growth of tumors of salivary gland and thymus than the RA large T antigen.     

Susceptibility and resistance to poliovirus-induced paralysis of inbred mouse strains.

Susceptibility to human poliovirus-induced disease in different inbred mouse strains was analyzed after intracerebral inoculation of two mouse-adapted type 2 polioviruses, the attenuated W-2 strain and the virulent Lansing strain. In contrast to inoculation with the Lansing strain, which was invariably lethal, inoculation with the W-2 strain defined three groups of mice with high, intermediate, or low disease incidence. Those in the high-disease-incidence group, the DBA/1J and DBA/2J mice, exhibited a high level of virus replication in the spinal cord by day 2 postinfection, with no detectable neutralizing-antibody response. Mice in the intermediate- and low-incidence groups had lower levels of virus replication in the spinal cord and/or produced neutralizing antibodies. No correlation was observed between H-2 haplotype and the extent of virus replication, production of neutralizing or enzyme-linked immunosorbent assay-detectable antibodies, or T-cell-proliferative response. However, mice of the H-2k haplotype manifested a low incidence of disease.     

Polyomavirus tumor induction in mice: influences of viral coding and noncoding sequences on tumor profiles.

We determined the DNA sequences of the noncoding regions of two polyomavirus strains that differ profoundly in their abilities to induce tumors in mice. Differences between strains were found, both on the late side of the replication origin in the region containing known enhancer elements and on the early side of the origin, affecting the number and location of large-T-antigen-binding sites. By constructing and analyzing recombinant viruses between these high- and low-tumor strains, we attempted to localize determinants which affect the frequency and histotype of tumors. Seven recombinants were constructed and propagated in vitro, and the tumor profile of each was established by inoculation into newborn C3H mice. Recombinants containing noncoding sequences from the high-tumor strain and coding sequences from the low-tumor strain behaved like the latter, inducing tumors at a low frequency and strictly of mesenchymal origin. Reciprocal recombinants with noncoding sequences of the low-tumor strain linked to structural determinants from the high-tumor strain induced several types of epithelial tumors typical of the high-tumor strain but at reduced frequency, in addition to mesenchymal tumors. A high frequency and full diversity of epithelial tumors required, in addition to structural regions from the high-tumor strain, noncoding sequences on the early side of the origin also present in this strain. A high-tumor profile thus resulted from the combined effects of structural and regulatory determinants in the high-tumor strain, with the former affecting primarily the tissue tropism and the latter affecting the frequency of tumors. No differential effects of the enhancer regions from the late side of the origin in the two virus strains were seen in this study.     

Genetic and structural analysis of a virulence determinant in polyomavirus VP1.

The LID strain of polyomavirus differs from other laboratory strains in causing a rapidly lethal infection of newborn C3H/Bi mice. This virulent behavior of LID was attenuated by dilution, yet at sublethal doses LID was able to induce tumors at a high frequency, like its parent virus PTA. By constructing and assaying LID-PTA recombinant viruses and by DNA sequencing, the determinant of virulence in LID was mapped to the major viral capsid protein, VP1. The VP1s of LID and PTA differed at two positions: at 185, LID has phenylalanine and PTA has tyrosine, and at 296, LID has alanine and PTA has valine. Results obtained with viruses constructed by site-directed mutagenesis showed that alanine at position 296 is sufficient to confer a fully virulent phenotype regardless of which amino acid is at position 185. However, with valine at position 296, an effect of phenylalanine at position 185 is apparent, as this virus possesses an intermediate level of virulence. A crystal structure of polyomavirus complexed with 3'-sialyl lactose previously indicated van der Waals contacts between the side chain of valine 296 and the sialic acid ring (T. Stehle, Y. Yan, T. L. Benjamin, and S. C. Harrison, Nature [London] 369:160-163, 1994). When this interaction was modeled with alanine, these contacts were greatly reduced. Direct confirmation that the substitutions in VP1 affected receptor binding was obtained by studying virus hemagglutination behavior. The ensemble of results are discussed in terms of the idea that a lower affinity of the virus for its receptor can result in more rapid spread and increased pathogenicity.     

Exchange of the VP5 of infectious bursal disease virus in a serotype I strain with that of a serotype II strain reduced the viral replication and cytotoxicity

Infectious bursal disease virus (IBDV), belonging to Avibirnavirus genus in the Birnaviridae family, consists of two segments of double-strand RNA. There are two distinct serotypes of IBDV, the pathogenic serotype I and the non-pathogenic serotype II. Comparison of the deduced amino acid sequences of a panel of VP5 genes retrieved from GenBank revealed a high identity among strains within the serotype I or serotype II group but a low identity between strains across two serotypes. In this study, we rescued two mosaic viruses, rGtGxVP5 and rGt2382VP5 by exchanging the VP5 gene of a cell culture-adapted serotype I Gt strain with its counterpart of the very virulent IBDV Gx strain, or a non-pathogenic 23/82 strain of the serotype II. In comparison to the parental strain rGt virus, the rGtGxVP5 showed the similar viral replication, cytotoxicity and the ability of inducing apoptosis; however, the other mosaic virus rGt2382VP5 had a lower titer and a reduced cytotoxicity. Although exchange of VP5 within serotype I group did not alter the viral replication and cytotoxicity of Gt strain, exchange of VP5 in the serotype I with that of a serotype II reduced the viral replication and cytotoxicity on chicken embryo fibroblast (CEF) cells. Therefore, the VP5 of serotype II may be one of the factors responsible for the distinct pathogenic features of two serotypes.     

Age-related differences in amplification of covalently closed circular DNA at early times after duck hepatitis B virus infection of ducks

Inoculation of 3-day-old (3D) or 3-week-old (3W) ducklings with duck hepatitis B virus results in chronic or transient infection, respectively. We previously showed that rapid production of neutralizing antibody following inoculation of 3W ducklings prevents virus from spreading in the liver and leads to a transient infection (Y.-Y. Zhang and J. Summers, J. Virol. 78:1195-1201, 2004). In this study we further investigated early events of viral infection in both 3D and 3W ducks. We present evidence that a lower level of virus replication in the hepatocytes of 3W birds is an additional factor that probably favors transient infection. We suggest that lower virus replication is due to a less rapid covalently closed circular DNA amplification, leading to lower viremias and a slower spread of infection in the liver, and that the slower spread of infection in 3W ducks makes the infection more sensitive to interruption by the host immune responses.     

A single-amino-acid substitution in polyomavirus VP1 correlates with plaque size and hemagglutination behavior.

The plaque size and hemagglutination characteristics of five cloned wild-type strains of polyomavirus were determined. The strains fell into two groups, those with large or small plaques, each with distinctive hemagglutination behavior at different temperatures and pHs. The nucleotide sequence of VP1, the major capsid protein of the virus, was determined for each of the viral strains. The PTA (large-plaque) and RA (small-plaque) strains differed only at residue 92 of VP1, where there is a glutamic acid or glycine, respectively (R. Freund, A. Calderone, C. J. Dawe, and T. L. Benjamin, J. Virol. 65:335-341, 1991). The same amino acid difference in VP1 correlated with plaque size and hemagglutination properties of the other sequenced viruses. Mutagenesis converting amino acid 92 from glutamic acid to glycine converted the plaque size and hemagglutination behavior of the large-plaque PTA strain to that of a small-plaque strain. Furthermore, PTA and RA VP1 proteins produced in Escherichia coli behaved as their parental viruses did in hemagglutination assays. These results demonstrate that amino acid residue 92 of VP1 is involved in determining the plaque size and hemagglutination behavior of polyomavirus and strongly suggest that this region of the VP1 polypeptide interacts directly with cell receptors.     

Herpes simplex virus 1 VP22 regulates translocation of multiple viral and cellular proteins and promotes neurovirulence

Herpes simplex virus 1 (HSV-1) protein VP22, encoded by the UL49 gene, is a major virion tegument protein. In the present study, we showed that VP22 was required for efficient redistribution of viral proteins VP16, VP26, ICP0, ICP4, and ICP27 and of cellular protein Hsc-70 to the cytoplasm of infected cells. We found that two dileucine motifs in VP22, at amino acids 235 and 236 and amino acids 251 and 252, were necessary for VP22 regulation of the proper cytoplasmic localization of these viral and cellular proteins. The dileucine motifs were also required for proper cytoplasmic localization of VP22 itself and for optimal expression of viral proteins VP16, VP22, ICP0, UL41, and glycoprotein B. Interestingly, a recombinant mutant virus with alanines substituted for the dileucines at amino acids 251 and 252 had a 50% lethal dose (LD(50)) for neurovirulence in mice following intracerebral inoculation about 10(3)-fold lower than the LD(50) of the repaired virus. Furthermore, the replication and spread of this mutant virus in the brains of mice following intracerebral inoculation were significantly impaired relative to those of the repaired virus. The ability of VP22 to regulate the localization and expression of various viral and cellular proteins, as shown in this study, was correlated with an increase in viral replication and neurovirulence in the experimental murine model. Thus, HSV-1 VP22 is a significant neurovirulence factor in vivo.     

Assembly of infectious Herpes simplex virus type 1 virions in the absence of full-length VP22

VP22, the 301-amino-acid phosphoprotein product of the herpes simplex virus type 1 (HSV-1) U(L)49 gene, is incorporated into the tegument during virus assembly. We previously showed that highly modified forms of VP22 are restricted to infected cell nuclei (L. E. Pomeranz and J. A. Blaho, J. Virol. 73:6769-6781, 1999). VP22 packaged into infectious virions appears undermodified, and nuclear- and virion-associated forms are easily differentiated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (J. A. Blaho, C. Mitchell, and B. Roizman, J. Biol. Chem. 269:17401-17410, 1994). As VP22 packaging-associated undermodification is unique among HSV-1 tegument proteins, we sought to determine the role of VP22 during viral replication. We now show the following. (i) VP22 modification occurs in the absence of other viral factors in cell lines which stably express its gene. (ii) RF177, a recombinant HSV-1 strain generated for this study, synthesizes only the amino-terminal 212 amino acids of VP22 (Delta212). (iii) Delta212 localizes to the nucleus and incorporates into virions during RF177 infection of Vero cells. Thus, the carboxy-terminal region is not required for nuclear localization of VP22. (iv) RF177 synthesizes the tegument proteins VP13/14, VP16, and VHS (virus host shutoff) and incorporates them into infectious virions as efficiently as wild-type virus. However, (v) the loss of VP22 in RF177 virus particles is compensated for by a redistribution of minor virion components. (vi) Mature RF177 virions are identical to wild-type particles based on electron microscopic analyses. (vii) Single-step growth kinetics of RF177 in Vero cells are essentially identical to those of wild-type virus. (viii) RF177 plaque size is reduced by nearly 40% compared to wild-type virus. Based on these results, we conclude that VP22 is not required for tegument formation, virion assembly/maturation, or productive HSV-1 replication, while the presence of full-length VP22 in the tegument is needed for efficient virus spread in Vero cell monolayers.     

Expression of major capsid protein VP-1 in the absence of viral particles in thymomas induced by murine polyomavirus

Thymomas induced by polyomavirus strain PTA in mice are known to express the major capsid protein VP-1. Since the expression of a late structural protein such as VP-1 is considered a sign of virus replication, the present work attempted to clarify the implication of the presence of this protein in tumor cells. Electron microscopy of tumors showed a striking absence of viral particles in the vast majority of the cells. However, immunoelectron microscopy of the same samples demonstrated intranuclear VP-1 in most cells despite the absence of viral particles. Very little infectious virus was recovered from tumors. A change in the electrophoretic mobility of VP-1 from thymomas was detected compared with VP-1 from productively infected cells. The data presented in this work prove that the expression of VP-1 in polyomavirus-induced tumors is not synonymous with the presence of infectious virus, suggesting a possible defect in viral encapsidation.     

Identification of a major T-cell epitope within VP3 amino acid residues 24 to 37 of Theiler's virus in demyelination-susceptible SJL/J mice.

Intracerebral inoculation of susceptible strains of mice with Theiler's murine encephalomyelitis virus (TMEV) results in a chronic, immunologically mediated demyelinating disease that shares many features with human multiple sclerosis. CD4+ T lymphocytes play a critical role in the pathogenesis of virus-induced demyelinating disease. We have identified a region within amino acid residues 24 to 37 of the VP3 capsid protein of TMEV (VP3(24-37)) that is recognized by T lymphocytes from the demyelination-susceptible SJL/J strain of mice. The T-cell response to VP3(24-37) represents a predominant Th-cell response against the virus from either TMEV-immunized or TMEV-infected SJL/J mice, and viral epitopes VP1(233-250), VP2(74-86), and VP3(24-37) account for most of the Th-cell response to TMEV.     

Radiation-Resistant and Radiation-Sensitive Forms of Host Resistance to Polyomavirus

Newborn mice of several inbred strains develop few or no tumors following inoculation with highly tumorigenic strains of polyomavirus. Here we show that such resistant strains can be divided into two groups based on the responses of adult mice to radiation followed by virus inoculation. Most strains show a radiation-sensitive form of resistance (R^r-s) and develop tumors following radiation and virus challenge. This type of resistance has previously been recognized as immunological, based on T-cell responses against virus-encoded neoantigen(s) expressed in tumor cells. Other strains exhibit a radiation-resistant form of resistance (R^r-r) and fail to develop tumors when treated in the same manner. Three additional properties of R^r-r mice distinguish them from R^r-s mice: (i) survival of newborns following inoculation with a highly virulent and usually lethal strain of virus, (ii) resistance to virus spread in newborns inoculated with either tumorigenic or virulent virus strains, and (iii) dominant or semidominant transmission of resistance in crosses with a highly susceptible strain. The R^r-r phenotype reflects a constitutive nonimmunological type of resistance that is targeted to the virus and blocks its dissemination.     

Infection immunity of piglets to either VP3 or VP7 outer capsid protein confers resistance to challenge with a virulent rotavirus bearing the corresponding antigen.

A single-gene substitution reassortant 11-1 was generated from two porcine rotaviruses, OSU (serotype 5) and Gottfried (serotype 4). This reassortant derived 10 genes, including gene 4 encoding VP3, from the OSU strain and only gene 9, encoding a major neutralization glycoprotein (VP7), from the Gottfried strain and was thus designated VP3:5; VP7:4. Oral administration of this reassortant to colostrum-deprived gnotobiotic newborn pigs induced a high level of neutralizing antibodies not only to Gottfried VP7 but also to OSU VP3, thus demonstrating that VP3 is as potent an immunogen as VP7 in inducing neutralizing antibodies during experimental oral infection. Gnotobiotic piglets infected previously with the reassortant were completely resistant to oral challenge with the virulent Gottfried strain (VP3:4; VP7:4), as indicated by failure of symptoms to develop and lack of virus shedding. Similarly, prior infection with the reassortant induced almost complete protection against diarrhea and significant restriction of virus replication after oral challenge with the virulent OSU strain (VP3:5; VP7:5). Thus, it appears that (i) the immune system of the piglet responds equally well to two rotavirus outer capsid proteins, VP3 and VP7, during primary enteric rotavirus infection; (ii) antibody to VP3 and antibody to VP7 are each associated with resistance to diarrhea; and (iii) infection with a reassortant rotavirus bearing VP3 and VP7 neutralization antigens derived from two viruses of different serotype induces immunity to both parental viruses. The relevance of these findings to the development of effective reassortant rotavirus vaccines is discussed.     

Non-invasive imaging of mouse hepatitis coronavirus infection reveals determinants of viral replication and spread in vivo

Bioluminescence imaging (BLI) is a powerful new method to study virus dissemination in the live animal. Here we used this method to monitor the spatial and temporal progression of mouse hepatitis coronavirus (MHV) infection in mice using luciferase-expressing viruses. Upon intranasal inoculation, virus replication could initially be observed in the nasal cavity and the cervical lymph nodes, after which the infection spread to the brain and frequently to the eyes. The kinetics of virus spread to and clearance from the brain appeared to depend on the inoculation dose. After intraperitoneal inoculation, virus replication was predominantly observed in the liver and occasionally in the intestines, but interestingly also in the tail and paws. BLI thus elucidated new anatomic locations of virus replication. Furthermore, MHV dissemination was shown to be critically depended on the viral spike protein, but also on the mouse strain used. Widespread dissemination was observed in mice lacking a functional type I interferon response. The importance of the type I interferon system in limiting viral spread was also demonstrated by the administration of type I interferons to mice. Our results provide new insights in coronavirus pathogenesis and demonstrate the potential of BLI to study coronavirus–host interactions in vivo .     

Role of VP2 amino acid 141 in tropism of Theiler's virus within the central nervous system.

Following intracranial inoculation, Theiler's virus causes either an acute encephalitis (strain GDVII) or a chronic demyelinating disease (strain DA). The DA strain sequentially infects the grey matter of the brain, the grey matter of the spinal cord, and, finally, the white matter of the spinal cord, where it persists in glial cells and causes demyelinating lesions. Analysis of the phenotype of recombinant viruses has shown that the viral capsid contains determinants for persistence and demyelination. Our previous studies showed that a Lys at position 141 of the VP2 capsid protein (VP2-141) could render a chimeric virus persistent. We also reported that another recombinant virus, virus R5, migrated from the grey matter of the brain to that of the spinal cord inefficiently and was unable to infect the white matter of the spinal cord. In this article, we report that introducing a Lys at position VP2-141 in virus R5 increases its ability to infect the white matter of the spinal cord. Our results indicate that this amino acid is important for the spread of the virus within the central nervous system.     

Recombinant mink cell focus-inducing virus and long terminal repeat alterations accompany the increased leukemogenicity of the Mo+PyF101 variant of Moloney murine leukemia virus after intraperitoneal inoculation.

We recently showed that different routes of inoculation affect the leukemogenicity of the Mo+PyF101 variant of Moloney murine leukemia virus (M-MuLV). Intraperitoneal (i.p.) inoculation of neonatal mice with Mo+PyF101 M-MuLV greatly enhanced its leukemogenicity compared with subcutaneous (s.c.) inoculation. We previously also suggested that the leukemogenicity defect of Mo+PyF101 M-MuLV when inoculated s.c. may result from the inability of this virus to form env gene recombinant (mink cell focus-inducing [MCF]) virus. In this study, virus present in end-stage tumors and in preleukemic animals inoculated i.p. by Mo+PyF101 M-MuLV was characterized. In contrast to s.c. inoculation, all tumors from i.p.-inoculated mice contained high levels of recombinant MCF virus. Furthermore, Southern blot analyses demonstrated that the majority of the tumors contained altered Mo+PyF101 M-MuLV long terminal repeats. The U3 regions from several tumors with altered long terminal repeats were cloned by PCR amplification. Sequence analyses indicated that the M-MuLV 75-bp tandem repeat in the enhancer region was triplicated. This amplification was also previously observed in mice infected s.c. with a pseudotypic mixture of Mo+PyF101 M-MuLV and Mo+PyF101 MCF virus. The enhancer triplication was an early event, and it occurred within 2 weeks postinfection. Recombinant MCF viruses were not detected by Southern blot analyses until 4 weeks postinfection. Thus, the M-MuLV enhancer triplication event was initially important for efficient propagation of ecotropic Mo+PyF101 M-MuLV. The increased leukemogenicity following i.p. inoculation could be explained if the triplication enhances Mo+PyF101 M-MuLV replication in the bone marrow and bone marrow infection is required for recombinant MCF virus formation.     

Amplification and novel locations of endogenous mouse mammary tumor virus genomes in mouse T-cell lymphomas.

Endogenous mouse mammary tumor virus genomes are amplified and located in novel cell DNA sequences in many mouse T-cell lymphomas. Transplanted tumors recovered from the same mouse strain and shown to be of independent origin by chromosomal analysis, by the presence of JH immunoglobulin gene rearrangements, or by the integration patterns of exogenous Moloney MuLV genomes frequently showed similar patterns of novel mouse mammary tumor virus-containing cell DNA fragments. This process of amplification and relocation can occur within a limited number of cell generations and in C57BL/6 mice does not lead to the synthesis of mature virus-encoded proteins. In some instances, amplified mouse mammary tumor virus genomes contained novel restriction cleavage sites in the gag-pol region. The restricted time course of occurrence, lack of synthesis of mature virion proteins, and apparent site specificity indicate that this process of retrovirus amplification differs significantly from virus replication after exogenous infection.     

Immunity in the female genital tract after intravaginal vaccination of mice with an attenuated strain of herpes simplex virus type 2.

Herpes simplex virus type 2 is a common human venereal pathogen which causes lethal neurological illness after intravaginal inoculation into BALB/cJ mice. To investigate whether an attenuated, nonlethal strain of this virus would confer immunity after inoculation of mice, we constructed a strain containing a partial deletion of the thymidine kinase gene, which is necessary for viral replication and spread in sensory ganglia. Unlike its wild-type counterpart, this deletion-containing strain of herpes simplex virus type 2 caused mild clinical disease and was not lethal when studied in an age-dependent murine model of intravaginal infection. Furthermore, after intravaginal infection, the deletion-containing strain could not be isolated from sensory ganglia at a time when wild-type virus was abundant. Of greater significance, intravaginal inoculation with the deletion-containing strain rendered mice completely resistant to rechallenge with a 10-fold 50% lethal dose of wild-type virus. These results suggest that a strain of herpes simplex virus type 2 containing a deletion of the thymidine kinase gene will be useful in studying the cellular basis of mucosal immunity in the genital tract.     

Altering the expression kinetics of VP5 results in altered virulence and pathogenesis of herpes simplex virus type 1 in mice

While many herpes simplex virus (HSV) structural proteins are expressed with strict-late kinetics, the HSV virion protein 5 (VP5) is expressed as a "leaky-late" protein, such that appreciable amounts of VP5 are made prior to DNA replication. Our goal has been to determine if leaky-late expression of VP5 is a requirement for a normal HSV infection. It had been shown previously that recombinant viruses in which the VP5 promoter was replaced with promoters of other kinetic classes (including a strict late promoter) exhibited no alterations in replication kinetics or virus yields in vitro. In contrast, here we report that alterations in pathogenesis were observed when these recombinants were analyzed by experimental infection of mice. Following intracranial inoculation, a recombinant expressing VP5 from a strict-late promoter (U(L)38) exhibited an increased 50% lethal dose and a 10-fold decrease in virus yields in the central nervous system, while a recombinant expressing VP5 from an early (dUTPase) or another leaky-late (VP16) promoter exhibited wild-type neurovirulence. Moreover, following infection of the footpad, changing the expression kinetics of VP5 from leaky-late to strict-late resulted in 100-fold-less virus in the spinal ganglia during the acute infection than produced by either the parent virus or the rescued virus. These data indicate that the precise timing of appearance of the major capsid protein plays a role in the pathogenesis of HSV infections and that changing the expression kinetics has different effects in different cell types and tissues.     

Targeting human glioblastoma cells: comparison of nine viruses with oncolytic potential

Brain tumors classified as glioblastomas have proven refractory to treatment and generally result in death within a year of diagnosis. We used seven in vitro tests and one in vivo trial to compare the efficacy of nine different viruses for targeting human glioblastoma. Green fluorescent protein (GFP)-expressing vesicular stomatitis (VSV), Sindbis virus, pseudorabies virus (PRV), adeno-associated virus (AAV), and minute virus of mice i-strain (MVMi) and MVMp all infected glioblastoma cells. Mouse and human cytomegalovirus, and simian virus 40 showed only low levels of infection or GFP expression. VSV and Sindbis virus showed strong cytolytic actions and high rates of replication and spread, leading to an elimination of glioblastoma. PRV and both MVM strains generated more modest lytic effects and replication capacity. VSV showed a similar oncolytic profile on U-87 MG and M059J glioblastoma. In contrast, Sindbis virus showed strong preference for U-87 MG, whereas MVMi and MVMp preferred M059J. Sindbis virus and both MVM strains showed highly tumor-selective actions in glioblastoma plus fibroblast coculture. VSV and Sindbis virus were serially passaged on glioblastoma cells; we isolated a variant, VSV-rp30, that had increased selectivity and lytic capacity in glioblastoma cells. VSV and Sindbis virus were very effective at replicating, spreading within, and selectively killing human glioblastoma in an in vivo mouse model, whereas PRV and AAV remained at the injection site with minimal spread. Together, these data suggest that four (VSV, Sindbis virus, MVMi, and MVMp) of the nine viruses studied merit further analysis for potential therapeutic actions on glioblastoma.