Unshackling the Links Between Reovirus Oncolysis, Ras Signaling, Translational Control and Cancer

Unshackling the links between reovirus oncolysis, Ras signaling, translational control and cancer

Maya Shmulevitz1, Paola Marcato1 and Patrick WK Lee*,1

1Department of Microbiology and Immunology, Dalhousie University, 7P Sir Charles Tupper Medical Building, 5850 College Street, Halifax, NS, Canada B3K 1X5

Reovirus has an inherent preference for replicating in cells cancer cells. Viruses have evolved intricate strategies to with dysregulated growth factor signaling cascades that ensure efficient replication in the host cell. Viral encoded comprise Ras activation. Precisely how reovirus exploits factors can alter the regulation of cellular processes the host cell Ras pathway is unclear, but there is evidence important for viral replication. Processes amicable to suggesting that activated Ras signaling is important for viral replication are provoked, while those that are efficient viral protein synthesis. Defining the molecular detrimental for replication are prevented. For example, mechanism of reovirus oncolysis will shed light on reovirus DNA viruses that require cell division for viral genome replication and important aspects of cellular transforma- replication have evolved strategies to ensure that cells tion, Ras signaling cascades and regulation of protein proliferate. Cellular responses that would clear viral translation. infections such as apoptosis, interferon (IFN)-induced Oncogene (2005) 24, 7720–7728. doi:10.1038/sj.onc.1209041 antiviral activities and presentation of viral antigens by multihistocompatibility class I (MHC-I) molecules are Keywords: cancer; oncolytic virus; oncolysis; reovirus; overcome to different extents by various viruses. In view translation; Ras of such highly evolved strategies to ensure efficient replication in cells, how can viruses be used specifically to kill cancer cells? With increased understanding of virus replication, viral-cell interactions and methods to manipulate viral Viruses with inherent or engineered preference for genomes, viruses have been generated that are no longer replication in cancer cells able to regulate specific host processes. The deletion of viral proteins important in regulating cellular processes Many exciting developments have been made in the last can, therefore, create viruses that are unable to replicate two decades with respect to the use of replication- in normal cells. Conversely, cellular events involved in competent viruses as specific oncolytic agents. A handful tumorigenesis, such as increased cell division and of oncolytic viruses have been identified. Some naturally dysregulated cell death programs, make tumor cells occurring viruses preferentially infect and kill trans- permissive to replication by these recombinant viruses. formed cells, such as vesicular stomatitis virus (VSV), ONYX-015 is an example of an engineered oncolytic herpes simplex virus (HSV), Newcastle disease virus virus, and is described thoroughly by Clodagh O’Shea (NDV) and reovirus (Norman and Lee, 2000; Stojdl and Frank McCormick in this edition of Oncogene et al., 2000; Sinkovics and Horvath, 2000; Farassati reviews. The engineered ONYX-015 virus is unable to et al., 2001). The Edmonston-B strain of measles virus make E1B-55K, an early adenoviral protein critical for (MV-Edm), originally attenuated for use as a live virus degrading p53 and thus overcoming p53-mediated cell vaccine, has potential oncolytic application (Grote et al., cycle arrest and apoptosis (Biederer et al., 2002). The 2001). Furthermore, several genetic modifications to selective oncolysis by ONYX-015 is likely in part due to adenoviruses, influenza viruses and herpes viruses have the dysregulation of p53 pathways in susceptible tumor been shown to confer selective oncolytic activity cells. Interestingly, it was recently discovered that (Martuza, 2000; Bergmann et al., 2001; Dobbelstein, following infection with ONYX-015 efficient export of 2004). A growing body of research demonstrates that late adenoviral RNA from the nucleus is restricted to these viruses can efficiently and selectively destroy tumor cells (O’Shea et al., 2004). This finding shows that transformed cell lines and also have antitumor proper- novel and unperceived differences between tumor and ties in vivo. Exciting developments have also been made normal cells can affect virus replication. in understanding the molecular bases of viral oncolysis. Remarkably, reoviruses show an inherent prefe- The use of replication-competent viruses for oncolysis rence for replication in many transformed cells. depends on their ability to selectively infect and destroy Possible explanations for why such a preference may have developed through the evolution of reovirus are discussed in this review. Reoviruses, the prototype *Correspondence: PWK Lee; E-mail: [email protected] member of the Reoviridae family, were first isolated Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7721 from the intestinal tracts of apparently healthy indivi- replication, while transfection of EGFR into these cells duals. In fact, most people have been exposed to proved significantly advantageous for reovirus infection reovirus by the age of 5 with little or no manifestation (Strong et al., 1993). Interestingly, the involvement of of symptoms. Hence, there are fewer concerns about the EGFR in reovirus replication was found to be safety of using reovirus in cancer treatment. As this dependent on signaling pathways initiated through the review will describe in detail, an understanding of the tyrosine kinase receptor rather than receptor binding cellular events that contribute to the conditional specificity. A mutated EGFR devoid of signaling replication of reovirus is emerging. Growth factor capabilities was unable to confer permissiveness to receptor signaling is intimately related with tumor reovirus replication, while the v-erbB oncoprotein, progression. Genetic changes resulting in constitutive which lacks the extracellular ligand-binding domain activation of growth factor signaling pathways result in but has a constitutively active kinase domain, was uncontrolled proliferation, differentiation and/or me- sufficient to permit reovirus infection (Strong et al., tastasis, and are associated with most if not all human 1993; Strong and Lee, 1996). Therefore, what started as cancers. Intriguingly, reovirus replication is sensitive to a search for a secondary receptor for reovirus entry that the status of signaling cascades downstream of specific could permit selective replication in transformed cells tyrosine kinase receptors such as the epidermal growth turned into a very interesting connection between the factor receptor (EGFR). The pathways downstream of status of intracellular signaling pathways and the tyrosine kinase receptors are not only numerous, but are reovirus life cycle (Figure 1). complicated by extensive cross-talk. Untangling the large networks of interactions coupled with cell signal- Involvement of Ras/RalGEF/p38 signaling pathways in ing and phenotypic alterations associated with cancer is reovirus replication a challenge. Therefore, not only is reovirus a promising oncolytic agent but also reovirus replication offers a The interaction of growth factors with their cognate unique approach to understanding the components and receptors results in the activation of a cascade of outcomes of signal transduction pathways aberrantly intracellular biochemical events leading to a cohort of regulated in cancer. cellular functions. Which cellular functions can affect reovirus replication? Which components of the signaling pathway are important? The EGFR signaling cascade initiates with receptor oligomerization and subsequent Intracellular signaling pathways involved in selective activation of tyrosine kinase activity resulting in replication of reovirus autophosphorylation of specific tyrosine residues (Ull- How the link between selective reovirus replication and rich and Schlessinger, 1990; Fantl et al., 1993; Weiss and cancer was discovered Schlessinger, 1998). Adaptor molecules with phospho- tyrosine binding, src homology 2 (SH2) domains are Differences between virally or spontaneously trans- recruited to activated EGFR and in turn assemble a formed cell lines and primary or untransformed cells cohort of proteins particular to one of many downstream with respect to their susceptibility to reovirus cytotoxi- cascades (Pawson, 1994; Pawson and Scott, 1997; Buday, city were first noted in 1977 (Hashiro et al., 1977). 1999). The Ras signaling cascade is a major pathway Likewise, human lung fibroblast cells (WI-38) showed downstream of EGFR (Campbell et al., 1998). Ras enhanced permissibility to reovirus replication when proteins form a subfamily of small GTP-binding proteins they were transformed with the simian virus 40 T- involved in regulation of a wide variety of cellular antigen (Duncan et al., 1978). Nevertheless, it was not function such as cell growth, differentiation and cell until the 1990s that clues were obtained regarding the survival. Membrane-anchored Ras proteins cycle be- molecular basis of the selective replication of reovirus in tween the inactive GDP-bound state and the active GTP- transformed cells. bound state. The activation of Ras proteins is promoted Receptor specificity often dictates the tropism of by guanine nucleotide exchange factors such as the viruses (Cohen et al., 1988; Schneider-Schaulies, 2000; Son of Sevenless (Sos), which are recruited by the Sieczkarski and Whittaker, 2005). When research on the previously described adaptor molecules to activated EGFRs determinants of selective oncolysis by reovirus began, it (Overbeck et al., 1995). With such an important role in was known that the fibrous tail of the reovirus cell cellular responses to activated EGFR, the Ras pathway attachment protein, s1, bound sialic acid (Armstrong became a logical target for possible involvement in et al., 1984; Pacitti and Gentsch, 1987; Dermody et al., conferring reovirus permissiveness. 1990; Chappell et al., 1997, 2000). The ubiquitous nature Thorough analysis showed that an activated Ras of sialic acid, however, suggested that it could not signaling pathway plays an integral role in reovirus account for gross differences in cell susceptibility. Since infectivity. Cell lines expressing constitutively active aberrations affecting tyrosine kinase receptors are often oncogenes of Sos or Ras were found to be permissive to associated with transformation, the use of these reovirus infection (Strong et al., 1998). Importantly, receptors for reovirus entry propounded a potential link when activated Ras was placed under a zinc-inducible between transformation and permissiveness to reovirus promoter, a productive infection by reovirus was infection. EGFR-minus mouse cell lines (NR6 and B82) only found in the presence of ZnSO4. This suggested were found to be relatively resistant to reovirus that the activated Ras protein itself, but not secondary

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7722

Figure 1 Signaling pathways and effectors involved in determining the outcome of reovirus infection. Proteins determined to correlate with reovirus infectivity are red, while those shown to be dispensable for reovirus replication are yellow. Upregulated expression of IFNb and PKR is indicated by multiple copies. The role of IFN, IFN recepetor and Grb2 (gray) in reovirus oncolysis has not been tested. Dashed lines indicate pathways proposed to be linked but through unknown intermediate components

consequences of prolonged transformation, was suffi- downstream components of Ras. Interestingly, previous cient to permit reovirus infection. observations that reovirus replication is enhanced in Of over 18 downstream effectors of Ras, the best cells subjected to stressful conditions such as UV characterized are the Raf kinase, the phosphatidylino- prompted analysis on the role of stress-activated protein sitol 3-kinase (PI3-kinase) and the guanine exchange kinases. Studies showed that p38, but not c-Jun-N- factors (GEFs) for the small G protein Ral pathways. terminal kinase (JNK), participates in establishing Descriptions of these pathways are beyond the scope of reovirus infectivity. An inhibitor of p38 activation this review, but have been provided by numerous other (SB203580) suppressed reovirus replication in cells publications (Reuther and Der, 2000; Campbell and transfected with activated Ras and RalGEF (Norman Der, 2004; Repasky et al., 2004). Other laboratories et al., 2004). The potential role of p38 in transformation studying the implications of Ras activation have created has become important in the last few years (Martin- many useful tools for distinguishing between down- Blanco, 2000). The data obtained with the p38 inhibitor stream pathways (White et al., 1995, 1996; Khosravi-Far suggests that p38 is a downstream effector of RalGEF, a et al., 1996; Rodriguez-Viciana et al., 1997). Constitu- relationship previously proposed (Ouwens et al., 2002). tively active Ras mutants (RasV12) with additional Alternatively, RalGEF and p38 may act as components mutations deleterious for binding to specific effector of separate pathways that must converge for a molecules PI3K, RalGEF or Raf showed that reovirus commonly effect. Missing links between RalGEF infection was independent of signaling through Raf or activation, p38 activity and reovirus replication remain PI3-kinase cascades (Norman et al., 2004). Further- to be found. What effectors act to mediate the RalGEF– more, an active mutant of RalGEF permitted efficient p38 connection? What cascades lie downstream of reovirus replication in the absence of activated Ras. A RalGEF and p38? dominant-negative mutant of Ral, the target protein for RalGEF-mediated GDP/GTP exchange, rendered Ras- Reovirus replication as a method of understanding the role transformed cells nonpermissive to reovirus. Therefore, of Ras signaling cascades in cancer there is significant data implicating the Ras/RalGEF pathway in selective reovirus replication. Aberrations in Ras proteins have been implicated in cell The signaling pathways downstream of Ras are proliferation, transformation, invasion and metastasis immensely complicated by the high degree of cross-talk (Malumbres and Barbacid, 2001, 2003). Mutations in and dependency on factors such as cell origin and Ras proteins are associated with approximately 30% of extracellular environment. Since reovirus replication is all human cancers (Bos, 1989). This number under- dependent on one or more arms of the Ras signaling estimates the true prevalence of activated Ras pathways cascade, it provides a useful readout for determining in cancer, as mutations in upstream activators and

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7723 downstream effectors of Ras are also frequently protein expression and the formation of newly formed associated with transformation. For example, over- reovirus particles. What remains to be determined is expression of receptor tyrosine kinases (including the precise mechanism by which the Ras pathway EGFR) is commonly found in cancers, and results in impacts the reovirus life cycle. Figure 2 provides a persistent Ras activation (Janes et al., 1994; Tzahar simplified diagram of reovirus structure and life and Yarden, 1998). Of the many downstream effectors, cycle from numerous studies (Tyler, 2001). Reovirus Raf, PI3K and RalGEF have been clearly demons- contains 10 double-stranded RNA segments encapsu- trated to mediate the oncogenic properties of Ras lated by two concentric protein coats. In cell culture, (Marshall, 1996; Wolthuis and Bos, 1999; Feig and upon receptor binding and endocytosis, reovirus under- Buchsbaum, 2002; Feig, 2003; Campbell and Der, 2004). goes sequential uncoating and is released from endo- Studies using effector domain mutants specific to one of somes as a core particle (a single protein shell Raf, PI3K and RalGEF pathways suggest that although surrounding the dsRNA). Positive sense RNA is each pathway can induce individual biological re- made within the reovirus core, released into the sponses, all three pathways cooperate for complete cytoplasm, and subsequently used for protein produc- cellular transformation. Although Ral, the target of tion using the cellular translation machinery. The RalGEF activity, has been associated with vesicle primary stage of reovirus replication produces sufficient sorting, gene expression and cellular proliferation, other proteins and RNA to form subviral particles (cores) downstream effectors of RalGEF involved in cancer within localized cytoplasmic subdomains called viral remain to be found. factories. A second round of mRNA and protein With reovirus replication as the readout, we hope to production causes a gross amplification of viral particles delineate essential components in Ras signaling and that completely assemble and exit the cell upon lysis. transformation and thereby provide a novel means of Which step(s) within the reovirus replication cycle is understanding the implication of signaling cascades in affected by the state of Ras signaling? cancer. Understanding the connection between Ras Reovirus undergoes equivalent binding, entry and signaling cascades and reovirus replication will also primary transcription in both Ras transformed and permit most effective use of reovirus as a cancer untransformed cells (Strong et al., 1998; Norman et al., therapeutic, as it will suggest which cancers are 2004). Conversely, protein expression monitored susceptible to reovirus oncolysis. through 35S-methionine labeling and immunofluorescent microscopy detection with reovirus specific antibodies suggest that the expression of reovirus proteins is largely inefficient in untransformed cells. The increased expres- Which step(s) of the reovirus life cycle is blocked in sion of reovirus proteins in Ras-transformed cells nonpermissive cells, and how? correlates with increased virus titer. Present analysis therefore suggests that the block in reovirus replication Reovirus translation is blocked in untransformed cells resides between primary transcription and protein The selective replication of reovirus in Ras-transformed expression. From both a virology and a cancer cells is evident by assessing the degree of reovirus perspective, it would be exciting to understand the

Figure 2 Reovirus translation is blocked in untransformed cells. Stages of the reovirus life cycle include: (1) cell attachment via s1 interactions with sialic acid and/or JAM receptors on target cells; (2) receptor-mediated endocytosis induces shedding of s1 and s3; (3) m1c cleaved by endocytic proteases disassociates endosome; (4) primary transcription occurs within reovirus cores and capped mRNAs are released; (5) primary translation of all 10 mRNA using host translation machinery; (6) viral proteins accumulate in viral factories produced by viral nonstructural proteins; (7) new cores assemble; (8) synthesis of minus-strand RNA; (9) secondary transcription occurs within new viral cores; (10) secondary translation ampliﬁes viral proteins; (11) complete assembly of outer capsid; (12) release follows cells lysis; (A) intestinal proteases produce intermediate subviral particles (ISVPs); (B) ISVPs penetrate directly through the cell membrane. PKR and other unidentiﬁed translational control elements are proposed to block primary and/or secondary reovirus protein translation. Ras activation is suggested to release translational blocks in transformed cells

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7724 precise process(es) blocked during reovirus replication in other effectors could play a role in altering the efﬁciency the absence of Ras activation. of reovirus mRNA translation?

Undetermined role of IFN in establishing resistance to Effects of PKR activation on reovirus translation reovirus A principal mechanism for prohibiting translation As previously mentioned, PKR is a critical translation initiation in response to viral infections and environ- regulator of the type I IFN-induced antiviral response mental stress involves phosphorylation of the alpha (Samuel, 2001). Are the effects of PKR activity on subunit of eukaryotic initiation factor-2 (eIF2) by reovirus translation inhibition attributed to the basal double-stranded-RNA-dependent protein kinase PKR population or PKR induced through secondary (PKR) (Wek, 1994; de Haro et al., 1996; Brostrom signaling cascades such as IFN? Preliminary data using and Brostrom, 1998). Although PKR is always present IFN-specific antibodies suggest that reovirus protein at low levels, its expression is upregulated in response to expression is blocked in untransformed cells despite IFN released by virally infected cells. Binding to double- inhibition of autocrine IFN responses. Conversely, since stranded RNA results in PKR dimerization, autopho- reovirus protein expression in Ras transformed cells sphorylation and activation. Phosphorylation of eIF2 does not commence until 6–12 h postinfection, there is by activated PKR precludes the formation of GTP- time for additional signaling molecules such as IFN to bound eIF2 and thus prevents loading of Met-tRNA impact the resolution of virus infection. and formation of the 43S initiation complex (as is Conditional oncolysis by NS1-mutated influenza further described by Ian Mohr in this issue). virus, NDV and VSV has been found to be modulated The possible involvement of PKR in reovirus replica- by the state of IFN signaling (Pecora et al., 2002; tion was indicated by two observations: the reovirus S1 Balachandran and Barber, 2004; Muster et al., 2004). segment mRNA was previously shown to be a potent Interestingly, additional changes in translation regula- activator of PKR (Bischoff and Samuel, 1989), and tion were shown to cooperate with an ablated IFN translation of reoviral mRNA was impaired in untrans- pathway for highest efficiency of VSV replication in formed cells relative to cells transformed with activated transformed cells. It is possible that alternative path- Ras. Using a standard in vitro kinase assay to detect ways converge on similar effects, such as translational activation of PKR, Strong et al. (1998) found that PKR control, which in turn determine the outcome of viral was phosphorylated in untransformed NIH3T3 cells in infection. response to reoviral replication. Lack of PKR phosphorylation in reovirus infected Ras-transformed cells Connections between PKR and cancer suggested that reovirus translation is spared in transformed cells because PKR is not activated. These studies While PKR was first associated with the IFN-mediated did not determine whether PKR activation is a cause of, antiviral response, research has since suggested that it or a consequence of, cellular changes beneficial for serves additional purposes within the cell. The connec- reovirus infection. A relatively specific chemical inhibition between PKR, translational control and cancer has tor of PKR phosphorylation restored reovirus transla- been reviewed thoroughly (Clemens, 2004). Since tion in untransformed cells, providing evidence for a translational control permits a rapid response to direct role of PKR in determining resistance to reovirus environmental conditions, it offers a critical step in replication. Most convincingly, reovirus protein expres- regulating cellular processes. Translation initiation is sion was deficient in normal (PKR þ / þ ) mouse embryo thought to be the rate-limiting step of protein expres- fibroblasts (MEFs) but efficient in PKRÀ/À MEFs. A sion. The activities of proteins that regulate translation link between PKR inactivation and Ras signaling has initiation, including PKR, affect the status of protein already been suggested, and may be compounded by expression, and are therefore often altered during these findings on selective reovirus replication transformation. PKR is involved in cell growth control (Mundschau and Faller, 1994). At present, observations and differentiation (Chong et al., 1992; Koromilas suggest that PKR activation plays a role in determining et al., 1992; Donze et al., 1995). Dominant-negative resistance to reovirus infection. PKR mutants were found to induce tumorigenic Many questions remain to be resolved with respect to transformation of mouse fibroblasts (Meurs et al., the relationship between PKR activation and reovirus 1993; Barber et al., 1995). When overexpressed, PKR replication. What is the mechanism of increased PKR was found to induce apoptosis and has been suggested activity in normal cells: increased PKR expression, to act as a tumor suppressor (Jagus et al., 1999; Donze increased PKR phosphorylation, and/or decreased et al., 1999). PKR is also involved in signaling pathways activity of phosphatases that act on PKR and PKR that stimulate transcription of more than a hundred substrates? What role, if any, do IFNs and dsRNA specific genes (Kumar et al., 1997; Cuddihy et al., 1999). activation play in determining the outcome of reovirus Although significant findings link PKR dysfunction with infection in transformed and untransformed cells? What cancer, the presence of elevated levels of PKR kinase is the molecular connection between Ras/RalGEF/p38 activity in some human cancers has stirred debate signaling pathways and PKR activation? Is PKR the regarding the role of PKR in transformation (Nuss- sole determinant of efficient reovirus replication? What baum et al., 2003). Through our studies on reovirus

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7725 oncolysis, we hope to further understand the connection ism for specific translational silencing of reovirus between PKR and Ras signaling pathways and the role transcripts. Poorly understood specific regulatory sys- of PKR in transformation. tems have been proposed to function in separating cellular translatable (polysomal) and untranslatable free Additional cellular contributors to the block in reovirus mRNA/protein complexes (Minich and Ovchinnikov, translation? 1992; Spirin, 1994; Matsumoto et al., 1998; Mazumder et al., 2001; Skabkin et al., 2004). Translational masking While current research supports the role for PKR in as a specific antiviral response against cytoplasmic viral establishing conditional reovirus replication, we are also transcripts, however, has not yet been discovered. investigating the involvement of other protein expres- Whether translation of reovirus RNAs are actively and sion determinants that coincide with cellular transfor- specifically silenced or inherently weak relative to host mation and Ras signaling. Growth and proliferation mRNA remains to be elucidated. Whatever mechanism often correspond with multiple effects that result accounts for differential translational suppression of in increased translation initiation of proteins. In addition reovirus and cellular mRNA, the fact that it is absent in to the decrease in phosphorylation of eIF2-alpha and Ras-transformed cells suggests that it may be an PKR, translation is upregulated through increased important event following growth factor receptor activity of eIF2B (exchanges GTP for GDP on eIF2- signaling. alpha), increased phosphorylation of eIF4E (binds the 50 methylated guanosine cap structure of mRNA and recruits remaining initiation factors) and increased Why is reovirus inherently sensitive to the status of Ras availability of eIF4E through phosphorylation of the and PKR? 4E-binding protein (binds and inhibits the activity of eIF4E) (Clemens and Bommer, 1999). Interestingly, the The natural portal of entry for reovirus may explain why stress-activated p38 cell signaling cascade is associated a preference for replication in Ras-transformed cells has with increased eIF4E phosphorylation and activation not been overcome through evolution. Reovirus natu- (Waskiewicz et al., 1997). The effects of increased rally infects intestinal epithelial cells lining the digestive translation initiation during growth and transformation tract (Rubin et al., 1985). These highly proliferative cells are not seen universally, but rather impact a select migrate along the villus and undergo continual lysis and group of mRNA. Several characteristics make specific exfoliation. Since Ras signaling pathways respond to growth-dependent mRNA weakly translated in untrans- positive growth conditions and instigate cell cycle formed cells and sensitive to the changes in transla- progression, it is possible that untransformed intestinal tion initiation associated with transformation (Willis, epithelial cells have sufficient Ras activity and PKR 1999; Pickering and Willis, 2005). At a glance, it does suppression to permit reovirus infection. The ability of not appear that reovirus transcripts share the character- s3 to reduce PKR activation to some extent may be istics known to give conditional translation initia- sufficient to avoid PKR activation in the context of tion. However, it is possible that features of reoviral intestinal epithelial cells (Beattie et al., 1995). In mRNA such as the absence of 30 polyadenylation comparison, normal cells do not undergo rapid cell (and hence lack of poly(A) binding protein-mediated division, have low Ras activity and therefore remain mRNA circularization) and very short 50 UTRs make resistant to reovirus infection. The unique intestinal them poor targets for translation initiation in the environment may provide additional unforeseen expla- absence of Ras transformation. What alternative cellu- nations for why activated Ras signaling pathways are lar activities could contribute to differential transla- advantageous for reovirus replication. tion of reovirus proteins in Ras-transformed and When used as a cancer therapeutic, reovirus is untransformed cells? administered directly into the tumor or blood system, Another interesting observation is that the translation and therefore encounters barriers distinct from the of cellular proteins is spared in untransformed cells natural infection. In the context of reovirus infection despite the inhibition of reoviral protein translation. The in cells outside the intestine, this natural distinction differential treatment of viral and cellular mRNA could between cells with and without activated Ras signaling be explained by compartmentalization of reovirus pathways can be exploited to permit killing of cancer mRNA and activated PKR. If reovirus proteins are but not normal cells. translated within or around viral factories, for example, localized PKR inactivation could account for the block in translation of only reovirus transcripts. Alternatively, since reoviral mRNAs are not polyadenylated, have very From in vitro data to in vivo application in cancer therapy short 50 and 30 untranslated regions (UTRs) and are made in the cytoplasm in the absence of nuclear RNA Reovirus was first assessed in vivo as an oncolytic binding proteins, they may be distinguished from agent in 1998. Regression of tumors established from cellular transcripts. Again, mRNA degradation cannot v-erbB-transformed NIH-3T3 cells was found in explain these phenomena since transcripts appear in 65–80% of immune deficient mice with a single equivalent amounts in both transformed and untrans- intratumoral injection of reovirus (Coffey et al., 1998). formed cells. Instead, this possibility requires a mechan- More importantly, regression of tumors in immune

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7726 competent mice was also significant following repeated injection of reovirus into subcutaneous tumors has no injections of reovirus. Strikingly, reovirus was also toxic effects. Although reovirus replication and tumor shown to eliminate distal or metastatic tumors in regression looked promising in this study, these results immune competent mice (Hirasawa et al., 2003). The are not statistically significant. Reovirus oncolysis is efficacy of reovirus oncolysis was enhanced by che- presently in phase I clinical trials for intravenous motherapeutic agents. When mice were exposed to inoculation and phase II trials for recurrent malignant reovirus prior to tumor implantation, which better gliomas. mimics the natural scenario in humans, combined reovirus and chemotherapy was able to overcome the immune-mediated clearing of reovirus. These studies underscore the potential application of reovirus for Conclusions cancer therapy. As mentioned previously, mutations in the Ras proto- This review on reovirus oncolysis aims to demons- oncogene and other components in the Ras signaling trate how the study of oncolytic viruses and cancer pathway are frequently associated with cancer. Thus, converge. Understanding determinants of transforma- since the initial experiments performed on murine Ras- tion permits identification and/or engineering of onco- transformed cells in the mouse animal model, several lytic viruses. In turn, investigations on the mechanisms human cancer cell lines have been evaluated for their of selective viral oncolysis help unravel the complex selective sensitivity to reovirus oncolysis. Several human cellular pathways involved in cellular transformation, colorectal cancer and ovarian carcinoma cell lines identify new targets for cancer therapy and suggest showed both increased Ras activity and sensitivity to which cancer backgrounds are compatible with viral reovirus oncolysis as compared with normal ovarian and oncolysis strategies. colon cell lines in vitro (Hirasawa et al., 2002). Tumor The outcome of viral infection relies heavily on xenografts established from these cancer cell lines cellular processes. Therefore, viruses with an evolved showed significant regression following both intraneo- or engineered preference for conditions established in plastic and intravenous inoculation of reovirus. Success- cancer cells provide a powerful tool for selective cancer ful oncolysis in additional studies using breast-tumor, mining. Innumerable cellular changes are required for glioblastoma and medulloblastoma (common pediatric the establishment of tumorigenesis. The key to the brain tumor) derived cells extends the potential human development and use of viruses as oncolytic agents is, targets for reovirus therapy against tumors and metas- therefore, to understand the unique conditions in tases (Coffey et al., 1998; Norman et al., 2002; Yang transformed cells that prove significantly advantageous et al., 2003, 2004). to virus replication. One potential obstacle to the use of viruses in clearing Extensive progress has been made regarding the cancer is that such a treatment may select for cancer mechanism(s) of selective viral oncolysis. The diversity cells resistant to viral infection. Reovirus is predomi- of cellular and viral factors found to contribute to nantly associated with a productive, lytic infection. In conditional virus replication in transformed cells is very cell culture, however, a few cells become persistently intriguing. While the status of p53 pathways and nuclear infected with reovirus (Ahmed et al., 1981; Kauffman export of viral mRNA have been associated with et al., 1983; Dermody et al., 1993). Although the efficient replication of E1B-55K deleted adenovirus prevalence of persistent infections in vivo is not known, (ONYX-015), other recombinant adenoviruses contain- it is important to assess the potential outcome of ing deletions in E1A have been shown to preferentially reovirus persistence during viral oncolysis. Unpublished replicate in tumor cells due to aberrances in the data by Alain et al. show that the few Burkitt lymphoma retinoblastoma protein (Heise et al., 2000; McCormick, (Raji) cells resistant to killing due to a persistent 2000; O’Shea et al., 2004). Selective replication of infection with reovirus were no longer tumorigenic in oncolytic measles virus has very recently been associated mice. This fascinating finding suggests that persistent with increased CD46 receptor density on tumor cells infections, if they do occur in vivo, would not be (Anderson et al., 2004). At the same time, activated Ras detrimental to the use of reovirus as a cancer therapeu- signaling cascades and/or dysregulated translational are tic. Furthermore, when persistently infected cells were involved in selective oncolysis by several viruses cured of reovirus infection, they regained tumorigenic discussed in this review, suggesting an important role properties but were once again sensitive to reovirus for these processes in virus replication and cellular oncolysis in vivo. This finding suggests that persistent transformation. reovirus infections do not pose a threat to the efficacy of Why is dysregulated translation a common emerging reovirus cancer therapy, and suggests an interesting theme in viral oncolysis? Protein synthesis control is a novel relationship between reovirus persistence and major cellular process that is altered during transforma- tumorigenicity. tion. Changes in translation regulation are not only a Promising results with reovirus oncolysis in mouse consequence of transformation, but can cause tumor- models gave way to the testing of reovirus safety in both igenesis as well. Therefore, it is not surprising that nonprimate hosts and a phase I human clinical trial antiviral mechanisms lead to blocked viral protein (Yang et al., 2004). Results from the first phase I clinical synthesis in normal cells, and that in transformed cells, trial on 18 terminal cancer patients showed that direct dysregulated translation and compromised antiviral

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7727 mechanisms converge to allow successful virus replica- cellular processes for efﬁcient replication. Research on tion. Forthcoming research will delineate the speciﬁc reovirus oncolysis may also reveal additional common- mechanisms by which reovirus exploits malfunctioning alities among oncolytic viruses.

References

Ahmed R, Canning WM, Kauffman RS, Sharpe AH, Hallum Feig LA. (2003). Trends Cell Biol., 13, 419–425. JV and Fields BN. (1981). Cell, 25, 325–332. Feig LA and Buchsbaum RJ. (2002). Curr. Biol., 12, R259–R261. Anderson BD, Nakamura T, Russell SJ and Peng KW. (2004). Grote D, Russell SJ, Cornu TI, Cattaneo R, Vile R, Poland Cancer Res., 64, 4919–4926. GA and Fielding AK. (2001). Blood, 97, 3746–3754. Armstrong GD, Paul RW and Lee PW. (1984). Virology, 138, Hashiro G, Loh PC and Yau JT. (1977). Arch. Virol., 54, 37–48. 307–315. Balachandran S and Barber GN. (2004). Cancer Cell, 5, 51–65. Heise C, Hermiston T, Johnson L, Brooks G, Sampson- Barber GN, Jagus R, Meurs EF, Hovanessian AG and Katze Johannes A, Williams A, Hawkins L and Kirn D. (2000). MG. (1995). J. Biol. Chem., 270, 17423–17428. Nat. Med., 6, 1134–1139. Beattie E, Denzler KL, Tartaglia J, Perkus ME, Paoletti E and Hirasawa K, Nishikawa SG, Norman KL, Alain T, Kossa- Jacobs BL. (1995). J. Virol., 69, 499–505. kowska A and Lee PW. (2002). Cancer Res., 62, 1696–1701. Bergmann M, Romirer I, Sachet M, Fleischhacker R, Garcia- Hirasawa K, Nishikawa SG, Norman KL, Coffey MC, Sastre A, Palese P, Wolff K, Pehamberger H, Jakesz R and Thompson BG, Yoon CS, Waisman DM and Lee PW. Muster T. (2001). Cancer Res., 61, 8188–8193. (2003). Cancer Res., 63, 348–353. Biederer C, Ries S, Brandts CH and McCormick F. (2002). Jagus R, Joshi B and Barber GN. (1999). Int. J. Biochem. Cell J. Mol. Med., 80, 163–175. Biol., 31, 123–138. Bischoff JR and Samuel CE. (1989). Virology, 172, 106–115. Janes PW, Daly RJ, deFazio A and Sutherland RL. (1994). Bos JL. (1989). Cancer Res., 49, 4682–4689. Oncogene, 9, 3601–3608. Brostrom CO and Brostrom MA. (1998). Prog. Nucleic Acid Kauffman RS, Ahmed R and Fields BN. (1983). Virology, 131, Res. Mol. Biol., 58, 79–125. 79–87. Buday L. (1999). Biochim. Biophys. Acta, 1422, 187–204. Khosravi-Far R, White MA, Westwick JK, Solski PA, Campbell PM and Der CJ. (2004). Semin. Cancer Biol., 14, Chrzanowska-Wodnicka M, Van AL, Wigler MH and Der 105–114. CJ. (1996). Mol. Cell. Biol., 16, 3923–3933. Campbell SL, Khosravi-Far R, Rossman KL, Clark GJ and Koromilas AE, Roy S, Barber GN, Katze MG and Sonenberg Der CJ. (1998). Oncogene, 17, 1395–1413. N. (1992). Science, 257, 1685–1689. Chappell JD, Duong JL, Wright BW and Dermody TS. (2000). Kumar A, Yang YL, Flati V, Der S, Kadereit S, Deb A, J. Virol., 74, 8472–8479. Haque J, Reis L, Weissmann C and Williams BR. (1997). Chappell JD, Gunn VL, Wetzel JD, Baer GS and Dermody EMBO J., 16, 406–416. TS. (1997). J. Virol., 71, 1834–1841. Malumbres M and Barbacid M. (2001). Nat. Rev. Cancer, 1, Chong KL, Feng L, Schappert K, Meurs E, Donahue TF, 222–231. Friesen JD, Hovanessian AG and Williams BR. (1992). Malumbres M and Barbacid M. (2003). Nat. Rev. Cancer, 3, EMBO J., 11, 1553–1562. 459–465. Clemens MJ. (2004). Oncogene, 23, 3180–3188. Marshall CJ. (1996). Curr. Opin. Cell Biol., 8, 197–204. Clemens MJ and Bommer UA. (1999). Int. J. Biochem. Cell Martin-Blanco E. (2000). Bioessays, 22, 637–645. Biol., 31, 1–23. Martuza RL. (2000). J. Clin. Invest., 105, 841–846. Coffey MC, Strong JE, Forsyth PA and Lee PW. (1998). Matsumoto K, Wassarman KM and Wolffe AP. (1998). Science, 282, 1332–1334. EMBO J., 17, 2107–2121. Cohen JA, Williams WV and Greene MI. (1988). Microbiol. Mazumder B, Seshadri V, Imataka H, Sonenberg N and Fox Sci., 5, 265–270. PL. (2001). Mol. Cell Biol., 21, 6440–6449. Cuddihy AR, Li S, Tam NW, Wong AH, Taya Y, Abraham McCormick F. (2000). Semin. Cancer Biol., 10, 453–459. N, Bell JC and Koromilas AE. (1999). Mol. Cell. Biol., 19, Meurs EF, Galabru J, Barber GN, Katze MG and Hovanessian 2475–2484. AG. (1993). Proc. Natl. Acad. Sci. USA, 90, 232–236. de Haro C, Mendez R and Santoyo J. (1996). FASEB J., 10, Minich WB and Ovchinnikov LP. (1992). Biochimie, 74, 1378–1387. 477–483. Dermody TS, Nibert ML, Bassel-Duby R and Fields BN. Mundschau LJ and Faller DV. (1994). Biochimie, 76, 792–800. (1990). J. Virol., 64, 5173–5176. Muster T, Rajtarova J, Sachet M, Unger H, Fleischhacker R, Dermody TS, Nibert ML, Wetzel JD, Tong X and Fields BN. Romirer I, Grassauer A, Url A, Garcia-Sastre A, Wolff K, (1993). J. Virol., 67, 2055–2063. Pehamberger H and Bergmann M. (2004). Int. J. Cancer, Dobbelstein M. (2004). Curr. Top. Microbiol. Immunol., 273, 110, 15–21. 291–334. Norman KL, Coffey MC, Hirasawa K, Demetrick DJ, Donze O, Dostie J and Sonenberg N. (1999). Virology, 256, Nishikawa SG, DiFrancesco LM, Strong JE and Lee PW. 322–329. (2002). Hum. Gene Ther., 13, 641–652. Donze O, Jagus R, Koromilas AE, Hershey JW and Norman KL, Hirasawa K, Yang AD, Shields MA and Lee Sonenberg N. (1995). EMBO J., 14, 3828–3834. PW. (2004). Proc. Natl. Acad. Sci. USA, 101, 11099–11104. Duncan MR, Stanish SM and Cox DC. (1978). J. Virol., 28, Norman KL and Lee PW. (2000). J. Clin. Invest., 105, 444–449. 1035–1038. Fantl WJ, Johnson DE and Williams LT. (1993). Annu. Rev. Nussbaum JM, Major M and Gunnery S. (2003). Cancer Lett., Biochem., 62, 453–481. 196, 207–216. Farassati F, Yang AD and Lee PW. (2001). Nat. Cell Biol., 3, O’Shea CC, Johnson L, Bagus B, Choi S, Nicholas C, Shen A, 745–750. Boyle L, Pandey K, Soria C, Kunich J, Shen Y, Habets G,

Oncogene Molecular mechanism of reovirus oncolysis M Shmulevitz et al 7728 Ginzinger D and McCormick F. (2004). Cancer Cell, 6, Spirin AS. (1994). Mol. Reprod. Dev., 38, 107–117. 611–623. Stojdl DF, Lichty B, Knowles S, Marius R, Atkins H, Ouwens DM, de Ruiter ND, van der Zon GC, Carter AP, Sonenberg N and Bell JC. (2000). Nat. Med., 6, 821–825. Schouten J, van der BC, Kooistra K, Bos JL, Maassen JA Strong JE, Coffey MC, Tang D, Sabinin P and Lee PW. and van DH. (2002). EMBO J., 21, 3782–3793. (1998). EMBO J., 17, 3351–3362. Overbeck AF, Brtva TR, Cox AD, Graham SM, Huff SY, Strong JE and Lee PW. (1996). J. Virol., 70, 612–616. Khosravi-Far R, Quilliam LA, Solski PA and Der CJ. Strong JE, Tang D and Lee PW. (1993). Virology, 197, (1995). Mol. Reprod Dev., 42, 468–476. 405–411. Pacitti AF and Gentsch JR. (1987). J. Virol., 61, 1407–1415. Tyler KL. (2001). Fields Virology, 4th edn. Knipe DM and Pawson T. (1994). Adv. Cancer Res., 64, 87–110. Howley PM (eds). Lippincott-Raven: Philadelphia, PA, pp. Pawson T and Scott JD. (1997). Science, 278, 2075–2080. 1729–1945. Pecora AL, Rizvi N, Cohen GI, Meropol NJ, Sterman D, Tzahar E and Yarden Y. (1998). Biochim. Biophys. Acta, 1377, Marshall JL, Goldberg S, Gross P, O’Neil JD, Groene WS, M25–M37. Roberts MS, Rabin H, Bamat MK and Lorence RM. (2002). Ullrich A and Schlessinger J. (1990). Cell, 61, 203–212. J. Clin. Oncol., 20, 2251–2266. Waskiewicz AJ, Flynn A, Proud CG and Cooper JA. (1997). Pickering BM and Willis AE. (2005). Semin. Cell Dev. Biol., EMBO J., 16, 1909–1920. 16, 39–47. Weiss A and Schlessinger J. (1998). Cell, 94, 277–280. Repasky GA, Chenette EJ and Der CJ. (2004). Trends Cell Wek RC. (1994). Trends Biochem. Sci., 19, 491–496. Biol., 14, 639–647. White MA, Nicolette C, Minden A, Polverino A, Van AL, Reuther GW and Der CJ. (2000). Curr. Opin. Cell Biol., 12, Karin M and Wigler MH. (1995). Cell, 80, 533–541. 157–165. White MA, Vale T, Camonis JH, Schaefer E and Wigler MH. Rodriguez-Viciana P, Warne PH, Khwaja A, Marte BM, (1996). J. Biol. Chem., 271, 16439–16442. Pappin D, Das P, Waterﬁeld MD, Ridley A and Downward Willis AE. (1999). Int. J. Biochem. Cell Biol., 31, 73–86. J. (1997). Cell, 89, 457–467. Wolthuis RM and Bos JL. (1999). Curr. Opin. Genet. Dev., 9, Rubin DH, Kornstein MJ and Anderson AO. (1985). J. Virol., 112–117. 53, 391–398. Yang WQ, Lun X, Palmer CA, Wilcox ME, Muzik H, Shi ZQ, Samuel CE. (2001). Clin. Microbiol. Rev., 14, 778–809, table. Dyck R, Coffey M, Thompson B, Hamilton M, Nishikawa Schneider-Schaulies J. (2000). J. Gen. Virol., 81, 1413–1429. SG, Brasher PM, Fonseca K, George D, Rewcastle NB, Sieczkarski SB and Whittaker GR. (2005). Curr. Top. Johnston RN, Stewart D, Lee PW, Senger DL and Forsyth Microbiol. Immunol., 285, 1–23. PA. (2004). Clin. Cancer Res., 10, 8561–8576. Sinkovics JG and Horvath JC. (2000). J. Clin. Virol., 16, 1–15. Yang WQ, Senger D, Muzik H, Shi ZQ, Johnson D, Brasher Skabkin MA, Kiselyova OI, Chernov KG, Sorokin AV, PM, Rewcastle NB, Hamilton M, Rutka J, Wolff J, Dubrovin EV, Yaminsky IV, Vasiliev VD and Ovchinnikov Wetmore C, Curran T, Lee PW and Forsyth PA. (2003). LP. (2004). Nucleic Acids Res., 32, 5621–5635. Cancer Res., 63, 3162–3172.

Oncogene