Therapy (2005) 12, S111–S117 & 2005 Nature Publishing Group All rights reserved 0969-7128/05 $30.00 www.nature.com/gt CONFERENCE PAPER Cellular factors influencing Semliki Forest vector biology

E Piver1, C Collin2, A Diatta1, P Vaudin1 and J-C Page`s1 1Universite´ Franc¸ois Rabelais, The Vector Group, Faculte´ de Me´decine, Tours, France; and 2Biochemistry, Trousseau University Hospital, Tours, France

Viral vectors are currently the best tools for gene delivery in transduction, and transgene expression, in view of the cellular a therapeutic setting, especially for in vivo use. , factors that affect each individual step. Classical virology will a family of positive singlestranded RNA , have been benefit from the knowledge accumulated studying vectors, and engineered to allow the formation of a highly efficient replicon. such work will shed light on crosstalk between intruding Using these replicons, it is possible to generate recombinant viruses and their hosts. Ultimately, these data should help the particles. Parental viruses and recombinant vectors share design of vectors adapted to specific target cells. certain pathways while interacting with their target cells. In this (2005) 12, S111–S117. doi:10.1038/ review, we describe the consecutive events leading to sj.gt.3302625

Keywords: VLP; endocytosis; TLR; IFN and SFV vectors

Introduction noteworthy that vectors would take advantage of a precise and complete knowledge of the cell cycle. Virally driven tumour cytotoxicity is an expanding field A companion paper by Lundstrom10 gives a useful in antitumoral research.1–6 Technically, it relies on two overview of the main steps of the alphavirus life cycle, exclusive modus operandi, tumour-restricted cytolytic introducing the essential features needed to understand replicative viruses or suicidal recombinant viral vectors. alphavirus vectors. In addition, we will comment on Whereas several viral genera have been employed for the use of an unconventional trans-complementation both approaches, only adenovirus based agents are approach for the production of SFV vectors.16 We have currently in the clinic.7,8 recently described an efficient procedure for the mobi- Among the viral vectors, those derived from alpha- lization of SFV replicons by the means of virus-like viruses occupy a specialized niche.9,10 Semliki Forest particles (VLP) (Figure 1). These particles are made of a Virus (SFV) and (SV) are the main two membrane-anchored vesicular stomatitis virus glyco- alphaviruses from which vectors have been obtained.11,12 protein (VSV-G) protein forming a lipo-protein shell In agreement with their biology, these vectors are encapsulating a full-length SFV replicon.16 More recently predominantly adapted to high levels of transient we have shown that the biological behaviour of the VLP transgene expression. This results from an optimized resembles that of viral particles in several ways, leading cytoplasmic genomic replication combined with the to cell entry (Diatta et al unpublished data). We also cytotoxic effect induced by a number of virally encoded observed unexpected features that triggered a series proteins.9 The clinical applications of SFV-derived of experiments dedicated to the understanding of the vectors are therefore restrained to situations where such relationship established between the SFV replicon and its a pattern of expression appears suitable, namely, genetic- cellular environment. based vaccination and the eradication of an invasive cell As noticed in other viral taxa, such as , proliferation, as observed in . Moreover, recent viruses and cells have evolved several molecular mechan- experimental reports showed that alphavirus-derived isms that rule their biological association.17 In the present vectors harbour an unexpected specificity to replication review, we will concentrate on alphavirus-derived vectors in tumour cells.13–15 Nevertheless, numerous features of biology with a specific interest on cellular factors. Host SFV biology remain open to clarification. This knowledge gene products exert a dual effect; on the one hand, some might lead to technological improvements and extend proteins positively participate in viral replication and, the use of SFV-derived vectors. conversely, when induced upon infection, other cellular In order to develop better tools, two topics are the focus products aim to hamper viral propagation. of ongoing studies, biosafety issues as well as the understanding of cancer cell replication specificities. It is Replicon is the basic structure of SFV vectors Correspondence: J-C Page`s, Universite´ Franc¸ois Rabelais, EA 3856-IFR 136, The Vector Group, Faculte´ de Me´decine 10 Bd Tonnelle´, 37000 Tours, Shortly after the genetic organization and sequence of France alphavirus was determined, several vector systems were Cellular factors and SFV vectors E Piver et al S112

Figure 1 Representation of the different steps in VLP production process. At day 1, the transcomplementing, envelope expressing plasmid is transfected; followed, at day 2, by a DNA form of the replicon. Viral harvest is performed at day 3. CMV: immediate-early cytomegalovirus promoter. VSV-G: vesicular stomatitis virus glycoprotein. NsP: nonstructural proteins. R.S: retroviral sequence. IRES: internal ribosome entry site. EGFP: enhanced GFP.

engineered to promote gene delivery and expression vectors must preserve some features of the regular viral from various strains of alphavirus, including SFV.18,19 cycle while eliminating others. The most striking The most striking feature of an SFV-derived vector is its difference between nonreplicating viral vectors and the genetic structure, which allows the generation of a self- wild type parental virus, resides in the late events replicating element, the replicon. A SFV replicon is governing particle formation, which are nondesirable composed of a single RNA that contains all the genetic most of the time and are therefore hampered in the case information needed for self-amplification and transgene of vectors. Conversely, early events that allow efficient expression. There are two ways by which the replicon transgene delivery must be indistinguishable in the two could be delivered to a target cell.4,20–22 The most species. It is noteworthy that each step of the viral cycle common system is based on RNA or DNA transfection. is susceptible to target-cell-specific modulation. For in vivo use, the formation of recombinant particles is possible through trans-complementation. However, in Entry the case of recombinant particle formation, safety issues Classically, the sequential events leading to delivery of must be considered since recombination events are the viral genomic material are divided into entry, frequent and lead to fully replicative autonomous 23 including decapsidation, which is followed by the particles. Genetic modification of the trans-comple- cytoplasmic addressing of the viral .27 At the mentation vectors has recently improved the safety of 23 end of the latter process, the viral genome is addressed to SFV vectors. Considering the biosafety, the VLP offer the cellular compartment where it will ultimately a relevant alternative to recombinant SFV. One key replicate and drive expression of the viral gene products. advantage conferred by the VLP comes from the Efficient, virally promoted, cellular entry is the raison heterologous origin of the trans-complementing materi- 16 d’eˆtre for the use of recombinant viruses in gene transfer. al. Of note, in this approach, replicon packaging is not All viruses, including alphaviruses, enter a target cell based on an interaction between the SFV and using a defined pathway that involves a virally encoded a sequence contained within the nonstructural protein protein expressed at the particle surface and a cellular (nsP) 2 open reading frame of the full length SFV RNA. receptor at the plasma membrane.27 First, we shall Thus, we had to find a substitutive method to achieve the consider the envelope protein density at the particle exclusive packaging of a complete replicon. Indeed, Rolls 24,25 surface, together with the receptor exposed at the cell et al showed that alike VLP were able to package surface. For retroviruses, these two features are known genomic RNA as efficiently as sub-genomic. The sub- to directly influence entry.28–30 In the case of SFV vectors, genomic RNAs that don’t have auto-replicative capabil- the high amount of envelope proteins generated through ities are not suitable for gene transfer approaches. sub genomic translation appears to prevent a deficient Therefore, we have introduced several mutations into envelope filling of the particles. Conversely, receptor the internal 26S promoter. In the absence of sub-genomic based restriction has been described for SFV.31 While the RNA synthesis, we obtained an exclusive packaging of precise receptor for SFV is still unknown, it is acknowl- the fully autonomous genomic RNA. To obtain success- edged to be widely expressed.32 Therefore, cells resistant ful transgene expression, it was hence necessary to use to transduction might express the protein but fail to an internal translation initiation, which can efficiently be make it available. In some circumstances, this defect in obtained by using a picorna virus internal ribosome 26 receptor accessibility could be circumvented by the use entry sequence (IRES). of nonspecific chemical agents that promote enveloped virus–cellular membrane interaction.31 Moreover, as for other enveloped virus pseudotyping, a simple procedure Molecular events along the process that allows receptor shift can provide a means to 33–35 of cell transduction strengthen vector–cell affinity. For SFV vectors, the VSV-G VLP achieves some kind of pseudotyping, We will now concentrate on the cellular events observed leading to a broad entry tropism.16,24,25,36 Of note, other when using recombinant SFV particles or VLP (Figure 2). procedures involving the formation of heterologous To get efficient transgene delivery, recombinant viral particles are possible to expose various envelopes at

Gene Therapy Cellular factors and SFV vectors E Piver et al S113

Figure 2 Schematic cartoon of the cellular and virus interaction during infection. the surface of SFV replicon containing particles (E Piver is also the case for the derived vector particles. Upon and JC Page`s unpublished data). acidification, a conformational envelope change pro- Following receptor–envelope interaction, the complex motes an endosome disruption and viral delivery. Of has several possible routes to deliver the viral genomic note, the membrane cholesterol content appears to material within the target cell. For these postbinding influence the efficacy of SFV entry.38–40 Concerning the events, it is common to distinguish a pH-dependent and VSV-G VLP, we have shown that entry promoted by the a pH-independent pathway.27,37 The latter process allows VSV-G was also dependent on pH lowering, as for a direct fusion at the cell surface, while the former the parental virus. Moreover, nystatin C, a cholesterol- involves an endosome-mediated internalization. SFV is a sequestering agent, has no obvious effect on VLP prototypic virus for endosome-dependent entry and this mediated replicon delivery (A Diatta unpublished

Gene Therapy Cellular factors and SFV vectors E Piver et al S114 results). We have also shown a partial inhibition of gens.45 According to the wide diversity of pathogenic transduction when the target cell was preincubated agents, the innate cellular defence response is triggered with brefeldine A, a compound that displays a pleio- by nucleic acids, lipo-polysaccharide or proteinaceous tropic action on the vesicular traffic within cells.41 These material. The two most studied systems include the IFN- data confirm the early endosome-mediated mode of induced response and the Toll-like receptor-driven (TLR) entry of the VLP and suggest that the reduction, molecular cascade.46,47 Of note, these two pathways are in replicon expression observed using brefeldin A might overlapping, the TLR pathway being able to trigger IFN result from impairment in cytoplasmic vacuole (CPV) (a/b) synthesis. The molecular effectors involved in IFN formation. and TLR response are broadly distributed throughout Following entry, cytoplasmic SFV replicons are asso- cell types, despite some specificity. Of the two classes of ciated with membrane structure, plasma membrane as IFN, only IFN I is active on most cell types, IFN g well as endosomal and CPV derivative.42 These cellular essentially being restricted to antigen-presenting cells structures are formed from endosomes and lysosomes and NK. In the perspective of vector use in vaccination 3 h after an SFV infection, and appear crucial for plus and cancer therapy, we need to understand the different strand RNA synthesis.43 The nsP 1 has an essential role in cellular modifications induced by these molecules upon determining membrane association (Figure 3).42–44 This replicon expression. Anticancer or immune responses are protein contains two membrane-targeting signals, a expected to be maximal if the transduced cells express stretch of 19 charged and hydrophobic amino acids the replicon at high levels. Therefore, we have to avoid interacting with phospholipids, and three palmitoylated delivery of the vector into a cell type where replication, cysteins. It is noteworthy that the membrane association and therefore cytotoxicity, is severely hampered by the of nsP 1 regulates the enzymatic activity of the protein.43 innate response. Conversely, during the vector-produ- Whereas nsP 1 missing the palmitoylation signal is still cing step, we would like to take advantage of a weakly associated with the plasma membrane, replicons comprehensive modulation of the replicon-induced harbouring this mutation are severely impaired in their cytotoxicity to obtain high recombinant titre. This is replication (E Piver unpublished data and A Merits specifically the case when using the VSV-G VLP, where personal communication). Therefore, membrane target- the VSV-G is produced by the RNA Pol II-dependent ing might be essential to SFV replication. cellular machinery.11,16 Any reduction in the trans- expression of the VSV-G directly affects the titre and Innate immunity and virally induced cytotoxicity thereby the utilization opportunities of the produced Cells have evolved a series of molecular mechanisms particles. devoted to the elimination of foreign intruding patho- We shall now consider the proteins involved in the two pathways. The double-stranded RNA-dependent protein kinase (PKR) is controlled by the IFN I family of proteins. In SFV biology, PKR protein has proved to be involved in the downregulation of cellular mRNA translation. The PKR harbours an autocatalytic kinase activity that converts an inactive monomeric into a fully active dimeric phosphorylated enzyme that will phosphorylate various cellular factors. The presence of cytoplasmic ds RNA acts as an essential co-factor that controls the activity status of the PKR. When activated, the PKR phosphorylates several targets including the elongation factor 2 a (elF2a) at position 51.48,49 Upon phosphorylation, elF2a is inactivated and therefore induces a broad translation inhibition. During vector production, it might therefore be suitable to control this response. The use of a negative transdominant mutant of PKR has been exploited to modulate SFV cytotoxicity. Alternatively, cells expressing a mutated phosphorylation defective elF2a could be advantageous. Nevertheless, a recent report indicated that not all the protein synthesis shutoff is explained by the PKR activation.50 Other inhibiting pathways remain to be characterized. Among the IFN I responsive , the MxA gene product has been shown to inhibit SFV replication through protein–protein interaction.51 While it is obvious that the cellular localization of SFV/MxA interaction is cytoplasmic, the precise molecular mechan- isms explaining the inhibition is presently not clear. It is noteworthy that using a transdominant negative mutant of the MxA protein, we have not been able to significantly improve SFV vector expression (A Diatta Figure 3 Localization of the VSV-G and the SFV replication complex in and JC Page`s unpublished data). The last IFN producing cells. VSV-G in green upper panel in red middle picture. I-dependent factor that could negatively influence SFV Overlay lower panel. replicon function is the RNase L, which is involved in

Gene Therapy Cellular factors and SFV vectors E Piver et al S115 mRNA and rRNA hydrolysis.52 Interestingly, several Expression viruses have selected a series of proteins in order to Classically, transgene expression from SFV vectors is neutralize each of these factors. It could be worth obtained using one or more internal 26S promoters.10,22 envisioning the use of such proteins to increase SFV- High levels of subgenomic RNA or transgene expression derived vector titre in producing cells. is achieved late in the infection cycle. At the time of The TLR family of proteins has a central role in innate transgene expression, most of the cellular functions are immunity and contains 10 members.47,53 The system subject to virally induced extinction. Therefore, it has to functions at different cellular levels crucial for virus be kept in mind that, if the transgene needs to be entry, by using compartmentalized receptors sensitive to expressed together with a cellular cofactor, this product a specific type of agent. Despite this diversity of sensors, should probably be also vectorized to be in sufficient the transduction pathways activated upon TLR binding amount. Here, the use of poorly cytotoxic-derived are convergent and share several of the proteins vectors is advantageous. However, as mentioned above, transducing the information. While several TLR have the relevancy of attenuated SFV vectors is less evident. been reported to respond to either ss or ds RNA, to An alternative to the use of an internal promoter is to date, SFV replication appears not to switch on a TLR clone the transgene 30 to an IRES.16 An IRES offers the pathway. This is surprising since the TLR 3 recognizes advantage of an expression arising earlier in the SFV life ds RNA, which is actively produced in SFV infected cycle. Nevertheless, an IRES needs certain cellular factors cells. However, other viruses with ds RNA inter- for optimal activity.63 The lack of availability of these mediate proved to replicate with the same efficiency factors might lead to reduced transgene expression, as in TLR3+/+ or TLR3À/À mice.54 One explanation for this has been reported in HeLa cells. The length of the polyA observation could be that the viral mode of entry tail also has an influence on gene expression.64 More employed by SFV does not allow TLR3 activation, sophisticated systems of translational control are now despite the fact that TLR3 is found within the endocytic available and could be beneficially exploited.65 compartment. Recently, ZAP, a zinc-finger protein has been shown to Safety modulate infection of retroviruses by interacting with Safety concerns arise when using recombinant SFV 17,55,56 their transcription products. Furthermore, this particles. This essentially results from the generation of protein also proved to exert a potent inhibitory effect autonomous particles at the stage of vector production. on alphaviruses. Considering the exclusive cytoplasmic As SFV has been rarely implicated in severe infection in life cycle of the SFV, the nuclear–cytoplasm shuttling of humans, any contamination must be prevented. Cyto- 57 ZAP might explain this observation. Now, bearing in plasmic RNA recombination leads to the reconstitution mind the dramatic effect of ZAP on alphavirus replica- of complete viral when producing recombinant tion, the protein is certainly absent within the cells particles using two complementary RNAs. If they exist, classically used to produce SFV recombinant particles. cellular determinants driving recombination are not yet ZAP might therefore have a negative role only in target identified. As BHK21 cells are the most widely used for cells. Together with its mode of action, it will be vector production, no cell type differences have been interesting to unravel the factors controlling ZAP reported concerning the likelihood of recombination. It is expression. also possible that the recombination events arise through Apoptosis, programmed cell death, is a common event template switching promoted by the replicase, as this is induced by several cytolytic viruses. Likewise, alpha- the case for retroviruses.66,67 The use of split genomes, virus-infected cells have been shown to undergo apop- together with the introduction of conditional mutations 58 tosis. Among viral proteins, nsP 2, which contains a into structural genes, has strongly reduced the likelihood nuclear localization signal, is thought to actively parti- of these recombination events.23,68 However, a risk still 13,59,60 cipate in the apoptosis of replicon-expressing cells. persists. The recent advent of an easy to handle system, This is, at least partially, confirmed by the study of dedicated to the control of translation, will certainly naturally arising variants, which show a milder pheno- provide an excellent tool to reinforce the safety of type when injected into mice. Upon sequence analysis, recombinant viral-vector production.65 For example, it the largest part of these mutants proved to harbour nsP 2 could be worth introducing, within the trans-comple- 58 61 modifications. Similarly, Perri et al have described a menting constructs, a constitutive nsP shutdown RNA. procedure for the in vitro selection of replicons with no Following transfection of the complementary plasmid, as cytotoxicity. Whereas single mutations have been un- in the standard procedure to generate recombinant successful in giving rise to long-term in vivo expression, a particles, nsP translation inhibition would be relieved combination of mutations has allowed the generation of through drug addition. If an autonomous recombinant 20,62 a long-lasting replicon. It is noteworthy that the species is generated, it should always be silent in the variants show a cell type-dependent phenotype and that absence of the defined drug. Moreover, to get rid of the the level of transgene expression is lower than that of silencing sequence, one supplementary recombination 16 wild-type replicons. Furthermore, such inferior replica- event is needed. The strategy therefore reduces the odds tion could be detrimental to the recombinant vector of autonomous particle formation. production process using the VSV-G. Finally, as for innate immunity, the development of an apoptosis-free replicon is of limited interest for the main SFV vector Conclusion applications, cancer therapy and vaccination. The use of exogenous apoptosis regulators seems more relevant, In the case of persistent nonlytic viruses, sometimes allowing a transient control of cell death only during the leading to a state of latency, host/virus interactions are production steps. generally complex. The lentivirus HIV-1 is one of the

Gene Therapy Cellular factors and SFV vectors E Piver et al S116 most studied models revealing continuous crosstalk- 11 Boorsma M et al. Alphavirus cDNA-based expression vectors: between the virus and its host. Conversely, less is known effects of RNA transcription and nuclear export. Biotechnol about host factors that modulate the phenotype of Bioeng 2003; 81: 553–562. cytopathogenic viruses. In the case of alphaviruses, there 12 Ehrengruber MU. Alphaviral gene transfer in neurobiology. is no doubt that certain aspects of their life cycle Brain Res Bull 2002; 59: 13–22. are influenced by host factors. This is highlighted 13 Tseng JC et al. In vivo antitumor activity of Sindbis viral vectors. by the description of some cell specificity of cytotoxicity. J Natl Cancer Inst 2002; 94: 1790–1802. A significant example for this comes from the unique 14 Tseng JC et al. Systemic tumor targeting and killing by Sindbis profile of viral replication. Within mosquitoes, which are viral vectors. Nat Biotechnol 2004; 22: 70–77. Epub 2003 Nov 2030. the first essential hosts for virus propagation, replication 15 Tseng JC et al. Using Sindbis viral vectors for specific detection appears less deleterious than in where replica- and suppression of advanced ovarian cancer in animal models. Cancer Res 2004; 64: 6684–6692. tion leads to pathology. In addition, the spontaneous 16 Dorange F et al. Vesicular stomatitis virus glycoprotein: a arising of nonpathogenic viral strains indicates that such transducing coat for SFV-based RNA vectors. J Gene Med 2004; crosstalk could give rise to equilibrium between the 6: 1014–1022. parasite and the host. The use of vector in animal models 17 Goff SP. restriction factors. Mol Cell 2004; 16: 849–859. has suggested that any means by which we can control 18 Higgs S, Powers AM, Olson KE. Alphavirus expression systems: SFV cytotoxicity might be useful. Accordingly, forced applications to vector studies. Parasitol Today 1993; 9: evolution of vectors has allowed the generation of poorly 444–452. cytolytic SFV variants, allowing long-lasting expression. 19 Liljestrom P. Alphavirus expression systems. Curr Opin Biotech- Nevertheless, the direction toward which we have to nol 1994; 5: 495–500. push vector adaptation is complex since it depends on 20 Lundstrom K, Abenavoli A, Malgaroli A, Ehrengruber MU. the desired effect. Whereas it is important to moderate Novel Semliki Forest virus vectors with reduced cytotoxicity and cytotoxicity within producing cells, there is a need for temperature sensitivity for long-term enhancement of transgene cytotoxicity for some applications, such as cancer gene expression. Mol Ther 2003; 7: 202–209. therapy. A further exciting challenge is to achieve tight, 21 Polo JM et al. Stable alphavirus packaging cell lines for Sindbis cell type-specific, replication. Currently, we are far from virus and Semliki Forest virus-derived vectors. Proc Natl Acad Sci achieving this aim with vectors that exclusively replicate USA 1999; 96: 4598–4603. in the cytoplasm. 22 Schlesinger S. Alphavirus vectors: development and potential therapeutic applications. Expert Opin Biol Ther 2001; 1: 177–191. 23 Smerdou C, Liljestrom P. Two-helper RNA system for produc- tion of recombinant Semliki Forest virus particles. J Virol 1999; Acknowledgements 73: 1092–1098. 24 Rolls MM, Haglund K, Rose JK. Expression of additional genes We thank GJ Towers for careful reading of the manu- in a vector derived from a minimal RNA virus. Virology 1996; script and the AFM and VLM for their financial support. 218: 406–411. 25 Rolls MM, Webster P, Balba NH, Rose JK. Novel infectious particles generated by expression of the vesicular stomatitis virus glycoprotein from a self-replicating RNA. Cell 1994; 79: References 497–506. 26 Martinez-Salas E, Fernandez-Miragall O. Picornavirus IRES: 1 Everts B, Poel HG. Replication-selective oncolytic viruses in the structure function relationship. Curr Pharm Des 2004; 10: treatment of cancer. Cancer Gene Ther 2004; 8:8. 3757–3767. 2 Fielding AK. as a potential . Rev Med 27 Sieczkarski SB, Whittaker GR. Viral entry. Curr Top Microbiol Virol 2004; 16: 16. Immunol 2005; 285: 1–23. 3 Larin SS, Georgiev GP, Kiselev SL. Gene transfer approaches in 28 Pages JC, Bru T. Toolbox for retrovectorologists. J Gene Med 2004; cancer immunotherapy. Gene Therapy 2004; 11 (Suppl 1): S18–S25. 6 (Suppl 1): S67–S82. 4 Leitner WW et al. Enhancement of tumor-specific immune 29 Bachrach E et al. Efficient cell infection by Moloney murine response with plasmid DNA replicon vectors. Cancer Res 2000; leukemia virus-derived particles requires minimal amounts of 60: 51–55. envelope glycoprotein. J Virol 2000; 74: 8480–8486. 5 Smyth JW, Fleeton MN, Sheahan BJ, Atkins GJ. Treatment of 30 Bachrach E et al. Effects of virion surface gp120 density on rapidly growing K-BALB and CT26 mouse tumours using infection by HIV-1 and viral production by infected cells. Semliki Forest virus and its derived vector. Gene Therapy 2004; Virology 2005; 332: 418–429. 16: 16. 31 Arudchandran R et al. Polyethylene glycol-mediated infection of 6 Zhang J, Frolov I, Russell SJ. Gene therapy for malignant glioma non-permissive mammalian cells with Semliki Forest virus: using Sindbis vectors expressing a fusogenic membrane glyco- application to signal transduction studies. J Immunol Methods protein. J Gene Med 2004; 6: 1082–1091. 1999; 222: 197–208. 7 Morley S et al. The dl1520 virus is found preferentially in tumor 32 Strauss JH, Strauss EG. The alphaviruses: gene expression, tissue after direct intratumoral injection in oral carcinoma. Clin replication, and evolution. Microbiol Rev 1994; 58: 491–562. Cancer Res 2004; 10: 4357–4362. 33 Ohno K et al. Cell-specific targeting of Sindbis virus vectors 8 Shah AC, Benos D, Gillespie GY, Markert JM. Oncolytic viruses: displaying IgG-binding domains of protein A. Nat Biotechnol clinical applications as vectors for the treatment of malignant 1997; 15: 763–767. gliomas. J Neurooncol 2003; 65: 203–226. 34 Perri S et al. An alphavirus replicon particle chimera derived 9 Rheme C, Ehrengruber MU, Grandgirard D. Alphaviral from venezuelan equine and sindbis viruses is a cytotoxicity and its implication in vector development. Exp potent gene-based delivery vector. J Virol 2003; 77: Physiol 2004; 12: 12. 10394–10403. 10 Lundstrom K. Biology and application of alphaviruses in gene 35 Sandrin V, Russell SJ, Cosset FL. Targeting retroviral and therapy. Gene Therapy 2005; 12 (suppl 1): S92–S97. lentiviral vectors. Curr Top Microbiol Immunol 2003; 281: 137–178.

Gene Therapy Cellular factors and SFV vectors E Piver et al S117 36 Lebedeva I, Fujita K, Nihrane A, Silver J. Infectious particles 52 Terenzi F et al. The antiviral PKR and RNase L suppress derived from Semliki Forest virus vectors encoding murine gene expression from viral and non-viral based vectors. Nucleic leukemia virus envelopes. J Virol 1997; 71: 7061–7067. Acids Res 1999; 27: 4369–4375. 37 Pelkmans L, Helenius A. Insider information: what viruses tell 53 Matsumoto M, Funami K, Oshiumi H, Seya T. Toll-like receptor us about endocytosis. Curr Opin Cell Biol 2003; 15: 414–422. 3: a link between toll-like receptor, interferon and viruses. 38 Vashishtha M et al. A single point mutation controls the Microbiol Immunol 2004; 48: 147–154. cholesterol dependence of Semliki Forest virus entry and exit. 54 Edelmann KH et al. Does Toll-like receptor 3 play a biological J Cell Biol 1998; 140: 91–99. role in virus infections? Virology 2004; 322: 231–238. 39 Waarts BL, Bittman R, Wilschut J. Sphingolipid and cholesterol 55 Gao G, Guo X, Goff SP. Inhibition of retroviral RNA production dependence of alphavirus membrane fusion. Lack of correlation by ZAP, a CCCH-type zinc finger protein. Science 2002; 297: with lipid raft formation in target liposomes. J Biol Chem 2002; 1703–1706. 277: 38141–38147. 56 Guo X et al. The zinc finger antiviral protein directly binds to 40 Lu YE, Cassese T, Kielian M. The cholesterol requirement for specific viral mRNAs through the CCCH zinc finger motifs. sindbis virus entry and exit and characterization of a spike J Virol 2004; 78: 12781–12787. protein region involved in cholesterol dependence. J Virol 1999; 57 Liu L, Chen G, Ji X, Gao G. ZAP is a CRM1-dependent 73: 4272–4278. nucleocytoplasmic shuttling protein. Biochem Biophys Res 41 Hunziker W, Whitney JA, Mellman I. Brefeldin A and the Commun 2004; 321: 517–523. endocytic pathway. Possible implications for membrane traffic 58 Li ML, Stollar V. Alphaviruses and apoptosis. Int Rev Immunol and sorting. FEBS Lett 1992; 307: 93–96. 2004; 23: 7–24. 42 Kujala P et al. Biogenesis of the Semliki Forest virus RNA 59 Fazakerley JK, Boyd A, Mikkola ML, Kaariainen L. A single replication complex. J Virol 2001; 75: 3873–3884. amino acid change in the nuclear localization sequence of the 43 Kaariainen L, Ahola T. Functions of alphavirus nonstructural nsP2 protein affects the neurovirulence of Semliki Forest virus. proteins in RNA replication. Prog Nucleic Acid Res Mol Biol 2002; J Virol 2002; 76: 392–396. 71: 187–222. 60 Glasgow GM et al. The Semliki Forest virus vector induces 44 Laakkonen P, Ahola T, Kaariainen L. The effects of palmitoyla- p53-independent apoptosis. J Gen Virol 1998; 79: 2405–2410. tion on membrane association of Semliki forest virus RNA 61 Perri S et al. Replicon vectors derived from Sindbis virus and capping enzyme. J Biol Chem 1996; 271: 28567–28571. Semliki forest virus that establish persistent replication in host 45 Katze MG, He Y, Gale Jr M. Viruses and interferon: a fight for cells. J Virol 2000; 74: 9802–9807. supremacy. Nat Rev Immunol 2002; 2: 675–687. 62 Jeromin A et al. A modified Sindbis vector for prolonged gene 46 Sarkar SN, Sen GC. Novel functions of proteins encoded by viral expression in neurons. J Neurophysiol 2003; 90: 2741–2745. stress-inducible genes. Pharmacol Ther 2004; 103: 245–259. 63 Choi K et al. Identification of cellular proteins enhancing 47 Malmgaard L. Induction and regulation of IFNs during viral activities of internal ribosomal entry sites by competition with infections. J Interferon Cytokine Res 2004; 24: 439–454. oligodeoxynucleotides. Nucleic Acids Res 2004; 32: 1308–1317, 48 Sudhakar A et al. Phosphorylation of serine 51 in initiation factor Print 2004. 2 alpha (eIF2 alpha) promotes complex formation between eIF2 64 Thoma C et al. Enhancement of IRES-mediated translation of the alpha(P) and eIF2B and causes inhibition in the guanine c-myc and BiP mRNAs by the poly(A) tail is independent of nucleotide exchange activity of eIF2B. Biochemistry 2000; 39: intact eIF4G and PABP. Mol Cell 2004; 15: 925–935. 12929–12938. 65 Bayer TS, Smolke CD. Programmable ligand-controlled ribor- 49 Kaufman RJ. Regulation of mRNA translation by protein folding egulators of eukaryotic gene expression. Nat Biotechnol 2005; 23: in the endoplasmic reticulum. Trends Biochem Sci 2004; 29: 152–158. 337–343. 50 Gorchakov R et al. PKR-dependent and -independent mechan- 66 Moumen A et al. Evidence for a mechanism of recombination isms are involved in translational shutoff during Sindbis virus during reverse transcription dependent on the structure of the infection. J Virol 2004; 78: 8455–8467. acceptor RNA. J Biol Chem 2003; 278: 15973–15982. 51 Landis H et al. Human MxA protein confers resistance to Semliki 67 Negroni M, Buc H. Mechanisms of retroviral recombination. Forest virus and inhibits the amplification of a Semliki Forest Annu Rev Genet 2001; 35: 275–302. virus-based replicon in the absence of viral structural proteins. 68 Fayzulin R et al. Sindbis virus with a tricomponent genome. J Virol 1998; 72: 1516–1522. J Virol 2005; 79: 637–643.

Gene Therapy