Hybrid Alphavirus–Rhabdovirus Propagating Replicon Particles Are Versatile and Potent Vaccine Vectors

Hybrid alphavirus–rhabdovirus propagating replicon particles are versatile and potent vaccine vectors

Nina F. Rose, Jean Publicover, Anasuya Chattopadhyay, and John K. Rose*

Department of Pathology, Yale University School of Medicine, New Haven, CT 06510

Edited by Robert A. Lamb, Northwestern University, Evanston, IL, and approved February 20, 2008 (received for review January 11, 2008) Self-propagating, infectious, virus-like particles are generated in infected cells (12) and could be precursors involved in formation animal cell lines transfected with a Semliki Forest virus RNA of the infectious particles containing VSV G (6). replicon encoding a single viral structural protein, the vesicular Experimental SFV particle-based vaccines are normally de- stomatitis virus (VSV) glycoprotein. We show here that these rived from a complementation/packaging system in which SFV infectious particles, which we call propagating replicons, are po- replicons encoding foreign antigenic proteins are packaged into tent inducers of neutralizing antibody in animals yet are nonpatho- SFV-like particles by SFV structural proteins expressed in trans genic. Mice vaccinated with a single dose of the particles generated (5). Such a complementation system is required for alphavirus high titers of VSV-neutralizing antibody and were protected from vector systems because of the strict size limit for encapsidation a subsequent lethal challenge with VSV. Induction of antibody of viral genomic RNA. Unless the structural genes are deleted, required RNA replication. We also report that additional genes there is no space for inclusion of genes expressing foreign (including an HIV-1 envelope protein gene) expressed from the antigens. Because these complemented particles do not encode propagating replicons induced strong cellular immune responses SFV structural proteins, they replicate for only a single cycle to the corresponding proteins after a single inoculation. Our when inoculated into animals. studies reveal the potential of these particles as simple and safe The hybrid SFV/VSV propagating replicon particles that we vaccine vectors inducing strong humoral and cellular immune described infect and propagate in certain cell lines (6) with VSV responses. G as the only viral structural protein. However, the immunogenicity of these particles (designated SFVG particles) had not Semliki Forest virus ͉ vesicular stomatitis virus ͉ HIV-1 been tested in an animal model. Here we have examined the potential of these particles as a vaccine vector in a mouse model. NA replicons from alphaviruses including Semliki Forest virus We found that the particles induced a potent neutralizing R(SFV) have been developed and used for transient expression antibody response to VSV in mice. Mice vaccinated with these of foreign proteins in mammalian cells and also as experimental particles were protected from all weight loss and from a lethal vaccine vectors (1–4). The alphavirus genome is a capped and encephalitis caused by a high dose of wild-type VSV given polyadenylated positive-strand RNA molecule Ϸ12 kb in length. intravenously. The genomic RNA itself is an mRNA that encodes the viral We have also tested the immunogenicity of SFVG particles replicase. A subgenomic mRNA copied from the antigenomic expressing HIV-1 envelope (Env) or VSV nucleocapsid (N) RNA after replication encodes the alphavirus structural proteins. proteins behind a second SFV promoter. These vectors generate RNA transcribed from SFV cDNA can initiate viral RNA strong primary CD8 T cell responses to the foreign proteins as replication following transfection into cells (5). well as memory T cell responses that can be recalled to high Our laboratory previously tested an SFV replicon developed levels after boosting. by Liljestro¨m and Garoff (5) for expression of the vesicular Results stomatitis virus (VSV) glycoprotein (G) (6). The starting SFV Induction of Neutralizing Antibodies to VSV G Protein in Mice Inocu- RNA replicon was derived from a DNA copy of SFV from which lated with SFVG Particles Requires Vector Replication. To determine the genes for the SFV structural proteins were removed. The whether the propagating replicon particles were able to induce VSV G gene was inserted in place of genes encoding the SFV antibody responses to VSV G protein in animals and whether structural proteins. VSV G protein is the single transmembrane replication was required for such induction, we inoculated mice by glycoprotein of the prototype rhabdovirus VSV. VSV G medi- the intramuscular (i.m.) route with 6 ϫ 103 infectious units (i.u.) of ates both virus binding and membrane fusion to allow viral entry SFVG particles that were either untreated or inactivated with UV (7, 8). When this SFVG replicon RNA expressing only the SFV

light to prevent RNA replication. After 1 month, serum- IMMUNOLOGY replication proteins and VSV G protein was transfected into neutralizing antibody titers to VSV were determined (Fig. 1 Left). BHK-21 cells, it initially replicated in the small fraction of These results showed 100% neutralization of VSV at serum dilu- transfected cells. Surprisingly, it also produced infectious, low tions of 1:160 for SFVG particles but no detectable neutralizing Ab density, membrane-enveloped particles lacking a nucleocapsid (Ͻ1:20) in animals given the UV-inactivated particles. These results protein that budded from the cells and infected and killed all indicate that the incoming G protein on the particles was not cells in the culture within 2–3 days. These infectious particles present in sufficient amounts to induce VSV-neutralizing antibody could be propagated (passaged) indefinitely in tissue culture, and that G protein must be synthesized in infected cells to generate and their infectivity was inactivated by a VSV-neutralizing such responses. antibody that binds the VSV G protein (6, 9). Although the precise mechanism of generation of the SFVG infectious particles remains unknown, it clearly involves release of vesicles Author contributions: N.F.R., J.P., A.C., and J.K.R. designed research; N.F.R., J.P., and A.C. containing VSV G protein and SFV RNA. The replication of all performed research; N.F.R., J.P., A.C., and J.K.R. analyzed data; and N.F.R., J.P., A.C., and positive-strand RNA viruses including SFV occurs in association J.K.R. wrote the paper. with cellular membranes (10). SFV replication occurs in asso- The authors declare no conﬂict of interest. ciation with cytopathic vacuoles containing invaginations called This article is a PNAS Direct Submission. spherules, which are probably the sites of SFV RNA synthesis *To whom correspondence should be addressed: E-mail: [email protected]. (11–13). These spherules are also seen on the surface of SFV- © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800280105 PNAS ͉ April 15, 2008 ͉ vol. 105 ͉ no. 15 ͉ 5839–5843 Downloaded by guest on September 26, 2021 r 105 10000 t

1000 y d 100

e 100 g

al 95 SFVG imm

10 ch

r Naïve

%P 1 3 1 90

VSV neutralizing antibody tite SFVG SFVG(UV) SFVG VSV 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 6 x 103 i.u. 105 i.u. Days post challenge Fig. 2. Vaccination with SFVG particles protects mice against pathogenesis Fig. 1. Replication is required for induction of neutralizing antibody by SFVG caused by wild-type VSV. Twelve BALB/c mice were immunized with 5 ϫ 105 i.u. particles. Mice were inoculated i.m. with 6 ϫ 103 i.u. of SFVG untreated or of SFVG particles by i.m. injection. At 36 days after immunization, these mice treated with 100 mJ of UV light (UV) or with 105 i.u. of SFVG or 105 pfu of VSV were challenged with 5 ϫ 107 pfu of wild-type VSV by the i.v. route. Twelve as indicated. Pooled sera from groups of three mice were assayed for neu- nonimmunized BALB/c mice were challenged as controls. After challenge, tralizing titers to VSV on day 28 after inoculation. UV inactivation of SFVG mice were weighed daily for up to 14 days and observed for signs of patho- particles before inoculation abolished generation of anti-VSV G-neutralizing genesis. Any animal exhibiting paralysis or distress during this period was antibody. The neutralizing antibody titer to VSV in sera from mice inoculated killed. The graph shows the average weights of the mice Ϯ one standard with (105 i.u.) of SFVG was equivalent to that in sera from mice inoculated with deviation. Numbers above the x axis indicate the number of mice in the control 105 pfu of VSV (1:5,120). group that died on the corresponding day.

We next determined whether the strength of the antibody SFVG Replicon Particles Are Not Pathogenic in Mice. After i.m. response to VSV G was dose-dependent. We inoculated mice injections of SFVG particles, we had not seen any signs of with 105 i.u. of SFVG particles or with 105 plaque-forming units pathogenesis in mice. To determine whether there was any (pfu) of VSV. VSV serum-neutralizing titers were determined at detectable pathogenesis caused by infection by other potentially 28 days after infection by using pooled serum from each group. more pathogenic routes, we gave the SFVG particles by both the The neutralizing antibody responses to VSV in sera from mice 5 inoculated with the high dose (105 i.u.) of SFVG was 1:5,120, i.v. and the intranasal routes (10 i.u.). We then weighed the mice 32-fold higher than that induced in mice inoculated with the daily for 2 weeks and then observed the mice for 60 days and saw lower dose (6 ϫ 103 i.u.) of SFVG. Remarkably, the neutralizing no signs of pathogenesis caused by the particles. titer from the animals receiving 105 i.u. of SFVG particles was equivalent to that generated by inoculation of the same titer of Generation of SFVG Replicons Expressing HIVgp140. To evaluate the VSV (Fig. 1 Right). In an additional experiment, we inoculated ability of infectious SFVG particles to generate cell-mediated mice with 2.5 ϫ 105 SFVG particles and found no increase in the immune responses, we generated particles expressing the HIV-1 VSV-neutralizing antibody titers. (IIIB) gp140 gene. This gene encodes a secreted form of HIV Env protein lacking the transmembrane and cytoplasmic por- Vaccination with SFVG Particles Protects Mice from Pathogenesis tions of gp41 (14). There is an immunodominant CD8 T cell After VSV Challenge. i.v. injection of wild-type VSV in BALB/c (p18) epitope (15, 16) in this gp140 protein (BALB/c mice), and mice causes severe weight loss over 4–5 days and also causes we have an MHC I tetramer that recognizes T cells specific for lethal encephalitis in 20–40% of the animals. To determine this epitope, allowing precise quantitation of the CD8 T cell whether immunization with SFVG particles was sufficient to response (17). The gp140 gene was inserted into the pSFVdpG-X protect mice from such pathogenesis, 12 mice were immunized vector (18) downstream from a second SFV promoter. To i.m. with SFVG particles and then challenged 36 days later with generate the replicons, RNA transcribed in vitro from this vector 5 ϫ 107 pfu of wild-type VSV by the i.v. route. Following was used to transfect BHK-21 cells, and infectious particles were challenge, the mice were weighed daily to follow pathogenesis. recovered after 28 h as described in Materials and Methods. The SFVG-immunized mice maintained the same or higher than Infectious SFVG-gp140 particles derived from pSFVdpG-gp140 prechallenge body weights following challenge and showed no were expected to encode VSV G and HIV gp140 from separate signs of pathogenesis. (Fig. 2). In contrast, all 12 age-matched mRNAs. naive control mice showed dramatic weight loss following the identical challenge, along with other signs of pathogenesis Coexpression of VSV G and HIVgp140 Proteins in Infected Cells. To including ruffled fur and hunched posture. In addition, 4/12 of determine whether the SFVG-gp140 particles expressed both VSV the control mice developed the severe hind–limb paralysis G and gp140 proteins, we infected BHK-21 cells with these particles indicative of VSV encephalitis and died or were killed on days for 20 h. Cells were then fixed, and expression of both gp140 and 5 and 6. There was a significant difference (p ϭϽ0.05, Mann– VSV G was detected by indirect immunofluorescence (Fig. 3). VSV Whitney test) in weight loss between the SFVG-immunized G protein was expressed predominantly on the cell surface (red, group and the control group, through day 7 after challenge. After Fig. 3A), whereas HIVgp140 was expressed in a pattern typical of day 7, the remaining animals in the control group began to the endoplasmic reticulum (green, Fig. 3B) in a focus of infection. recover to normal weight. The protection from paralysis The merged image (Fig. 2C) shows that cells expressing VSV G also (encephalitis) was also statistically significant (P ϭ 0.047, expressed HIV gp140. The differential interference contrast image Fisher’s exact test) between the immunized and control groups. of the same field shows that some cells in the periphery of the focus We also checked VSV-neutralizing antibody titers in individ- of infection (white arrows) were not yet infected and expressed ual immunized animals at day 30, 6 days before challenge. They neither G nor gp140 (Fig. 2D). ranged from 1:640 to 1:5120 in the 12 vaccinated animals. The For direct visualization of the sizes of proteins expressed by control animals had undetectable VSV-neutralizing antibody the SFVG-gp140 particles, we also performed metabolic labeling titers. The high titer antibodies in the vaccinated animals are of infected cells. BHK-21 cells were infected with SFVG, SFVG- consistent with the complete protection observed. gp140, or a VSV recombinant expressing gp140 (14) and labeled

5840 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800280105 Rose et al. Downloaded by guest on September 26, 2021 SFVG SFVG-gp140 A 4 4 Avnti G- An VIH-it nE 10 10 A B A 0.035 0.2 2.21 0.41 3 3 10 10

2 2 10 10 FL2-H FL2-H

1 1 10 10

0 12.6 87.2 0 13.1 84.3 10 10 0 1 2 3 4 10 10 10 10 10 10 0 10 1 10 2 10 3 10 4 FL1-H FL1-H Tetramer SFVG +boost SFVG-gp140 +boost 4 4 10 10 3.9 0.88 21 1.6 Env 3 egreM DIC 10 10 3 C D 102 10 2 FL2-H FL2-H

1 101 10

39.2 56 0 34.9 42.6 100 10 0 1 2 3 4 10 0 101 102 103 104 10 10 10 10 10 FL1-H FL1-H CD62L

B lo 25 Fig. 3. Coexpression of VSVG protein and gp140 in cells infected with 20

SFVG-gp140 particles. (A) Indirect immunoﬂuorescence microscopy of BHK-21 , CD62L cells infected with SFVG-gp140 particles for 20 h, ﬁxed, and stained by using + 15 antibodies recognizing VSV G protein on the cell surface followed by an Alexa 10

Fluor-594 secondary antibody. (B) The same ﬁeld of cells permeabilized and CD8 T cells 5 stained by using sheep anti-gp120 antibody followed by FITC-conjugated 0 SFVG SFVG-gp140 VSV-gp140 Prime

secondary antibody. (C) Merge of A and B.(D) Differential interference Env tetramer contrast image of cells with uninfected cells noted by white arrows. - + - + - + Boost (vPE16)

with [35S]methionine for 1 h. Cell lysates were prepared and Fig. 5. SFVG-gp140 vaccination generates primary, Env-speciﬁc CD8 T cell responses that are readily recalled upon boosting. (A) Representative FACS plots either fractionated directly by SDS/PAGE or immunoprecipi- of CD8 T cells from spleens of individual BALB/c mice inoculated i.m. with 105 pfu tated by using antibodies to VSV or HIV Env before PAGE as of SFVG or SFVG-gp140 and analyzed at 7 days after inoculation (Upper). CD8ϩ, indicated (Fig. 4). VSV proteins (L, G, N, P, and M) as well as Env tetramerϩ, and CD62Llo cells are in the upper left quadrants (0.035% SFVG gp140 are easily seen without immunoprecipitation because of and 2.21% SFVG-gp140). (Lower) The same analysis done on CD8ϩ T cells from the effective shut-off of host protein synthesis (lane 1). Anti- individual mice inoculated with SFVG or SFVG-gp140, boosted at day 29 with VSV antibody precipitated all VSV proteins (lane 2), whereas vPE16, and then analyzed at day 35 (3.9% SFVG ϩ boost and 21.0% SFVG-gp140 ϩ ϩ ϩ lo anti-Env antibody precipitated gp140 (lane 3) from the VS- boost). (B) The average CD8 , Env tetramer , and CD62L cells from multiple mice primed or primed and boosted as in A (n ϭ 4–5 for all groups except SFVG Vgp140 lysate. Cells infected with SFVG particles expressed prime group only where n ϭ 2). Background responses (CD8ϩ, Env tetramerϩ, VSV G but not gp140 (lanes 4 and 5), whereas cells infected with and CD62Llo) determined from VSV vaccinated mice were subtracted and were SFVG-gp140 expressed both G and gp140 (lanes 6 and 7). Յ0.06%. Error bars represent one standard deviation. Mock-infected cells were used as controls (lanes 8 and 9).

SFVG-gp140 Particles Elicit Env-Specific CD8 T Cell Responses. To substantial population of activated Env-specific CD8 T cells Ϯ lo ϩ determine whether the SFVG vector expressing HIV Env gp140 (2.3 0.3% CD62L , tetramer , CD8 T cells, Fig. 5 A and B). was able to induce CD8 T cell responses to Env, we vaccinated The population elicited by SFVG-gp140 was similar to the primary response elicited by the vaccinia vectors (3.5 Ϯ 0.5%, mice i.m. with SFVG-gp140 particles and used an MHC I ϩ tetramer assay employing an H-2 Dd tetramer loaded with the CD62Llo , tetramer CD8 T cells) in mice that had previously immunodominant peptide p18–110 (16) from HIV IIIb Env seen only the control SFVG vector (Fig. 5 A and B). This protein (17). response is also the same as that generated in naive mice given Seven days after vaccination with SFVG-gp140, mice had a the vaccinia vector expressing HIV Env (17). The primary response to Env elicited by SFVG-gp140 was 4–5-fold lower than that elicited by VSVgp140 (Fig. 5B). 041pgVSV FS VG S 1pg-GVF 40 kcoM Because of the ability of SFVG-gp140 particles to elicit a ydobitnA oN I P VvVS nE VvVS En SV V vnE VvVS nE strong primary T cell response, we next examined the recall of memory cells after a boost with vaccinia expressing HIV Env. On L day 29 after prime, mice were boosted with vaccinia virus

041pg (vPE16) expressing the HIV Env protein (19). Recall Env- IMMUNOLOGY specific CD8 T cell responses were measured 6 days after boost at day 35. Mice primed with VSVgp140 or SFVG-gp140 elicited G a strong recall response after vPE16 boost. In fact, the Env- specific CD8 T cell response was equivalent after boost when P,N primed with either the VSV or the SFV vector (Fig. 5B; 21.2% Ϯ 1.1% and 21.6% Ϯ 3.2%, respectively). To examine the versatility of our vector system, we also tested an M SFVG-N vector that expresses the VSV nucleocapsid protein (18) and analyzed its ability to initiate cellular immune responses. We 54321 9876 vaccinated mice i.m. with SFVG-N and looked at the primary and recall responses to VSV N. For these experiments, we used an Fig. 4. Protein expression by SFVG and SFVG-gp140. BHK-21 cells were H-2Kb tetramer containing an immunodominant peptide from infected with SFVG particles, with SFVG-gp140 particles, with a VSV recombi- VSV N (20, 21). nant expressing gp140, or left uninfected. Metabolic labeling with [35S]methionine was between 5 and 6 h after infection. Cell lysates were prepared and We found a defined population of N-specific CD8 T cells in the lo either run directly on a 10% PAGE (VSVgp140) or immunoprecipitated by spleens of animals vaccinated with SFVG-N (0.5% CD62L , ϩ using antibodies to VSV or HIV Env as indicated. tetramer ), which were boosted to high levels (12% CD62Llo,

Rose et al. PNAS ͉ April 15, 2008 ͉ vol. 105 ͉ no. 15 ͉ 5841 Downloaded by guest on September 26, 2021 tetramerϩ) with a vaccinia recombinant [v38 (22)] expressing boosting with the same vector. However, this block can be VSV N protein. overcome by switching to a G protein from a different VSV serotype or a different vesiculovirus in the boosting vector (23). Discussion Presumably, this same strategy of switching G proteins would Vaccine vectors based on live viruses or viral replicons are work for SFVG vectors also. The extensive repertoire of avail- typically potent inducers of long-lasting immune responses in able vesiculovirus glycoproteins could therefore greatly expand animals. The Semliki Forest virus replicon has been used the potential of this vaccine strategy. The G protein-based extensively as an effective single-cycle vaccine vector (1). This propagating replicon strategy could also likely be extended to vector is normally packaged by using SFV capsid protein and other alphavirus replicon systems (4) to further extend vaccine envelope glycoproteins expressed in trans. In the current study, applications. we have taken advantage of our earlier finding that expression of the VSV G protein from the SFV replicon results in budding Materials and Methods of infectious membrane-enveloped particles containing VSV G Plasmid Construction. To construct pSFVG-gp140, a 2022-bp DNA fragment protein and the SFV replicon. We show here that these particles encoding the HIVgp140 protein (IIIB strain) was amplified by PCR with VENT induced potent antibody responses to the VSV G protein in mice polymerase (NEB) from pBSEnvG709 (24) by using the forward primer 5Ј- and could protect mice from pathogenesis including lethal GATCGATCG GGCCCAACAT GAGAGTGAAG GAGAAATATC AGC-3Ј and the encephalitis caused by VSV. They also induced strong cellular reverse primer 5ЈATCTGGCT ACGGGCCCTC AACTTGCCC ATTTATCTAATTCC- Ј immune responses to other proteins such as an HIV Env protein 3 . Both of the primers contained an ApaI site. The PCR product was digested with ApaI, purified, and ligated into the pSFV1-Gdp vector linearized with expressed from a second transcription unit added to the SFVG ApaI (18). The correct sequence of the gp140-insert was verified (Yale Keck replicons. Facility). The hybrid–virus vaccine platform we have described here could have significant advantages over traditional alphavirus- Transcription of RNA and Transfection to Generate Infectious Particles. To based vectors. Because the VSV G protein is expressed directly generate the infectious RNA genome of the propagating replicons, pSFV1-G from the replicon, there is no requirement for expression of and pSFVG-gp140 plasmids were linearized with SpeI and transcribed for 2 h packaging proteins in trans as in other alphavirus systems. In at 37°C in a 40-␮l reaction mixture. The reaction was a modification of the these complementation systems, there is also the potential of Ampliscribe SP6 transcription kit (Epicentre technologies) containing SP6 reconstituting wild-type alphaviruses through recombination. reaction buffer, 5 mM each of ATP, CTP, and UTP, 1 mM GTP, 4 mM m7G(ppp)G Because none of the SFV structural protein genes are present in RNA cap analog (NEB S1404L), 10 mM DTT, and 2 ␮l of SP6 polymerase. The Ϫ the SFVG vector, reconstitution of wild-type SFV is not possible. transcription reactions were stored at 80°C. Also, the relatively nonspecific packaging of the genomes into Transfection of cells for growing stocks of propagating replicons was performed as follows: 4 ϫ 106 BHK-21 cells were plated the day before infectious vesicles in the absence of a nucleocapsid (6) makes it transfection on 10-cm diameter plates. They were then transfected with 60 ␮l likely that there will not be a strict packaging limit for the RNA, of transcription reaction in 9 ml of serum-free DMEM containing 90 ␮lofa as there is in alphaviruses or other vectors with well-defined cationic liposome reagent containing dimethy–dioctadecyl ammonium bro- capsid structures. The SFVG particles expressing foreign anti- mide (25) as described (6). The cells were scraped into the medium at 28 h after gens could be produced in cell lines already approved for vaccine transfection and sonicated by using a Branson 450 sonicator to release infec- production without any requirement for modification to express tious particles. After sonication, cell debris was removed by centrifugation for complementing proteins in trans. 8 min at 625 ϫ g in a table-top centrifuge, and the supernatants were We initially used the rather cumbersome method of transcrib- transferred to new tubes. For concentration of the stocks, supernatants were ing capped SFVG vector RNA in vitro and then transfecting the transferred to Beckman ultraclear tubes and centrifuged at 40,000 rpm for 1 h RNA onto cells to generate the propagating replicon particles. in a Beckman an SW50.1 rotor. The infectious particles were resuspended in a volume of PBS that concentrated the particles 40-fold. More recently, we have tested a DNA-launched version of the SFVG vector by using the pBK-T-SFV1 vector with a CMV Immunofluorescence Microscopy. For titration of the stocks, BHK-21 cells promoter (1) to drive expression of the SFVG-gp140 RNA in plated on coverslips were infected for 21 h with different dilutions of the virus, cells. This system bypasses the in vitro transcription step and fixed with 3% paraformaldehyde, and incubated with a 1:200 dilution of greatly simplifies production of infectious particles. monoclonal antibodies (26) to VSV G protein followed by Alexa Fluor 488 goat Importantly, we found that the SFVG particles were non- anti-mouse IgG (HϩL) (Invitrogen) diluted 1:250. Green fluorescent areas of pathogenic in mice even when given by the i.v. route, a route that infected cells or plaques were counted on an Olympus CK40 microscope allows widespread dissemination in the animal. Lack of patho- equipped with a ϫ10 objective, and titers were calculated. Infectious particle genesis probably results from inefficient particle production in titers in the range of 1–5 ϫ 107 per ml were obtained, depending on the the absence of a nucleocapsid protein. Pathogenic animal viruses construct. all have capsid or nucleocapsid proteins to allow efficient For visualization of both VSVG and HIVgp140 proteins after transfection with pSFVG-gp140 or infection with SFVG-gp140 particles, BHK-21 cells were fixed and packaging of their nucleic acids. This low efficiency of packaging incubated with anti VSV-G antibody as above followed by Alexa Fluor 594 goat of the particles probably prevents spread of infection from anti-mouse IgG (1:500) secondary antibody. The cells were then permeabilized initially infected cells and thus rapidly limits the infection. with 1% Triton X-100 and incubated with a 1:100 dilution of polyclonal sheep Certainly more extensive studies of potency and potential patho- anti-HIVgp120 antiserum (NIH AIDS Research and Reference Reagent Program) genesis will have to be undertaken in non-human primate and followed by incubation with FITC-conjugated donkey anti-sheep serum diluted other animal models before these vectors could be used in 1:50. Cells were observed with a Nikon Eclipse 80i fluorescence microscope clinical trials. equipped with a Nikon Plan Apochromat ϫ60 oil objective and a Photometrics Vaccine vectors based on normal human pathogens are prob- CoolSnap camera. lematic because of the preexisting immunity to the vector in the human population that can limit vector effectiveness. For both [35S]Methinonine Labeling, Immunoprecipitation, and SDS/PAGE. BHK-21 cells Ϸ VSV and SFV, there is negligible preexisting immunity in grown to 50% confluency on 35-mm-diameter plates were infected with humans. However, because viral vaccine vectors express viral SFVG or SFVG-gp140 particles at a multiplicity of infection of one. Infected cells were incubated at 37°C for 5 h. The medium was removed, and cells were proteins in addition to the immunizing antigen, they are poten- washed twice with methionine-free Dulbecco’s modified Eagle’s medium tially limited for reuse or for effective boosting because immu- (DMEM). Then 1 ml of methionine-free DMEM containing 100 ␮Ci (1 Ci ϭ 37 nity is generated to the vector proteins in the initial application. GBq) of [35S]methionine was added to each plate for 1 h at 37°C. To prepare For example, with vectors based on live-attenuated VSV, there labeled cell extracts, the medium was removed, and the cells were washed is a strong antibody response to VSV G protein that prevents twice with PBS and lysed in 500 ␮l of detergent solution (1% Nonidet P-40,

5842 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800280105 Rose et al. Downloaded by guest on September 26, 2021 0.4% deoxycholate, 50 mM EDTA, 10 mM Tris⅐HCl, pH 7.8) on ice for 5 min. The hind leg in a total volume of 50 ␮l. At 36 days after immunization, these mice cell lysates were collected into 1.5 ml of Eppendorf tubes, and cell debris was were challenged with 5 ϫ 107 pfu of wild-type VSV (Indiana serotype, San Juan removed by centrifugation for two minutes at 16,000 ϫ g. strain) by the i.v. route in a total volume of 100 ␮l per mouse. Twelve naive Immunoprecipitation of VSVG and HIVgp140 proteins from the labeled BALB/c mice were also challenged as controls. Viral stocks were diluted to cell lysates was carried out as follows. The lysates were incubated with appropriate titers by using serum-free DMEM. Following challenge, mice were polyclonal rabbit anti-VSVserum or a polyclonal sheep anti-HIV gp120 weighed daily for up to 14 days and observed for signs of pathogenesis for a serum for1hat37°C. Protein A–Sepharose (Zymed Laboratories) was total of 60 days. Any animals exhibiting paralysis or distress during this period added, and samples were then incubated for 30 min at 37°C. The Sepharose were killed. was washed three times with radioimmune precipitation assay buffer (1% ⅐ Nonidet P-40, 1% deoxycholate, 0.1% SDS, 150 mM NaCl, 10 mM Tris HCl, Tetramer Assay. The tetramer assay was performed on fresh splenocytes as pH 7.8). Labeled immunoprecipitated proteins were analyzed by electro- described previously (27). Splenocytes were obtained 7 days after the primary phoresis on an SDS-10% polyacrylamide gel. vaccination in all experiments and on day 35 after primary vaccination (day 6 after boost) in boosting experiments. Responses to HIV Env were measured in Inoculation of Mice. Eight-week-old female BALB/c and C57BL/6 mice were BALB/c mice by using the Env tetramer (MHC class I Dd) described previously obtained from Jackson Laboratories and kept for at least 1 week before and containing the Env peptide N-RGPGRAFVTI-C (16). Responses to VSV N experiments were initiated. Mice were housed in microisolator cages in a were measured in vaccinated C57BL/6 mice by using the N tetramer (MHC class biosafety-level-2-equipped animal facility. Viral stocks were diluted to appro- IKb) described previously and containing the N peptide N-RGYVYQGL-C (20, priate titers in serum-free DMEM. Mice were vaccinated by i.m. injection in the 21). Tetramers were obtained from the National Institute of Allergy and ␮ right hind leg with VSV or SFVG in a total volume of 50 l. Vaccinia boosts were Infectious Diseases Tetramer Facility. Cells that were tetramerϩ, activated 5 performed via the i.p. route of infection with 1 ϫ 10 pfu of virus. The ϩ (CD62Llo), and CD8 were identiﬁed by using ﬂow cytometry as described Institutional Animal Care and Use Committee of Yale University approved of previously (27). To determine background levels of tetramer binding, spleno- all animal experiments done in this study. cytes from naive (N tetramer assay) or VSV vector vaccinated mice (Env tetramer assay) were used. VSV Neutralization Assay. Blood was obtained from mice at 28 days after vaccination. Serum was collected and heat-inactivated, and neutralization Statistical Analysis. GraphPad Prizm software, version 4.0, was used for all assays were performed as described previously (23). Mouse sera were analyses. serially diluted on 96-well plates and incubated with 100 pfu of VSV per well for1hat37°C. BHK-21 cells were added to each well, and plates were incubated at 37°C for 2–3 days. Each assay was performed in duplicate. ACKNOWLEDGMENTS. We thank Drs. P. Liljestro¨m, G. Karlsson, and H. Garoff (Karolinska Institute, Stockholm) for providing materials and advice and mem- Neutralization titers are given as the highest dilutions that showed com- bers of J.K.R.’s laboratory for helpful suggestions throughout the course of this plete inhibition of VSV infection and cytopathic effect. work. This work was supported by National Institutes of Health Grants R37AI40357 and AI057158. Polyclonal sheep anti-HIVgp120 antiserum was ob- Immunizations and VSV Challenge. For challenge experiments, 12 BALB/c mice tained through the National Institutes of Health AIDS Research and Reference were immunized with 5 ϫ 105 pfu of SFVG particles by i.m. injection in the right Reagent Program.

1. Karlsson GB, Liljestro¨m P (2004) Delivery and expression of heterologous genes in 15. Shirai M, Kozlowski S, Margulies DH, Berzofsky JA (1997) Degenerate MHC restriction mammalian cells using self-replicating alphavirus vectors. Methods Mol Biol 246:543– reveals the contribution of class I MHC molecules in determining the fine specificity of 557. CTL recognition of an immunodominant determinant of HIV-1 gp160 V3 loop. JIm- 2. Olkkonen VM, Dupree P, Simons K, Liljestro¨m P, Garoff, H (1994) Expression of munol 158:3181–3188. exogenous proteins in mammalian cells with the Semliki Forest virus vector. Methods 16. Shirai M, Pendleton CD, Berzofsky JA (1992) Broad recognition of cytotoxic T cell Cell Biol 43(Pt A):43–53. epitopes from the HIV-1 envelope protein with multiple class I histocompatibility 3. Pushko P, et al. (1997) Replicon–helper systems from attenuated Venezuelan equine molecules. J Immunol 148:1657–1667. encephalitis virus: Expression of heterologous genes in vitro and immunization against 17. Haglund K, et al. (2002) High-level primary CD8(ϩ) T-cell response to human immu- heterologous pathogens in vivo. Virology 239:389–401. nodeficiency virus type 1 gag and env generated by vaccination with recombinant 4. Rice CM (1996) Alphavirus-based expression systems. Adv Exp Med Biol 397:31–40. vesicular stomatitis viruses. J Virol 76:2730–2738. 5. Liljestro¨m P, Garoff H (1991) A new generation of animal cell expression vectors based 18. Rolls MM, Haglund K, Rose JK (1996) Expression of additional genes in a vector derived on the Semliki Forest virus replicon. Bio/Technology 9:1356–1361. from a minimal RNA virus. Virology 218:406–411. 6. Rolls MM, Webster P, Balba NH, Rose JK (1994) Novel infectious particles generated by 19. Earl PL, Moss B, Doms RW (1991) Folding, interaction with GRP78-BiP, assembly, expression of the vesicular stomatitis virus glycoprotein from a self-replicating RNA. and transport of the human immunodeficiency virus type 1 envelope protein. Cell 79:497–506. J Virol 65:2047–2055. 7. Florkiewicz RZ, Rose JK (1984) A cell line expressing vesicular stomatitis virus glyco- 20. Masopust D, Vezys V, Marzo AL, Lefrancois L (2001) Preferential localization of effector protein fuses at low pH. Science 225:721–723. memory cells in nonlymphoid tissue. Science 291:2413–2417. 8. Riedel H, Kondor-Koch C, Garoff H (1984) Cell surface expression of fusogenic vesicular 21. Van Bleek GM, Nathenson SG (1990) Isolation of an endogenously processed immu- stomatitis virus G protein from cloned cDNA. EMBO J 3:1477–1483. nodominant viral peptide from the class I H-2Kb molecule. Nature 348:213–216. 9. Kelley JM, Emerson SU, Wagner RR (1972) The glycoprotein of vesicular stomatitis virus 22. Mackett M, Yilma T, Rose JK, Moss B (1985) Vaccinia virus recombinants: Expression of is the antigen that gives rise to and reacts with neutralizing antibody. J Virol 10:1231– VSV genes and protective immunization of mice and cattle. Science 227:433–435. 1235. 23. Rose NF, Roberts A, Buonocore L, Rose JK (2000) Glycoprotein exchange vectors based on 10. Salonen A, Ahola T, Kaariainen L (2005) Viral RNA replication in association with vesicular stomatitis virus allow effective boosting and generation of neutralizing antibod- IMMUNOLOGY cellular membranes. Curr Top Microbiol Immunol 285:139–173. ies to a primary isolate of human immunodeficiency virus type 1. J Virol 74:10903–10910. 11. Froshauer S, Kartenbeck J, Helenius A (1988) Alphavirus RNA replicase is located on the 24. Owens RJ, Rose JK (1993) Cytoplasmic domain requirement for incorporation of a cytoplasmic surface of endosomes and lysosomes. J Cell Biol 107:2075–2086. foreign envelope protein into vesicular stomatitis virus. J Virol 67:360–365. 12. Grimley PM, Berezesky IK, Friedman RM (1968) Cytoplasmic structures associated with 25. Rose JK, Buonocore L, Whitt MA (1991) A new cationic liposome reagent mediating an arbovirus infection: Loci of viral ribonucleic acid synthesis. J Virol 2:1326–1338. nearly quantitative transfection of animal cells. BioTechniques 10:520–525. 13. Kujala P, et al. (2001) Biogenesis of the Semliki Forest virus RNA replication complex. 26. Lefrancios L, Lyles DS (1982) The interaction of antibody with the major surface J Virol 75:3873–3884. glycoprotein of vesicular stomatitis virus. I. Analysis of neutralizing epitopes with 14. Publicover J, Ramsburg E, Rose JK (2005) A single-cycle vaccine vector based on monoclonal antibodies. Virology 121:157–167. vesicular stomatitis virus can induce immune responses comparable to those generated 27. Publicover J, Ramsburg E, Rose JK (2004) Characterization of nonpathogenic, live, viral by a replication-competent vector. J Virol 79:13231–13238. vaccine vectors inducing potent cellular immune responses. J Virol 78:9317–9324.

Rose et al. PNAS ͉ April 15, 2008 ͉ vol. 105 ͉ no. 15 ͉ 5843 Downloaded by guest on September 26, 2021