Vaccination of Mice with Replication-Defective Human

ARTICLE

doi:10.1016/j.ymthe.2006.02.021

Vaccination of Mice with Replication-Defective Human Immunodeficiency Virus Induces Cellular and Humoral Immunity and Protects against Vaccinia Virus-gag Challenge Christopher S. Baliga,1 Marc van Maanen,2 Michael Chastain,3 and Richard E. Sutton2,* 1

Department of Pediatrics, Section of Allergy and Immunology, and 2Department of Molecular Virology and Microbiology, Baylor College of Medicine, Houston, TX 77030, USA; 3 Department of Vaccine and Biologics Research, Merck Research Laboratories, West Point, PA 19486, USA

*To whom correspondence and reprint requests should be addressed. Fax: +1 713 798 3586. E-mail: [email protected].

Available online 19 May 2006

Here we describe as a potential vaccine candidate a replication-defective HIV that encodes multiple viral genes in addition to a cassette that includes both truncated cyclin T1 and an autofluorescent protein. After confirming functionality of the cyclin T1, we immunized mice intramuscularly once or twice with the replication-defective HIV vector pseudotyped with vesicular stomatitis virus (VSV) G protein (RD HIV), a plasmid encoding CMV-driven gag (gag DNA), or adenovirus gag (Ad5-gag). Capsid-specific antibody titers following RD HIV immunization were N106/ml and approximately equivalent to those induced by gag DNA and Ad5-gag. Antibodies against the autofluorescent protein and VSV G were also detected. After RD HIV immunization ELISpot assays demonstrated Gag-specific interferon-g (IFN-g) SFU equivalent to that of Ad5-gag and fourfold greater than that of gag DNA. HIV polymerase-specific IFN-g SFU values were similar, and boosting increased both antibody titers and the IFN-g response. Challenge using vaccinia virus (VV)-gag demonstrated significantly lower recoverable VV for RD HIV-immunized mice compared to controls. No significant differences were observed in vaccinated mice challenged with wild-type VV. This study demonstrates the efficacy of RD HIV in conferring HIV-specific immunity and protection in mice and suggests its potential use in humans as either a prophylactic or a therapeutic vaccine. Key Words: replication-defective human immunodeficiency virus, AIDS vaccine, cellular immune response, humoral immune response, vaccinia virus

INTRODUCTION A safe and effective prophylactic vaccine will likely be the most cost-effective means of controlling human immunodeficiency virus (HIV) worldwide [1]. Several strategies for developing HIV vaccines have been employed, including subunit vaccines, killed virus vaccines, liveattenuated vaccines, replicons, DNA vaccines, and viruslike particles [2]. Varying vaccination strategies induce variable levels of cellular and humoral immunity. Subunit vaccines (purified protein) and inactivated virus vaccines induce mostly antibody responses, live-attenuated vaccines usually induce both humoral and cellular immunity, and replication-defective viral vectors can elicit a cytopathic T lymphocyte (CTL) response that is typically stronger than the antibody response [3]. Live-attenuated simian immunodeficiency virus (SIV) vaccines were able to control infection in nonhuman

432

primates, even those due to nonhomologous virulent strains of SIV [4–7]. Pathogenicity in neonatal [8,9] and adult animals [10], coupled with concerns regarding reversion to wild-type virus [11,12], has prevented their advancement [13], although it is conceivable that further attenuated viruses will be safer and more palatable yet retain their efficacy [14,15]. The ideal prophylactic vaccine would result in long-lived sterilizing immunity. Failing that, in the near term a prophylactic vaccine that elicits an immune response capable of controlling any subsequent infection could be considered a more economical and tolerable version of highly active antiretroviral therapy. Several vaccine candidates, including naked DNA boosted with various forms of interleukin-2 [16], naked DNA boosted with modified vaccinia Ankara [17], replication-competent vesicular stomatitis virus [18], and replication-defective adenovirus [19], each encoding various

MOLECULAR THERAPY Vol. 14, No. 3, September 2006 Copyright C The American Society of Gene Therapy 1525-0016/$30.00

doi:10.1016/j.ymthe.2006.02.021

ARTICLE

FIG. 1. Truncated cycT1 increases HIV protein expression in mice. (A) MEFs that had previously been transduced with the MLV vector with or without cycT1 were subsequently transduced with increasing amounts of HIV-IRES-eYFP (VSV G) and analyzed by flow cytometry. Fold increase in MFI ratio is indicated above gray bars. (B) At top is a schematic of HIV, with gene products indicated. At bottom is vector pHIV-cycT1-IRES-eYFP. Gene products are not to scale. (C) MEFs were transduced with increasing amounts of the indicated HIV vectors and analyzed by flow cytometry. Transduction efficiency is indicated above bars.

lentiviral genes, have demonstrated differing abilities both to control viral replication and to maintain peripheral CD4+ T cell counts in rhesus macaques challenged by a variety of routes with SIV–HIV (SHIV) chimeras. For some of these, CTL responses have correlated with low viral loads and absence of pathologic changes [17], and the most promising analogous HIV vaccines are entering clinical trials. Vectors based upon HIV have shown tremendous potential for gene transfer in nondividing or terminally differentiating cells, such as myocytes, pancreatic islet cells, neurons, and hematopoietic stem cells [20]. The most advanced of these vectors do not encode any viral gene products such that the transduced cells should not elicit any antiviral immune response. However, it may

MOLECULAR THERAPY Vol. 14, No. 3, September 2006 Copyright C The American Society of Gene Therapy

also be possible to include a limited number of viral open reading frames without seriously compromising safety and yet at the same time stimulating both humoral and cellular immune responses by foreign antigen cellular expression and antigen presentation, respectively. In that sense, replication-defective (RD) or single-cycle HIV would behave similar to a live-attenuated HIV, but with a much lower risk of reversion to wild-type or fully replication-competent HIV and consequent pathogenicity. As such, limited data have suggested that RD HIV can induce humoral immunity in mice [21]. Recently, Evans et al. demonstrated that single-cycle SIV stimulates diverse immune responses and reduces viral loads after challenge with a highly pathogenic strain [22].

433

ARTICLE

doi:10.1016/j.ymthe.2006.02.021

RESULTS

We wished to see whether incorporation of cycT1 into HIV-IRES-eYFP would increase expression of eYFP in a similar manner. In pHIV-CycT1-IRES-eYFP cycT1 is translated as the upstream ORF in a bicistronic mRNA also encoding eYFP (Fig. 1B). We transduced MEFs using increasing m.o.i. of 0.3, 3, and 30 with either HIV-CycT1-IRES-eYFP (VSV G) or HIVIRES-eYFP (VSV G) and analyzed them by flow cytometry 5 days later (Fig. 1C). At each m.o.i. the transduction efficiency was approximately the same irrespective of vector, but the MFI of cells transduced with HIV-CycT1-IRES-eYFP was 100– 300% greater compared to that of those transduced with HIV-IRES-eYFP. This suggests that cycT1 can function in the cis configuration.

Cyclin T1 Functions in the cis Configuration We first wished to determine whether truncated human cyclin T1 (cycT1) would increase HIV gene product expression in mouse cells. We transduced mouse embryo fibroblasts (MEFs) with a murine leukemia virus (MLV)-based empty vector or one encoding cycT1, positively selected transduced cells, and confirmed cycT1 expression by immunoblotting using hemagglutinin-specific antisera. We then transduced MEFs with increasing amounts of HIV-eYFP (VSV G) and evaluated transduction efficiency and mean fluorescence intensity (MFI) by flow cytometry. The presence of cycT1 dramatically increased the expression of enhanced yellow fluorescent protein (eYFP), which was especially evident at lower multiplicities of infection (m.o.i.) (Fig. 1A). At an m.o.i. of 12.5, the fold increase in MFI ratio (MFI of transduced cells divided by that of untransduced cells) in the presence of cycT1 was 14, whereas it was reduced 6.1 and 2.6 at m.o.i. of 25 and 62.5, respectively. This is due to the MFI ratio increasing with higher m.o.i. in the absence of cycT1 and is consistent with previous results demonstrating the positive HIV transcriptional effects of cycT1 in mouse cells.

Immunized Mice Generate Antibody Responses Against CA Pseudotyped HIV particles could be used to generate protective immune responses, especially if viral genes are expressed in the transduced cells. Limited evidence had suggested that mice immunized with HIV vectors were capable of mounting an antibody response [21]. To investigate this we immunized mice intramuscularly (im) (N = 6 per group) with concentrated RD HIV, DNA encoding gag, or Ad5-gag. At 4 weeks, we measured serum CA-specific IgG titers by ELISA. All vaccinated groups of mice were able to generate measurable anti-CA titers, and antibody titers increased after boosting the mice with homologous immunogen (Fig. 2). As a control, we heatinactivated (HI) RD HIV, which reduced infectious titer by 106-fold. Anti-CA antibody titers generated by HI RD HIV were ~1000-fold less compared to those of untreated RD HIV, suggesting that infectivity was required for maximal immunogenicity. gag DNA, 107 and 108 IU of Ad-gag, and 108 IU of RD HIV all elicited approximately equivalent CA antibody titers, which exceeded 106/ml.

Here we demonstrate the efficacy of a RD HIV, pseudotyped with vesicular stomatitis virus G protein (VSV-G), in inducing both cellular and humoral HIVspecific immunity in the mouse as well as the ability to withstand a challenge with recombinant vaccinia virus (VV) containing gag. The immune responses generated were superior to those generated by naked DNA and roughly comparable to those elicited by adenovirus. These results suggest that RD HIV could be an attractive vaccine candidate for prophylactic and possibly therapeutic use.

FIG. 2. CA titers after immunization. Mice (N = 6 per group) were immunized as shown and CA titers per milliliter determined by ELISA and end-point dilution. Closed bars, no boost; open bars, boost at 4 weeks. Down arrow indicates titer b100/ml.

434


doi:10.1016/j.ymthe.2006.02.021

Anti-CA titers after 107 IU of RD HIV were 10-fold lower than those after 108 IU. Because the RD HIV was prepared by plasmid transient transfection of 293T cells, we measured the amount of plasmid DNA by semiquantitative PCR using HIV-specific primers, and it did not exceed 200 pg in 108 IU. Thus, we feel it is unlikely that the antibody responses observed were due to the presence of this DNA, especially given the poor results with HI RD HIV (Fig. 2). Immunized Mice Generate Humoral Responses Against Other Vector Components It is well established that neutralizing antibodies against VSV G are generated after VSV infection of mice [23]. To examine the anti-VSV G antibody responses we preincubated mouse sera with MLV encoding eYFP that were pseudotyped with either VSV G or amphotropic 4070A envelope glycoprotein (Ampho). We then titered these

ARTICLE

particles on human cells and measured transduction efficiency 3 days later by epifluorescence microscopy. In a dose-dependent manner sera from mice that had been immunized with RD HIV neutralized MLV pseudotyped with VSV G (Fig. 3A), without effect on Ampho-pseudotyped MLV (Fig. 3B). Sera from mice that had been immunized with 108 IU of RD VSV G had greater neutralizing activity than sera from mice immunized with 107 IU (Fig. 3A). HIV-IRES-eYFP (Ampho) was not neutralized by any sera (Fig. 3C). Serum from mice immunized with HI RD HIV had no neutralizing activity (Fig. 3A), although that serum did recognize denatured VSV G, especially after boosting (Fig. 4A, compare lanes 7 and 8). Serum from mice immunized with RD HIV also recognized denatured VSV G, again in a dose-dependent manner (Fig. 4A, lanes 3–6). Only trace amounts of eYFP protein were present in concentrated vector supernatants, based upon immunoblotting. To determine whether immunized mice pro-

FIG. 3. Neutralization of VSV G-pseudotyped particles by antisera. (A) Sera were added to MLV vector pBABE-IRES-eGFP (VSV G) 1 h prior to addition to target human cells. Titer was determined 3 days later by epifluorescence microscopy. Paired columns 1, 5 Al of sera from mice that had received 108 IU Ad-gag with boost; 2, 0.5 Al 108 IU Ad-gag with boost; 3, 5 Al 107 RD-HIV; 4, 0.5 Al 107 RD-HIV; 5, 5 Al 107 RD-HIV with boost; 6, 0.5 Al 107 RD-HIV with boost; 7, 5 Al 108 RDHIV; 8, 0.5 Al 108 RD-HIV; 9, 5 Al 108 RD-HIV with boost; 10, 0.5 Al 108 RD-HIV with boost; 11, 5 Al HI 108 RD-HIV; 12, 0.5 Al HI 108 RD-HIV; 13, 5 Al HI 108 RDHIV with boost; 14, 0.5 Al HI 108 RD-HIV with boost. Gray bars, 20,000 IU input pBABE-IIRES-eGFP (VSV G); black bars, 200,000 input IU. Downward arrow indicates near absence of transduced cells. (B) Similar to (A) except pBABE-IRES-eGFP (Ampho) was used, with same order and amount of sera. Gray bars, 1000 IU input pBABE-IRES-eGFP (Ampho); black bars, 8000 input IU. (C) Similar to (A) except HIV-IRES-eYFP (Ampho) was used, with same order and amount of sera. Gray bars, 800 IU input HIV-IRES-eYFP (Ampho); black bars, 8000 input IU.


435

ARTICLE

doi:10.1016/j.ymthe.2006.02.021

antibody being rabbit anti-mouse conjugated to horseradish peroxidase (HRP) (Fig. 4B). Sera from mice immunized with RD HIV reacted against GST–eGFP in a dosedependent manner (Fig. 4B, lanes 3–6) but not GST alone (lane 1), whereas serum from mice immunized with HI RD HIV had minimal reactivity, comparable to serum from mice immunized with Ad-gag (compare lane 2 to lanes 7 and 8). These results are consistent with eYFP being expressed after im injection of RD HIV in transduced mouse tissues and subsequently being recognized as a foreign antigen. FIG. 4. Reactivity of sera against VSV G and eYFP. (A) 293T cells were transfected with VSV G expression construct and cell lysates separated by SDS–PAGE, transferred to nitrocellulose, probed with 1:1000 diluted pooled mouse sera followed by anti-mouse–HRP secondary antibody, and developed using ECL. Lanes 1, monoclonal anti-VSV G P5D4 (Sigma); 2, sera from mice that had received 108 IU Ad-gag; 3, 107 RD-HIV; 4, 107 RD-HIV with boost; 5, 108 RD-HIV; 6, 108 RD-HIV with boost; 7, 108 IU HI RD-HIV; 8, 108 IU HI RDHIV with boost. Arrow indicates 70-kDa VSV G protein. Reactivity was not observed using lysates from parental 293T cells. (B) E. coli-purified GST (lane 1) or GST–eGFP (lanes 2–8) was size-separated by SDS–PAGE, transferred to nitrocellulose, probed with 1:1000 diluted pooled mouse sera followed by anti-mouse–HRP secondary antibody, and developed using ECL. Lanes 1, sera from mice that had received 108 IU RD HIV with boost; 2, 108 IU Ad-gag; 3, 107 RD-HIV; 4, 107 RD-HIV with boost; 5, 108 RD-HIV; 6, 108 RD-HIV with boost; 7, 108 IU HI RD-HIV; 8, 108 IU HI RD-HIV with boost. Arrowhead denotes position of GST alone; arrow indicates GST–eGFP fusion. Note that strips were slightly misaligned.

duced antibody directed against eYFP, we immunoblotted a purified glutathione S-transferase (GST)–eGFP fusion protein using various mouse sera, with the secondary

Immunized Mice Generate Cellular Responses We then quantified both Gag- and Pol-specific cellular responses by measuring interferon-g (IFN-g) production from harvested splenocytes, determined by two separate enzyme-linked immunospot (ELISpot) assays (Table 1). The cellular response data roughly paralleled the antibody titers in terms of vaccine efficacy. Boosted mice always produced higher responses than nonboosted animals. HI RD HIV at 108 IU resulted in at most 50% of the Gag-specific spot-forming units (SFU) produced by 107 IU of untreated RD HIV, and it was only slightly higher following the boost. The gag DNA was able to induce a higher IFN-g immune response than HI RD HIV, but it was always less than half the values elicited by RD HIV. The SFU improved following a boost with gag DNA, but remained inferior to RD HIV or Ad5-gag. IFN-g production was predictably greater when we used 108 IU of either RD-HIV or Ad5-gag to vaccinate the mice than when we used 107 IU.

TABLE 1: ELISpot assays of immunized micea Vaccine

Media

Assay 1 Sham immunized RD HIV 107 IU RD HIV 108 IU Heat-inactivated RD HIV gag DNA (100 Ag) Ad5-gag 107 IU Ad5-gag 108 IU

35 15 90 60 14 2 20

Assay 2 Sham immunized RD HIV 107 IU RD HIV 108 IU Heat-inactivated RD HIV gag DNA (100 Ag) Ad5-gag 107 IU Ad5-gag 108 IU

1 28 26 19 2 1 1

a b c d e

Pol

Media

Prime and boostb Gag

35 216 842 134 82 320 685

2 146 478 102 NDe ND ND

140 682 185 9 50 48

862 1010 330 408 1210 1972

496 N500 235 ND ND ND

1 234 378 35 67 315 306

1 65 110 11 ND ND ND

66 76 80 1 7 N500

467 750 179 234 687 N500

134 N500 84 ND ND ND

Prime only Gagc

d

Pol

Table entries represent number of IFN-g-producing cells per 106 splenocytes (3 mice pooled for each assay). Boosts were performed 4 weeks after initial immunization. Represents pooled Gag peptides. Represents pooled Pol peptides. ND, not determined.

436


ARTICLE

doi:10.1016/j.ymthe.2006.02.021

The HIV Gag-induced IFN-g production was consistently three times greater than the Pol-induced IFN-g SFU (Table 1). RD HIV at 108 IU was equivalent to 107 and 108 IU of Ad5-gag in terms of Gag-specific SFU. We identified Pol-specific IFNg-secreting cells in all three RD HIV vaccine groups. We performed the ELISpot experiments twice due to unanticipated high background SFU in the RD HIV-immunized mice. Once repeated, the background was much lower except in the boosted 108 IU Ad5-gag-vaccinated mice (Table 1, assay 2). However, the pattern of cellular immune responses was still evident and consistent between the two separate analyses of the ELISpot assay. Use of Heterologous VSV-G Does Not Improve Immune Response When immunizing with a live vector, it is conceivable that the immune response generated will neutralize any benefit from a booster dose of the same vaccine. To address this issue experimentally, we produced RD HIV pseudotyped with either the Indiana strain of VSV G (IVSV G) or the Chandipura strain (C-VSV G). Both vectors were of equivalent quality in terms of titer on adherent human cells and CA content/IU. We demonstrated the absence of cross-neutralization of sera of mice immunized with RD HIV (I-VSV G), using HIV-IRES-eYFP (I-VSV G)- and HIV-IRES-eYFP (C-VSV G)-pseudotyped particles (Table 2). We immunized mice using various prime–boost combinations of RD HIV pseudotyped with the two strains of VSV G and also Ad5-gag (Fig. 5). While boosted vaccinees achieved higher antibody titers than nonboosted mice, there was no clear advantage to priming and then boosting with RD HIV pseudotyped with heterologous VSV G in terms of CA antibody titers (Fig. 5A). Nor were the CA antibody titers higher if we boosted Ad5-gag homologously versus heterologously with either RD HIV. No matter how Ad5-gag was boosted antibody titers exceeded 106/ml (Fig. 5A). TABLE 2: Homologous VSV G neutralizationa Input IUb

PBSc

Add

HIe

HI-boostf

HIV (VSV G)g

Indiana challenge 4000 b5 400 b5

b5 b5

b5 b5

b5 b5

N95 N95

Chandipura challenge 4000 b5 400 b5

b5 b5

b5 b5

b5 b5

b5 b5

a Entries represent percentage inhibition of either Indiana or Chandipura HIV-IRES-eYFP (VSV G)-pseudotyped particles, with readout being epifluorescence microscopy 3 days after transduction. b Number of eYFP infectious units used to transduce HOS cells. c Phosphate-buffered saline. d Sera from mice that had received 108 IU of Ad-gag vector. e Sera from mice that had received heat-inactivated HIV (VSV G). f Sera from mice that had received boosted heat-inactivated HIV (VSV G). g Sera from mice that had received 107 IU of HIV (VSV G), Indiana strain.


FIG. 5. Immune responses after heterologous prime–boost. (A) Mice (N = 6 per group) were primed and boosted as indicated and mean geometric CA titers per milliliter were determined 4 weeks later. (B) ELISpot results from a subset of the groups shown in (A). Downward arrow indicates undetectable SFU. VSV G denotes RD HIV; ch, Chandipura pseudotypes; in, Indiana pseudotypes.

ELISpot assays of the boosted mice demonstrated no clear differences between heterologous VSV G boosts and homologous ones (Fig. 5B), although IFN-g SFUs of Ad5-gag followed by RD HIV pseudotyped with either I-VSV G or CVSV G were ~65% higher compared to homologous Ad5-gag boosting (Ad5-gag/Ad5-gag) or priming and then boosting with heterologous VSV G without Ad5-gag (Fig. 5B). Immunized Mice are Protected Against VV-gag After immunization with mock, gag DNA, or RD HIV (with or without a boost) we challenged female Balb/c mice with VVgag or a control virus containing an irrelevant peptide casette, VV-sc (Fig. 6). All groups challenged with VV-sc had recoverable VV titers that were not statistically different from each other by the Kruskal–Wallis test ( P = 0.0812) (Fig. 6A). Similarly, we found no significant differences when we compared individual vaccine groups to the sham-immunized group by the Wilcoxon rank sums test (Fig. 6A). We challenged separately immunized mice with VV-gag (Fig. 6B). There was an ~30-fold decrease in recoverable VVgag from the mice immunized with RD HIV, which was

437

ARTICLE

doi:10.1016/j.ymthe.2006.02.021

FIG. 6. Vaccinia virus challenge of immunized mice. (A) Immunized mice (N = 6 per group) were challenged with 1 107 IU vaccinia virus-sc ip. Amount of recovered VV is indicated on a per mouse basis. P N 0.05 for all comparisons. (B) Immunized mice (N = 10 per group) were challenged with 1 107 IU VV-gag ip. P b 0.05 for boosted RD HIV vs sham and P N 0.05 for DNA vs sham. Bars represent the geometric mean of recoverable VV PFU and HIV denotes RD HIV.

significant compared to the sham-immunized mice by the Wilcoxon rank sums test ( P b 0.001) (Fig. 6B). At the higher challenge dose, gag DNA-vaccinated mice demonstrated a trend toward a significant difference versus the shamimmunized mice ( P = 0.058). For comparison, boosted Ad5-gag im followed by intraperitoneal (ip) challenge with VV-gag resulted in no recoverable virus at 1:1000 dilutions in five of five mice, with two of five sham-immunized mice having undetectable virus.

DISCUSSION Although live-attenuated SIV has succeeded in protecting from homologous SIV challenge in rhesus macaques, the risks may outweigh the benefits of its use [13]. A potential alternative that has yet to be used in animals is a conditionally replicating virus [24–27]. Other vaccine candidates have been proposed, and only recently has a modicum of success been achieved in controlling subsequent SHIV infection and

438

consequent immune system decline. A candidate that has many of the benefits of live-attenuated virus without all of the associated safety concerns is RD HIV. Here we demonstrate that RD HIV is capable of inducing both humoral and cellular responses in mice. The humoral responses measured were roughly equivalent to those of naked DNA and replication-defective adenovirus. The cellular responses obtained, based on IFN-g ELISpot assays, were superior to DNA and comparable to adenovirus. Heterologous prime–boosting modestly increased both the humoral and the cellular responses. In addition, RD HIV afforded relative protection to a VV-gag challenge that was again superior to DNA (and likely equivalent to adenovirus). Immunization with VSV G-pseudotyped viral particles resulted in a neutralizing antibody response against the glycoprotein. Boosting with particles bearing a different VSV G did not improve the observed immune response, whereas boosting with the heterologous Ad vector appeared to increase both cellular and humoral immun-


doi:10.1016/j.ymthe.2006.02.021

ity. An obvious advantage of using VSV G-pseudotyped particles is that for VSV the seroprevalence rate in the human population of industrialized countries is low [28], whereas for many serotypes of Ad preexisting immunity may be problematic for a prophylactic or therapeutic vaccine [29]. We did not observe any benefit of priming and boosting with nonheterologous strains of VSV G even in the presence of demonstrable neutralizing antibodies. This would suggest that having prior immunity to the Indiana strain of VSV may not adversely influence the efficacy of RD HIV as a vaccine. There are multiple blocks to HIV replication in the mouse. Murine and other rodent cells are much more poorly transduced by HIV-based vectors compared to human and simian cells [30], and even when cycT1 is provided in trans the level of viral gene expression is less than that observed in human cells (our unpublished results). In addition, Gag is inefficiently processed to CA in murine cells, with little CA release. Thus, although cycT1 was encoded within the vector here, the in vivo transduction rate and the absolute amounts of viral proteins found inside and secreted from cells were likely low. Taking this into account underscores the magnitude of the immune responses observed here. RD HIV has its own potential pitfalls. Like all retroviruses, RD HIV integrates essentially irreversibly into the host genome. Transcriptionally active genes are favored as integration site targets [31,32]. As recently demonstrated by the X-linked SCID phase I gene therapy trial, insertional oncogenesis may occur with retroviral vectors. At least two patients developed acute lymphoblastic leukemia as a result of increased expression of the LIM domain only 2 (LMO2) transcription factor due to integration of the MLV vector into that gene [33]. It should be noted, however, that MLV has a predilection for integration near the 5V end of transcriptional units [31,34], whereas HIV integrates without bias throughout genes [31,32]. Other factors (specifically the common interleukin receptor g chain) likely played supportive roles in the oncogenic process in the SCID trial [33]. It may be possible to increase the safety of RD HIV by reducing integrase activity, either by introducing point mutations or coadministering a small molecule inhibitor [35,36]. Some of the immunogenic properties of the vector would likely be maintained without genomic integration, especially since certain viral gene products can be expressed in the absence of integration [37,38]. Inducible suicide genes could also be included within the vector such that any transduced cells die from apoptosis or necrosis in the presence of a small molecule. This may be particularly advantageous in stimulating the immune response [39]. Insulators and related sequences could be placed within the LTRs to reduce further the chances of insertional oncogenic activation [40]. Another potential concern is the presence of replication-competent lentivirus (RCL), especially of altered


ARTICLE

host range, within vector supernatants. At present, however, there is no FDA-approved assay for detection of RCL and currently there is no consensus opinion with regard to screening. An obvious trade-off exists between number of encoded gene products, generation of RCL, and induction of an immune response. Novel approaches may further reduce the chances of generation of RCL during vector production [41]. A mixture of single-cycle SIVs expressing multiple glycoproteins but not Nef was recently used to inoculate rhesus macaques repeatedly [22]. Cellular immune responses were easily detected, and after challenge with highly pathogenic SIVmac239 a majority of the animals had controlled viremia, supporting further preclinical evaluation of this vaccine candidate. Here, env was not included in the vector for safety reasons and also because we were most interested in generating cellular immunity against Gag for the challenge model. It should be possible, however, to produce mixed pseudotyped particles by using one or more HIV envelopes in trans, along with VSV G. This approach may allow the generation of neutralizing antibodies against diverse viral clades and strains. Full-length, intact reverse transcriptase (RT) was included in the vector, mainly for convenience, as was nef. Future studies could employ catalytically inactive RT (trans-complemented in the producers with active RT) along with a nef allele deficient for MHC class I down-regulation. This would thus maximize the number of T cell epitopes encoded within the vector and yet maintain a margin of safety. Use of RD HIV as a prophylactic vaccine would depend upon the HIV seroprevalence rate. Use of RD HIV as a therapeutic vaccine involves less obvious risks. First, most antiretroviral therapy would need to be discontinued for at least a few days, which could lead to viral rebound and drug resistance. After cellular transduction, the vector used here maintained all cisacting sequences. Thus, if the same cell were to be infected with replication-competent HIV, there is a chance of vector mobilization. This could amplify the immune response and there would also be the chance of recombination with wild-type HIV. Given the nature of the vector used here, it is unlikely that any recombinant virus would be more virulent or pathogenic than wildtype HIV. It is recognized that the challenge model here was based upon vaccinia virus, whose replication strategy and ability to generate an immune response differ greatly from HIV. The results, however, suggest that RD HIV can elicit a Gag-specific protective cellular response, which is the foundation of many cellular immunity HIV vaccines. While the results presented here are applicable only to a murine model and may not hold true in primate studies, they serve as proof-of-concept that RD HIV appears to be at least as efficacious as Ad5-gag and thus warrants further investigation in nonhuman primate models as a potential

439

ARTICLE

doi:10.1016/j.ymthe.2006.02.021

therapeutic or prophylactic vaccine, possibly in prime– boost combination with other leading candidates.

MATERIALS

AND

METHODS

Preparation of plasmids and vectors. The MLV vector pBabe-puroHACycT1 was constructed as described [42] and encodes a hemagglutinin-tagged, truncated form of human cyclinT1 (amino acids 1–303, kind gift from Kathy Jones) [43]. The MLV vector pBABE-IRES-eGFP was obtained from Garry Nolan. pHIV-CycT1-IRES-eYFP was constructed by blunt-end ligating the 0.9-kb truncated cyclinT1 into the unique SmaI site just upstream of the internal ribosome entry site (IRES) of pHIV-IRES-eYFP [44]. This HIV vector, based upon strain NL4-3, encodes viral genes gag, pol, tat, and rev and has deletions in nef, vif, vpu, vpr, and env. Pseudotyped MLV vector supernatants were produced as described [45] by triple transfection of 293T cells with the appropriate envelope expression construct, pHIT60, and MLV vector. Pseudotyped RD HIV vector supernatants were produced by cotransfection using pME VSV G (Indiana strain) or pCI-Chandipura (1.6-kb cDNA of Chandipura strain of VSV G ligated into pCI from Promega) and pHIV-CycT1-IRES-eYFP and titered by end-point dilution on HOS cells [45]. For each HIV vector, titers (as measured by epifluorescence microscopy) ranged from 1.0 to 3.0 107 IU/ml (average of 2.0 107 IU/ml) before ultracentrifugation and 1.0 109 IU/ml afterward (~50% recovery of IU). Ten million International Units of VSV G-pseudotyped particles contained approximately 2 Ag of CA, as measured by p24 ELISA (Coulter–Beckman). These were stored at 808C or in liquid nitrogen and thawed immediately prior to use. Contaminating HIV plasmid DNA was measured by semiquantitative PCR using HIVspecific primers [45]. Aliquots of concentrated vector supernatant (108 IU) were tested for the presence of replication-competent HIV by a modified marker rescue assay [44] and none was detected. Where indicated vector supernatant was heat-inactivated by incubation at 558C for 60 min. Functionality of truncated cyclinT1. p53 null ( / ) mouse embryo fibroblasts (gift from Larry Donehower) were transduced at low m.o.i. with VSV G-pseudotyped pBabe-puro-HACycT1 or empty vector and positively selected using 20 Ag/ml puromycin. These MEFs were then transduced with HIV-IRES-eYFP(VSV G) using increasing m.o.i. Transduction efficiencies and MFIs of transduced cells were determined by flow cytometry. Unmodified MEFs were transduced using either unconcentrated HIVCycT1-IRES-eYFP (VSV G) or HIV-IRES-eYFP (VSV G) using increasing m.o.i. (as determined on HOS cells). Both MFIs and transduction efficiencies were measured by FACS analysis, collecting at least 10,000 events. Humoral immunity in mice. Female Balb/c mice ages 8–10 weeks were obtained from The Jackson Laboratory and housed at Merck Research Laboratories under standard conditions. Groups of N = 6 mice were immunized im with 2.0 107 or 1.0 108 IU (equivalent to 0.02 or 0.1 ml of concentrated vector stocks) of HIV-CycT1-IRES-eYFP (VSV G), 1.0 108 IU of heat-inactivated HIV-CycT1-IRES-eYFP (VSV G), 100 Ag of plasmid encoding HIV gag, or 1.0 107 or 1.0 108 IU of adenovirus encoding HIV gag (Ad5-gag) [46]. Where indicated, mice were boosted 4 weeks later. Serum ELISAs were performed against HIV capsid (p24) 4 weeks after the first immunization or 3 days after boosting. ELISA plates were prepared by coating Nunc Maxisorb plates with recombinant HIV-1 p24 at 1 mg/ml (Austral Biochemicals) and incubating overnight at 48C. The plates were washed five times with PBS containing 0.05% Tween 20 (PBS-T) and then incubated for 2 h at room temperature with blocking buffer (PBS-T plus 1% BSA). Mouse serum was obtained via retro-orbital bleed or exsanguination. Serial threefold dilutions were made from each serum in PBS-T and incubated on the plates for 2 h at room temperature. Plates were then washed five times with PBS-T. Rabbit anti-mouse IgG–HRP (0.1 ml of 1:1000 dilution, from Zymed) was added to each well and incubated at room temperature for 1 h. Plates were again washed with PBS-T. o-Phenylenediamine dihydrochloride (0.1 ml of 0.67 mg/ml with H2O2) was added to each well and incubated for 10–30 min at room temperature, followed by 0.1 ml of 0.5 M H2SO4 to stop the reaction. Optical density (OD) for each well was read at 492 nm on an ELISA plate reader. Antibody titer (units/ml) was defined as the reciprocal of the

440

highest dilution that gave an OD N 2 standard deviations above the mean of the control wells [47]. Reactivity against denatured VSV G was tested by transiently transfecting 293T cells with pME VSV G alone, preparing a detergent cell lysate 72 h later, performing SDS–PAGE immunoblotting using various mouse sera at 1:1000 as primary followed by goat anti-mouse antibody conjugated to HRP at 1:1000, and developing using enhanced chemiluminescence (ECL; Amersham). Neutralizing activity against VSV G was tested by preincubating sera for 1 h with two different amounts of MLV or HIV pseudotyped particles and then transducing 5 105 HOS cell targets with the mix. Resultant titer was determined by epifluorescence microscopy 72 h later. Reactivity against eYFP was tested by purifying GST or a GST–eGFP fusion protein from Escherichia coli nonionic detergent lysate by use of glutathione Sepharose beads (Pharmacia). Approximately 2 Ag of purified protein was size-separated by SDS–PAGE and immunoblotted as described above. Cellular immunity in mice. After sacrifice, splenocytes were isolated aseptically from each mouse spleen and pooled in groups of three per immunization group. The ELISpot assay was used to measure HIV-specific IFN-g-secreting cells [48]. Multiscreen opaque plates (Millipore, Molsheim, France) were coated with 5 Ag/ml purified rat anti-mouse IFN-g IgG1, clone R4-6A2 (Pharmingen, San Diego, CA, USA) in PBS at 48C overnight. Subsequently, 1 105 to 5 105 cells in 0.1 ml of complete T cell medium were added to each well in triplicate along with either HIV Gag or Pol peptides. Background wells had cells incubated with complete medium only. Plates were incubated for 24 h at 378C in a water-jacketed, 5% CO2 incubator, following which they were washed and incubated at 48C overnight with 1.25 Ag/ml of biotin-conjugated rat anti-mouse IFN-g monoclonal antibody (clone XMG1.2 from Pharmingen). The plates were developed as described [48]. The number of spots in each well were manually counted using a dissecting microscope and normalized to an input of 106 cells. Vaccinia virus challenge. Both the control vaccinia virus encoding hgalactosidase (VV-sc) and the vaccinia virus encoding HIV gag (VV-gag) [46] were plate-amplified in HOS-TK cells. Infected cells, associated debris, and supernatant were collected and centrifuged at 600g for 5 min. Using a Branson Sonifier 450, pellets were sonicated in an ice slurry until the solution was homogeneous with no visible cell debris. The resulting lysate was titered on HOS TK cells by triplicate, serial 10-fold dilutions and scored for PFU after crystal violet staining of fixed cells. Resulting VVsc and VV-gag stock titers were adjusted to be ~1 109 PFU/ml. Female Balb/c mice ages 8–10 weeks were housed in microisolator cages at the Center for Comparative Medicine at the Baylor College of Medicine. In the first series of experiments, mice were immunized im with 100 Ag of sterile plasmid DNA encoding gag, pol, tat, and rev (HIV-PV [49]) (N = 12), 5 106 IU of HIV-CycT1-IRES-eYFP (VSV G) (N = 24), or 0.2 ml of sterile PBS (N = 16). Mice that received plasmid DNA were boosted with the same amount of plasmid 4 weeks later, and 12 of the mice that received RD HIV were also boosted with the same amount of RD HIV. Mice were subsequently challenged with 1.0 106 PFU of VV-gag or 1.0 107 PFU VV-sc injected ip. In the second challenge experiment mice were immunized im with 100 Ag of plasmid DNA as above (N = 10), 5.0 106 IU RD HIV (N = 10), or 0.2 ml PBS (N = 10). Mice receiving plasmid were boosted with the same amount of DNA 4 weeks later, and mice receiving RD HIV were boosted with the same amount of IU 4 weeks later. One week later mice were challenged ip with 1.0 107 IU of VV-gag. Three days postchallenge, all mice were sacrificed and ovaries collected bilaterally under aseptic conditions. Single cells were isolated by mechanical disruption of the ovaries through a sterile nylon mesh into 0.5 ml of complete DMEM with 1% penicillin and streptomycin. The amount of recoverable VV was titered as described above. Statistical analyses. Differences in recoverable amounts of VV between the vaccinated mice and the sham-immunized mice following either VVgag or VV-sc challenge were analyzed by the Kruskal–Wallis test for


doi:10.1016/j.ymthe.2006.02.021

differences between the groups as a whole and by the Wilcoxon scores (rank sums) for differences between sham-immunized mice and immunized groups.

ACKNOWLEDGMENTS This work was supported in part by NIH T32 AI07456. We thank Drs. Larry Donehower (Baylor), Jack Rose (Yale), Garry Nolan (Stanford), and Kathy Jones (Salk Institute) for generous reagent gifts. We acknowledge Stanley Cron of the Baylor/UT–Houston CFAR Design and Analysis Core Facility for provision of statistical expertise. R.E.S. was an Edward Mallinckrodt, Jr., Foundation Scholar. RECEIVED FOR PUBLICATION AUGUST 27, 2005; REVISED FEBRUARY 6, 2006; ACCEPTED FEBRUARY 23, 2006.

REFERENCES 1. Piot, P., et al. (2001). The global impact of HIV/AIDS. Nature 410: 968 – 973. 2. Nabel, G. J. (2001). Challenges and opportunities for development of an AIDS vaccine. Nature 410: 1002 – 1007. 3. Robinson, H. L. (2002). New hope for an AIDS vaccine. Nat. Rev. Immunol. 2: 239 – 250. 4. Daniel, M. D., et al. (1992). Protective effects of a live attenuated SIV vaccine with a deletion in the nef gene. Science 258: 1938 – 1941. 5. Wyand, M. S., et al. (1999). Protection by live, attenuated simian immunodeficiency virus against heterologous challenge. J. Virol. 73: 8356 – 8363. 6. Almond, N., et al. (1995). Protection by attenuated simian immunodeficiency virus in macaques against challenge with virus-infected cells. Lancet 345: 1342 – 1344. 7. Johnson, R. P., and Desrosiers, R. C. (1998). Protective immunity induced by live attenuated simian immunodeficiency virus. Curr. Opin. Immunol. 10: 436 – 443. 8. Baba, T. W., et al. (1999). Live attenuated, multiply deleted simian immunodeficiency virus causes AIDS in infant and adult macaques. Nat. Med. 5: 194 – 203. 9. Baba, T. W., et al. (1995). Pathogenicity of live, attenuated SIV after mucosal infection of neonatal macaques. Science 267: 1820 – 1825. 10. Hofmann-Lehmann, R., et al. (2003). Live attenuated, nef-deleted SIV is pathogenic in most adult macaques after prolonged observation. AIDS 17: 157 – 166. 11. Sawai, E. T., et al. (2000). Pathogenic conversion of live attenuated simian immunodeficiency virus vaccines is associated with expression of truncated Nef. J. Virol. 74: 2038 – 2045. 12. Chakrabarti, L. A., et al. (2003). A truncated form of Nef selected during pathogenic reversion of simian immunodeficiency virus SIVmac239Deltanef increases viral replication. J. Virol. 77: 1245 – 1256. 13. Ruprecht, R. M. (1999). Live attenuated AIDS viruses as vaccines: promise or peril? Immunol. Rev. 170: 135 – 149. 14. Wyand, M. S., et al. (1996). Vaccine protection by a triple deletion mutant of simian immunodeficiency virus. J. Virol. 70: 3724 – 3733. 15. Desrosiers, R. C., et al. (1998). Identification of highly attenuated mutants of simian immunodeficiency virus. J. Virol. 72: 1431 – 1437. 16. Barouch, D. H., et al. (2000). Control of viremia and prevention of clinical AIDS in rhesus monkeys by cytokine-augmented DNA vaccination. Science 290: 486 – 492. 17. Amara, R. R., et al. (2001). Control of a mucosal challenge and prevention of AIDS by a multiprotein DNA/MVA vaccine. Science 292: 69 – 74. 18. Rose, N. F., et al. (2001). An effective AIDS vaccine based on live attenuated vesicular stomatitis virus recombinants. Cell 106: 539 – 549. 19. Shiver, J. W., et al. (2002). Replication-incompetent adenoviral vaccine vector elicits effective anti-immunodeficiency-virus immunity. Nature 415: 331 – 335. 20. Quinonez, R., and Sutton, R. E. (2002). Lentiviral vectors for gene delivery into cells. DNA Cell Biol. 21: 937 – 951. 21. Tung, F. Y., Jr.Rinaldo, C. R., and Montelaro, R. C. (1998). Replication-defective HIV as a vaccine candidate. AIDS Res. Hum. Retroviruses 14: 1247 – 1252. 22. Evans, D. T., et al. (2005). Immunization of macaques with single-cycle simian immunodeficiency virus (SIV) stimulates diverse virus-specific immune responses and reduces viral loads after challenge with SIVmac239. J. Virol. 79: 7707 – 7720.


ARTICLE

23. Rose, N. F., Roberts, A., Buonocore, L., and Rose, J. K. (2000). Glycoprotein exchange vectors based on vesicular stomatitis virus allow effective boosting and generation of neutralizing antibodies to a primary isolate of human immunodeficiency virus type 1. J. Virol. 74: 10903 – 10910. 24. Smith, S. M., et al. (2001). Constitutively dead, conditionally live HIV-1 genomes: ex vivo implications for a live virus vaccine. J. Biol. Chem. 276: 32184 – 32190. 25. Marzio, G., Verhoef, K., Vink, M., and Berkhout, B. (2001). In vitro evolution of a highly replicating, doxycycline-dependent HIV for applications in vaccine studies. Proc. Natl. Acad. Sci. USA 98: 6342 – 6347. 26. Xiao, Y., et al. (2000). Dox-dependent SIVmac with tetracycline-inducible promoter in the U3 promoter region. Virology 269: 268 – 275. 27. Das, A. T., Baldwin, C. E., Vink, M., and Berkhout, B. (2005). Improving the safety of a conditional-live human immunodeficiency virus type 1 vaccine by controlling both gene expression and cell entry. J. Virol. 79: 3855 – 3858. 28. Reif, J. S., et al. (1987). Epizootic vesicular stomatitis in Colorado, 1982: infection in occupational risk groups. Am. J. Trop. Med. Hyg. 36: 177 – 182. 29. Schmitz, H., Wigand, R., and Heinrich, W. (1983). Worldwide epidemiology of human adenovirus infections. Am. J. Epidemiol. 117: 455 – 466. 30. Baumann, J. G., et al. (2004). Murine T cells potently restrict human immunodeficiency virus infection. J. Virol. 78: 12537 – 12547. 31. Mitchell, R. S., et al. (2004). Retroviral DNA integration: ASLV, HIV, and MLV show distinct target site preferences. PLoS Biol. 2: 1127 – 1137. 32. Schroder, A. R., et al. (2002). HIV-1 integration in the human genome favors active genes and local hotspots. Cell 110: 521 – 529. 33. Hacein-Bey-Abina, S., et al. (2003). LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 302: 415 – 419. 34. Wu, X., Li, Y., Crise, B., and Burgess, S. M. (2003). Transcription start regions in the human genome are favored targets for MLV integration. Science 300: 1749 – 1751. 35. Hazuda, D. J., et al. (2004). Integrase inhibitors and cellular immunity suppress retroviral replication in rhesus macaques. Science 305: 528 – 532. 36. Hazuda, D. J., et al. (2000). Inhibitors of strand transfer that prevent integration and inhibit HIV-1 replication in cells. Science 287: 646 – 650. 37. Wu, Y., and Marsh, J. W. (2001). Selective transcription and modulation of resting T cell activity by preintegrated HIV DNA. Science 293: 1503 – 1506. 38. Wu, Y., and Marsh, J. W. (2003). Early transcription from nonintegrated DNA in human immunodeficiency virus infection. J. Virol. 77: 10376 – 10382. 39. Matzinger, P. (2002). The danger model: a renewed sense of self. Science 296: 301 – 305. 40. Burgess-Beusse, B., et al. (2002). The insulation of genes from external enhancers and silencing chromatin. Proc. Natl. Acad. Sci. USA 99(Suppl. 4): 16433 – 16437. 41. Evans, D. T., Bricker, J. E., and Desrosiers, R. C. (2004). A novel approach for producing lentiviruses that are limited to a single cycle of infection. J. Virol. 78: 11715 – 11725. 42. Coskun, A. K., and Sutton, R. E. (2005). Expression of glucose transporter 1 confers susceptibility to human T-cell leukemia virus envelope-mediated fusion. J. Virol. 79: 4150 – 4158. 43. Wei, P., et al. (1998). A novel CDK9-associated C-type cyclin interacts directly with HIV-1 Tat and mediates its high-affinity, loop-specific binding to TAR RNA. Cell 92: 451 – 462. 44. Segall, H., Yoo, E., and Sutton, R. E. (2003). Characterization and detection of artificial replication-competent lentivirus of altered host range. Mol. Ther. 8: 118 – 129. 45. Sutton, R. E., Reitsma, M. J., Uchida, N., and Brown, P. O. (1999). Transduction of human progenitor hematopoietic stem cells by human immunodeficiency virus type 1 vectors is cell-cycle dependent. J. Virol. 73: 3649 – 3660. 46. Casimiro, D. R., et al. (2002). Vaccine-induced immune responses in rodents and nonhuman primates by use of a humanized human immunodeficiency virus type 1 pol gene. J. Virol. 76: 185 – 194. 47. Fan, J., et al. (2004). Preclinical study of influenza virus A M2 peptide conjugate vaccines in mice, ferrets, and rhesus monkeys. Vaccine 22: 2993 – 3003. 48. Miyahira, Y., et al. (1995). Quantification of antigen specific CD8+ T cells using an ELISPOT assay. J. Immunol. Methods 181: 45 – 54. 49. Sutton, R. E., et al. (1998). Human immunodeficiency virus type 1 vectors efficiently transduce human hematopoietic stem cells. J. Virol. 72: 5781 – 5788.

441