Inhibition of Human Immunodeficiency Virus Type 1 (HIV-1) Replication by HIV-1-Based Lentivirus Vectors Expressing Transdominant Rev

Abstract

Retrovirus vectors expressing transdominant-negative mutants of Rev (TdRev) inhibit human immunodeficiency virus type 1 (HIV-1) replication by preventing the nuclear export of unspliced viral transcripts, thus inhibiting the synthesis of Gag-Pol, Env, and genomic RNA. The use of HIV-1–based vectors to express TdRev would have the advantage of allowing access to nondividing hematopoietic cells. It would also provide additional levels of protection by sequestering the viral regulatory proteins Tat and Rev, competing for encapsidation into wild-type virions, and inhibiting reverse transcription. Here we describe HIV-1-based vectors that express TdRev. These vectors contain mutations in the splicing signals or replacement of the Rev-responsive element by the simian retrovirus type 1 constitutive transport element, making them less sensitive to the inhibitory effects of TdRev. In addition, overexpression of Rev and the use of an HIV-1 helper plasmid that drives high levels of Gag-Pol synthesis were used to transiently overcome the inhibition by TdRev of the synthesis of Gag-Pol during vector production. SupT1 cells transduced with these vectors were more resistant to HIV-1 replication than cells transduced with Moloney murine leukemia virus-based vectors expressing TdRev. Furthermore, we show that these vectors can be mobilized by the wild-type virus, reducing the infectivity of virions escaping inhibition and conferring protection against HIV-1 replication to previously untransduced cells.

Antiviral gene therapy for human immunodeficiency virus type 1 (HIV-1) aims to reconstitute the immune system with genetically altered cells that resist HIV-1 infection. Initial experiments toward achievement of this goal have used ex vivo transduction of CD4⁺ T cells or CD34⁺ progenitor cells with Moloney murine leukemia virus (MoMLV)-based retroviruses expressing anti-HIV-1 genes, followed by reinfusion of the altered cells into recipients. This approach has been shown to inhibit HIV-1 replication in vitro and to prolong T-cell survival in vivo (9, 18, 19, 25, 32, 33, 41, 42). The advantages of using murine retrovirus vectors as anti-HIV-1 delivery vehicles are that they can efficiently transduce hematopoietic cells, they have a proven clinical safety record, and neither the integrated vector nor the packaging cell line is sensitive to the action of the anti-HIV-1 genes expressed by the vector. However, MoMLV-based vectors can integrate only into dividing cells, such as activated T cells or ex vivo-cultured CD34⁺ cells, and cannot access quiescent T cells or the macrophage reservoir, which may play an important role in the maintenance of HIV-1 infection (11, 12).

The use of HIV-1–based conditionally replicating vectors expressing anti-HIV-1 genes for the treatment of AIDS has several theoretical advantages over the current vector systems being used. These vectors can effectively transduce both dividing and nondividing cells, such as resting T cells, macrophages, or dendritic cells, which have a central role in the HIV-1 life cycle. Even in the absence of specific inhibitory genes, these vectors would be able to inhibit HIV-1 replication by binding the HIV-1 regulatory proteins Tat and Rev and by competing for packaging into HIV-1 virions, facilitating the spread of the vector to unprotected cells in vivo (2, 6, 8, 10). Therefore, in vivo mobilization of the vector by patients' HIV-1 may result in transfer of the vector to the viral reservoirs that may represent sites of long-term maintenance of wild-type HIV-1. Furthermore, this inhibition could be enhanced by the expression of anti-HIV-1 genes.

Among the most powerful inhibitors of HIV-1 replication are transdominant-negative mutants of Rev (TdRev) (20, 31, 40). TdRev acts by inhibiting the nuclear export of Rev through the formation of inactive multimers (15).

The expression of TdRev in an HIV-1–based vector packaging system causes a problem, as the expression of this protein is expected to inhibit the production of HIV-1 Gag-Pol encoded by the packaging plasmid as well as the transport of unspliced vector genomes to the cytoplasm. The inhibition by TdRev of the vector genomic RNA is a problem not only during vector production but also in the target cell, where the ability of the vector genomic RNA to act as a defective interfering particle (DIP) would be greatly reduced if the vector were sensitive to TdRev. In order to provide the most effective inhibition of HIV-1 replication while favoring its own packaging and spread to secondary target cells, the vector must have a competitive advantage for packaging over the wild-type viral mRNA. As genomic RNA needs to be exported to the cytoplasm in order to be packaged, the use of a vector that is not sensitive to TdRev would favor its nuclear export and subsequent packaging over the wild-type viral genomic RNA.

In a previous work, we studied the effect of modifications that rendered HIV-1–based lentivirus vectors less sensitive to the inhibitory action of TdRev (23). The two strategies tested involved mutation of the splicing signals flanking the Rev-responsive element (RRE) present in the vector and replacement of the RRE by a simian retrovirus type 1 (SRV-1) constitutive transport element (CTE). These vectors could be packaged with almost the same efficiency as classic Rev-dependent HIV-1–based vectors and showed increased levels of cytoplasmic unspliced mRNAs as well as a moderate but significant decrease in sensitivity to TdRev.

Building on those studies, we tested whether lentivirus vectors directly expressing TdRev could be packaged with a Rev-dependent HIV-1 helper at titers sufficient for ex vivo transduction. As the inhibition by TdRev of the synthesis of Gag-Pol need only be overcome transiently during vector packaging, we tested the effect of overexpressing Rev as an alternative to increasing vector titers. Here we demonstrate that TdRev-expressing lentivirus vectors can be efficiently produced and that they confer potent inhibition of HIV-1 replication in vitro. Furthermore, we show that these vectors can be mobilized by the wild-type virus, reducing the infectivity of virions escaping inhibition and conferring protection against HIV-1 replication to previously untransduced cells.

MATERIALS AND METHODS

Plasmids.

TdRev (mutant RD3) was isolated by PCR from plasmid pGCsapSL3RD3 using the primers 5′-CTCGGATCCGCCATGGCAGGAAGAAGC-3′ and 5′-TCCGGATCCATGCATCGACTATTCTTTAGCTCC-3′. The fragment containing the RD3 gene in pGCsapSL3RD3 was previously cloned from plasmid LR^DLDDSN (31). The PCR fragment containing the RD3 gene was cloned into plasmid pPUR (Clontech), digested with SmaI and PvuII, to create plasmid pSRD3. In this plasmid, the expression of RD3 is driven by a simian virus 40 (SV40) promoter.

Plasmids pcSR7 and pcSR9 were created by cloning a PvuII-XbaI fragment containing the SV40-RD3 expression cassette from pSRD3 into the StuI site of plasmids pNL7 and pNL9, respectively (23). Plasmid pcSR9–3C was created by amplifying the SV40-RD3 expression cassette from pSRD3 with primers that introduced flanking SpeI and XbaI sites. This amplified fragment was used to replace the SpeI fragment containing the cytomegalovirus (CMV)-enhanced green fluorescent protein (EGFP) expression cassette present in pcCG9–3C (21).

The construction of plasmids pcSN7 and pHSN2 has been previously described (23). HIV-1 helper packaging vectors pCMVΔR8.2 (27) and pCD/NLBH (24) were used to package these vectors as virus particles. Plasmid pLTR-G (34) codes for the vesicular stomatitis virus envelope glycoprotein, driven by the HIV-1 long terminal repeat (LTR) promoter. Plasmid pCMVRev was obtained from the AIDS Research and Reference Reagent Program, National Institute of Allergy and Infectious Diseases, National Institutes of Health. GC-SN_DC is an MoMLV-based vector that confers neomycin resistance (7), and GC-Rev^TDSN (anti-TAR_DC) is another MoMLV-based vector that expresses TdRev plus an antisense sequence complementary to HIV-1 trans-activating response element (TAR) driven by an RNA polymerase III expression cassette, cloned in a double-copy configuration (40).

HIV-1 inhibition determined by a transient transfection assay.

293T cells (2 × 10⁵) (28) were plated on six-well poly-d-lysine-coated plates and transfected with 2 μg of vector plasmid DNA, 0.5 μg of plasmid pNL4–3 (vector/HIV-1 molar ratio, ≈8:1), and 0.5 μg of the luciferase-encoding plasmid pGL3 (Promega) using Superfect transfection reagent (Qiagen). Cells were washed 3 h after transfection and cultured for 48 h in Dulbecco modified Eagle medium (DMEM)–10% fetal calf serum (FCS). The amount of virus in the supernatant was determined by a p24 antigen capture assay (Coulter) and was subsequently normalized to the transfection efficiency as determined by luciferase activity.

Vector production and DNA transfection.

Virus vectors were produced by transient transfection of 4 × 10⁶ 293T cells plated on poly-d-lysine-coated 100-mm dishes (Becton Dickinson). Cells were cultured in DMEM supplemented with 10% FCS, 2 mM glutamine, and 100 μg of penicillin/ml in a 5% CO₂ incubator for 24 h. The culture medium was changed 4 h prior to transfection. The calcium phosphate DNA coprecipitation method was used to transfect a total of 22 μg of plasmid DNA per dish: 7 μg of transfer vector plasmid, 8 μg of packaging helper plasmid pCD/NLBH, 3.5 μg of pLTR-G, and 5 μg of pCMVRev. After 16 h, cells were washed three times with phosphate-buffered saline, 10 ml of new medium was added, and cells were incubated for another 48 h. Supernatants were harvested and filtered through 0.45-μm-pore-size cellulose acetate filters. Virus preparations were concentrated ∼40-fold by ultracentrifugation at 50,000 × g for 90 min and resuspended in 0.5 ml of RPMI medium.

Titer determination.

Physical titer determination for vectors cSR7 and cSR9–3C relative to cSN7 (23) was performed by comparing the DNA-free RNA levels present in the vector supernatants with the RNA levels obtained for vector cSN7, whose titer can be functionally measured (see below). RNA purification was carried out by adding 100 μg of carrier yeast tRNA and 1 μg of tracer plasmid DNA (pEGFP-N1; Clontech) to 2.5 ml of filtered nonconcentrated vector supernatant. RNA was extracted with 7.5 ml of Trizol LS (Gibco BRL) and 2 ml of chloroform. The aqueous phase was precipitated with isopropanol, and the RNA was digested with RNase-free DNase. RNA was reextracted with Trizol LS-chloroform and precipitated. The RNA concentration was determined, and equal amounts of RNA (5, 1.5, and 0.5 μg) for each sample were dot blotted on duplicate nylon membranes. One membrane was hybridized with a probe corresponding to the LTR-gag region common to all vectors analyzed, and the other membrane was hybridized with an EGFP probe to control for contaminating DNA. Quantification of the hybridization signals was performed by phosphorimaging (Fuji BAS-1500). Additionally, physical vector titers were also determined by measuring p24 by an antigen capture assay (Coulter).

For titer determination for the reference vector cSN7, 10⁵ TE671 cells (ATCC CRL-8805)/well were infected in triplicate with serial dilutions of filtered vector supernatant prepared in DMEM–10% FCS supplemented with 8 μg of Polybrene/ml. After 12 to 14 days under selection in medium supplemented with G418 at 1 mg/ml, antibiotic-resistant colonies were stained and counted.

Transduction of SupT1 cells.

SupT1 cells (2 × 10⁵ to 3 × 10⁵) were transduced with 1 ml of concentrated supernatant (∼5,000 ng of p24/ml; multiplicity of infection [MOI], 20) from vectors encoding TdRev. Transduction was performed with fibronectin-coated 12-well plates in the presence of 5 μg of protamine sulfate/ml for 16 h. Cells transduced with vector cSN7 or HSN2 were maintained under selection in RPMI medium–10% FCS supplemented with G418 at 1 mg/ml for 3 weeks.

Southern and Northern blot analyses.

The structures of the proviruses and the expression of the RD3 transgene in SupT1 cells were determined by Southern and Northern blot analyses by standard procedures. Total RNA was purified using RNeasy columns (Qiagen), electrophoresed on 1% agarose–0.66 M formaldehyde gels, transferred to Hybond N⁺ membranes, and hybridized with a radiolabeled RD3 DNA probe. The percentages of SupT1 cells that were transduced with vectors cSR7 and cSR9–3C were estimated by quantitative Southern blotting. Standard genomic DNAs were purified from different cell mixtures of untransduced SupT1 cells and SupT1 cells stably transduced with cSN7, which were considered to be 100% transduced and arbitrarily to have an average of at least one vector copy per cell. Genomic DNAs (10 μg) purified from the standards and from SupT1 cells transduced with each vector were digested with EcoRV, electrophoresed through 1% agarose gels, transferred to nylon membranes, and hybridized with a probe complementary to the LTR-gag region common to the RD3-positive vectors and pcSN7. Quantification of the hybridization signals was performed by phosphorimaging.

Western blotting of TdRev.

Protein extracts obtained from SupT1 cells transduced with vector cSR7 or cSR9–3C as well as from 293T cells transiently transfected with plasmid pcSR7, pcSR9–3C, pSRD3, or pCMVRev were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and subsequent immunoblotting using a sheep anti-HIV-1 Rev polyclonal antibody (dilution, 1:5,000; Fitzgerald, Concord, Mass.) as a primary antibody and horseradish peroxidase–polyclonal donkey anti-sheep immunoglobulin G as a secondary antibody (dilution, 1:10,000; Bethyl Laboratories, Montgomery, Tex.). Chemiluminescence detection was performed using ECL Plus detection reagents (Amersham-Pharmacia). Ten nanograms of a purified recombinant HIV-1 Rev protein (Bachem, San Diego, Calif.) was used as a positive control.

HIV-1 infection.

Strain HIV-1_NL4–3 (10^6.4 50% tissue culture infective doses/ml) was obtained from the AIDS Research and Reference Reagent Program. Transduced SupT1 cells (10⁶/4 ml of RPMI medium–10% FCS plus 4 μg of Polybrene/ml) were infected with HIV-1_NL4–3 at MOIs ranging from 0.001 to 0.1 in six-well plates. After 4 h of incubation at 37°C, cells were washed with 10 ml of phosphate-buffered saline, resuspended in 5 ml of RPMI medium–10% FCS, and plated in six-well plates. Aliquots of culture medium (2 ml) were removed every 2 days and replaced with fresh medium. Samples (0.2 ml) of these aliquots were stored in 96-well plates at −80°C for subsequent reverse transcriptase (RT) assays; the remaining 1.8 ml was reserved for infectivity experiments. Cell viability was determined by trypan blue exclusion every 5 days.

For reinfection of naive SupT1 cells with supernatants collected from the primary challenges, 2-ml aliquots of supernatants were frozen at −80°C, and a 200-μl aliquot from each supernatant was saved for p24 analysis. After p24 determination, supernatants were filtered and diluted to achieve the same p24 concentrations.

PCR detection of vector and virus sequences.

The primers used for PCR amplification of a 272-bp RD3-specific fragment were as follows: 5′-GGACCCGACAGGCCCGAAGGAATA-3′ and 5′-TCCGGATCCATGCATCGACTATTCTTTAGCTCC-3′. For the amplification of a 912-bp HIV-1_NL4–3-specific fragment, the primers used were as follows: 5′-GAAGAAGCGGAGACAGCGACGAAGAG-3′ and 5′-GCAAAACCAGCCGGGGCACAAT-3′. Cycling conditions were as follows: 15 min at 95°C; 34 cycles of denaturation (30 s at 94°C), annealing (30 s at 56°C), and elongation (1 min at 72°C); and a final extension cycle of 5 min at 72°C. A 200-ng sample of genomic DNA in a 50-μl reaction volume was used for each PCR.

RESULTS

Vector design and effect of TdRev on p24 synthesis.

In a previous work, we demonstrated that vectors containing mutations in the splicing signals flanking the RRE sequence expressed higher levels of cytoplasmic unspliced RNA in the absence of Rev expression and showed decreased sensitivity to the expression of TdRev supplied in trans without affecting vector titers (23). We also showed that lentivirus vectors in which RRE is replaced by the SRV-1 CTE have similar properties (21, 23). Based on these results, these vector designs were used to clone expression cassettes encoding TdRev under the control of the SV40 promoter (Fig. 1).

FIG. 1.

Diagrams of retrovirus vectors and plasmids used in this study. HIV-1–based vectors are indicated with white LTRs, and MoMLV-based vectors are indicated with gray LTRs. Δgag, truncation of HIV-1 gag; cP, central polypurine tract and central termination sequence; SD, splice donor; SA, splice acceptor; SV40-P, SV40 immediate-early promoter; Neo, neomycin phosphotransferase; asT, antisense complementary to HIV-1 TAR driven by an RNA polymerase III expression cassette, cloned in double-copy configuration; pA, polyadenylation signal.

As a preliminary way to test whether the expression of TdRev, in the context of these vectors, was able to inhibit HIV-1 Gag-Pol synthesis, we performed a one-step inhibition assay by transiently cotransfecting 2 μg of each vector with 0.5 μg of plasmid pNL4–3, which has a full proviral copy of HIV-1_NL4–3 (1). As a negative control we used the plasmid pBluescript (Stratagene). Figure 2 shows that transfection of a plasmid encoding TdRev (pSRD3) produces 50% inhibition of the synthesis of p24. A similar inhibitory effect is observed after transfection of vectors pcSN7 and pcSN9–3C, which express a neo gene instead of TdRev. This inhibitory effect has been previously attributed to the sequestration of the regulatory proteins Tat and Rev, which upregulate the expression of Gag-Pol (8). We next tested the inhibitory properties of the TdRev-containing vectors pcSR7 and pcSR9–3C. In the context of a lentivirus vector backbone, the expression of TdRev by the SV40 promoter in vectors pcSR7 and pcSR9–3C produces an additional inhibitory effect, leading to an almost complete block of the synthesis of p24 in this assay.

FIG. 2.

Inhibition of p24 synthesis encoded by HIV-1_NL4–3 in a transient transfection assay. The indicated plasmids were cotransfected at an 8:1 molar ratio with plasmid pNL4–3 into 293T cells. After 48 h, p24 levels in the supernatant were determined. Luciferase activity (as relative light units [RLU]), provided by inclusion of plasmid pGL3 in the transfection mixture, was measured to normalize for transfection efficiency. Error bars represent the standard deviations of triplicate transfections.

Partial rescue of Gag-Pol synthesis by Rev overexpression.

As the inhibition of Gag-Pol synthesis likely will limit lentivirus vector production, we tested the inhibitory effect of increasing doses of a vector expressing TdRev on the synthesis of Gag-Pol and tested whether this inhibition could be alleviated by coexpression of increasing amounts of Rev. We tested two different Gag-Pol helper plasmids that have been shown to express different levels of p24 after transfection (23). Plasmids pCMVΔR8.2 (27) and pCD/NLBH (24) differ in the size of the env gene deletion and in the presence of an SV40 origin of replication inserted within the 1.2-kb deleted env region of the latter construct. Both plasmids code for all HIV-1 accessory proteins.

We cotransfected each helper plasmid with different amounts of vector pcSR9 coding for TdRev and increasing amounts of the wild-type Rev expression plasmid pCMVRev (Fig. 3). In the absence of TdRev expression, cotransfection of 3 to 5 μg of pCMVRev can double the amount of p24 synthesized by helper plasmid pCMVΔR8.2. Conversely, p24 synthesis driven by this helper plasmid suffers an ∼20-fold reduction after transfection of only 1 μg of plasmid expressing TdRev. This inhibition can be partially overcome by coexpression of Rev, although p24 values are still below baseline (∼20 ng/ml) even at the lowest dose of the input TdRev-expressing vector. In contrast, helper plasmid pCD/NLBH shows only a fourfold reduction in the synthesis of p24 after cotransfection of 8 μg of vector coding for TdRev. In this situation, the coexpression of 5 μg of pCMVRev can moderately increase p24 expression (approximately twofold). An interesting observation is that even at the highest dose of TdRev-expressing vector and without Rev overexpression, p24 levels expressed by pCD/NLBH (120 ng/ml) are twofold higher than p24 levels expressed by pCMVΔR8.2 in the absence of TdRev expression (65 ng/ml). Due to the differences in sensitivity to TdRev expressed by these two HIV-1 helper plasmids, we chose to package the TdRev-expressing vectors with helper plasmid pCD/NLBH.

FIG. 3.

Effects of TdRev and Rev expression on the synthesis of p24 driven by helper plasmids pCVMΔR8.2 and pCD/NLBH. Each helper plasmid (6 μg) was cotransfected with 0, 1, 4, or 8 μg of vector pcSR9 coding for TdRev and 0, 0.5, 1, 3, or 5 μg of plasmid pCMVRev by the calcium phosphate coprecipitation method. The molar ratios of pcSR9 to helper plasmid were 0, 0.3, 1.3, and 2.7, respectively. Plasmid pGL3 (1 μg) expressing luciferase was included in all transfections to control for equal transfection efficiencies. Plasmid pBluescript was included as a carrier to yield equal amounts of plasmid DNA in each transfection. The p24 levels were determined 48 h after transfection.

Packaging of TdRev-expressing lentivirus vectors.

To package the TdRev-expressing vectors into lentivirus particles, we performed a three-plasmid cotransfection into 293T cells. The transfection mixture consisted of 7 μg of either vector, 8 μg of helper pCD/NLBH, and 3.5 μg of vesicular stomatitis virus envelope glycoprotein-encoding plasmid. To test the effect of Rev overexpression on vector titers, we performed transfections with or without 5 μg of pCMVRev. As these vectors lack a selectable marker, to allow functional evaluation of titers we estimated the physical vector titers by molecular methods (22; see also Materials and Methods). We quantified the amounts of vector RNA and capsid protein p24 present in the vector supernatants and correlated these data with the functional vector titers for the reference vector cSN7. The functional titers (determined by G418 resistance) for vector cSN7 were 3 × 10⁵ and 5 × 10⁵ transducing units (tu)/ml in the absence and presence of Rev overexpression, respectively. The estimated physical vector titers for vectors cSR7 and cSR9–3C are shown in Table 1. It can be seen that overexpression of Rev has only a moderate effect in improving vector titers measured as vector RNA, producing a maximum increase of 1.5-fold for cSN7 (P < 0.07) and cSR9–3C (P < 0.01). Physical titers estimated by p24 measurements showed a consistent twofold increase in the presence of Rev for all vectors tested (P < 0.01). These results are in accordance with the experiment shown in Fig. 3 for plasmid pCD/NLBH, where the p24 values were increased twofold after overexpression of Rev. A 40-fold concentration of these vector supernatants by ultracentrifugation yielded estimated titers of up to 2 × 10⁷ tu/ml.

TABLE 1.

Estimation of physical vector titers^a

Vector	− Rev		+ Rev
	vRNAb	p24c	vRNAb	p24c
cSN7	3.0 × 105	3.2 × 105	5.0 × 105	6.5 × 105
cSR7	2.5 × 105	3.6 × 105	2.9 × 105	6.7 × 105
cSR9-3C	1.9 × 105	1.1 × 105	3.1 × 105	3.0 × 105

^aFunctional vector titers were measured by endpoint dilution for reference vector cSN7 (n = 3). Physical vector titers were determined by measuring either vector RNA (vRNA) or p24 present in the supernatant (n = 2). Average values were compared to those obtained for vector cSN7 and are expressed as transducing units per milliliter. 

^bSee Material and Methods for details. 

^cTiters based on p24 were estimated assuming 500 tu/ng of p24. This value was obtained by measuring functional titers (transducing units per milliliter) and p24 concentrations (nanograms per milliliter) in nonconcentrated supernatants for vector cSN7. 

Transduction of SupT1 cells.

SupT1 cells were transduced with 1 ml of concentrated supernatant from vector cSR7 (14 μg of p24; MOI, ∼36) or cSR9–3C (6 μg of p24; MOI, ∼15) or with 1 ml of nonconcentrated supernatant from vector cSN7 (MOI, 1). SupT1 cells transduced with control vector cSN7 or the MoMLV-based vectors GC-SN_DC and GC-Rev^TDSN (anti-TAR_DC) (40) were selected in G418 and therefore are considered to be 100% transduced. To estimate the fraction of cells transduced with the TdRev-expressing vectors, we performed quantitative Southern blotting. As a quantitative standard, we used genomic DNA purified from mixtures of SupT1 and SupT1/cSN7 cells at different ratios. Equal amounts of genomic DNA were subjected to Southern blot hybridization using a probe corresponding to the LTR-gag region, which is common to cSN7 and the TdRev-expressing vectors. Quantification of the hybridization signals allowed the estimation of the average vector copy number per cell. The results indicated that cell populations with average vector copy numbers of 1.4 and 0.74 were obtained after transduction with vectors cSR7 and cSR9–3C, respectively. The sizes of the restriction fragments observed by Southern blotting also indicated that the vectors had not suffered major genomic rearrangements during the process of packaging, reverse transcription, and integration (data not shown).

To analyze the level of expression of TdRev mRNA in these populations of cells, total RNA was purified and subjected to Northern blot hybridization with a probe complementary to the RD3 gene. Expression of RD3 mRNA in SupT1/cSR7 cells was twofold higher than that in SupT1/cSR9–3C cells, consistent with the approximately twofold-higher average vector copy number per cell in the former (data not shown). Only one band of the expected size was observed in the Northern blots, indicating the lack of induction of the LTR promoter in the absence of Tat expression. Additionally, we performed Western blot analysis to verify TdRev protein expression in these transduced populations of cells. TdRev expression was approximately the same for SupT1/cSR9–3C and SupT1/cSR7 cells (data not shown). Taking into account the differences in average vector copy number between these two cell populations, the expression of TdRev from the cSR9–3C vector appeared to be higher than that from the cSR7 vector. The difference in TdRev expression levels between these two vectors was also apparent in transiently transfected 293T cells, where at least fivefold-higher TdRev protein levels were observed for cSR9–3C (data not shown).

Challenge of transduced SupT1 cells with HIV-1.

To test the degree of inhibition of HIV-1 replication conferred by lentivirus vectors expressing TdRev, we compared these vectors to the MoMLV-based vector GC-Rev^TDSN (anti-TAR_DC), which expresses TdRev and a TAR antisense RNA and which was shown to inhibit HIV-1 replication in vitro (40). Control SupT1 cells and SupT1 cells transduced with vectors cSR7 and cSR9–3C were infected with HIV-1_NL4–3 at an MOI of 0.001. Figure 4 demonstrates that lentivirus vectors cSR7 and cSR9–3C conferred a high degree of protection against HIV-1 replication during the tested period. As previously described, vector GC-Rev^TDSN (anti-TAR_DC) decreased the RT activity detected in the supernatant with respect to that in nonprotected cells. These results indicate that the protection conferred by lentivirus vectors expressing TdRev was higher than that obtained with MoMLV-based vectors expressing the same anti-HIV-1 gene. These differences were not likely to be due to the higher MOIs used for transduction with the lentivirus vectors, as the final average vector copy number per cell was lower for cSR9–3C (0.74) than for GC-Rev^TDSN (anti-TAR_DC) (∼1).

FIG. 4.

Comparison of MoMLV- and HIV-1–based vectors expressing TdRev for inhibition of HIV-1 replication. SupT1 cells (10⁶) transduced with vector GC-SN_DC (♦), GC-Rev^TDSN (anti-TAR_DC) (▪), cSR7 (∗), or cSR9–3C (●) were infected with HIV-1_NL4–3 at an MOI of 0.001. Virus replication was measured by an RT assay (expressed in phosphorimager phosphostimulated luminescence [PSL] units).

To further test the protective effect that could be achieved with vectors cSR7 and cSR9–3C, we repeated the infections with HIV-1_NL4–3 at higher MOIs (Fig. 5). In these infections, we included vectors GC-SN_DC, cSN7, and HSN2 as alternative negative controls. Vector cSN7 is similar to cSR7, while HSN2 has wild-type RNA splice donor and acceptor signals flanking the RRE fragment and lacks the central polypurine tract/central termination sequence fragment.

FIG. 5.

Inhibition of HIV-1 replication by vectors pcSR7 and pcSR9–3C. SupT1 cells (10⁶) transduced with vector GC-SN_DC (♦), HSN2 (▪), cSN7 (∗), cSR9–3C (•), or cSR7 (▴) were infected with HIV-1_NL4–3 at an MOI of 0.002 (A and C) or 0.1 (B and D). Wild-type virus replication was measured by an RT assay (A and B) (expressed in phosphorimager phosphostimulated luminescence [PSL] units), and cell viability was determined by trypan blue exclusion (C and D).

Vector cSN7 exhibited some inhibition of HIV-1 replication at low MOIs (Fig. 5A). However, this protective effect was not seen at higher MOIs and was always lower than the protection provided by the analogous vector HSN2. This result suggests that the mutation of splicing signals in the vector backbone of cSN7 reduces its capacity to act as a DIP (see Discussion).

The population of SupT1 cells transduced with vector cSR9–3C showed a delay in the onset of infection, but replication of the wild-type virus was not prevented (likely caused by the fraction of nontransduced cells in this population). SupT1 cells transduced with vector cSR7 conferred protection against HIV-1_NL4–3 replication at MOIs of 0.002 and 0.1 (Fig. 5A and B). Even at the highest MOI of 0.1, no virus production was detectable by measurement of RT activity, and cells displayed high viability during the first 25 days postinfection (d.p.i.) (Fig. 5D). In this unselected cell population, HIV-1 replication was inevitable, and capsid protein became detectable in a p24 enzyme-linked immunosorbent assay (<140 ng/ml) by 30 d.p.i.; however, no syncytium formation or cytopathic effects were visible in these cultures by up to 45 d.p.i. (data not shown). The observation of p24 synthesis but no cytopathic effects suggests that little or no HIV-1 envelope gene expression is occurring (the envelope proteins gp120 and gp41 are the mediators of syncytium formation) or that the lack of inhibition of p24 synthesis may be taking place in a subset of cells of the population.

Infectivity of viral particles that escape TdRev-mediated inhibition.

To investigate whether HIV-1 particles released from transduced cells had reduced infectivity, naive SupT1 cells were infected with equal amounts of p24 from the supernatants used in the previous experiment. Viruses released from SupT1/GC-SN_DC, SupT1/HSN2, SupT1/cSR7, and SupT1/cSR9–3C challenges by 9, 11, 30, and 13 d.p.i., respectively, were applied to SupT1 cells. Rechallenges were performed using 100, 20, or 4 ng of p24 from each supernatant. The kinetics of virus replication were monitored for 19 days and evaluated by measurement of RT activity (Fig. 6).

FIG. 6.

Infectivity of virions released from SupT1/cSR7 and SupT1/cSR9–3C infected with HIV-1. SupT1 cells (10⁶) were infected with the equivalent of 100 ng (A) or 20 ng (B) of p24 from supernatants obtained from SupT1 cells transduced with vector GC-SN_DC (♦), HSN2 (▪), cSR7 (•), or cSR9–3C (▴) and infected with HIV-1_NL4–3. Assuming 5,000 infectious units per ng of p24, the estimated MOIs were ∼0.5 and 0.1 for 100 and 20 ng of p24, respectively. The supernatants used for infection were collected at the initial phase of the productive infection, to ensure the collection of fresh virion particles. Wild-type virus replication was measured by an RT assay (expressed in phosphorimager phosphostimulated luminescence [PSL] units).

In comparison to virus obtained from MoMLV vector-transduced cells, virus released from cells transduced with HIV-1–based vectors showed delayed progression of the infection. The greatest reduction in infectivity was obtained with virus that escaped from the G418-selected cell population transduced with vector HSN2. From a comparison of Fig. 6A and B, it can be seen that a fivefold reduction in the amount of infectious particles (100 ng to 20 ng of p24) was translated as a 2-day delay in the maximum peak of RT activity for the same supernatant. From these data, it can be estimated that the presence of vector cSR7 or cSR9–3C in infected cells produced an approximate eightfold reduction in the infectivity of the resulting virus. The amount of RT activity was lower in both cases (Fig. 6) for vector cSR9–3C.

Mobilization of proviral copies of TdRev-expressing vectors by HIV-1.

To determine if vectors cSR7 and cSR9–3C could be mobilized by wild-type HIV-1 to secondary target cells, genomic DNA was purified at different times from SupT1 cells infected with supernatants obtained from HIV-1_NL4–3-infected SupT1/cSR7 or SupT1/cSR9–3C cells. Genomic DNA was then analyzed by PCR for the presence of vector and wild-type viral sequences (Fig. 7). Results from this experiment demonstrated that vector mobilization was detected in a dose- and time-dependent manner. Although the PCR amplifications shown here are not quantitative, the intensity of the amplification products allows some semiquantitative observations to be drawn. The intensity of the PCR vector-specific bands at any time point is higher when higher doses of p24 are used to establish the secondary infection (i.e., 100 ng versus 20 or 4 ng). During the course of the secondary infection, the intensity of the vector-specific bands increases over time (Fig. 7, top panel, lanes 1, 4, and 7), correlating with a higher level of amplification of the virus-specific bands (Fig. 7, lower panel). This finding can be explained either by an enrichment of TdRev-expressing cells due to preferential killing of nontransduced cells or by an active process of vector mobilization driven by ongoing viral replication.

FIG. 7.

Mobilization of cSR7 and cSR9–3C vector sequences by HIV-1_NL4–3. SupT1 cells (10⁶) were infected with supernatants obtained from HIV-1_NL4–3-infected SupT1 cells transduced with vector cSR7, cSR9–3C, or GC-SN_DC using the indicated amounts of p24. After the indicated number of days postinfection, genomic DNA was purified and analyzed by PCR for the presence of sequences specific for TdRev or HIV-1. The negative control was genomic DNA from noninfected SupT1 cells (lane SupT1) or no DNA (lane No DNA). Positive controls for vector amplification were genomic DNA from SupT1 cells transduced with vector cSR7 (lane SupT1/cSR7) or cSR9–3C (lane SupT1/cSR9–3C).

Taken together, the results shown in Fig. 5, 6, and 7 suggest that inhibition of viral replication is driven by the expression of TdRev, as well as by vector encapsidation and mobilization to secondary cells. The last mechanism seems to reduce the infectivity of the viral particles that escape the inhibition mediated by TdRev.

DISCUSSION

In this report, we describe the development of new anti-HIV-1 lentivirus vectors expressing TdRev. Several reports have shown effective inhibition of HIV-1 replication mediated by expression of TdRev in T-cell lines, primary T cells, CD34⁺-derived macrophage or monocytes, or CD34⁺-derived T cells (3–5, 29, 31, 39, 40). Clinical trials using transduced or transfected primary T cells have also shown a protective advantage for cells expressing TdRev (33, 41). These approaches used MoMLV-based vectors to express TdRev, as they have the advantage of not being sensitive to TdRev, thus permitting high-titer vector preparations to be obtained. In spite of the observed protection against HIV-1 replication, high levels of TdRev expression are required to maintain the inhibition (29), and in most instances protection is reflected only by a transient delay in the onset of viral replication. The lack of complete protection against viral replication suggests that additional mechanisms of inhibition should be incorporated in the vector to enhance its protective potential.

Lentivirus vectors have been shown to inhibit HIV-1 replication by several mechanisms, including sequestration of Tat and Rev (8), competition for packaging into virions (6), and interference with reverse transcription (2, 6). Therefore, expression of anti-HIV-1 genes in HIV-1–based vectors would have higher inhibitory potential than that in MoMLV-based vectors. The disadvantage of expressing TdRev in an HIV-1–based vector system is that TdRev inhibits the expression of helper-encoded HIV-1 Gag-Pol, leading to a significant reduction in vector titers.

Several strategies have been tested to express HIV-1 Gag-Pol in a Rev-independent fashion (13, 16, 17, 35, 36, 37, 38, 43). Some studies have focused on the use of the CTE to increase the expression of Gag-Pol from helper plasmids lacking the RRE while using transfer vectors that are dependent on the RRE-Rev mRNA transport mechanism. In these cases, the synthesis of Gag-Pol was reduced 7-fold (13) to 50-fold (16) relative to the results obtained when using helpers that employ the RRE-Rev transport mechanism, resulting in vector titers of 10⁴ to 10³ CFU/ml for each of these studies. Other strategies involve the introduction of silent mutations into the coding region of Gag-Pol in order to eliminate the instability sequences that are responsible for the Rev-dependent character of the mRNAs coding for Gag-Pol (17, 30, 36, 44). In a recent report, a Rev-independent version of Gag-Pol in combination with a Rev-dependent or a Rev-independent transfer vector yielded titers of ∼5 × 10⁵ and ∼1 × 10⁵ tu/ml, respectively (17). A further solution to this problem could be to use a mutant RRE isolated from viral strains that are resistant to TdRev (14) or to express TdRev in a conditional and regulated fashion (in such a way that its expression is suppressed in the packaging cell line and becomes activated only in the target cell). Finally, as this inhibition has to be overcome only transiently during vector packaging, overexpression of Rev might partially alleviate the negative effects of TdRev. This approach has been recently described for helper vector pCMVΔR8.2 (26). The results obtained by Mukhtar et al. (26) are in accordance with the results presented here: a twofold increase in Gag-Pol synthesis by overexpression of Rev, with a maximum p24 level of 22 ng/ml. However, no estimation of vector titers or transduction efficiencies in T-cell lines was reported.

The data reported here demonstrate that the use of vectors with splice signal mutations or the replacement of an RRE by a CTE can yield functional TdRev-expressing lentivirus vectors with relatively high vector titers (3 × 10⁵ tu/ml). Packaging was additionally facilitated by use of an HIV-1 helper plasmid that expresses large amounts of Gag-Pol and by overexpression of Rev during packaging. Helper plasmid pCD/NLBH shows less sensitivity to TdRev than does helper plasmid pCMVΔR8.2. This difference may be attributable to the presence of the SV40 origin of replication cloned within the deleted env region of helper plasmid pCD/NLBH. The presence of the SV40 origin of replication may facilitate a higher plasmid copy number within transiently transfected 293T cells, which express the SV40 large T antigen.

After transduction into a human T-cell line, these vectors conferred more protection against HIV-1 replication than did previously described MoMLV vectors expressing TdRev. The increased titers of vector cSR7 were translated in a higher fraction of protected cells than were those of vector cSR9–3C. Consequently, the inhibition of HIV-1 replication caused by vector cSR7 was the strongest and most durable. Cells protected by this vector did not suffer a reduction in viability for over 45 days, even at the high MOI of 0.1. For all vectors, the presence of a lentivirus vector backbone in infected cells decreased the infectivity of the virus that escaped inhibition and correlated with vector mobilization to secondary target cells.

To be an effective DIP, in an environment where TdRev is being expressed, the vector RNA should have a competitive advantage over the wild-type viral genomic RNA for transport and packaging. For this purpose, the vectors used in this study contained structural modifications that made them less sensitive to TdRev and allowed them to express more unspliced cytoplasmic RNA in cells expressing TdRev (23). The data obtained with neo gene-containing vectors cSN7 and HSN2 apparently contradict this idea. In the challenge experiments shown in Fig. 6, vector HSN2 inhibits HIV-1 replication better than vector cSN7. As the percentages of transduced cells in these populations are the same (100%) and as the RRE and TAR decoy sequences present in the vectors are the same, the differences in the splicing signals are likely to be the cause of the observed results. We and other investigators have described increased cytoplasmic genomic mRNA levels in vectors having mutations in the splice donor sequence (17, 23). Although we did not see a decrease in vector titers, Kotsopoulo et al. (17) described equal or decreased splice donor-negative (SD⁻) vector titers that resulted in a reduced packaging ability of these vectors. This reduced packaging ability can be caused by direct effects of the SD mutation on the packaging signal or by an alteration of the nuclear export pathway that might result in a different subcytoplasmic localization, where packaging is not favored.

Taking all these observations into consideration, the results observed for vectors HSN2 and cSN7 may be explained as follows. In a cell where there is no expression of TdRev, both SD⁺ vectors (such as HSN2) and SD⁻ vectors (such as cSN7) will express high levels of cytoplasmic genomic RNA, but an SD⁺ vector will be a better DIP than an SD⁻ vector due to its putative enhanced competitive advantage for packaging. On the contrary, in a cell where TdRev is expressed, the cytoplasmic levels of genomic RNA will be higher for an SD⁻ vector (such as cSR7) than for an SD⁺ vector, overcoming the putative deficiency in packaging and making it a better DIP than a Rev-dependent SD⁺ vector. This model would also explain the relative difference in infectivity of the virus that escapes TdRev-mediated inhibition. Figure 6 shows that an SD⁺ vector (HSN2) produces a larger reduction in the infectivity of the escaped virus than an SD⁻ vector, such as cSR7 or cSR9–3C. This result can be interpreted as a superior competitive advantage for packaging of HSN2 over cSR7 or cSR9–3C.

The experiments designed to evaluate the overall inhibitory strengths displayed by vectors cSR9–3C and cSR7 demonstrated consistently less HIV-1 replication in cells transduced with cSR7. The lower level of protection observed with cSR9–3C than with cSR7 was not caused by lower levels of TdRev expression, as the former vector shows increased expression of TdRev protein, but may be correlated with lower vector titers. The higher level of expression of TdRev protein from cSR9–3C in transiently transfected 293T cells is likely the main cause of the reduced titers for this vector, which ultimately are reflected as a lower percentage of transduced and protected cells. Therefore, the reduced inhibition of HIV-1 replication displayed by cSR9–3C with respect to cSR7 might have been caused by the higher fraction of unprotected cells which subsequently were infected by HIV-1. Alternatively, the differences observed could have been caused by diminished transport of cSR9–3C unspliced genomic RNA to the appropriate packaging sites in the cytoplasm. This situation would result in a reduced competitive advantage for vector packaging and spread relative to the wild-type viral genomic RNA. A direct comparison of the inhibitory capacity of these two vectors would require working with populations of cells having the same average vector copy numbers per cell. Such work could be accomplished by use of selectable markers coexpressed with TdRev. However, previous experiments that included neomycin or puromycin selectable markers within similar lentivirus vectors led to very low titers (<10² CFU/ml; data not shown).

The potent inhibition of virus replication caused by cSR7 is likely to be the result of several independent mechanisms. The most important contribution seems to be the expression of TdRev, as the cotransfection of pcSR7 or pcSR9–3C with pNL4–3 produces a large decrease in the level of synthesis of p24 (Fig. 2). A second mechanism of inhibition may be the sequestration of Tat and Rev by the TAR and RRE sequences present in the vector backbone. This effect is consistent with the results seen in the transient cotransfection of pcSN7 or pcSN9–3C with pNL4–3. A third possible mechanism is the competition of vector genomic RNA for packaging and its interference with reverse transcription in heterodimeric virions. This mechanism is predicted to decrease the infectivity of the outcoming virions and also to cause mobilization of the vector genome to secondary cells, thus decreasing the rate of virus replication in unprotected populations of target cells. The two latter possibilities are supported by the data presented for vectors cSR7 and cSR9–3C in Fig. 6 and 7. The observation of vector mobilization by the wild-type virus does not imply that the vector completely failed to inhibit HIV-1 replication. This result indicates that the process of inhibition is a complex and dynamic event where the vector may permit low levels of Gag-Pol synthesis while competing for encapsidation into virions. Thus, the end result may be a reduction in the magnitude of wild-type virion formation and the mobilization of the vector, which potentially amplifies the inhibitory process.

When comparing the inhibition of HIV-1 replication caused by different HIV-1–based vectors expressing TdRev, it is important to note that the final inhibition is the result of several independent factors and mechanisms. The vector structural features will define the relative competitive advantage of the vector to act as a DIP in an environment where TdRev is being expressed. The internal promoter will influence the level of TdRev expression, which will determine the degree of inhibition of wild-type virus replication in the target cells. However, high levels of TdRev expression will have a detrimental effect during vector production and will also hamper subsequent vector mobilization by the wild-type virus. High levels of TdRev expression, with a concomitant reduction in helper-encoded Gag-Pol synthesis, will be translated as lower vector titers, which in turn will result in a smaller fraction of transduced cells. In the absence of selection after transduction, the percentage of transduced cells in the population is a critical factor that will determine the final outcome of a challenge experiment, regardless of the inhibition manifested by each individual cell. Virus replication in the nontransduced fraction of the cell population will result in passive recruitment of the protected cells into syncytia or in superinfection of the protected cells, resulting in cytopathic effects regardless of the lack of virus replication in the protected cells. Therefore, a delicate equilibrium must be reached, in which the expression of TdRev is sufficient to allow efficient packaging of the vector to achieve high vector titers and great enough to inhibit HIV-1 replication in target cells although still permitting some synthesis of Gag-Pol to allow vector spread to nontransduced cells.

Different avenues can be explored to improve the titers and the general performance of these vectors. The use of conditional promoters to suppress the expression of TdRev in the packaging cell line while allowing expression in the target cell line and combining TdRev expression with ribozymes to confer a more selective advantage to vector RNA are alternatives that deserve to be considered in future work.