HUMAN IMMUNODEFICIENCY VIRUS TYPE I AS A TARGET FOR GENE THERAPY

Magnús Gottfredsson and Paul R. Bohjanen

Division of Infectious Diseases and International Health, Department of Medicine, Duke University Medical Center, Durham, NC 27710

Received 11/17/97 Accepted 11/24/97

4. TARGETS FOR GENE THERAPY

Increases in our understanding of the HIV-1 life cycle has lead to the identification of numerous potential targets for gene therapy. Of these potential targets, the interaction between Tat and TAR and the interaction between Rev and RRE have been studied the most extensively, although other viral targets have also been evaluated. This section summarizes the research that has been done to characterize viral molecules as targets for gene therapy.

4.1. Tat and TAR

Because the Tat/TAR interaction is essential for activation of HIV-1 transcription, several gene therapy approaches that target either Tat or TAR have been evaluated. Analysis of the effect of mutant Tat polypeptides on HIV-1 transcriptional activation has lead to the identification of trans-dominant negative Tat mutants that suppress HIV-1 gene expression (87-90) Trans-dominant negative Tat mutants likely function by blocking an essential interaction between the wild-type Tat/TAR complex and cellular cofactor(s). They may do this by binding to wild-type Tat, TAR or essential cellular cofactors. A subset of functional trans-dominant negative Tat polypeptides (87, 88) contain mutations in an activation domain that is necessary for Tat function and appears to bind to a cellular cofactor (91, 92). Over-expression of a Tat polypeptide with a mutation in this region was shown to inhibit activation from viral latency and to inhibit viral replication in a human cell line (87). Other trans-dominant negative Tat polypeptides that suppress HIV-1 gene expression (89, 90, 93) contain mutations in the basic domain of Tat which is involved in TAR-binding (49, 94). A potential problem with overexpressing trans-dominant Tat mutants as an approach to gene therapy is that these mutants may bind to and sequester important cellular factors and thereby disrupt normal cellular functions.

Intracellular immunization is another approach to gene therapy that has been applied to inhibiting the function of the Tat protein. Expression of a single chain intracellular antibody (or intrabody) against Tat effectively inhibited HIV-1 replication in stably transfected cells (95). This intrabody appeared to function by binding to and sequestering Tat in the cytoplasm, thereby preventing it from regulating HIV-1 transcription in the nucleus (95).

Gene therapy strategies that target tat or TAR RNA have also been studied. For example, expression of anti-sense tat RNA (96, 97) or expression of ribozymes designed to cleave tat RNA (97, 98) have been shown to inhibit HIV-1 replication in T cells lines. In one study, it appeared that expression of anti-sense tat RNA was superior to an anti-tat ribozyme for inhibiting viral replication in a T cell line (97). Expression of anti-sense RNA targeted to the TAR sequence has also been shown to inhibit HIV-1 replication in cell lines (99).

RNA decoys containing the HIV-1 TAR structure have been studied extensively as inhibitors of HIV-1 transcription (56, 100-102). TAR decoys specifically inhibit Tat-activated HIV-1 transcription in a cell-free transcription system (100, 101), and appear to function by interacting with Tat-containing RNA polymerase II transcription complexes that assemble on the HIV-1 promoter (56). Overexpression of TAR decoys rendered a T cell line resistant to HIV-1 replication (103, 104). In addition to inhibiting HIV-1 replication, TAR decoys also inhibited HIV-1 RNA expression as assessed by northern blot (102). Polymerization of TAR decoys by expressing multiple tandem copies of the TAR sequence on a single RNA molecule may increase the efficiency with which TAR decoys inhibit HIV-1 transcription (105) and viral replication (106, 107). A potential problem regarding expression of RNA decoys within cells is a relatively short half-life due to the action of ribonucleases. The stability of TAR decoys in cellular extracts increases dramatically when TAR decoys are circularized, yet circular TAR decoys retain their ability to inhibit HIV-1 transcription (101). Circular TAR decoys, however, have not yet been tested within cells for their ability to inhibit viral replication.

In comparison to single gene therapy approaches, combination approaches that target the Tat/TAR interaction can be more efficient at inhibiting viral replication. For example, expression of a combination of anti-sense tat RNA and polymeric TAR decoys blocked HIV-1 gene expression and viral replication more efficiently than expressing either alone (108). A dual function anti-tat gene that produced a polymeric TAR RNA linked to an anti-sense tat RNA sequence was shown to efficiently inhibit HIV-1 replication in cell lines as well as in primary human peripheral blood mononuclear cells (109). Expression of this anti-tat gene was regulated by the HIV-1 long terminal repeat such that it was only expressed in HIV-1-infected cells (109). This gene was effective at inhibiting replication of laboratory strains of HIV-1 as well as patient isolates (109, 110). It also inhibited HIV-1 replication when it was transduced ex vivo into peripheral blood mononuclear cells isolated from patients with AIDS (111). Transduction of this gene into human CD34+ hematopoietic progenitor cells derived from cord blood resulted in the differentiation of CD4+ T cells that were resistant to HIV-1 infection (112). This type of genetic manipulation of stem cells such that they produce T cell progeny that are resistant to HIV-1 infection could be a powerful gene therapy approach to combating HIV-1 infection.

4.2. Rev and RRE

The binding of Rev to the RRE and the subsequent Rev-dependent transport of RRE-containing mRNA to the cytoplasm is essential for viral replication. Therefore numerous approaches to inhibit this binding have been studied. Trans-dominant Rev mutants (113-121), chimeric Rev molecules which inhibit nuclear export of RNA (122), antisense RRE sequences (123-125), RNA decoys (126-131), ribozymes which selectively cut rev (98, 132) and intracellular antibodies which bind to Rev (133-136) have all been evaluated.

Malim and colleagues developed RevM10, the first trans-dominant Rev mutant in 1989 (113). Subsequently a number of studies have demonstrated that expression of trans-dominant Rev mutants inhibited HIV-1 replication in T cell lines and primary T cells challenged with primary HIV isolates (114-118, 121). A recent report described transduction of RevM10 into CD34-enriched hematopoietic progenitor cells from umbilical cord blood or peripheral blood. These transduced cells gave rise to mature T-cells in SCID-hu mice that were resistant to HIV-1 replication (137). A phase I clinical trial in AIDS patients is underway to study the safety of administering lymphocytes expressing RevM10 to patients (119, 120).

Recently, a chimeric protein consisting of the wild-type Rev protein covalently linked to a mutated NS1 protein of the Influenza A virus (NS1RM-Rev), was found to inhibit nuclear export of HIV-1 mRNA (122). Because oligomerization of Rev seems to be important for Rev function the chimeric NS1RM-Rev protein may display improved oligomerization compared to trans-dominant Rev mutants such as RevM10. In comparative studies the NS1RM-Rev protein has been shown to be at least as effective as the trans-dominant mutant RevM10 (122).

Inhibition of HIV-1 by antisense oligodeoxynucleotide sequences to the RRE has been described as an effective method for antiviral therapy (123, 124). However, the virus typically breaks through after growth in culture. Nevertheless, at low multiplicity of infection, stable expression of antisense RRE has been shown to confer resistance to breakthrough growth of virus for up to 5 weeks (125).

Another approach to achieve Rev suppression is to overexpress the RRE target sequence. Lee and colleagues demonstrated that overexpression of a 45-nucleotide chimeric tRNA-RRE transcript, suppressed HIV-1 in more than 90% of expressing T cells (127). However, due to concerns that viral RNA sequences may bind and interfere with the function of essential cellular factors a "minimal" 13 nucleotide RNA decoy, corresponding to the authentic Rev binding domain has been constructed and demonstrated to confer resistance to HIV-1 replication in T cell lines (128) as well as primary T-cells (131). Transient protection has also been shown in T-cells transduced with adeno-associated virus (AAV) vectors expressing RRE decoys (129). Recently, CD34+ cells have been transduced with the minimal 13 nucleotide RRE decoy. The macrophage progeny of these progenitor cells were challenged with HIV and shown to be resistant to replication of macrophage-tropic strains of HIV-1 (126).

Intracellular immunization against Rev has also been studied. Expression of an anti-Rev single-chain antibody (SFv) inhibited HIV-1 replication in human cells (133-135). This intrabody localized to the cytoplasm where it appeared to exert its effects by sequestering wild-type Rev (134) and accelerate its degradation (136). Delivery of SFv to primary T-cells and alveolar macrophages by an AAV vector was shown to effectively inhibit infection with clinical and laboratory isolates of HIV-1, including a multidrug-resistant isolate. A combination approach using SFv and RRE decoys lead to synergistic inhibition (138) and a clinical trial has been proposed based on these findings.

A ribozyme which targets the rev/env coding region of HIV-1 has been tested in T cells. This construct protected the T cells from selected strains of HIV-1 (139). Furthermore, ribozymes targeting the rev gene of HIV has been successfully used for temporary HIV inhibition in T cell lines (98, 132).

4.3. Nef

The Nef protein, an early gene product which is important for maintenance of high viral loads and the development of AIDS (74), functions to downregulate CD4+ and MHC-I expression on infected cells through endocytosis, thus evading the protective effect of CTLs (75, 78, 79). Clinical observations on patients infected with HIV-1 strains that contain deletions in their nef gene show that these individuals have remained asymtomatic with very low viral loads for 10-14 years (140). This deletion therefore seems to render the virus far less pathogenic, suggesting that nef might be a good target for gene therapy. Larsson and co-workers used antisense oligonucleotides and ribozymes to target a conserved 14-nucleotide region in the nef gene. They demonstrated a dose-dependent repression in HIV-1 replication in peripheral mononuclear blood cells after administration of the antisense oligonucleotide to the cell culture medium (141). A hammerhead ribozyme targeted at the same site was also shown to effectively repress HIV growth in T-cell lines (141). The conserved nef site thus seems to be an attractive new target for HIV-1 gene therapy.

4.4. The pol gene

Ribozymes with different pol targets have been described. Yu and colleagues constructed a hairpin ribozyme targeted to cleave a conserved region in the pol gene. T cells transduced with a retroviral construct encoding this ribozyme gene exhibited significant inhibition of different strains of HIV. Greater protection was shown when the ribozyme expression was driven by a pol III promoter as compared to a pol II promoter (142).

4.4.1. Reverse transcriptase

The RT enzyme is effectively inhibited by pharmacologic agents such as nucleoside and non-nucleoside analogues. In addition intracellular immunization against HIV-1 has been accomplished by intracellular expression of antibody fragments directed against RT (143). Cells expressing the antibody fragments were resistant to infection with laboratory and clinical isolates of HIV-1 and HIV-2. The potential advantage of this approach is that it inhibits viral replication prior to integration into the host genome, thereby allowing possible resolution of the infected state (143).

Another method is being developed for inhibiting RT by using antisense oligonucleotides. Results indicate that antisense oligonucleotides to the U5 region of the viral RNA inhibits RT activity in vitro (144). Studies on the efficacy of this approach in cells have not yet been published.

4.4.2. Integrase

In order for the viral cDNA to be transcribed it must be integrated into the genome of the host cell, a process that requires viral integrase, an enzyme encoded within the pol gene. This integration is dissimilar to any reaction known to be involved in the normal functioning of the host cell, making this enzyme an attractive target for viral inhibition (145). Inhibition of integrase expression has been accomplished in Escherichia coli by using a specific ribozyme molecule which recognizes a GUC sequence in the integrase RNA (146). Intracellular expression of single-chain variable antibody fragments which bind to integrase prevented productive HIV-1 infection in both T-cell lines as well as peripheral blood mononuclear cells (147) presumably by specifically neutralizing the enzyme prior to integration (147).

4.4.3. Protease

The HIV protease, a product of the pol gene, is essential for the life cycle of the virus by cleaving the Gag-Pol polyprotein into structural proteins and viral enzymes. Recent success in pharmacologic inhibition of HIV-1 protease has sparked interest in this enzyme and its potential inhibition by using gene therapy. Trans-dominant mutant protease has been constructed and shown to prevent protease activation and virion maturation (148). Junker and co-workers showed that T-cell lines which constitutively expressed the mutant protein had dramatically reduced HIV replication when compared to cells expressing the wild-type protein and this approach was also shown to be effective against protease-resistant isolates (149). Another approach to inhibit the protease through competitive inhibition has been taken by Serio and colleagues who overexpressed a chimeric Vpr molecule which contained a protease cleavage site (150) (see 4.7, Vpr) .

4.5. The gag gene

The products of the gag gene exist in a highly multimerized form in the mature HIV virions. Therefore, they may represent attractive targets for inhibition by trans-dominant mutants (151). A number of mutant Gag proteins have been generated that can interfere with the production of infectious viral particles. High-level cellular expression of such Gag mutants confers partial resistance to HIV-1 replication when cells are challenged with a wild-type virus (151). Other methods of gag-suppression have been studied, such as antisense sequences complementary HIV-1 sequences in the 5' leader-gag region (152, 153). In T cell lines expressing the antisense sequence, HIV-1 replication was inhibitied for 10-20 days. Despite stable expression of the protective sequence, the antiviral effect was not lasting (153). However, by expressing a longer anti-gag sequence, a more effective and durable suppression in HIV replication was demonstrated in primary CD4+ lymphocytes (154).

Ribozymes targeted to HIV-1 gag transcripts have been developed and transfected into cell lines (155). Subsequent challenge with HIV-1 revealed cleavage of gag RNA and substantial reduction of viral production. Importantly, the ribozyme expression had no effect on cell replication and viablity, suggesting that it was not toxic to cells.

A novel antiviral strategy has been designed in which the retroviral gag genes are fused to nuclease genes, creating fusion proteins that are destructive to viral components within the assembling virion. This strategy has been referred to as capsid-targeted viral inactivation (156, 157). Schumann and co-workers studied RNase HI from Escherichia coli fused to Moloney murine leukemia virus (MMLV) coat protein. They demonstrated that the fusion protein is encapsidated into virions where it is cleaved by the MMLV protease, releasing a functional RNA-cleaving enzyme. This approach reduced virus output by 97-99% in MMLV infected cells (157), but has yet not been studied in HIV-1 infection.

4.6. The env gene

The possibility of generating novel gene delivery vectors with tropism for HIV-infected cells may open novel strategies for treatment of persistent and latent infections. Two recent reports of selective targeting of HIV-infected cells expressing viral envelope proteins warrant description (158, 159). Mebatsion et al. generated a rabies virus with altered receptor specificity. By pseudotyping the virus with CD4 and CXCR4-derived proteins they were able to demonstrate that it selectively infected cells expressing HIV-1 envelope proteins (gp120 and gp41). Antibodies to CD4 or the envelope protein blocked entry of this modified rabies virus (158). Similarly, Schnell and co-workers constructed a recombinant vesicular stomatitis virus expressing CD4 and CXCR4. This virus was unable to infect normal cells but infected propagated on and killed HIV-infected cells that expressed the HIV-1 membrane fusion protein (gp120/gp41) on their cell surface. Moreover, this killing of HIV-1 infected cells controlled the infection in a T-cell line and reduced titers of infectious HIV-1 in the culture by 300-fold to 10,000-fold (159). Conversely, recent reports describe selective targeting of CD4-expressing cells by pseudotyping murine leukemia virus such that it expresses the HIV-1 envelope proteins (160).

Intracellular immunization to HIV-1 has been studied by using a intracellular single chain antibody (intrabody) to gp120 (sFv105) (161). This intrabody reacts with the nascent folded envelope protein (gp160) within the endoplasmic reticulum and prevents cleavage to gp120 thereby inhibiting subsequent transit of the envelope protein to the cell surface. The infectivity of the cells expressing this intrabody was substantially reduced (161), but viral replication was only delayed temporarily (5 days).

Multitarget ribozymes which target several different sites in the gp120 sequence have recently been developed. Some of these constructs target up to nine different conserved regions in the RNA (nonaribozymes) (162). Results from cotransfection experiments in cell lines indicate that antiviral activity can be substantially increased by using several hammerheads with catalytic activity on the same molecule (see 5, combinations of antiviral strategies).

4.7. Vpr

The Vpr accessory protein can be seen as an attractive target for gene therapy because the nuclear localization signal of Vpr mediates transfer of the preintegration complex to the cell nucleus (19), enabling HIV to infect nondividing cells (20-22). The Vpr protein also arrests the infected cells in G2 of the cell cycle, thereby increasing viral production of each infected cell (22). The Vpr protein is packaged into the HIV-1 virions through specific interactions with the Gag polyprotein precursor. The specificity of Vpr for the viral particle can be exploited to deliver antiviral elements and thereby interfere with virus maturation. It has been shown that foreign proteins can be targeted to HIV-1 particles via fusion with Vpr. Wu and colleagues fused a staphylococcal nuclease gene to Vpr, generating a fusion protein which, when coexpressed with its cognate Gag (p55) protein was incorporated into the viral particles where it was enzymatically active (163). Another recent report describes the use of chimeric Vpr proteins containing HIV-specific protease cleavage sites added to the C terminus of Vpr. Interestingly, the chimeric Vpr containing the cleavage sites completely abolished virus infectivity (150). The authors suggested that the mechanism of inhibition was by overwhelming protease activity by a competitive mechanism (150).

4.8. The 5' leader sequence

Hairpin ribozymes which target the HIV-1 5' leader sequence (U5) (164, 165) have been shown to inhibit HIV-1 in T cells lines (166-168). Primary lymphocytes expressing this ribozyme have been shown to be protected from HIV replication (169). Theoretically, all HIV-1 RNAs should be susceptible to the action of this ribozyme since it cleaves the cap structure away from the RNA and capless RNAs cannot be translated into proteins. This ribozyme was recently transduced with a retroviral vector into CD34+ cells from fetal cord blood. Subsequently, macrophage-like progeny derived from these CD34+ cells were shown to resist infection by a macrophage-tropic isolate of HIV-1 (170).

4.9. Packaging sequence y

For retroviruses in general, including HIV-1, the efficient packaging of genomic RNA into viral particles requires a specific sequence, termed the y (psi) packaging sequence (84, 85). In HIV-1, this region has been shown to be highly conserved, suggesting its feasibility as a target for gene therapy. Deletion of 19 base pairs encompassing y between the 5' LTR and the gag gene initiation codon results in a virus that is markedly attenuated for replication in human T-cells (85). Sun and colleagues showed that y antisense constructs suppressed viral production. They further constructed a ribozyme which cleaved y, rendering T-cell clones less susceptible to viral replication when challenged with HIV-1 (171).

4.10. HIV-regulated suicide genes

Another novel approach to achieve selective killing of HIV-infected cells involves the introduction of suicide genes to the genome. The rationale for the use of suicide genes is to eliminate infected cells selectively and thus inhibit the propagation of virus. Thus, to be useful for gene therapy the expression of the suicide genes needs to be limited to infected cells. The herpes simplex shutoff gene (vhs) encodes a protein which nonspecifically accelerates the degradation of mRNA molecules to facilitate shutoff of host protein synthesis during lytic infections. Hamouda et al. recently tested the ability of vhs expressed from the viral LTR promoter to inhibit HIV-1 replication (172). They demonstrated that the protein inhibited HIV-1 replication more than 44,000-fold when compared to a nonfunctional mutant protein.

Another potential suicide gene is the herpes simplex type 1 thymidine kinase gene (HSV-TK) under the transcriptional control of HIV-1 LTR (173) or modified HIV-2 promoters (174). Acyclovir was added to the medium and upon infection with HIV-1 the HSV-TK gene was transcribed and thus phosphorylated acyclovir to its toxic metabolite which lead to cell death. This approach has been shown to halt the spread of HIV-1 in cell cultures upon addition of acyclovir (173) or ganciclovir (174) in concentrations routinely achieved in the clinical setting. Similar studies have been conducted by expressing this gene from replication-defective adenoviral vectors (175). Other genes have also been studied for selective killing of HIV-infected cells, such as the Diphteria Toxin A under the transcriptional regulation of the Tat and Rev proteins (176, 177).

A hybrid molecule of the human CD4 and a modified version of the Pseudomonas exotoxin A (CD4-PE40) has been designed to kill HIV-infected cells (178). The molecule is designed to bind to infected cells by CD4-gp120 interaction at the cell surface. After uptake by the cell, the exotoxin inhibits protein synthesis, leading to cell death. CD4-PE40 has been studied in cell lines infected with laboratory strains of HIV-1 and has been shown to be efficatious. The CD4-PE40 is less active in the presence of primary isolates, however, since continuous presence of the molecule is required to inhibit HIV-1 replication (178).