Despite thousands of researchers worldwide working on a cure for HIV infection, none of the modalities have been completely successful. Currently, four classes of anti-retroviral drugs are available: nucleoside/nucleotide analogs, non-nucleoside reverse transcriptase inhibitors, protease inhibitors, and fusion (or entry) inhibitors. These drugs, used in various combinations to treat HIV, form what is known as highly active antiretroviral therapy (HAART). However, HAART is expensive, has high toxicity rates, and must be administered lifelong, i.e. it is not curative. In addition to the above problems, the rate of emergence of resistant strains is high post-HAART. In studies conducted in the United States and Europe, over 50% of patients experienced virologic failure (viremia) while on antiretroviral therapy, and approximately 80% of these patients showed drug resistant HIV genotypes 910. One long-term study found that by six years, approximately 80% of patients had their medications switched repeatedly due to drug resistance, resulting in an overall cumulative failure rate of 38% 11, placing these patients in danger of exhausting their treatment options 12. Transmission of drug resistant HIV mutants is also an increasing problem. In a study among newly infected individuals, 14% of patients were infected with HIV that already had one or more key drug resistance mutations 13. For these reasons, there is an increasing urgency to find a cure for HIV infection.

With the advent of the molecular and genetic age of medicine, research to create gene therapy for HIV has been on the rise. Since the 1980's, researchers have explored the possibility of using gene therapy to cure HIV-positive patients. In 1988, David Baltimore used the term 'intracellular immunization' to describe this treatment approach 14. Initial in vitro experiments were successful and now scientists are applying some of these methods in clinical trials.

Figure 1 is a schematic representation of the life cycle of HIV showing the various stages at which genetic therapy could be applied. Therapy could also be aimed at any one of the many target cells for HIV infection in vivo, including immune cells such as CD4⁺ and CD8⁺ T cells, dendritic cells, monocytes, macrophages, hematopoietic stem cells (HSCs), brain cells, and other cells from the gastrointestinal tracts that could serve as host cells for HIV. Since T cells are the major cell population implicated in HIV infection and its progression to AIDS, making these cells immune to infection is a very important aspect of therapy. Even more desirable are the HSCs. These self-replicating progenitor cells give rise to all other members of the lymphoid and myeloid lineages and have the capability of repopulating the immune system with a potentially HIV-resistant phenotype.

A variety of viral or cellular components could serve as targets for anti-HIV gene therapy. Targeting viral factors is currently the most prevalent method. A major problem with this strategy is that HIV can quickly form resistant strains to these genetic modifications due to high mutation rates. Targeting cellular factors makes the occurrence of resistant strains less likely but raises the issues of adverse effects on normal cell function. HIV receptors and co-receptors are attractive targets, but there have been recent reports of liver toxicity with CCR5 antagonists in HIV-infected patients 15. A summary of the main anti-viral approaches is as follows:

The early success of in vitro studies paved the way for clinical trials. All the above-mentioned strategies for gene therapy have shown good anti-HIV activity in vitro. However, not all of them have been tested in clinical trials. As reported in the literature, the trials conducted to date have used a suicide gene (HyTK), transdominant negative proteins (Rev M10, TdRev, or huM10), a chimeric receptor (CDζ), an RNA decoy (RRE), and ribozymes (anti-tat/vpr ribozyme and tat ribozyme, RRz2). A trial using siRNA was started recently and the results have not yet been published in literature. Following is an account of these clinical trials and their results (see Table 1 for summary).

In 1996, Riddell et al. published one of the earliest trials, involving the use of a suicide gene 16. This study enrolled six HIV-seropositive patients. Autologous CD8⁺ T cells were genetically modified using a retrovirus-mediated gene transfer technique. The retrovirus, designated HyTK, comprised the hygromycin phosphotransferase gene (Hy) and the herpes virus thymidine kinase gene (TK) as a fusion gene under the control of the murine leukemia virus (MLV) long terminal repeat (LTR). Hygromycin was used to positively select for transduced autologous cells in vitro prior to infusion into the patient. Ganciclovir could have been used, if necessary, to negatively select (kill) the transduced cells in the patient. In four increasing doses, researchers transfused autologous HyTK-transduced CD8⁺ cells at 14-day intervals. There were no significant side effects. However, five of six patients developed a CTL response to the foreign protein and thus rejected the modified CD8⁺ cells, which cleared in response to each subsequent transfusion. The results of this trial suggested other strategies that would make modified cells less susceptible to the immune response and thus inspired further research 16.

In the same year, Woffendin et al. published the results of a pilot trial involving a transdominant negative protein approach 36. For genetic modification of CD4⁺ T cells, they used Rev M10 and a deletion mutant of Rev M10, which showed no antiviral activity. Woffendin and colleagues transduced the T cells using a non-viral vector and showed that following transfusion, there was a preferential survival of Rev M10 modified CD4⁺ T-cells, as compared to cells that received the mutant. They also detected Rev M10 until 2 months post-infusion. Though there was increased survival of Rev M10-expressing cells, the overall numbers of transduced cells were low in vivo 36.

Trying to improve upon this trial, they conducted another pilot study in which they transduced CD4⁺ cells with Rev M10, but this time they used retroviral vectors instead of a nonviral vector 37. They detected Rev M10 for an average of 6 months post-infusion. In addition, cells transduced with Rev M10 survived longer than those transduced with the negative control vector. There were no detectable immune responses to the 'foreign proteins' (Rev M10 or MLV gp70 envelope protein). Though these studies with Rev M10 showed an improved efficacy of gene delivery with a retroviral vector, there was no effect on the patients' viral loads 37.

In 2002, Kang et al. published the results of another trial that used a transdominant negative mutant Rev protein (TdRev) 38. This study had only two subjects. Both were HIV positive and had malignancies. One had leukemia and the other had refractory Hodgkin's lymphoma. Fourteen days prior to receiving gene therapy, the researchers stopped their cyclosporine and HAART medications. Six days prior to gene therapy, the patients received fludarabine-cyclophosphamide containing regimen (non-myeloablative) for five days and then cyclosporine and HAART was restarted. Each patient had an HIV negative sibling who was an HLA-compatible donor for bone marrow cells. The donor underwent blood apheresis followed by isolation of CD34⁺ cells. The cells were genetically modified by using either TdRev or a control vector encoding human GP91phox. Following transplantation with these modified CD34⁺ syngeneic cells, the patients developed CMV antigenemia, which was treated. By day 96 post-transplantation, both patients showed 100% transfer of either gene into lymphoid and myeloid lineages. Both patients developed acute graft versus host reaction beyond day 100, which was successfully treated. The patient with Hodgkin's disease died 12 months after transplantation due to relapse of the disease unrelated to any complication of the gene therapy 38. In a follow-up paper published by the same group, the second patient showed persistence of TdRev at three years post-treatment and continued to be in remission. However, this might be due to the effect of HAART and not TdRev alone 39.

A 2005 study involving the transdominant negative protein approach was focused on the pediatric age group of HIV-positive patients 40. Two children were enrolled in this study, a nine-year-old and an eight-year-old who were both on HAART. CD34⁺ bone marrow cells from the participants were transduced with two retroviral vectors, one encoding a "humanized" dominant negative REV protein (huM10) and one encoding an internal control for gene marking (FX) that is not translated. A humanized protein is one in which the codon usage has been optimized for mammalian expression. Following infusion of the modified cells, huM10 and FX could be detected in peripheral blood mononuclear cells (PBMC) for 1–3 months. During a two-year follow-up period, levels of huM10 and FX expression dropped to at or below the limit of detection. In one patient, during a period of non-compliance to HAART regimen, PBMCs containing huM10 reappeared suggesting a selective increase in survival for PBMCs containing huM10 during periods of high viral loads 40.

Using a chimeric receptor approach, Walker et al. (2000) investigated the effect of CD4ζ-modified syngeneic T-cells in HIV-positive patients 41. The study was conducted using sets of twins, one of whom was HIV-positive and the other HIV-negative. The HIV-negative twin acted as the donor for syngeneic T-cells (either CD8⁺ or CD4⁺), while the HIV-positive twin was the recipient. T cells from the donor were genetically modified ex vivo to express CD4ζ and then transfused into the recipient twin. The study was performed in two phases.

In the first phase, 27 patients were enrolled and received one to six cell infusions every eight weeks. Three patients received 10⁷ modified cells, while the remaining 24 patients were randomly assigned to four groups, receiving either 10⁸, 10⁹, or 10¹⁰ modified cells or 10¹⁰ unmodified cells. The researchers observed that one of three subjects receiving 10⁷ cells and four of six subjects receiving 10⁸ cells showed low levels of CD4ζ expression. In three of these five CD4ζ-positive patients, CD4ζ was no longer detected after one to three days, but in two of them, CD4ζ could be detected for up to 2 and 24 weeks, respectively. In 11 of 12 patients receiving higher doses and multiple infusions of modified CD8⁺ cells, CD4ζ could be detected for up to 15 to 40 weeks post-infusion 41.

In the second phase of the Walker et al. trial, 33 patients were enrolled, 25 from the previous phase and 8 new participants 41. Of these, three did not receive any cell infusions. Patients received modified CD8⁺ cell infusions either alone or with IL-2. The group receiving IL-2 with the modified CD8⁺ cells showed higher levels of persistence of CD4ζ as compared to the group receiving no IL-2 supplementation. In order to test whether the IL2 was substituting for HIV-specific CD4⁺ T-cell help, 17 of the 30 participants received modified CD4⁺ cells along with modified CD8⁺ cells in a second series of infusions. Modified CD8⁺ and CD4⁺ cells were detected in the peripheral blood of these 17 patients for at least one year post-infusion, indicating that co-administration of the CD4⁺ cells may have increased survival of the modified CD8⁺ cells 41.

In the same year that Walker et al. reported their findings, Mitsuyasu and fellow researchers published the results of a similar chimeric receptor study 42. Patients enrolled in this study were receiving anti-retroviral therapy (ART) and had viral loads between 1,000 and 100,000 copies/ml and CD4⁺ T cell counts greater than 50 per microliter. Following cell infusions, patients were followed for eight weeks. Eleven patients received CD4ζ-modified T cells along with IL-2, and thirteen patients received CD4ζ-modified T cells alone. In contrast to the previous study by Walker et al., administration of IL-2 did not increase the survival of the modified T cells. However, an increase in cell number was observed at eight weeks post-infusion; an average increase of 73 CD4⁺ cells per microliter was observed in the group receiving IL-2 as compared to the group that did not receive IL-2. They detected CD4ζ in 1 to 3% of the peripheral blood mononuclear cells (PBMCs) at eight weeks and 0.1% at one year post-infusion. CD4ζ-modified T-cells were also isolated from bulk rectal tissue and/or lamina propria lymphocytes in five of five patients at 14 days and two of three patients at one year, showing good tissue distribution. In addition, there was a transient decrease of greater than 0.5 log mean in rectal tissue-associated HIV RNA for at least 14 days 42.

Encouraged by data from this trial, the same group conducted a phase II randomized trial of CD4ζ gene-modified versus unmodified T cells in adult male HIV-positive patients in 2002 43. All participants were on combination ART. In 37 patients, there were no measurable viral loads (<50 copies per ml). In three patients, low levels were detected (53, 57 and 65 copies per ml). Only 40 of 42 enrolled patients proceeded to receive the study treatment. The researchers found good expression of CD4ζ for at least 24 weeks in all patients. However, no significant difference was found between patients receiving gene therapy versus the control group when the six following parameters were analyzed: plasma viral load, HIV co-culture on PBMCs, blood HIV DNA, rectal biopsy DNA, rectal biopsy RNA, and blood CD4⁺ cell count 43.

A clinical trial involving RNA decoys was conducted in a pediatric population using an RRE decoy to modify CD34⁺ bone marrow cells. Kohn et al. (1999) showed that retroviral mediated CD34⁺ cell transduction had no significant adverse effects and that leukocytes containing an RRE decoy could be isolated from peripheral blood, even one year post-infusion; however, the numbers of transduced cells were extremely low 44. Four HIV-positive patients, three teenagers, and one eight-year-old were enrolled in this pilot study. Bone marrow cells positive for CD34 were isolated from these patients and transduced with Moloney murine leukemia (MoMuLV) vector virus carrying the RRE decoy sequences. No change in the HIV viral load was observed among the participants 44.

In 2004, Amado et al. demonstrated long-term maintenance of a therapeutic transgene in a phase I clinical trial involving ribozymes 45. They used MoMLV vector virus transduction of CD34⁺ HSCs for introduction of a ribozyme targeted to highly conserved regions in the HIV-1 tat and vpr genes. Ten patients participated in the study and researchers could detect the vector in naïve T cells for up to three years, the last time-point evaluated. There was an average increase of 10 CD4⁺ T cells per microliter from the beginning of the trial until the third year. In six patients, viral loads decreased by an average of 2.25 logs. Three patients had undetectable viral loads. One patient showed an increase of one log. The researchers found no correlation between the changes in viremia or CD4⁺ T cell counts with vector expression or detection in any cell type. However, during this trial, all patients were on ART, and the researchers attribute the changes in viral load as well as the cell numbers to individual viral susceptibility to ART 45.

More recently in 2005, Macpherson et al. published the results of a phase I clinical trial on ribozymes involving identical twins with discordant HIV status 46. Again, one twin acted as the donor (HIV-negative) and the other was the recipient (HIV-positive) of genetically engineered CD4⁺ T cells expressing a ribozyme. Specifically, the cells were transduced with an anti-HIV-1 tat ribozyme (RRz2) and a control LNL6 retroviral vector (for cell marking). Patients were followed initially for 24 weeks and then at regular intervals over a 4-year period. PBMCs containing both RRz2 and LNL6 were detected consistently. However, the effect of this therapy on HIV-1 viral load or the CD4 count was not specifically addressed.

Lastly, Morgan et al. (2005) published data from a clinical trial involving anti-sense RNA in conjunction with a trans-dominant negative protein 47. This study employed 10 pairs of twins. Like the earlier study involving twins with discordant HIV status, one twin served as the donor (HIV-negative) and the second twin as the recipient (HIV-positive). Lymphocytes from the donors were transduced to express a control gene (neo gene) or anti-HIV gene(s); a transdominant mutant Rev protein (TdRev) was used alone or with an anti-sense element directed against the HIV-1 TAR sequence on the same construct. Polymerase chain reaction demonstrated increased survival of modified lymphocytes in the initial weeks post-infusion in 9 of 10 recipients. In six of six recipients followed for approximately two years, T cells containing anti-HIV genes could be consistently detected and there was preferential survival of modified cells in one patient during a period of high HIV load 47.