Introduction

Highly active anti-retroviral therapy (HAART) was a major breakthrough in the treatment of human immunodeficiency virus (HIV) infection as it can effectively reduce viral load and support regeneration of cellular immunity, thereby considerably prolonging survival of HIV-infected patients. However, despite the effective suppression of virus replication, HIV persists, integrated into the host genome, and rebounds as soon as treatment is interrupted or drug-resistant virus emerges. Even with the most effective antiviral drug combinations, it has not been possible to "cure" HIV infection, and life-long antiviral therapy is required to prevent progression of immunodeficiency. This vital long-term treatment is limited by drug toxicity and viral resistance, and the number of patients for whom HAART fails is increasing.^1,2 Moreover, even prolonged periods of successful HAART have failed to restore HIV-specific immune responses.^3,4,5 Thus, novel therapeutic approaches are still urgently required.

Several therapeutic strategies involving the transfer of antiviral genes have been developed for HIV-1 infection. In clinical trials, T cells and hematopoietic stem cells have been targeted (as reviewed in refs. 6, 7). Both approaches have led to long-term engraftment of gene-modified cells. However, levels of gene modification in vivo have been low and no overall therapeutic benefit has been reported. Recently, Levine and colleagues observed a transient decline of viral load in two, and more sustained virus suppression in one, of five patients treated. In this trial, autologous T cells transduced with a conditionally replicating lentiviral vector that encodes an antisense RNA against the HIV envelope were infused.⁸ The mechanism of the antiviral effect is, however, not yet clear, as gene-marking levels were below 1%.

For the trial presented here, T cells were also chosen for genetic modification, primarily on the grounds of safety. Retroviral gene transfer into hematopoietic stem cells can cause insertional activation of oncogenes and leukemia under certain circumstances. Thus, stem cell gene therapy is currently restricted to life-threatening conditions that lack therapeutic options (reviewed in ref. 9). In T cells, leukemogenesis as a result of retroviral vector insertional mutagenesis, although theoretically possible, has never been reported. Safety considerations also precluded the transfer of cells modified with a control vector in addition to the cell population engineered to express the antiviral gene. Such control-engineered cells have been used in previous clinical trials to study the potential relative survival advantage provided by the therapeutic gene; however, they carry an additional risk of insertional leukemogenesis.

The antiviral gene used in this study is expressed from a retroviral vector (M87o) and encodes the membrane-anchored antiviral peptide C46 (maC46).^10,11 C46 comprises 46 amino acids, is derived from the second heptad repeat of the HIV-1 envelope glycoprotein gp41, and effectively inhibits fusion of the viral and cellular membranes during virus entry.¹² The 36 C-terminal amino acids of C46 correspond to the fusion-inhibitory peptide C36 (T-20/enfurvitide), the first HIV fusion inhibitor approved for clinical use. maC46 is expressed as a fusion protein with an N-terminal signal peptide, which targets the peptide through the endoplasmic reticulum to the cell surface and a C-terminal linker followed by a membrane-spanning domain (Figure 1). Expression of maC46 was shown to result in effective inhibition of a broad range of HIV isolates, including viruses resistant to C36/T-20.^10,13 In HIV-1-infected cell cultures, T cells that express maC46 have a strong selective advantage, rapidly accumulate, and gain prevalence. The accumulated cells do not carry an HIV provirus.¹⁰ This is a clear advantage of early over post-integration inhibitors, as the latter are expected to support the outgrowth of gene-modified cells that harbor a suppressed HIV-1 provirus (discussed in detail in refs. 6, 14). In clinical trials, only post-integration inhibitors have been used so far. Thus, our approach for the first time targets HIV-1 entry in a clinical gene therapy trial with the potential for non-infected, gene-protected T cells to accumulate over time in the patient.

Figure 1.

Map of the therapeutic gamma-retroviral vector M87o expressing the membrane-anchored antiviral peptide C46 (maC46). LTR, long terminal repeat; maC46 orf, open reading frame encoding the membrane-anchored C46 peptide; MSD, membrane-spanning domain; RRE, rev-responsive element; SP, signal peptide; wPRE, woodchuck hepatitis virus post-transcriptional regulatory element.

Full figure and legend (3K)

The gene encoding maC46 was introduced by retroviral gene transfer (M87o) into autologous T cells of ten HIV-1-infected patients with severe immunodeficiency and HAART failure. No major toxicity was observed. Although T-helper-cell counts rose significantly, viral load was not affected. Gene-modified cells could be detected throughout the 1-year follow-up, but the levels were too low to account for the marked rise seen in CD4 counts.

Results

Study design

Fifteen patients were screened and thirteen were enrolled into the study. The patients' baseline characteristics and HAART regimen are summarized in Table 1. All patients were on HAART and had CD4 cell counts below 200 cells/l or a CD4 percentage of less than 15% of CD3 cells, a viral load above 5,000 RNA copies/ml, and resistance to, or incompatibility with, at least one drug within each class of antivirals. Two patients were excluded from the study before T-cell infusion because T-cell processing failed. One patient quit of his own volition before infusion. The remaining ten patients received T-cell infusions of between 3 10⁹ and 17 10⁹ CD8-depleted T lymphocytes (see Supplementary Table S3) modified with the retroviral vector M87o expressing maC46 (Figure 1). This corresponded to a mean of 1 10⁹ gene-modified cells. All patients completed the 1-year follow-up. HAART was continued throughout the duration of the study, but the regimen was changed (as allowed by the study protocol) in seven patients between 4 and 9 months after infusion. The manufacturing data are summarized in Supplementary Table S3.

Table 1 - Baseline patient characteristics.

Full table

Toxicity

No serious adverse events related to the cell infusions were observed. At the time of writing, all patients were alive and well at between 24 and 31 months after infusion. Mild adverse events (Grades 1 and 2), which were classified as probably or possibly related to the cell infusions, were seen in six of ten patients. The most frequent were flu-like symptoms, mild fever, fatigue, and headache during and directly after infusions, which subsided within 24 hours and were most likely attributable to the infusion of limited amounts of dimethyl sulfoxide with the cell product.

Long-term survival of gene-modified cells in vivo

The kinetics of the gene-modified cells in vivo were observed using quantitative polymerase chain reaction. Analysis by flow cytometry was not possible owing to the low levels of gene modification in the patient. Within the first hour after infusion, gene-modified cells peaked at 0.8% of total peripheral blood leukocytes and then dropped to levels between 0.005 and 0.56% at 60 minutes after infusion, followed by a further sharp decline to levels between <0.0001 (non-detectable) and 0.3% 1 week after infusion (Figure 2a). Marking levels varied within this range, with a slight increase in the average level between weeks 6 and 12, followed again by a slow decline. Patients 4, 7, 11, and 15 received a relatively high number of total T cells (mean of 1.3 10¹⁰ cells), and M87o-positive cells could be detected throughout the first 9–12 months (Figure 2b). Patients 1, 14, and 16 received a lower T-cell number (mean of 0.6 10¹⁰ cells), and gene marking was detected only in the first 2–3 months (Figure 2c). For Patients 6, 12, and 17, who received the lowest T-cell doses (mean of 0.36 10¹⁰ cells), only one or two sporadic samples were positive for each patient during the 1-year follow-up. Thus, a higher infused cell dose (Supplementary Table S3) was associated with a higher level of gene marking. This positive correlation reached significance for all time points analyzed in the first 2 months and for several time points later during follow-up (P < 0.05, R = 0.68–0.86). A similar level of correlation and significance was reached when percentage gene marking was correlated with the total dose of gene-modified cells. Patient 16 would then fall into the high-dose group plotted in Figure 2b.

Figure 2.

Gene-marked cells redistributed rapidly within the first hour and the level of long-term gene marking depended on the cell dose. The M87o vector copy number per 100 cells (genomes) of total peripheral blood leukocytes was determined by quantitative polymerase chain reaction for several time points (a) during the first hour after infusion of gene-marked cells and regularly during the first year of follow-up in (b) the high-dose group (mean cell dose 1.3 10¹⁰) and (c) the intermediate-dose group (mean cell dose 0.6 10¹⁰). The three patients in the lowest-dose group (mean cell dose 0.36 10¹⁰, Patients 6, 12, 17) showed only sporadic marking and are not included.

Full figure and legend (23K)

As most patients received highly purified CD4 cell preparations, the level of gene modification within the CD4 cell compartment can be calculated. The data shown in Figure 2 for total leukocytes correspond to levels of gene marking in the CD4 compartment of up to 12% immediately after infusion and of up to 0.4% during follow-up (data not shown).

A total of seven lymph node samples and one bone marrow sample were analyzed for gene-modified cells by quantitative polymerase chain reaction (Supplementary Table S4). M87o-marked cells were detected in three of seven lymph nodes (3–6 months after transfusion) and in the bone marrow sample (14 months after transfusion). Levels of gene marking were usually similar to the values for corresponding blood samples from the same patient at the same time point. Thus, there was no indication of a specific homing pattern of transfused T cells.

Virologic and immunologic effects

During the first 4 months of follow-up, there was no significant change in mean plasma viral load at any time point (Figure 3a). Thereafter, antiviral therapy was changed in seven of ten patients (Table 1). In the three patients who remained on their pre-infusion HAART regimen (Patients 1, 12, and 15), no change in mean plasma viral load was seen throughout the 1-year follow-up. Of the seven patients who changed antiviral therapy, four showed a significant (>1 log) decline in plasma viral load (Patients 4, 6, 11, and 17; Figure 3b). Interestingly, within the group with good levels of gene marking after 3 months (Patients 4, 7, 11, 15), only the two patients who remained viremic throughout the 12 months' follow-up still had significant levels of gene-modified cells after 9–12 months. This could indicate an in vivo selective pressure on the gene-modified cells.

Figure 3.

Plasma viral load. (a) Plasma viral load was stable during the first 4 months after cell infusion. (b) Thereafter, four of the seven patients who changed antiviral therapy showed a significant (>1 log) decline in plasma viral load (Patients 4, 6, 11, and 17).

Full figure and legend (22K)

Figure 4 shows the change in absolute CD4 cell counts from baseline levels, defined as the CD4 count on the day of T-cell infusion. Not unexpectedly for a cohort of patients with high viral load, on average T-helper-cell counts declined during the period between the first documented visit before treatment (screening) and infusion (baseline). After infusion, there was an increase in CD4 counts relative to baseline at all time points during follow-up, which was significant (P < 0.01) up to day 45 (absolute: 41 per mm³, relative: 37%) and again became significant after month 5. Patients with higher baseline CD4 counts responded better, with a stronger absolute and relative increase in T-helper-cell numbers after treatment (R = 0.8 at month 3). A change in the HAART regimen may well have been responsible for the increase in CD4 counts at later time points of this study, but even the three patients who adhered to their pre-treatment combination of antivirals showed stable T-helper-cell counts above or around baseline levels despite persistent high viral loads (Patients 1, 12, 15). The individual time courses of T-helper-cell counts are shown in Figure 5.

Figure 4.

Although a significant decline in CD4 counts was observed during the weeks before cell infusion, CD4 levels rose significantly after infusion. The mean absolute change and 95% confidence intervals of CD4 cell counts relative to baseline at the day of cell infusion are shown.

Full figure and legend (12K)

Figure 5.

One year after cell infusion, the CD4 cell counts of five patients were clearly above baseline levels, and those of five patients around baseline levels (<20 cells/l deviation from baseline). The individual absolute changes in CD4 cell counts relative to baseline (day of cell infusion) are shown. ART, change in highly active anti-retroviral therapy regimen.

Full figure and legend (29K)

The increase in mean T-helper-cell count was accompanied by a significant increase in the mean stimulatory index for pokeweed mitogen (Supplementary Materials and Methods, Supplementary Table S2), indicating an improvement in functional T-cell responses. No significant changes in lympho-proliferative responses for other common antigens (tetanus, candida, cytomegalovirus, Mycobacterium tuberculosis purified protein derivative, toxoplasma, Flu) or for HIV antigens (p24, p17, gp160) were observed (Supplementary Table S2).

Interferon gamma (IFN-) as well as IL-4 responses of CD4 T cells were analyzed by enzyme-linked immunosorbent spot (ELISpot) assay at three different time points using recombinant proteins derived from either SEB (used as positive control) or recall antigens, such as tetanus toxoid and cytomegalovirus, and HIV-1-derived proteins including gag (p17, p24) and env (gp41/160). Whereas all patients displayed extremely weak IFN- responses in the CD4 population against recombinant HIV-1 proteins (Supplementary Figure S1), clear and also strong signals were detected upon engagement with SEB and against the recall antigens tetanus toxoid and cytomegalovirus. Moreover, five of ten patients showed increasing IFN- responses over time (Patients 4, 6, 7, 15, and 16), whereas three individuals (Patients 1, 14, and 17) showed hardly any INF- responder cells in this assay (Figure 6a). One patient (Patient 17), who possessed hardly any IFN--secreting cells (Figure 6a), displayed increasing numbers of IL-4-secreting cells over time for SEB and recall antigens, in contrast to the other nine individuals, who had weak or even undetectable IL-4 responses (Figure 6b). Thus, six out of ten subjects showed improved cytokine secretion within the CD4 subset after the transfer of gene-modified autologous T helper cells.

Figure 6.

For several patients, improvement of CD4 responses to recall antigens or CD8 responses to human immunodeficiency virus (HIV) antigens were observed. (a) Percentage spot-forming cells (SFCs) in an INF- enzyme-linked immunosorbent spot (ELISpot) assay measuring CD4 responses to the recall antigens Staphylococcus Enterotoxin B, tetanus toxoid (TT), and cytomegalovirus. (b) Percentage SFCs in an IL-4 ELISpot assay measuring CD4 responses to recall antigens. (c) Percentage SFCs in an INF- ELISpot assay measuring CD8 responses to HIV-1 antigens shown for the four patients with an improvement of responses after T-cell infusions. (d) Percentage SFCs in an INF- ELISpot assay measuring CD8 responses against HIV-1 antigens in the six patients who did not show improvement of responses after T-cell infusions.

Full figure and legend (30K)

Within the CD8 T-cell population, responses against HIV-1 gag/nef/pol/env peptides at three different time points at screening and after transfusion (day 45 and month 3) were analyzed using IFN- ELISpot. Four patients (Patients 4, 6, 7, and 16; Figure 6c) displayed increasing CD8 responses toward HIV-1 gag, nef, and/or pol over time, although Patient 6 had low baseline activities, whereas the remaining six individuals had comparatively low IFN- responder frequencies in their CD8 subset that tended to decline over time (Figure 6d).

Immune response to the transgene product

Humoral and cellular immune responses to maC46 were monitored. The antibody response was analyzed by flow cytometry using a maC46-expressing T-cell line. Results were confirmed by western blot (data not shown). Eight of ten patients had pre-existing antibodies to the M87o transgene product. Patients 6 and 12 were anti-maC46 negative. Titers remained stable or even declined during the first months after infusion, and the two maC46-seronegative patients did not seroconvert after therapy (Figure 7a). Interestingly, the two patients with no detectable antibodies to M87o protein had relatively low levels of gene marking, and Patient 4, who had the highest antibody titers, showed a relatively high level of gene-marked cells after transfusion (Figure 2). These results do not suggest that gene-modified cells were eliminated by humoral effector mechanisms.

Figure 7.

No clear immune response to maC46 protein was induced by the infusion of gene-modified cells. (a) The time course of antibody titers to maC46 in patient sera as determined by flow cytometry. (b) The time course of T-helper-cell responses measured as spot-forming cells to recombinant maC46 protein measured in (b) an INF- and (c) an IL-4 enzyme-linked immunosorbent spot assay for the patients treated with infusions of gene-modified T cells (Pt-1 to Pt-17) and for three healthy untreated, seronegative, control blood donors (C-1, C-2, C-3).

Full figure and legend (20K)

None of the individuals displayed detectable cytotoxic T-lymphocyte responses toward C46-derived peptides at any time, as analyzed by IFN- ELISpot (data not shown). CD4 responses to the recombinant maC46 protein produced in Escherichia coli were weak, sporadic, and not clearly dose dependent (Figure 7b and c). However, two patients (Patients 6 and 14) displayed IFN-- as well as IL-4-secreting cells and could be regarded as "maC46-responders." These two patients indeed had a relatively low level of gene marking; however, both also had received relatively low cell doses. In Patient 6, no M87o-positive cells were detectable between week 1 and month 10 and CD4 responses to maC46 had already been detected before cell infusion. In Patient 14, the level of gene-marked cells dropped below the detection limit after month 4 and CD4 responses to maC46 appeared to increase in the first 3 months after cell infusion. These observations could indicate an immune-mediated elimination of gene-modified cells in these two patients. However, as sporadic and significantly higher CD4 responses to maC46 protein were also observed in control lymphocytes from HIV-seronegative donors (Figure 7b and c, controls C-1, C-2, C-3), the relevance of these responses in the study participants is difficult to interpret and may simply reflect an immune response to traces of E. coli antigens contaminating the recombinant maC46 protein preparation.

Discussion

Ten HIV-infected patients were treated with autologous CD8-depleted T cells that had been gene-modified to express the antiviral membrane-anchored peptide maC46. Low levels of gene marking were detected throughout the 1-year follow-up. During the first 3 months after infusion, however, a significant increase of T helper cells was observed, with no effect on viral load. Thereafter, several patients showed a strong decline in viral load, most likely owing to changes in their HAART regimens.

No major toxicity was observed during this phase I trial. Specific safety issues addressed were transgene-induced immuno- and genotoxicity. The potentially pre-existing immunity to the maC46 transgene product discussed below did not cause any clinically overt toxicity. An additional concern was the potential risk of leukemia caused by retroviral insertional mutagenesis. Vector-induced leukemias have been observed after stem cell gene therapies for X-linked severe combined immunodeficiency,although here the common gamma-chain transgene might have promoted leukemogenesis.^9,15,16 Leukemia as a result of retroviral gene transfer has never been observed after the infusion of gene-modified T cells and thus appears to be extremely unlikely. So far, there is indeed no indication of retroviral genotoxicity in T cells in the study presented here. A longer follow-up is required before drawing a final conclusion, however.

Gene-marking levels achieved were generally below 0.01% of total leukocytes beyond day 7 and clearly correlated with the infused T-cell dose. Stable levels of gene marking were achieved only in the group who received more than 8 10⁹ cells. Assuming that these immunodeficient patients do not have many more than 10¹¹ T helper cells, an infusion of between 10⁸ and 10⁹ gene-marked cells would be expected to lead to much higher levels of gene marking of above 0.1% T helper cells. A low level of gene modification relative to the number of infused cells has been observed in many similar previous trials.^{8,17,18,19,20,21} Homeostatic mechanisms most likely hinder engraftment of ex vivo–cultured T cells, and a substantial proportion of activated T cells are thought to undergo apoptosis after infusion. Higher levels of more than 1% gene-modified T cells in vivo have been reported only if the transgene was expected to promote cell expansion or a strong survival advantage and in T-cell-depleted recipients, e.g., after conditioning.^22,23

In vitro studies have clearly shown that the maC46 protein effectively protects cells against HIV infection and thus indeed confers a strong selective advantage with a rapid accumulation of gene-modified cells in HIV-infected cell cultures.¹⁰ However, a clear accumulation of gene-modified cells over time was not seen in the patients treated in this trial. There are several possible explanations for this discrepancy. First, the gene-modified cells may have down-regulated expression of the transgene in vivo. Antiviral efficacy of maC46 correlates strongly with the level of maC46 expressed on the cell surface¹⁰ (F. G. Hermann, H. Martinius, M. Egelhofer, T. Giroglou, R. Zahn, P. Schult-Dietrich et al., manuscript submitted). In addition, the infused cells may have lost their repopulation potential during ex vivo expansion and thus may per se have a selective disadvantage. Finally, the percentage of HIV-infected T cells and thus the total turnover of T helper cells are much lower in patients in vivo (<1%) than in HIV-infected T-cell cultures (>80%). Therefore, the selective advantage conferred by an antiviral transgene is relatively low in HIV-infected patients, most likely much lower than for the transgenes used to treat severe combined immunodeficiency (common gamma-chain and adenosine deaminase genes) in children. The latter have indeed effectively supported in vivo accumulation of gene-corrected cells.^24,25

Finally, maC46 may be immunogenic, thus limiting expansion of transgene-expressing cells in vivo. Eight of ten patients had pre-existing antibodies to maC46. The antibody titers remained stable or declined after therapy, and gene marking showed no correlation to antibody titers. Thus, there was no indication that gene-modified cells were eliminated by antibody-dependent immune effector mechanisms. No CD8 responses were detectable to C46 peptides, but a slight boost of T-helper-cell responses to maC46 was observed in the ELISpot assay in two patients (Figure 7, Patients 6 and 14). Thus, an elimination of gene-modified T lymphocytes by cellular immune effector mechanisms cannot be excluded but seems unlikely for several reasons. Lymphocytes from seronegative donors also showed reactivity to the recombinant maC46, which had been produced in E. coli. These low responses were thus most likely directed toward contaminating traces of E. coli antigens. Both patients did have low levels of long-term gene marking, but this can easily be explained by the low cell dose transfused. In Patient 14, gene-modified cells became undetectable on day 120 after an initially relatively good level of gene marking, which may be indicative of an immunologic elimination of the gene-modified cells. However, the time point for this response is unexpectedly late. In a previous study that provided evidence for the elimination of T cells expressing the hygTk transgene by the immune system, gene marking had already drastically declined within the first week after the second cell infusion, which was given 2 weeks after the first cell dose.²⁶

Most likely because of the insufficient level of gene-protected cells, no clear antiviral effect was observed in the first 4 months after cell infusion. After month 3, the protocol permitted a change in the patients' HAART regimens. Thus, the antiviral activity of the cell infusion could not be evaluated for later time points. Interestingly Four of seven patients who received new antiviral drugs showed a significant decline in viral load. Interestingly, four of five patients who changed to an enfuvirtide-containing regimen responded to therapy. Thus, there was no indication that the infusion of maC46-expressing cells decreased the susceptibility to enfuvirtide (T-20).

A continuous decline in T-helper-cell counts had been observed for most patients before cell infusion in this study. After cell therapy, this trend was reversed, and T-helper-cell counts increased significantly in the first weeks after therapy. This has been observed previously in patients after infusion of T cells, especially after ex vivo expansion using anti-CD3 and anti-CD28 antibodies immobilized on beads.^8,27 It has been unclear, however, whether this increase was due to massive in vivo expansion of the infused T cells or whether the resident lymphocytes in the recipient were induced to proliferate. Ex vivo–stimulated T cells do indeed produce cytokines that can promote T-cell growth. A prominent example is IL-2, which is known to augment T-helper-cell counts in HIV-infected patients.²⁸ The low percentages of gene-marked cells in the recipients after infusion seen in this trial clearly show that the infused cells contributed only insignificantly to the increase in T-helper-cell counts and that indirect mechanisms were more likely involved. The rise in total T-helper-cell numbers was associated with an increase in T-cell function in some of the assays, such as the pokeweed mitogen-lymphocyte transformation test, and the ELISpot assay detecting CD4 responses to recall antigens and CD8 responses to HIV-1 antigens. Although the clinical significance of these observations remains unclear owing to the lack of a control arm, they might suggest an immunologic benefit.

The overall clinical course of the patients in this study has been unexpectedly positive considering their initial HAART failure, high viral loads, and low T-helper-cell counts. One year after therapy, T-helper-cell counts were around or above baseline levels and the viral load was suppressed effectively in four of seven patients who had changed their thus-far-ineffective HAART regimen. Today, 2 years and longer after treatment, all patients are alive and well. Clearly, the treatment approach presented here appears to be safe, but further, larger controlled clinical trials would be required to prove a clear potential beneficial effect on the overall clinical course of HIV infection.

Materials and Methods

Patients. Ten male patients between 33 and 56 (median 51) years of age were enrolled into this phase I study after written informed consent. The protocol was approved by the local Ethics Committee of the University of Hamburg and conducted according to good clinical practice guidelines. All patients had been infected with HIV-1 for more than 9 (median 14.3) years and were in an advanced stage of immunodeficiency with symptomatic HIV disease. All but two patients fulfilled the clinical definition of Stage C disease (full-blown acquired immunodeficiency syndrome) according to the Centers for Disease Control and Prevention classification. According to the inclusion criteria, all patients had an unchanged HAART regimen for a minimum of 3 months before cell infusion, a viral load despite HAART of above 5,000 RNA copies/ml, an absolute CD4 count below 200/l or a relative proportion of CD4 cells below 15% of total T cells, a Karnofsky performance status of at least 80%, and a life-expectancy of at least 6 months at entry. All had been exposed to all three classes of anti-retrovirals (protease inhibitors, nucleoside reverse transcriptase inhibitors, and non-nucleoside reverse transcriptase inhibitors) and had proven primary mutations in the reverse transcriptase and protease genes conferring resistance to at least one drug from each class. The median CD4 count at baseline was 93 cells/l and the median viral load was 4.96 log₁₀ RNA copies/ml despite anti-retroviral combination therapy (see also Table 1). The patient code numbers were assigned in the order of recruitment. Two code numbers, 9 and 10, had been reserved for a second trial site, but this site did not recruit. Thus, these two numbers were not assigned.

Vector production and T-cell processing. Clinical-grade serum-free retroviral vector supernatant containing M87o was produced on multi-tray cell factories (Nunc, Wiesbaden, Germany) according to current good manufacturing practice guidelines as previously described.²⁹

Gene-modified CD4⁺ T cells were manufactured according to current good manufacturing practice guidelines. Patients underwent lymphapheresis, and a minimum of 1.0 10¹⁰ mononuclear cells were collected. After overnight storage, cells were washed with a CytoMate device (Baxter, Heidelberg, Germany) and incubated with magnetic beads labeled with anti-CD8 antibodies (Miltenyi Biotech, Bergisch-Gladbach, Germany) for 30 minutes. After a second wash step, CD8⁺ cells were depleted using the CliniMacs (Miltenyi Biotech). A maximum of 2.5 10⁸ CD3⁺ cells were then incubated with anti-CD3/anti-CD28-coated Xcyte Dynabeads (Xcyte Therapies, Seattle, WA) at a CD3⁺ cell to bead ratio of 1:3 for 30 minutes on a lab rotator. Labeled cells were then enriched via the MaxSep permanent magnet (Baxter) and carefully resuspended in X-Vivo 15 medium (Cambrex)complemented with 100U/ml rhIL-2 (Chiron, Munich, Germany), 2mM l-glutamine (Cambrex), 5% human serum (Cambrex), and 20mM HEPES (Invitrogen, Karlsruhe, Germany) at a cell density of 5 10⁵ cells/ml and seeded into tissue culture bags (Baxter). A mixture of antivirals (1M nelfinavir (Viracept), Roche, Basel, Switzerland 0.33M amprenavir (Agenerase), GlaxoSmithKline, Munich, Germany; 10g/ml T-20 (Fuzeon), Roche) was added to the cell suspension to avoid viral replication. After 4 days of culture at 37°C and 5% CO₂, Xcyte Dynabeads were removed from the cell suspension. Cells were then transduced in tissue culture flasks pre-loaded with retroviral vectors as described elsewhere.³⁰ Transduction was repeated the next day. After transduction, the cells were expanded for a maximum of 7 days in a static culture until the required cell number was achieved. Finally, the remaining Xcyte Dynabeads were removed and cells were harvested with a CytoMate device and cryopreserved in dimethyl sulfoxide (WAK Chemie, Steinbach, Germany), PlasmaLyte A (Baxter), Plasmasteril (6% hydroxyethyl starch; Fresenius Kabi, Bad Homburg, Germany), and human serum albumin (20%, Baxter) for long-term storage.

T-cell transfusions.Nine of ten patients received a test dose of 0.1 10⁹ cells followed by infusion of the cell product 6–8 hours later. For Patient 17, only 3.5 10⁹ cells could be manufactured, and this cell product was infused without a prior test dose.

Quantitative polymerase chain reaction and reverse transcriptase polymerase chain reaction. The presence of M87o proviral sequences in the genomic DNA isolated from patient samples was determined using the ABI PRISM 7700 Sequence Detection System (PE Applied Biosystems, Weiterstadt, Germany). In brief, primers t20-f (5'-GCGCCAGATCTTGGATGGA-3') and t20-r (5'-CTCGTTCTTCTCCTGCTGGTTCT-3') amplified a 95-base pair fragment of the M87o cDNA. Amplification was detected with the FAM-labeledprobe t20-p (5'-ACCGCGAGATCAACAACTACACCAGCC-3'). The human erythropoietin receptor gene was used as an internal control to quantify the M87o reaction. Primers hepo-f (5'-CTGCTGCCAGCTTTGAGTACACTA-3') and hepo-r (5'-GAGATGCCAGAGTCAGATACCACAA-3') amplified a 138-base pair fragment from exon 8 of the human erythropoietin receptor gene. Amplification was determined by the VIC-labeled probe hepo-p (5'-ACCCCAGCTCCCAGCTCTTGCGT-3'). Both reactions were carried out in a single tube. The amplification cycle was 15 seconds at 94°C, 1 minute at 60°C. In each experiment, the amplification of DNA generated from HT1080 cells containing a single copy of the M87o vector mixed into Wild-type HT1080 cells in defined ratios was used to quantify the percentage of M87o genome integrations per human genome.

Immunologic and virologic monitoring. Viral load and T-helper-cell subsets were measured at regular predefined intervals (weekly for the first 2 months and monthly thereafter) by multicolor flow cytometry after staining with the corresponding monoclonal antibodies (FACSCalibur, Becton-Dickinson, Heidelberg, Germany) and reverse transcriptase polymerase chain reaction (Roche Amplicor, version 3.0, Roche Diagnostics, Mannheim, Germany). Aliquots of peripheral blood mononuclear cells and sera were collected and cryopreserved at the above-mentioned intervals for later functional T-cell and antibody assays as described below.

HLA typing. Peripheral blood of the patients included in the study was analysed for HLA-A, HLA-B, and HLA-C loci using the reverse sequence-specific oligonucleotide line blot assay.³¹ Allelic assessment was performed using the Helmberg score software based on the most recent nomenclature report and library.^32,33,34

Peptide selection for CD8 ELISpot. From the amino acid sequence of (Y T S L I H S L I E E S Q N Q Q E K N E Q E L L E L D K W A S L W N W F) HLA-matched peptides were identified with HLA-peptide-matching programs (www.bimas.dcrt.nih.gov/molbio/hla_bind/ and www.syfpeithi.de/) and synthesized (EMC Microcollections, Tübingen, Germany). Peptide sequences and major histocompatibility complex class I haplotype scores for various class I alleles as defined at www.syfpeithi.de according to their scoring classification are given in the Supplementary Materials and Methods and Supplementary Table S1. HIV-1-specific CD8⁺ cells were analysed using optimal peptides based on the HIV-1 clade B consensus sequence (for review see also ref. 35).

CD8 responses byIFN- ELISpot.Immobilon-coated microtiter plates (MAIPS45, Millipore, Cambridge, UK) were coated with 5ng/well anti-IFN- (Mab 1-D1K, Mabtech, Stockholm, Sweden) overnight at 4°C. The plates were washed six times with phosphate-buffered saline (PBS) (Gibco, Paisley, UK) and blocked for 2 hours with 2% bovine serum albumin (Sigma-Aldrich, Munich, Germany) in PBS at 37°C and 5% CO₂. In the meantime, frozen peripheral blood mononuclear cells were thawed, washed in 50ml of ice-cold Roswell Park Memorial Institute medium (RPMI) for 8 minutes at 400 g, followed by two washes in pre-warmed (37°C) RPMI. The resulting pellet was resuspended in 10ml pre-warmed R10-Medium (RPMI, 100mM HEPES) and 1% penicillin/streptomycin, both from Gibco, and 10% fetal calf serum (Biochrom, Berlin, Germany) and cultured for 2–4 hours at 37°C and 5% CO₂ in 15-ml conical tubes (Falcon, Becton-Dickinson) before trypan blue staining (Gibco) and adjustment to 10⁶ cells/ml R10. The bovine serum albumin buffer was removed and the microtiter plates were coated with 30l R10/well, to which 0.2g of the desired peptide (stock: 2mg/dimethyl sulfoxide; synthesized by EMC Microcollections, 10-fold diluted in RPMI, 100mM HEPES, and 1% penicillin/streptomycin) was added. Analysis was performed in duplicate.

To each well, 10⁵ cells/100l R10 were added and cultured for 12–16 hours at 37°C and 5% CO₂. Cells were removed and plates washed six times with PBS, followed by incubation with biotinylated anti-IFN- detection monoclonal antibody (clone 7-B6; Mabtech, Nacka Strand, Sweden, 0.01g/well) for 1 hour at room temperature. After repeating the washing step, streptavidin-alkaline phosphatase was added (Mabtech, 5l/10 ml PBS; 50l/well) for a further 45 minutes at room temperature. Plates were washed and incubated for 15–20 minutes with 50l/well 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NBT) liquid substrate buffer system (Sigma-Aldrich).Washing the plates with tap water terminated the reaction. Plates were allowed to dry overnight and the number of dots was estimated using a dissecting microscope (Zeiss, Jena, Germany). The number of dots is expressed as percentage of CD8 spot-forming cells, which was calculated relative to the CD8⁺ T cells, as assessed by flow cytometry using the monoclonal antibodies described below.

CD4 responses by IFN- and CD4 ELISpot. Frozen peripheral blood mononuclear cells and, if available, lymph node cells were thawed and washed as described above. Cells were adjusted to 5 10⁵ to 5 10⁶/2ml cell culture medium containing RPMI, 10% fetal calf serum (Biochrom, Berlin, Germany), 1% penicillin/streptomycin (Gibco), 1% l-glutamine (Gibco), 1% non-essential amino acids (Gibco), and 0.05% 2-mercaptoethanol (Gibco) and cultured for 16–20 hours in 24-well plates (Greinet, Frickenhausen, Germany) in the absence (medium alone) or presence of one of the following antigens: SEB (1.5g; Sigma), tetanus toxoid (1.5g; Sigma-Aldrich), cytomegalovirus (strain AD169; ABI via Tebu, Offenbach, Germany), or recombinant maC46. The complete maC46 protein (including signal peptide, C46, hinge, and membrane anchor) was produced in E. coli. The bacterial expression plasmid coding for the recombinant maC46 protein (derived from pSW55, ref. 36) codes for the maC46 protein flanked by two 6-histidine tags for purification. The maC46 protein was purified by immobilized metal affinity chromatography according to standard procedures.

After exposure to the antigens, 1 ml/well cell culture was removed, placed into 12-ml Falcon tubes pre-filled with 5ml of 37°C warm RPMI, and washed at 400g at room temperature for 8 minutes. Cells were adjusted to 10⁵/ml in cell culture medium and 200l/well was placed into either anti-IL-4- or anti-IFN--coated (50ng/well; all monoclonal antibodies were obtained from Mabtech, Stockholm, Sweden) and bovine serum albumin–blocked MAIP S45 plates in triplicate. The plates were cultured for 16–20 hours at 37°C in a CO₂ incubator (5%) before development of cytokine spots and estimation of percentage of CD4 spot-forming cells, as described above for the CD8 ELISpot.

Flow cytometry. An aliquot of cells prepared for the CD4 or CD8 ELISpot was analysed by flow cytometry to determine T-cell subsets using the following antibodies/fluorochromes: CD3-PerCP, CD4-APC, and CD8-PE (all from Becton-Dickinson). Surface labeling was performed by incubation of the cells with a cocktail of three different antibodies for 30 minutes at 4°C, followed by wash and fixation in 0.5% paraformaldehyde(Merck, Darmstadt, Germany) in PBS. Viability was assessed using propidium iodide/annexin staining, and analysis was performed on a FACSCalibur using the Cell Quest acquisition and analysis program (Becton-Dickinson, version 3.2.1).

Detection of anti-M87o antibodies in patient plasma. To investigate a possible immune response to the transplanted M87o-expressing cells, the antibody titer in patient plasma was determined by flow cytometry. Blood samples were collected at four different time points in the course of the study: screening, day of re-infusion (day 0), and days 45 and 56 after re-infusion. Plasma was obtained using a Ficoll gradient and heat-inactivated by incubation at 56°C for 30 minutes. Then, 5 10⁵ PM1 cells expressing maC46 from the vector M87o were incubated with serial dilutions of the plasma, ranging from 1:10 to 1:1,280, in PBS plus 3% fetal calf serum for 30 minutes at room temperature. After washing, binding was detected using a PE-coupled goat anti-human antibody (Dianova, Hamburg, Germany). To quantify the number of anti-M87o antibodies, native PM1 cells were stained as control. The resulting EC₅₀ values are the plasma dilution at which 50% of PM1 M87o cells stain positive. The human monoclonal antibody 2F5, which recognizes a specific epitope on the M87o peptide, was used as positive control (kindly provided by H. Katinger and G. Stiegler³⁷). Background staining of human plasma to the M87o peptide was measured using plasma obtained from ten healthy individuals. EC₅₀ values at a dilution of 1–20 and greater were considered significant.

Statistical analysis. Statistical hypotheses were not formed in this study, and there was no formal sample size calculation for any hypothesis. All analyses were of an exploratory nature. However, during the study correlation analysis was performed to test for correlations between different variables. For some variables, a t-test for changes from baseline was performed.

References

REFERENCES

1. Little, SJ, Holte, S, Routy, JP, Daar, ES, Markowitz, M, Collier, AC et al (2002). Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med 347: 385–394. | Article | PubMed | ISI | ChemPort |
2. Palella, FJ Jr.,Deloria-Knoll, M, Chmiel, JS, Moorman, AC, Wood, KC, Greenberg, AE et al (2003). Survival benefit of initiating antiretroviral therapy in HIV-infected persons in different CD4⁺ cell strata. Ann Intern Med 138: 620–626. | PubMed | ISI |
3. Connors, M, Kovacs, JA, Krevat, S, Gea-Banacloche, JC, Sneller, MC, Flanigan, M et al (1997). HIV infection induces changes in CD4⁺ T-cell phenotype and depletions within the CD4⁺ T-cell repertoire that are not immediately restored by antiviral or immune-based therapies. Nat Med 3: 533–540. | Article | PubMed | ISI | ChemPort |
4. Gorochov, G, Neumann, AU, Kereveur, A, Parizot, C, Li, T, Katlama, C et al (1998). Perturbation of CD4⁺ and CD8⁺ T-cell repertoires during progression to AIDS and regulation of the CD4⁺ repertoire during antiviral therapy. Nat Med 4: 215–221. | Article | PubMed | ISI | ChemPort |
5. Malhotra, U, Huntsberry, C, Holte, S, Lee, J, Corey, L and McElrath, MJ (2006). CD4⁺ T-cell receptor repertoire perturbations in HIV-1 infection: association with plasma viremia and disease progression. Clin Immunol 119: 95–102. | Article | PubMed | ChemPort |
6. von Laer, D, Hasselmann, S and Hasselmann, K (2006). Gene therapy for HIV infection: what does it need to make it work? J Gene Med 8: 658–667. | Article | PubMed | ChemPort |
7. von Laer, D, Baum, C, Schambach, A, Kuehlcke, K, Zahn, R, Newrzela, S et al. (in press). Gene therapeutic approaches for immune modulation in AIDS. CMC-AIAA.
8. Levine, BL, Humeau, LM, Boyer, J, Macgregor, RR, Rebello, T, Lu, X et al (2006). Gene transfer in humans using a conditionally replicating lentiviral vector. Proc Natl Acad Sci USA 103: 17372–17377. | Article | PubMed | ChemPort |
9. Baum, C, Dullmann, J, Li, Z, Fehse, B, Meyer, J, Williams, DA et al (2003). Side effects of retroviral gene transfer into hematopoietic stem cells. Blood 101: 2099–2114. | Article | PubMed | ISI | ChemPort |
10. Egelhofer, M, Brandenburg, G, Martinius, H, Schult-Dietrich, P, Melikyan, G, Kunert, R et al (2004). Inhibition of human immunodeficiency virus type 1 entry in cells expressing gp41-derived peptides. J Virol 78: 568–575. | Article | PubMed | ISI | ChemPort |
11. Hildinger, M, Dittmar, MT, Schult-Dietrich, P, Fehse, B, Schnierle, BS, Thaler, S et al (2001). Membrane-anchored peptide inhibits human immunodeficiency virus entry. J Virol 75: 3038–3042. | Article | PubMed | ISI | ChemPort |
12. Melikyan, GB, Egelhofer, M and von Laer, D (2006). Membrane-anchored inhibitory peptides capture human immunodeficiency virus type 1 gp41 conformations that engage the target membrane prior to fusion. J Virol 80: 3249–3258. | Article | PubMed | ChemPort |
13. Lohrengel, S, Hermann, F, Hagmann, I, Oberwinkler, H, Scrivano, L, Hoffmann, C et al (2005). Determinants of human immunodeficiency virus type 1 resistance to membrane-anchored gp41-derived peptides. J Virol 79: 10237–10246. | Article | PubMed | ChemPort |
14. von Laer, D, Hasselmann, S and Hasselmann, K (2006). Impact of gene-modified T cells on HIV infection dynamics. J Theor Biol 238: 60–77. | Article | PubMed | ISI | ChemPort |
15. Hacein-Bey-Abina, S, von Kalle, C, Schmidt, M, Le Deist, F, Wulffraat, N, McIntyre, E et al (2003). A serious adverse event after successful gene therapy for X-linked severe combined immunodeficiency. N Engl J Med 348: 255–256. | Article | PubMed | ISI |
16. Woods, NB, Bottero, V, Schmidt, M, von Kalle, C and Verma, IM (2006). Gene therapy: therapeutic gene causing lymphoma. Nature 440: 1123. | Article | PubMed | ChemPort |
17. Buchschacher, GL, Jr. and Wong-Staal, F (2001). Approaches to gene therapy for human immunodeficiency virus infection. Hum Gene Ther 12: 1013–1019. | Article | PubMed | ISI | ChemPort |
18. Ranga, U, Woffendin, C, Verma, S, Xu, L, June, CH, Bishop, DK et al (1998). Enhanced T cell engraftment after retroviral delivery of an antiviral gene in HIV-infected individuals. Proc Natl Acad Sci USA 95: 1201–1206. | Article | PubMed | ChemPort |
19. Woffendin C, Ranga U, Yang Z, Xu L and Nabel, GJ (1996). Expression of a protective gene-prolongs survival of T cells in human immunodeficiency virus-infected patients. Proc Natl Acad Sci USA 93: 2889–2894. | Article | PubMed | ChemPort |
20. Morgan, RA, Walker, R, Carter, CS, Natarajan, V, Tavel, JA, Bechtel, C et al (2005). Preferential survival of CD4⁺ T lymphocytes engineered with anti-human immunodeficiency virus (HIV) genes in HIV-infected individuals. Hum Gene Ther 16: 1065–1074. | Article | PubMed | ChemPort |
21. Macpherson, JL, Boyd, MP, Arndt, AJ, Todd, AV, Fanning, GC, Ely, JA et al (2005). Long-term survival and concomitant gene expression of ribozyme-transduced CD4⁺ T-lymphocytes in HIV-infected patients. J Gene Med 7: 552–564. | Article | PubMed | ChemPort |
22. Gattinoni, L, Finkelstein, SE, Klebanoff, CA, Antony, PA, Palmer, DC, Spiess, PJ et al (2005). Removal of homeostatic cytokine sinks by lymphodepletion enhances the efficacy of adoptively transferred tumor-specific CD8⁺ T cells. J Exp Med 202: 907–912. | Article | PubMed | ISI | ChemPort |
23. Dudley, ME, Wunderlich, JR, Yang, JC, Hwu, P, Schwartzentruber, DJ, Topalian, SL et al (2002). A phase I study of nonmyeloablative chemotherapy and adoptive transfer of autologous tumor antigen-specific T lymphocytes in patients with metastatic melanoma. J Immunother 25: 243–251. | Article | PubMed | ISI | ChemPort |
24. Hacein-Bey-Abina, S, Le Deist, F, Carlier, F, Bouneaud, C, Hue, C, De Villartay, JP et al (2002). Sustained correction of X-linked severe combined immunodeficiency by ex vivo gene therapy. N Engl J Med 346: 1185–1193. | Article | PubMed | ISI | ChemPort |
25. Aiuti, A, Slavin, S, Aker, M, Ficara, F, Deola, S, Mortellaro, A et al (2002). Correction of ADA-SCID by stem cell gene therapy combined with nonmyeloablative conditioning. Science 296: 2410–2413. | Article | PubMed | ISI | ChemPort |
26. Riddell, SR, Elliott, M, Lewinsohn, DA, Gilbert, MJ, Wilson, L, Manley, SA et al (1996). T-cell mediated rejection of gene-modified HIV-specific cytotoxic T lymphocytes in HIV-infected patients. Nat Med 2: 216–223. | Article | PubMed | ISI | ChemPort |
27. Levine, BL, Bernstein, WB, Aronson, NE, Schlienger, K, Cotte, J, Perfetto, S et al (2002). Adoptive transfer of costimulated CD4⁺ T cells induces expansion of peripheral T cells and decreased CCR5 expression in HIV infection. Nat Med 8: 47–53. | Article | PubMed | ISI | ChemPort |
28. Stellbrink, HJ, van Lunzen, J, Westby, M, O'Sullivan, E, Schneider, C, Adam, A et al (2002). Effects of interleukin-2 plus highly active antiretroviral therapy on HIV-1 replication and proviral DNA (COSMIC trial). AIDS 16: 1479–1487. | Article | PubMed | ChemPort |
29. Schilz, AJ, Kuhlcke, K, Fauser, AA and Eckert, HG (2001). Optimization of retroviral vector generation for clinical application. J Gene Med 3: 427–436. | Article | PubMed | ISI | ChemPort |
30. Kuhlcke, K, Fehse, B, Schilz, A, Loges, S, Lindemann, C, Ayuk, F et al (2002). Highly efficient retroviral gene transfer based on centrifugation-mediated vector preloading of tissue culture vessels. Mol Ther 5: 473–478. | Article | PubMed | ISI | ChemPort |
31. Helmberg, W, Lanzer, G, Zahn, R, Weinmayr, B, Wagner, T and Albert, E (1998). Virtual DNA analysis—a new tool for combination and standardised evaluation of SSO, SSP and sequencing-based typing results. Tissue Antigens 51: 587–592. | PubMed | ChemPort |
32. Marsh, JC, Chang, J, Testa, NG, Hows, JM and Dexter, TM (1990). The hematopoietic defect in aplastic anemia assessed by long-term marrow culture. Blood 76: 1748–1757. | PubMed | ChemPort |
33. Marsh, SG, Albert, ED, Bodmer, WF, Bontrop, RE, Dupont, B, Erlich, HA et al (2005). Nomenclature for factors of the HLA system, 2004. Tissue Antigens 65: 301–369. | Article | PubMed | ISI | ChemPort |
34. Robinson, J, Waller, MJ, Parham, P, de Groot, N, Bontrop, R, Kennedy, LJ et al (2003). IMGT/HLA and IMGT/MHC: sequence databases for the study of the major histocompatibility complex. Nucleic Acids Res 31: 311–314. | Article | PubMed | ISI | ChemPort |
35. Brander, C and Goulder, PJ (2001). The identification of optimal HIV derived CTL epitopes in diverse populations using HIV clade specific consensus. In: Korber, BT, Haynes, BF, Koup, R, Kuiken, C, Moore, JP, Walker, BD et al. (eds). HIV Molecular Imunology Database. Los Alamos National Laboratory: Los Alamos, NM. 1–20.
36. Uherek, C, Fominaya, J and Wels, W (1998). A modular DNA carrier protein based on the structure of diphtheria toxin mediates target cell-specific gene delivery. J Biol Chem 273: 8835–8841. | Article | PubMed | ISI | ChemPort |
37. Stiegler, G, Armbruster, C, Vcelar, B, Stoiber, H, Kunert, R, Michael, NL et al (2002). Antiviral activity of the neutralizing antibodies 2F5 and 2G12 in asymptomatic HIV-1-infected humans: a phase I evaluation. AIDS 16: 2019–2025. | Article | PubMed | ChemPort |

Acknowledgements

This study was supported in part by research funding from Fresenius Biotech, Bad Homburg, to D.L., from Vision7 GmbH to J.L., and from the European Union (grant TRIoH LSHG-CT-2003-503480) to D.L. We thank Patricia Schult-Dietrich, Kristina Lenz, and Janine Mohn for excellent technical assistance and Silvia Koob for editorial work. We are also grateful to the study coordinators, Claudia Schlesner and Sophie Elena Enderwitz, for excellent patient care and for helping to conduct the study according to good clinical practice guidelines.