Male cynomolgus macaques (Macaca fascicularis), weighing between 3 and 6 kg were imported from Mauritius and housed in single cages within level 3 biosafety facilities, according to national institutional guidelines (Commission de génie génétique, Paris, France). All experimental procedures were carried out in accordance with European guidelines for primate experiments (Journal Officiel des Communautés Européennes, L358, December 18 1986).

Bone marrow mononuclear cells were obtained from the iliac crest or by aspiration from the humerus and isolated by standard Ficoll density-gradient centrifugation (MSL2000, Eurobio, Les Ulis, France). Cells were washed twice in phosphate-buffered saline (PBS, Eurobio, Les Ulis, France) and resuspended in 1% FCS (Fetal Calf Serum; Bio West, France) in PBS. The cellular fraction was then enriched in CD34⁺ cells by positive immunomagnetic selection, using beads coupled to a specific antibody (clone 561; Dynabeads M-450 CD34, Progenitor Cell Selection System, Dynal, Oslo, Norway), according to the manufacturer' s instructions. Immunoselected CD34⁺ cells were stained with a specific PE-conjugated anti-CD34 antibody (clone 563; Pharmingen, Becton Dickinson, California, USA) and analyzed by flow cytometry (LSR, Becton Dickinson, California, USA) to evaluate the level of enrichment. All preparations contained more than95% CD34⁺ cells, with a mean value of 97% ± 1% (n = 12) for in vitro assays and 96% ± 1% (n = 4) for in vivo assay.

Two SIV-derived vectors were produced, one for in vitro studies and the other for in vivo studies: 1) pRMES8 is a minimal packaging-competent SIVmac251-based vector34. It contains the enhanced green fluorescent protein (eGFP) marker gene under control of the mouse phosphoglycerate kinase (PGK) promoter, placed between the SIVmac251 LTRs and leader sequences. It carries the SIVmac251 RRE region and minimal sequences of the gag and pol genes encompassing central polypurine tract/central termination sequence (cPPT/CTS) regions (figure 1A). pRMES8 was used for in vitro assays investigating the susceptibility of CD34⁺ cells from primate bone marrow to transduction with SIVmac251-derived vectors. 2) For in vivo assays, we used pGASE; this plasmid is an optimised version of pRMES8, with a 3'-SIN-LTR for safety and insertion of an exon splicing enhancer (ESE) upstream the PGK promoter to increase titer 41

pSIV3⁺ is the packaging plasmid derived from the BK28 molecular clone of SIVmac251, as described elsewhere33. Briefly, the pSIV3⁺ gag/pol expression plasmid was obtained by replacing the 5' LTR of SIVmac251 (nucletotides 1 to 506) by the human cytomegalovirus (CMV) early-immediate promoter and enhancer region. The 5' half of the env gene (nt 6582 to 7981) was also removed, leaving the RRE (REV-responsive element) sequence and the 5' and 3' exons of the tat and rev regulatory genes intact. The 3' LTR (nt 9444 to 10249) was replaced by a SV40 polyadenylation sequence, resulting in deletion of the 3' end of the nef gene. Finally, the nef initiation codon was inactivated to prevent translation (figure 1B).

pGREV was used for pseudotyping. It is a bicistronic expression construct encoding the vesicular stomatitis virus glycoprotein (VSV-g) and the REV regulatory protein, linked by an EMCV IRES. Expression of this cassette, which contains the rabbit β-globin intron II and polyadenylation (pA) sequences (figure 1C), is driven by the constitutive CMV promoter.

293T cells were plated at a density of 4.0 × 10⁵ cells per well (in 6-well plates) on the day before transfection. Cells were transfected as previously described42. SIV vectors were produced by cotransfection with three plasmids: the SIV plasmid vector (pRMES8 or pGASE)(1.7 μg), the helper plasmid, pSIV3⁺, encoding Gag-Pol and regulatory proteins other than Env and Nef (1.7 μg) and the envelope glycoprotein-encoding plasmid pGREV (2.2 μg). The transfection medium was replaced after 16 hours of incubation. Virus-containing medium was collected 40 hours after transfection, clarified by centrifugation for 5 minutes at 800 g, and passed through a filter with 0.45 μm pores. For high-titer preparations, SIV vectors were concentrated by ultracentrifugation at 110,000 g for 2 hours. The viral pellet was resuspended by incubation for 2 hours at 4°C in phosphate-buffered saline supplemented with 1% glycerol34.

For determination of the infectious titer, sMAGI cells were seeded at a density of 4 × 10⁵ cells/ml in six-well plates one day before transduction in DMEM medium (Life Technologies Inc., Berlin, Germany) supplemented with 10% fetal bovine serum (FBS) (Gibco BRL, Grand Island, New York, USA), polybrene (6 μg/ml) (Sigma, Saint Louis, USA) and an antibiotic mixture (5 mg/ml penicillin; 5 mg/ml streptomycin; 10 mg/ml neomycin; Gibco BRL, Grand Island, New York, USA). The cells were cultured for one day, and we then added serial dilutions of virus preparations and incubated the plates for a further four hours. Cells were then washed in DMEM (Life Technologies Inc., Berlin, Germany). Transduction rates was determined 48 hours after infection, as the percentage of GFP-positive sMAGI cells (%GFP⁺c), by flow cytometry (FACScan, Becton Dickinson, San Jose, Mountain View, California, USA) after transducing 4 × 10⁵ cells with 1 ml of diluted viral supernatant (dilution factor = d). The infectious titer (IT), expressed as transducing units/ml, was calculated as: IT = %GFP⁺cells × 4 × 10⁵/100 × d.

Following immunoselection, CD34⁺ cells were cultured in a proliferation medium composed of Iscove's MDM supplemented with 1% bovine serum albumin (BSA), bovine pancreatic insulin (10 μg/ml), human transferrin (200 μg/ml), 2-mercaptoethanol (10^-4M) and L-glutamine (2 mM; Stemspan, Stem Cell Technologies, Meylan, France). The medium was supplemented with 50 ng/ml recombinant human (rh) SCF (Stem Cell Technologies, Meylan, France), 50 ng/ml rh Flt3-L (Stem Cell Technologies, Meylan, France), 10 ng/ml rh IL-3 (R&D Systems, Minneapolis, USA),10 ng/ml rh IL-6 (R&D Systems, Minneapolis, USA) and 4 μg/ml polybrene (Sigma, Saint Louis, USA) in plates coated with retronectin (Cambrex Bio Science, Paris, France). The CD34⁺cells were then transduced by 24 hours of coculture with the vector (multiplicity of infection (MOI) = 100).

Following the coculture of CD34⁺ cells with lentiviral vector, part of the cell culture was fixed in CellFix solution (Becton Dickinson, Erembodegem, Belgium) for evaluation of the rate of transduction of undifferentiated CD34⁺ cells. Part of the cell culture was cultured for 14 days in 35 mm Petri dishes containing semi-solid medium (Methocult GF H4434, Stem Cell Technologies, Meylan, France) composed of Iscove's MDM medium supplemented with 1% methylcellulose, 30% fetal bovine serum, 10^-4 M 2-mercaptoethanol, 2 mM L-glutamine, 50 ng/ml rhSCF, 10 ng/ml rhGM-CSF, 10 ng/ml rhIL-3 and 3 IU/ml rhEPO. Cells were cultured at a density of 10⁴ cells/ml (in triplicate) at 37°C, under an atmosphere containing 5% CO₂, to allow the myeloid differentiation of colony-forming cells (CFC).

The remaining cells were cocultured in 96-well plates for 35 days at 37°C, under an atmosphere containing 5% CO₂, on a layer of stromal cells of the murine fibroblastic cell line M2-10B4, in a medium composed of αMEM supplemented with 12.5% horse serum (HS), 12.5% FBS, 2 mM L-glutamine, 10^-4 M 2-mercaptoethanol, 0.16 M I-inositol and 16 μM folic acid (Myelocult H5100, Stem Cell Technologies, Meylan, France) and 10^-6 M hydrocortisone. Cells were cultured at a concentration of 10³ cells per well (24 wells per condition per monkey), to allow long-term culture-initiating cells (LTC-IC) to undergo myeloid differentiation to generate progenitor cells or CFC. The CFC were cultured for 14 days on semi-solid medium, as described above, to allow their myeloid differentiation into more mature cells.

CD34⁺ cells were treated with AZT before transduction, to inhibit transduction due to reverse transcription of the lentiviral vector genome. Immunoselected CD34⁺ cells were cultured overnight in the proliferation medium described above, with AZT concentrations of 0, 10^-7, 10^-6 and 10^-5 molar. The cells were washed twice and transduced with the lentiviral vector, according to the protocol described above. The real percentage of GFP-positive cells resulting from reverse transcription of the lentiviral vector was thus determined by subtracting the percentage of GFP-positive cells obtained after treatment with a saturating dose of AZT, from the percentage of GFP-positive cells obtained in the absence of AZT treatment.

After transduction and myeloid differentiation in semi-solid medium, the colonies formed by AZT-treated CFC were observed by fluorescence microscopy (Axiovert S100, Zeiss) using a magnification factor of 100. Fluorescence microscopy was used to detect GFP in each colony subtype, making it possible to determine the percentage of the colonies positive for GFP. We considered all colonies containing GFP-producing cells to be GFP-positive. Images were analyzed with Adobe Premiere and Adobe Photoshop software (Adobe Systems Inc., San Jose, CA, USA).

Eight animals were sedated with ketamine (Imalgène; 10 mg/kg, i.m.), Rhône-Mérieux, France) and placed in a restraint chair. They received myeloablative conditioning, in the form of total body exposure to ⁶⁰Co gamma rays with an anterior unilateral direction. A total midline tissue dose of 6 Gy was delivered at a rate of 25.92 cGy/minute. Dosimetry was performed, with 100 μL ionization chambers placed in paraffin wax cylindrical phantoms of a similar size and orientation to the seated animal.

After the coculture of CD34⁺ cells with the lentiviral vector, four animals underwent intramedullary infusion, of whole immunoselected CD34⁺ cells into both humeri (Table 1).

All animals received clinical support in the form of antibiotics and fresh irradiated whole blood, as required. An prophylactic antibiotic regimen was initiated when leukocyte count fell below 1,000/μl and continued daily until it exceeded 1,000/μl for three consecutive days: 1 ml/10 kg/day Bi-Gental^® (Schering-Plough Santé Animale) and 1 ml/10 kg Terramycin^® (Pfizer). Fresh, irradiated (25 Gy; ¹³⁷Cs gamma radiation) whole blood (approximately 50 ml/transfusion) from a random donor pool was administered if platelet count fell below 20,000/μl and hemoglobin concentration was less than 6 g/dl.

Peripheral blood and bone marrow mononuclear cells were incubated for 30 min at 4°C with 10 μl of selected monoclonal antibodies for single- or triple-color membrane staining. The following antibodies were used: APC-conjugated anti-CD3 (SP34-2, Becton Dickinson), PE-conjugated anti-CD14 (clone M5E2, BD Pharmingen), PE-conjugated anti-CD11b (BEAR-1, Beckman Coulter), PerCP-conjugated anti-CD20 (clone B9E9, Immunotech), PE-conjugated anti CD8 (clone RPA-T8, Becton Dickinson) and PerCP-conjugated antiCD4 (clone L200, BD Pharmingen). Cells were washed twice and fixed in CellFix solution (Becton Dickinson, Erembodegem, Belgium) for 3 days before analysis on a Becton Dickinson FACS apparatus with CellQuest Software (Becton Dickinson). eGFP fluorescence was detected in the isothiocyanate (FITC) channel. Negative controls from normal macaques were run with every experimental sample and were used to establish gates for eGFP quantification.

Cellular DNA was extracted from peripheral blood mononuclear cell (PBMC) samples, using the High Pure PCR Template Preparation Kit according to the manufacturer's instructions (Roche Mannheim, Germany). DNA was quantified by measuring optical density (Spectra Max 190; Molecular Devices, California, USA). The eGFP sequence was analyzed by quantitative real-time PCR on 250 ng of DNA run on an iCycler real-time thermocycler (Bio-Rad, California, USA). Primers were as follows: forward primer, 5'ACGACGGCAACTACAAGACC3'; reverse primer, 5'GCCATGATATAGACGTTGTGG3'. Amplification was performed in a final volume of 50 μl, with IQ™ SYBR^®Green Supermix (Bio-Rad, California, USA), in accordance with the manufacturer's instructions. Amplification was carried out over 40 cycles of denaturation at 95°C, annealing at 59°C and elongation at 72°C. Standard curves for the eGFP sequence were generated by serial 10-fold dilutions of duplicate samples of the eGFP plasmid in DNA from untransduced PBMC, with 250 ng of total DNA in each sample. Samples from animals were run in duplicate, and the values reported correspond to the means for replicate wells.

Paired and unpaired comparisons were performed using non parametric Kruskal Wallis, Wilcoxon rank and Mann & Whitney tests, respectively, both of which can be used for the analysis of small samples when normal distribution is uncertain or not confirmed. Tests were performed using StatView 5.01 sofware (Abacus Concepts, Berkeley, CA).

We first assessed, in vitro, the efficiency with which a SIVmac251-derived vector transduced CD34⁺ hematopoietic cells from macaque bone marrow (BM). We harvested BM cells from the iliac crests of 12 different animals. CD34⁺ cell preparations with a purity of 97% ± 1% were obtained by immunomagnetic purification. The CD34⁺ cells were then transduced by coculture for 24 h with the lentiviral vector (MOI = 100) in medium supplemented with SCF, Flt3-L, IL-3 and IL-6. The vector used (pRMES8) was derived from SIVmac251 and contains the eGFP reporter gene under control of the phosphoglycerate kinase (pGK) promoter (Figure 1). Transduction efficiency (Figure 2A and 2B), as evaluated by flow cytometry analysis of eGFP expression at 24 h, was 41% ± 9% on average (n = 12). After 24 hours of culture with the lentiviral vector, some of the purified CD34⁺ cells were cultured for 14 days in semi-solid medium containing SCF, GM-CSF, IL-3 and EPO to allow the myeloid differentiation of colony-forming cells (CFC), whereas some cells were cocultured for 35 days on a layer of murine fibroblasts of the M2-10B4 cell line and were then cultured for 14 days on semi-solid medium containing SCF, GM-CSF, IL-3 and EPO, for the identification of long-term culture-initiating cells (LTC-IC). Transduction had no effect on the clonogenic capacity of CD34⁺ cells: the mean number of colonies was 41 ± 10 for non transduced cells and 44 ± 12 for pRMES8-transduced cells (12 animals tested, P = 0.60 (Mann & Whitney test)). Similar results were obtained for LTC-IC, with 19 ± 3 colonies obtained for non transduced cells and 19 ± 3 for transduced cells (n = 12; P = 0.79 (Mann & Whitney test)). Transduction rates did not differ significantly between CFC and LTC-IC (P = 0.4884 (Wilcoxon test), n = 12), with 18% ± 7% and 19% ± 7% of colonies, respectively, eGFP-positive. However, in both cases, the percentage of eGFP-positive cells was significantly lower than that observed 24 hours after transduction (P < 0.0001 (Wilcoxon test)). This apparent discrepancy between analyses carried out at 24 h and analyses on CFC or LTC-IC may be due to the eGFP protein present in viral particles and incorporated into the cell cytoplasm during the coculture period. The proportion of cells producing eGFP shortly after transduction was reduced by 25% ± 15% (Figure 2C) if 10^-6 M AZT was added to cocultures of CD34⁺ BM cells and lentiviral vector (MOI = 100). Untreated CFC cultures gave percentages of eGFP-producing cells similar to those observed before differentiation (26% ± 5%) (Figure 2D). No fluorescence was detected after myeloid differentiation of the AZT-treated CFC (n = 3), confirming that eGFP detection resulted from the production of this protein from integrated vector.

Mosaicism was observed in eGFP gene expression in several colonies (Figure 3). Indeed, eGFP was detected in 56% ± 4% of colonies, whereas only 26% ± 5% of individual cells were eGFP-positive. These results suggest that, on average, only 47% of cells from a single colony contained the SIV vector.

We explored the capacity of autologous CD34⁺BM cells transduced ex vivo with a lentiviral vector to engraft efficiently into macaques after total body irradiation (TBI) with a gamma source at the sublethal dose of 6 Gy. Three groups of 4 animals were used: 1) In Group 1, macaque CD34⁺ BM cells (96% ± 1% pure on average) were obtained from the two humeri before gamma irradiation (Table 1). These cells were cocultured, as described above, with pGASE, which is an improved version of pRMES8. Indeed, a mean transduction efficiency of 72% ± 4% was obtained (n = 4) at 24 hours and 37% ± 10% of CFC produced eGFP. Two days after gamma irradiation, 1.4 × 10⁶ to 2.9 × 10⁶ CD34⁺ cells per kg were injected into both humeri of macaques (Table 1); 2) Group 2 included irradiated (6 Gy) macaques that did not undergo cell transplantation: 3) Group 3 included 4 non irradiated animals, which were used as controls, with a similar bleeding frequency.

Following total-body irradiation with 6 Gy, transfusion and an antibiotic regimen were required to ensure that all the animals survived. However, one animal from group 1 (7036) died on day 40 due to profound pancytopenia (Figure 4). This macaque received the smallest number of autologous and transduced CD34⁺ BM cells. All other animals from groups 1 and 2 were studied from days -1 to 471 after gamma irradiation. Controls were followed over the same period.

Radiation rapidly induced severe anemia in all animals (data not shown). A significant decrease in the number of polymorphonuclear cells in the periphery was observed, starting on day 1 after irradiation (Figure 4). No significant difference was observed between the animals of groups 1 and 2 in terms of the minimum number of cells (821 ± 226 cells/μl for group 1 and 658 ± 107 cells/μl for group 2, P = 0.3768 (Mann & Whitney test)) or the time at which that minimum occurred (6 ± 5 days for group 1 and 7 for group 2, P = 0.4795 (Mann & Whitney test)). Lymphocyte counts also decreased in all macaques by day 1 after gamma irradiation (Figure 4), falling to a minimum of 220 ± 107 lymphocytes/μl on day 18 ± 12 in group 2 and of 347 ± 62/μl on day 11 ± 12 in transplanted animals (group 1). Animals undergoing transplantation tended to display less severe lymphopenia, but no statistical difference was observed between the two groups of irradiated animals in terms of the day on which minimum lymphocyte count was reached (P = 0.1939 (Mann & Whitney test)) or the level of that minimum (P = 0.3805 (Mann & Whitney test)). A significant decrease in platelet counts, beginning by day 10 (Figure 4), was observed in all irradiated animals. Thrombocytopenia (platelet count < 20,000/μl) was characterized in non transplanted animals by a minimum value of 3.75 ± 2.49 × 10³ platelets/μl on day 18 ± 3. Thrombocytopenia tended to be less severe in transplanted animals, but this difference was not significant for the minimum number of platelets (10.33 ± 5.25 × 10³ platelets/μl; P = 0.1124 (Mann & Whitney test)) or for the day on which that minimum occurred (14.33 ± 0.94; P = 0.3123 (Mann & Whitney test)). This thrombocytopenia required one transfusion in all animals (other than animal 7036, which needed two transfusions) of both groups. However, platelet reconstitution seemed to be correlated with the dose of CD34⁺ cells infused, the speed of reconstitution increasing with the number of CD34⁺ cells injected (macaque 6653).

We determined the effects of CD34⁺ bone marrow cell transplantation following gamma irradiation on the ex vivo proliferation and differentiation of hematopoietic progenitors. Before gamma irradiation, a mean of 40 ± 9 and 38 ± 6 colonies was observed for groups 1 and 2, respectively (Figure 5). Colony number decreased significantly (P < 0.0001 (Wilcoxon test)) by day 7 in all animals. In both groups, clonogenic activity was detected by day 43 after gamma irradiation with reconstitution significantly better in the animals undergoing transplantation than in those that did not undergo transplantation (P = 0.0009 (Mann & Whitney test)).

Cells with integrated SIV-vector DNA were detected by PCR (Table 2) as early as day 3 after transplantation, in at least two animals (6653 and 6833). These two animals had received the largest numbers of transduced CD34⁺ bone marrow cells. Monkey 7036, which died within 40 days of gamma irradiation had very few transduced cells in the bone marrow and SIV-DNA was not detected in peripheral blood cells. In the three remaining animals, vector DNA was detected in peripheral blood cells (up to 500 copies per million cells) and in the bone marrow (up to 6250 copies per million cells) more than one year after transplantation (day 471).

Flow cytometry analysis demonstrated the presence of eGFP-producing cells among peripheral blood mononuclear cells in myeloid and lymphoid lineges of monkey 6896 (Figure 6). Peripheral blood cells were sorted on the basis of eGFP production, with the aim of characterizing the phenotype of populations of cells expressing the transgene in more detail. We found that 61% of eGFP-positive cells were CD11b-positive,5% of these cells appeared to be CD14+ monocytes, 14% were CD20⁺ B cells and 10% were CD3+ T cells, 23% of which expressed CD8 and 77% expressed CD4 (data not shown).