Major Subsets of Human Dendritic Cells Are Efficiently Transduced by Self-Complementary Adeno-Associated Virus Vectors 1 and 2

Abstract

Dendritic cells (DC) are antigen-presenting cells pivotal for inducing immunity or tolerance. Gene transfer into DC is an important strategy for developing immunotherapeutic approaches against infectious pathogens and cancers. One of the vectors previously described for the transduction of human monocytes or DC is the recombinant adeno-associated virus (rAAV), with a genome conventionally packaged as a single-stranded (ss) molecule. Nevertheless, its use is limited by the poor and variable transduction efficiency of DC. In this study, AAV type 1 (AAV1) and AAV2 vectors, which expressed the enhanced green fluorescent protein and were packaged as ss or self-complementary (sc) duplex strands, were used to transduce different DC subsets generated ex vivo and the immunophenotypes, states of differentiation, and functions of the subsets were carefully examined. We show here for the first time that a single exposure of monocytes (M_o) or CD34⁺ progenitors (CD34) to sc rAAV1 or sc rAAV2 leads to high transduction levels (5 to 59%) of differentiated M_o-DC, M_o-Langerhans cells (LC), CD34-LC, or CD34-plasmacytoid DC (pDC), with no impact on their phenotypes and functional maturation of these cells, compared to those of exposure to ss rAAV. Moreover, we show that all these DC subpopulations can also be efficiently transduced after commitment to their differentiation pathways. Furthermore, these DC subsets transduced with sc rAAV1 expressing a tumor antigen were potent activators of a CD8⁺-T-cell clone. Altogether, these results show the high potential of sc AAV1 and sc AAV2 vectors to transduce ex vivo conventional DC, LC, or pDC or to directly target them in vivo for the design of new DC-based immunotherapies.

Dendritic cells (DC) are antigen-presenting cells (APC) pivotal for regulating immune responses (43). They have a role in the activation and function of both innate and adaptive immune responses (2, 3). Furthermore, depending on the DC subset, their maturation stage, and the inflammatory stimuli that they encounter, DC can induce either immunity or tolerance (28, 44). At least three subsets of mature human DC are known: myeloid conventional DC (cDC), Langerhans cells (LC), and plasmacytoid DC (pDC), with evidence of functional plasticity for a given DC subset (36). Indeed, the subsets exhibit some differences in their abilities to regulate T-cell responses, to produce antiviral type I interferon, and to cross-present exogenous antigens to CD8⁺ T cells.

Human DC, LC, and pDC can be easily generated from monocytes or CD34⁺ progenitors, allowing the procurement, in considerable amounts, of these otherwise scarce cells, which represent approximately 0.2% of total white blood cells (5-8, 19, 21, 22, 32, 37, 39). The genetic modification of DC subpopulations offers great potential for the development of immune response-regulatory strategies for purposes ranging from active immunization to tolerance induction. This approach may provide long-lasting expression of the entire array of epitopes directly processed from a native protein and potentially allow their presentation by the complete set of human leukocyte antigen (HLA) molecules present on the APC.

Efficient gene delivery to human DC has been achieved mainly by using lentiviral vectors (LV) (40, 42, 48). One of the alternate vectors used to transduce monocytes or DC has been the recombinant adeno-associated virus (rAAV), with a genome conventionally packaged as single-stranded (ss) molecules (9, 27, 34), despite the fact that the cloning capacity of this vector is more limited than that of LV (4.7 kb instead of 8 to 9 kb) (13, 50). Indeed, rAAV is unique among viruses that are being developed for gene transfer in that the wild-type virus has never been shown to cause human disease. Like LV, rAAV presents the ability to transduce both dividing and nondividing cells, which may allow the transduction of DC from a broad range of sources and in various activation and maturation states (13, 50). Another advantage of rAAV is the absence of viral coding sequences, which may diminish the elimination of transduced DC by virus-specific cytolytic T cells (26, 55). So far, preclinical studies based on DC immunotherapies have been limited by the poor and variable levels of efficiency of DC transduction with ss rAAV (26, 34, 52).

Recently, AAV vectors have been diversified by alternate serotypes with different cell tropisms (51). It was recently shown that the ss AAV type 5 (ss AAV5) vector presents a higher degree of tropism for human DC than vectors of other serotypes, which had poor levels of efficiency (54). Moreover, some studies have demonstrated that the packaging of rAAV2 as a self-complementary duplex strand (sc) allows earlier and higher levels of transgene expression by bypassing the synthesis of the complementary strand of the AAV viral genome, which constitutes a limiting step for efficient cell transduction (30, 49). A recent study reported the efficient transduction of mouse bone marrow-derived DC with sc rAAV6 (1).

In this study, we compared the levels of efficiency and the time courses of transduction of human DC subsets (cDC, LC, and pDC) with ss rAAV1 and ss rAAV2 or with their self-complementary counterparts. Furthermore, we carefully examined the immunophenotypes, states of differentiation, and functionalities of these transduced DC subsets. We found that in contrast to exposure to ss rAAV, a single exposure of monocytes or CD34⁺ progenitors to sc rAAV1 or sc rAAV2 enables high levels of transduction of cDC, LC, or pDC, with no impact on their phenotypes and functional maturation. We further showed that all these DC subsets can be efficiently transduced after commitment to their differentiation pathways. Finally, our results indicate that these cDC, LC, and pDC derived from monocytes or CD34⁺ progenitors are potent activators of a specific CD8⁺-T-cell clone after transduction with sc AAV1 vector expressing a tumor antigen. Altogether, these results show the strong potential of sc AAV1 and sc AAV2 vectors for cDC-, LC-, or pDC-based immunotherapy applications.

MATERIALS AND METHODS

AAV vector construction and production.

Pseudotyped AAV vectors were generated by packaging AAV2-based recombinant genomes in AAV1 or AAV2 capsids. All the vectors used in the study were produced using the three-plasmid transfection protocol as described elsewhere (38). Briefly, HEK293 cells were tritransfected with the adenovirus helper plasmid pXX6 (53), a pAAV packaging plasmid expressing the rep and cap genes (pACG2.1 for AAV2 and pLT-RC02 for AAV1), and the relevant pAAV2 vector plasmid. ss AAV vectors were produced with the conventional pGG2 AAV2 vector plasmid expressing enhanced green fluorescent protein (E-GFP) under the transcriptional control of the cytomegalovirus (CMV) immediate early promoter associated with a simian virus 40 poly(A) signal. To produce sc AAV vectors, a modified pGG2-scCMV-GFP plasmid was constructed by deleting the D sequence and the terminal resolution site (trs) from one of the inverted terminal repeats by using MscI digestion, leaving the other repeat intact (49). To generate self-complementary vectors expressing the MART-1 peptide, the CMV GFP sequence of pGG2-scCMV-GFP was replaced with the MART sequence (ELAGIGLTV) (47) coupled with the phosphoglycerate kinase promoter. Recombinant vectors were purified by double CsCl₂ ultracentrifugation followed by dialysis against sterile phosphate-buffered saline. Physical particles were quantified by real-time PCR, and vector titers are expressed as viral genomes (vg) per milliliter. The proportion of sc AAV vector genomes was assessed by alkaline agarose electrophoresis.

Cell line.

The OP9 stroma cell line expressing human delta 1 (OP9-Del1) was kindly provided by A. Olivier (Genethon, Evry, France) and maintained as previously described (32).

Culture of peripheral blood monocytes and CD34⁺ cells.

Each DC subset was differentiated from monocytes or CD34⁺ progenitors from at least three different donors. Monocytes were generated from samples from normal volunteers after the elutriation of peripheral blood according to the Établissement Français du Sang procedures (J. Bernard, Reims, France). This method yielded purified (92.2% ± 5.1%) CD14⁺ CD45⁺ cells as assessed by flow cytometry. Briefly, cryopreserved monocytes were cultured in 6-well plates at a density of 10⁶ cells/ml in RPMI 1640 (Invitrogen Life Technology, Auckland, New Zealand) supplemented with 10% fetal calf serum (FCS; HyClone, Logan, UT) and 1% l-glutamine (Invitrogen). Monocytes were differentiated either into cDC (Mo-DC) in the presence of 50 ng/ml of recombinant human granulocyte-macrophage colony-stimulating factor (Novartis, Bâle, Switzerland) and 15 ng/ml of recombinant human interleukin-4 (IL-4; Tebu-bio, Le Perray, France) or into LC (Mo-LC) by the addition of 20 ng/ml of recombinant human transforming growth factor β (Abcys SA, Paris, France) (21). Maturation was induced in some experiments by the addition of lipopolysaccharide (LPS; 7 μg/ml [Sigma-Aldrich, St. Louis, MO]) at day 8 for 24 h (Fig. 1A).

FIG. 1.

Graphic depiction of the experimental protocols. d, day; CD34⁺, CD34⁺ progenitors; +, with; −, without.

LC were generated from cord blood CD34⁺ cells (CD34-LC) by following protocols previously described by Caux et al. (6) with some modifications. Briefly, CD34⁺ progenitor cells were separated from cord blood mononuclear cells by using the CD34 isolation kit (Miltenyi Biotech, GmbH, Bergisch Gladbach, Germany). Cryopreserved CD34⁺ progenitor cells were seeded at a density of 2 × 10⁵ cells/ml in 24-well tissue culture plates in RPMI 1640 supplemented with 10% FCS, 1% l-glutamine (GIBCO/BRL, Rockville, MD), 50 ng/ml of recombinant human granulocyte-macrophage colony-stimulating factor, 15 ng/ml of recombinant human IL-4, 15 ng/ml of recombinant human transforming growth factor β, and 20 ng/ml of recombinant human tumor necrosis factor alpha. At day 6, cells were collected and plated in new 12-well plates at a density of 8 × 10⁵ cells/ml. In some experiments, the maturation of CD34-LC was induced by the addition of LPS (7 μg/ml) at day 12 for 24 h (Fig. 1B).

pDC were generated from cord blood CD34⁺ cells (CD34-pDC) by following protocols previously described by Olivier et al. (32). CD34⁺ progenitors (2 × 10⁴) were added to OP9-Del1 cells seeded 1 day before in 24-well plates at 3 × 10⁴ cells/well. Cells were cultured in RPMI 1640 (Invitrogen) supplemented with 10% FCS (HyClone), 1% l-glutamine, and 1% penicillin-streptomycin (GIBCO) in the presence of recombinant human Fms-like tyrosine kinase 3 ligand (5 ng/ml) and rIL-7 (5 ng/ml; R&D Systems, Minneapolis, MN). The maturation of CD34-pDC was induced in some experiments by the addition of CpG oligodeoxynucleotide type A (ODN 2216 at 2 μM) at day 10 for 24 h (Fig. 1C). All cells were cultured in a humidified incubator at 37°C and 5% CO₂.

Flow cytometric analysis.

The DC and T-cell phenotypes were assessed using three-color immunostaining with biotinylated phycoerythrin-, Cy-Chrome-, and allophycocyanin-conjugated monoclonal anti-CD86 (FUN-1), anti-CD11c (B-Ly6), anti-HLA-DR (G46.6), anti-Langerin (DCGM4), anti-cutaneous leukocyte antigen (anti-CLA; HECA-452), anti-CD8 (RPA-T8), and anti-CD3 (HIT3a; all purchased from Becton-Dickinson, Mountain View, CA) and anti-CD123 (AC145; Miltenyi Biotech). Data were acquired using a FACSCalibur flow cytometer (Becton Dickinson), and data analyses were performed using the CellQuest program (Becton Dickinson).

Transduction of monocytes and CD34⁺ cells with rAAV.

After thawing, monocytes and CD34⁺ cells were transduced with 10⁵ vg/cell of ss and sc rAAV types 1 and 2 at a fixed concentration of 1 × 10⁷ and 1 × 10⁶ cells/ml, respectively, in RPMI 1640. After 3 h at 37°C, cells were cultured in complete medium as previously described and were analyzed by flow cytometry at different times, as indicated in Fig. 1.

Transduction of differentiated Mo-DC, Mo-LC, CD34-LC, and CD34-pDC with rAAV.

Monocytes and CD34⁺ progenitors were differentiated as described above into Mo-DC, Mo-LC, CD34-LC, or CD34-pDC. Semiadherent and nonadherent cells in culture were harvested at different times as depicted in Fig. 1 and transduced with ss and sc rAAV types 1 and 2 under the same conditions described above for monocytes and CD34⁺ cells. After 3 h at 37°C, cells were re-placed into the same complete medium and then cultured for 5 to 6 additional days (Fig. 1).

MLRs.

Enriched naïve CD2⁺ CD45⁺ T cells were recovered after the elutriation of monocytes. This method yielded purified (82.4% ± 6.1%) CD2⁺ CD45⁺ cells as assessed by flow cytometry. CD2⁺ CD45⁺ T cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) at a final concentration of 0.5 μM for 20 min at 37°C before being extensively washed. E-GFP-negative and -positive Mo-DC and Mo-LC were sorted on a MoFlow flow cytometer (Dako, Glostrup, Denmark). For mixed leukocyte reactions (MLRs), matured allogeneic Mo-DC and Mo-LC (Fig. 1A) were extensively washed and cultured in 96-well U-bottom plates at different densities with 1 × 10⁵ CFSE-labeled CD2⁺ CD45⁺ T cells. On day 4, cells were harvested, washed, labeled for T specificity with anti-CD3 antibody, and analyzed by flow cytometry. The percentage of dividing T cells was linearly correlated with the decrease in CFSE fluorescence.

Activation of a MART-1 CD8⁺-T-cell clone by transduced DC subpopulations.

Matured HLA-A2⁺ DC subpopulations were obtained after the transduction of cells with sc rAAV1 expressing a MART-1 peptide under the control of a CMV promoter, as depicted in Fig. 1. Nontransduced matured HLA-A2⁺ DC subpopulations were pulsed with the high-affinity MART-1 peptide (ELAGIGLTV) at 10 μM for 3 h and then washed extensively before coculture. These nontransduced, transduced, or pulsed DC subpopulations were cocultured in 96-well U-bottom plates at different ratios with a specific MART-1 HLA-A2-restricted CFSE-labeled CD8⁺-T-cell clone (LT12) at 10⁵ cells/well as described above for MLRs. On day 5, cells were harvested, washed, labeled with an anti-CD8 antibody, and analyzed by flow cytometry. The percentage of dividing T cells was linearly correlated with the loss of CFSE fluorescence.

Statistical analyses.

Results are presented as means ± standard deviations (SD). Student's t test for paired data was used to determine significant differences between two groups. A P value of <0.05 was considered statistically significant.

RESULTS

Transduction of monocytes and CD34⁺ progenitors with ss and sc rAAV1 and rAAV2.

We first compared the gene transfer efficiency with ss and sc rAAV1 and rAAV2 (i) into monocytes that were then differentiated into cDC or LC and (ii) into CD34⁺ progenitors that were further differentiated into LC or pDC with the appropriate cytokine cocktails as previously described (32, 48). These vectors expressing E-GFP driven by the ubiquitous CMV promoter were used under identical conditions. E-GFP expression is easily and accurately monitored by fluorescence-activated cell sorter analysis (FACS), allowing the enumeration of DC subpopulations on a single-cell basis. Preliminary experiments performed with different amounts of viral particles (5 × 10³ to 1.5 × 10⁴ vg/cell) at a fixed cell density showed that maximum transduction levels were reached with 10⁴ vg/cell, with no cellular toxicity (data not shown). E-GFP expression and FACS analysis for DC subset identification were monitored 8 days posttransduction for Mo-DC and Mo-LC and 5 to 6 days posttransduction for CD34-LC and CD34-pDC, since the latter are dividing cells and AAV vectors are mainly episomal (Fig. 1). Mo-DC were analyzed for CD11c expression, Mo-LC and CD34-LC were analyzed for CLA expression, and CD34-pDC were analyzed for CD123 expression (32, 48). A single exposure of both monocytes and CD34⁺ progenitors to ss rAAV1 or rAAV2 led to low levels of transduced Mo-DC, Mo-LC, CD34-LC, and CD34-pDC, with percentages of E-GFP-positive cells ranging from 1 to 8% (Fig. 2). In contrast, the transduction efficiencies of each DC subpopulation with sc rAAV under similar conditions were significantly increased: sc rAAV1 was 10-fold, sixfold, fourfold, and fourfold more efficient than its ss counterpart for Mo-DC, Mo-LC, CD34-LC, and CD34-pDC, respectively. Similarly, sc rAAV2 led to transduction levels 4.5-fold, 4.7-fold, 6.7-fold, and 6.6-fold higher than ss rAAV2 for Mo-DC, Mo-LC, CD34-LC, and CD34-pDC, respectively (Fig. 2). Of note, Mo-LC and CD34-LC, which were described as more immature DC (35), were more efficiently transduced with both serotypes of sc rAAV than Mo-DC and CD34-pDC (Fig. 2). Altogether, these results show that sc rAAV vectors of serotypes 1 and 2 (i) are very efficient for the transduction of monocytes and CD34⁺ progenitors compared to their ss counterparts and (ii) equivalent in transducing both cell types, further allowing for the generation of gene-modified cDC, LC, or pDC.

FIG. 2.

Transduction efficiencies of monocytes and CD34⁺ progenitors with ss and sc rAAV1 and rAAV2 serotypes. Monocytes or CD34⁺ progenitors were transduced with ss and sc rAAV1 and rAAV2 serotypes expressing E-GFP and then differentiated into cDC, LC, or pDC in the presence of the corresponding cytokine cocktail. Mo-DC, Mo-LC, CD34-LC, and CD34-pDC were analyzed 5 to 8 days posttransduction, as indicated in Fig. 1. For each subpopulation, the graphs show the percentages of E-GFP-expressing cells after transduction with the indicated vectors or without transduction (NTd). The percentages of E-GFP-expressing cells were obtained by flow cytometry. Cells were gated in forward scatter and side scatter; cDC, LC, and pDC were analyzed for E-GFP expression and markers characteristic of each DC subset, CD11c, CLA, and CD123, respectively. Results are expressed as the mean percentages of cells ± SD as determined in the numbers of independent experiments indicated (n). Note that the P values indicated are always greater than 0.05, which represents a nonsignificant difference between the considered groups.

Effects of DC differentiation on the efficiency of transduction with ss and sc rAAV1 and rAAV2.

We evaluated whether cells already committed toward cDC, LC, or pDC could also be transduced with sc and ss rAAV1 and rAAV2 expressing E-GFP. Committed cells were transduced 3 or 6 days after the induction of differentiation (Mo-DC and Mo-LC), 6 or 12 days after the induction of differentiation (CD34-LC), and 5 or 10 days after the induction of differentiation (CD34-pDC) and then analyzed 5 to 6 days posttransduction by flow cytometry to detect E-GFP expression (Fig. 1). Day-3 and -6 Mo-DC and Mo-LC were efficiently transduced with sc rAAV1 and rAAV2 at levels similar to those at day 0 (6 to 23%). In contrast, no transduction with the ss counterpart vectors was seen (Fig. 3). As observed at day 0, cells committed toward LC were more efficiently transduced with both vectors than cells committed toward DC. Levels LC transduction were 1.5- and 2.5-fold higher than those DC when transduced with sc rAAV1 and sc rAAV2, respectively, 3 days after the induction of differentiation. These differences reached 1.9- and 3.5-fold when the transduction occurred 6 days after the induction of differentiation (Fig. 3). In contrast to monocyte-derived cells, CD34⁺ progenitor-derived cells could still be transduced with both ss rAAV serotypes after differentiation, although with very low effectiveness (less than 9%). Furthermore, CD34-LC were efficiently transduced 6 days after the induction of differentiation but not at 12 days and CD34-pDC were efficiently transduced at both 5 and 10 days with both sc serotypes (14 to 56%) after commitment to their differentiation pathways, like CD34⁺ progenitors (Fig. 3). In addition, day-6 CD34-LC were 1.8- and 1.4-fold more efficiently transduced with sc rAAV1 and rAAV2, respectively, than CD34-pDC (Fig. 3). These results indicate that monocytes and CD34⁺ progenitors already committed toward cDC, LC, or pDC can be efficiently transduced by a single infection with sc rAAV1 and sc rAAV2 serotypes.

FIG. 3.

Transduction efficiencies of DC subpopulations with ss and sc rAAV1 and rAAV2 serotypes. Mo-DC and Mo-LC were transduced at day 0 (D0) or at day 3 or 6 after the induction of differentiation. CD34-LC and CD34-pDC were transduced at day 0, 6, or 12 and at day 0, 5, or 10 after the induction of differentiation, respectively (Fig. 1). Cells were transduced with ss or sc rAAV and rAAV2 and then cultured in the same medium in the presence of the appropriate cytokine differentiation cocktail. Cells in final cultures were analyzed for E-GFP expression by flow cytometry 5 to 6 days posttransduction. The percentage of E-GFP-expressing cells in each DC subpopulation was obtained as described in the legend to Fig. 2 (n = 3). Note that the P values indicated are always greater 0.05, which represents a nonsignificant difference between the considered groups.

Yields and immunophenotypical analysis of sc rAAV-transduced cDC, LC, and pDC.

In order to better understand the mechanisms involved in the immunogenicity of sc AAV vectors, it is interesting to investigate the effects of these vectors on DC phenotypes and maturation states. We characterized immunophenotypically the different nontransduced and sc rAAV-transduced DC subpopulations generated either from monocytes or from CD34⁺ progenitors. Nontransduced monocytes were differentiated into either cDC or LC, and final yields of total cells were 29% ± 10% or 32% ± 6% of the original cell populations, respectively (Fig. 4 A and B). The phenotypical analysis showed that more than 95% of the cells were Mo-DC, with a homogeneous population expressing high levels of HLA-DR and CD11c and low levels of the costimulatory molecule CD86 (Fig. 5A and B). More than 87% of the cells were Mo-LC, with homogeneous expression of CLA, heterogeneous expression of HLA-DR, and low expression of CD86 and Langerin (Fig. 5A and B). The transduction of nondividing monocytes with sc AAV vectors at day 0 did not significantly modify the yield of Mo-DC or Mo-LC (Fig. 4A and B) or their phenotypes or maturation stages (Fig. 5A and B) at the end of the culture period. Furthermore, an upregulation of the CD86 costimulatory marker was observed when sc rAAV-transduced Mo-DC and Mo-LC were activated by LPS for 24 h, showing that their response to LPS was unaltered (Fig. 5C).

FIG. 4.

Cell culture characterization of sc rAAV-transduced DC subpopulations. (A and B) Yields of total Mo-DC or Mo-LC that were nontransduced (NTd) or transduced at day 0 (Td d0) with sc rAAV1 or sc rAAV2 were measured at day 8. (C and D) The proliferation index of total CD34-LC that were nontransduced or transduced at day 0 or day 6 (Td d6) with sc rAAV1 or sc rAAV2 were determined at day 12. The proliferation index of total CD34-pDC that were nontransduced (NTd) or transduced at day 0 (Td d0) or day 5 (Td d5) with sc rAAV1 or sc rAAV2 were determined at day 10. Yields and proliferation index were determined by cell counting and flow cytometry analysis. Results are expressed as the mean percentage of cells or the index of cell proliferation ± SD determined in the numbers of independent experiments indicated (n).

FIG. 5.

Immunophenotypes of sc rAAV-transduced DC subpopulations. Phenotypes of Mo-DC and Mo-LC at day 8, after transduction with sc rAAV at day 0, and phenotypes of CD34-LC and CD34-pDC at days 12 and 10, after transduction with sc rAAV at days 6 and 5, respectively. (A) Flow cytometry analysis was performed to monitor the expression of markers characteristic of cDC, LC, and pDC. Cells gated in forward scatter-side scatter parameters were analyzed by double staining for the expression of HLA-DR and CD11c for Mo-DC, CLA and Langerin for Mo-LC and CD34-LC, and HLA-DR and CD123 for CD34-pDC. Values indicate the percentages of cells in the corresponding quadrant. Overlay histograms show the expression of relevant antigens for nontransduced cells (thin black lines), total transduced cells (thick black lines), and E-GFP-positive gated cells (green lines) versus those for isotype-matched controls (dotted lines). (B) Comparative phenotypes of nontransduced and transduced DC subpopulations. Overlay histograms show the expression of the CD86 costimulatory molecule for nontransduced (thin black lines), total transduced (thick black lines), and E-GFP-positive gated (green lines) DC subpopulations versus those for isotype-matched controls (dotted lines). (C) Comparative phenotypes of transduced DC in the absence or presence of LPS or CpG for 24 h. Overlay histograms show the levels of expression of CD86 by immature (thin black lines) and total mature (thick black lines) and E-GFP-positive gated mature (green lines) transduced DC subpopulations versus those of isotype-matched controls (dotted lines). The results are representative of those from at least three experiments.

CD34⁺-derived cells were transduced at day 0, 5, or 6 and analyzed at the end of the culture period (Fig. 1). Results are expressed as the index of proliferation instead of yields, since CD34⁺ progenitors are proliferating cells. The proliferation index for total nontransduced cells in CD34-LC and CD34-pDC cultures were 8% ± 1% and 3.7% ± 1% (Fig. 4C and D). The phenotypical analysis of CD34-LC transduced at day 6 showed that more than 77% of cells were LC, with the expression of HLA-DR and CLA and a low level of expression of CD86, and the phenotypical analysis of CD34-pDC transduced at day 5 showed that more than 30% of cells were pDC, with the expression of HLA-DR, a high level of expression of CD123, and homogeneous expression of CD86 (Fig. 5A and B). It should be noted that the proliferation index for CD34-LC decreased slightly when cells were transduced with both sc rAAV serotypes at day 0 or at day 6 (Fig. 4C). The phenotypes and maturation states of CD34-LC and CD34-pDC were not significantly changed when transduced with either sc rAAV vector (Fig. 5A and B). As previously, we observed an upregulation of the CD86 costimulatory marker when transduced CD34-LC and CD34-pDC were activated by LPS and CpG, respectively (Fig. 5C). Transduction of DC at later time points in the differentiation pathways did not affect the yields or phenotypes of these cells (data not shown). Lastly, infection with ss rAAV vectors was poorly efficient and did not affect the yields and the phenotypes of the DC subpopulations (data not shown).

Altogether, our results indicate that sc rAAV1 and sc rAAV2 can efficiently transduce cells already committed to the different DC differentiation pathways without inducing phenotypic and maturation changes.

Functional properties of DC subpopulations transduced with sc rAAV1 and rAAV2.

We evaluated the ability of transduced monocyte-derived cells to stimulate allogeneic T cells in MLRs. Monocytes transduced with sc rAAV1 or sc rAAV2 expressing E-GFP were differentiated into cDC or LC as previously described, matured in LPS for 24 h, and then sorted by flow cytometry on the basis of E-GFP expression. Nontransduced, E-GFP-negative and -positive sorted cDC and LC were used to stimulate CFSE-labeled allogeneic T cells. As shown in Fig. 6A, the E-GFP-negative and -positive Mo-DC and Mo-LC displayed allostimulatory capacities comparable with those of nontransduced DC. The capacities of the transduced DC subpopulations to activate a CD8⁺-T-cell clone were then evaluated. Nonproliferating HLA-A2⁺ monocytes transduced at day 0 were differentiated into DC or LC, whereas proliferating HLA-A2⁺ CD34-LC or CD34-pDC were transduced at day 6 with sc rAAV1 expressing the MART-1 peptide. The different DC subpopulations obtained were efficient in activating a specific CD8⁺-T-cell clone (Fig. 6B). The capacities of the different transduced DC subpopulations to secrete characteristic cytokines, such as IL-12, IL-10, and tumor necrosis factor alpha for cDC and LC and alpha interferon for pDC, were then assessed. For the four subpopulations analyzed, no change in the secretion profiles compared to those of the nontransduced cells was seen (data not shown).

FIG. 6.

Functions of sc rAAV-transduced DC subpopulations. (A) The alloantigen presentation capacities of Mo-DC and Mo-LC were assessed 8 days posttransduction after maturation in LPS for 24 h and cell sorting on an E-GFP expression basis. Total nontransduced cells (Ntd) and E-GFP-negative and E-GFP-positive sorted Mo-DC (left graphs) and Mo-LC (right graphs), transduced with either sc rAAV1 (upper graphs) or sc rAAV2 (lower graphs), were incubated with allogeneic T cells stained with CFSE. After 5 days of coculture, percentages of CD8⁺ dividing T cells measured by flow cytometry were linearly correlated with the loss of CFSE fluorescence. The data are shown as the means ± SD of triplicate results and represent one out of three independent experiments. (B) In vitro antigen presentation capacities of sc rAAV1-transduced HLA-A2 DC subpopulations. Cells were transduced with sc rAAV1 expressing the MART-1 peptide, as depicted in Fig. 1, or pulsed for 3 h with the peptide the day of the coculture. Then mature nontransduced (NTd), transduced (Td), or peptide-pulsed (Pulsed) DC were cocultured with the HLA-A2-restricted CD8⁺-T-cell clone specific for the MART-1 peptide (LT12) stained with CFSE. After 5 days of coculture, percentages of CD8⁺ dividing T cells measured by flow cytometry were linearly correlated with the loss of CFSE fluorescence. The data are representative the results of one of two independent experiments.

Altogether, these results indicate that the functional properties of Mo-DC, Mo-LC, CD34-LC, and CD34-pDC are not altered by transduction with sc AAV1 and AAV2 vectors. Furthermore, all these transduced DC were able to activate a CD8⁺-T-cell clone.

DISCUSSION

DC are the professional APC of the immune system, fully equipped to initiate primary immune responses, including tolerogenic responses (3, 43, 44). For these reasons, DC are an attractive target for genetic manipulation of the immune system to increase otherwise insufficient immune responses in cases of infectious disease and cancer. Immunization with ex vivo-generated DC has proven feasible and permits the enhancement as well as the dampening of antigen-specific immune responses in humans. Nevertheless, DC-based vaccines have yet to be improved since clinical responses are rarely complete and long lasting (33). Therefore, a future vaccine generated ex vivo will probably be heterogeneous and composed of several DC subsets, each presenting specific targeting to a given immune effector.

Direct DC transduction with antigen-expressing vectors offers potential advantages over regular peptide- or protein-loading protocols: (i) it ensures long-lasting expression of the antigen and production of an entire array of epitopes presented by the autologous HLA molecules, and (ii) antigens are delivered to both endogenous major histocompatibility complex class I and class II antigen presentation pathways (2, 23).

In this study, we investigated the potential of AAV1 and AAV2 vectors, with the conventional ss genome or sc genome, to transduce human Mo-DC, Mo-LC, CD34-LC, and CD34-pDC. We showed that a single exposure of monocytes or CD34⁺ progenitors to sc rAAV1 or sc rAAV2 results in high levels of transduced Mo-DC, Mo-LC, CD34-LC, or CD34-pDC (5 to 59%) without impacting the phenotypes, functions, and abilities of these cells to respond to inflammatory signals. We previously showed that cells committed toward LC, but not DC, can be efficiently transduced with LV without any facilitating agents such as Polybrene (48). Here, we showed that not only LC but also cDC and pDC can be efficiently transduced with sc AAV1 and sc AAV2 vectors after commitment to their differentiation pathways, offering a great advantage over transduction with LV. Furthermore, it appears clear that even though most rAAV genomes remain episomal (11, 31, 41), they permit long-term persistence and stable transgene expression (12, 25, 56) in nondividing cells. Gene delivery to human DC has also been achieved with other viral vectors such as adenovirus- and vaccinia virus-derived vectors, but the applicability of these vectors in vivo is rather limited since most humans have preexisting immunity to these viruses (24). The preexistence of humoral immunity to some wild-type AAV in humans has been described previously (10, 16). Nevertheless, rAAV may overcome these limitations due to the development of alternative serotypes which are derived from a variety of human and nonhuman wild-type AAVs (18, 38).

We further showed that 8-day posttransduction Mo-DC and Mo-LC transduced at day 0 of the induction of differentiation and 12-day posttransduction CD34-LC and CD34-pDC transduced at day 5 or 6 with sc rAAV1 were potent activators of a CD8⁺-T-cell clone specific for the tumor antigen MART-1. Our results showed that sc rAAV1 and sc rAAV2 are efficient tools for ex vivo DC-based immunotherapy. Furthermore, these ex vivo evaluations suggest that these vectors may also be good candidates for the in vivo targeting of specific DC subsets. This finding would be an important step toward the development of a DC-based immunotherapy where the induction of a given type of immune response is required. For this purpose, the identification of promoters driving the expression of specific DC subset cell surface molecules, such as DC-specific intercellular adhesion molecule-3-grabbing nonintegrin expressed by myeloid cDC (20), Langerin expressed by LC (45, 46), and blood DC antigen 2 expressed by pDC (14, 15), will allow specific targeting of each of these DC subsets. One of the limitations of sc rAAV is the limited packaging capacity, which cannot exceed 2.3 kb. However, given the fact that for immunization purposes, only a fraction of the coding sequence can be used to generate an immunoreactive epitope, this restricted of the expression cassette should not be a limitation.

Other reports have described the efficient transduction of human Mo-DC with ss rAAV2, but large numbers of vectors or repeated exposures to the vectors were necessary (9, 27, 34). We found here that monocytes, CD34⁺ progenitors, and cells committed to differentiation into one of the DC subsets were very poorly transduced by using a single exposure to ss rAAV1 and ss rAAV2 (less than 9%), suggesting, among other explanations such as tropism, that the second-strand synthesis may be a limiting step for transgene expression in human DC subsets. Several studies have reported that the potential benefit of sc rAAV is the achievement of earlier and higher-level transgene expression in various cell types, such as muscle, brain, and liver cells, related to a bypass of the complementary strand synthesis (1, 17, 30, 49). In our experiments, we did not observe significant transduction levels with ss rAAV1 and ss rAAV2 even when a longer time for expression was allowed (data not shown). Our results clearly demonstrate that the major step limiting DC transduction with AAV vectors is the conversion of the single-stranded viral genome.

To date, gene therapy clinical trials performed to correct monogenic diseases such as cystic fibrosis and hemophilia B have been based on conventional ss AAV vectors (4). Results have shown the safety of these vectors but also have been somewhat disappointing in terms of efficacy. Furthermore, the immune response against the AAV capsid imposes some limitations (29). The use of sc AAV vectors represents an opportunity to increase gene transfer efficiency by providing transcriptionally active recombinant viral genomes to the cell. However, using sc rAAV for in vivo gene therapy applications may increase the risk of immune responses against the transgene product through the transduction of DC. To counteract this highly unwanted consequence, the use of promoters specific to the targeted cells will be an absolute requirement.

Altogether, we show here for the first time that sc AAV type 1 and type 2 vectors allow for the efficient transduction of monocytes, CD34 progenitors, and their derivative DC subsets, before and after commitment to their differentiation pathways, without alteration of their phenotypes and functional maturation capabilities. Furthermore, sc rAAV-transduced DC subsets are able to present a specific peptide and to activate a CD8⁺-T-cell clone, demonstrating the potential use of this approach for a DC based-immunotherapy aiming at inducing active immunization or peripheral tolerance. Finally, these results also suggest that sc AAV vectors can be of high value for in vivo strategies in which a specific DC subset must be targeted.