Human CD34-Derived Myeloid Dendritic Cell Development Requires Intact Phosphatidylinositol 3-Kinase–Protein Kinase B–Mammalian Target of Rapamycin Signaling

Dendritic cells (DCs) are professional APCs that play a crucial role in immunity, ranging from tolerogenic to immunogenic responses (1). The type of the immune response is determined by the DC maturation state, the life span of the Ag-bearing DC, and the DC subtype (2–5). Based on surface markers, localization, functional abilities, and ontogeny, a large variety of DC subsets can be recognized. The myeloid migratory DC subtypes, including Langerhans cells (LCs) and interstitial DCs (intDCs), are known for their superior T cell priming ability compared with, for example, plasmacytoid DCs.

DCs develop from CD34⁺ hematopoietic progenitor cells (HPCs) via subset-specific precursors. These precursors are either present as direct precursor DCs (pre-DCs) or as immune effector cells with their own function in immunity, such as monocytes, giving rise to DCs when necessary (6, 7). Whereas the cytokines responsible for the differentiation and survival of DCs are becoming increasingly clear (7), the molecular mechanisms regulating DC development remain poorly understood. Furthermore, the majority of studies examining the role of specific signaling pathways and transcription factors have focused only on DC development in mice (7).

A likely candidate for the regulation of DC development is phosphatidylinositol 3-kinase (PI3K), which regulates multiple cellular processes in a variety of hematopoietic cells, including erythrocytes, neutrophils, eosinophils, and T and B lymphocytes (8–10). Signaling downstream of activated PI3K includes the activation of protein kinase B (PKB; also called c-Akt) and regulation of several downstream substrates, such as mammalian target of rapamycin (mTOR), glycogen synthase kinase-3β, and Forkhead box O transcription factors (10, 11). In myeloid DCs, PI3K signaling has been shown to be involved in the regulation of phenotypic and functional maturation as well as activation-induced survival (12–16). However, the importance of PI3K and its downstream substrates for human myeloid DC development is still largely undefined, and the few available studies showing effects on differentiation or yields are mainly limited to monocyte-derived DCs (moDCs) (17–19).

DC differentiation from monocytes is the most widely used model to study DC biology (20). To investigate the development of myeloid DC subsets from a less committed progenitor, CD34⁺ HPC can be used. Upon culture with GM-CSF, TNF-α, and stem cell factor (SCF), the two functionally different myeloid subsets, intDC and LC, develop via independent pathways and can therefore be studied separately (21, 22). The differentiation in two phases, with pre-DCs developing prior to terminal differentiation, enables detailed analysis of CD34-derived myeloid DC development by this in vitro system. In addition, CD34-derived DC development can also be studied in vivo, by xenotransplantation of human HPCs into immune-deficient mice (23–25).

In this study, we investigated the importance of PI3K signaling for the development of functional intDCs and LCs from human cord blood CD34⁺ HPC. PI3K-induced mTOR activation, mediated by PKB, was shown to be crucial for human myeloid CD34-derived DC development in vitro and in vivo. Precursor proliferation and survival strongly depended on this signaling module. Although inhibition of this pathway did not affect the acquisition of a DC phenotype, DCs generated under PI3K or mTOR inhibition were functionally impaired. In contrast to the requirement for PI3K signaling during development, the survival of terminally differentiated CD34-derived myeloid DCs was not dependent on PI3K or mTOR activity. However, this signaling module was still required for other important processes associated with DC function. These findings demonstrate the importance of the PI3K–PKB–mTOR signaling module during CD34-derived myeloid DC development, providing possibilities to manipulate DC subset distribution and function to regulate immunity.

Umbilical cord blood samples were obtained ex utero according to legal guidelines. CD34⁺ hematopoietic progenitor cells were isolated and cultured as described previously (22, 26). CD34⁺ cells were isolated from mononuclear fractions through positive selection using anti-CD34–coated microbeads and MS separation columns (Miltenyi Biotec, Bergish Gladbach, Germany) to a purity of 85–98%. After cryopreservation to standardize differentiation, cells were cultured in complete medium containing RPMI 1640 (Invitrogen, Breda, The Netherlands) supplemented with 8% heat inactivated FCS (Hyclone; Thermo Fisher Scientific, Etten-Leur, The Netherlands), 10 mM Hepes (Invitrogen), 2 mM l-glutamine (Invitrogen), 50 μM β-mercaptoethanol (Merck, Darmstadt, Germany), and penicillin/streptomycin (Invitrogen). From day 0 to 6, the cells were cultured in complete medium supplemented with 100 ng/ml GM-CSF (Schering-Plough, Houten, The Netherlands), 25 ng/ml SCF (PeproTech, London, U.K.), 2.5 ng/ml TNF-α (R&D Systems, Abingdon, U.K.) and 5% heat inactivated AB⁺ pooled human serum (four different donors; Sanquin, Rotterdam, The Netherlands). Then, cells were harvested, washed and further cultured in complete medium containing only GM-CSF (100 ng/ml). Where indicated, LY294002 (LY; 2 μM unless indicated otherwise; Biomol, Plymouth Meeting, PA) or Rapamycin (Rapa; 20 nM unless indicated otherwise; Biomol) was added to the cultures.

Terminally differentiated DCs were activated by addition of LPS (100 ng/ml; Invivogen, San Diego, CA). Alternatively, cells were cocultured with CD40L transfected L cells (27) in a DC:L cell ratio of 4:1. Nontransfected L cells served as control cells. During activation, cells were incubated in the presence of 100 ng/ml GM-CSF.

A bicistronic retroviral DNA construct, coexpressing the genes encoding myristylated PKB (myrPKB) and enhanced green fluorescence protein (eGFP), with an internal ribosomal entry side (IRES) preceding eGFP (LZRS-myrPKB-IRES-eGFP; provided by Dr. H. Spits, Academic Medical Center, Amsterdam, The Netherlands) was used to transduce CD34⁺ cells. A second vector expressing only IRES followed by eGFP (LZRS-IRES-eGFP) was used as a negative control. The retroviral packaging cell line Phoenix-ampho (28) was stably transfected with 10 μg DNA by calcium phosphate coprecipitation. Twenty-four hours before transfection, cells were plated into 6-cm dishes. DMEM medium (Invitrogen) was refreshed 16 h after transfection. After an additional 24 h, cells were cultured in selection medium containing 10 μg/ml puromycin (Sigma-Aldrich, St. Louis, MO) until confluent. Selection medium was then replaced by CD34⁺ culture medium. After 24 h, viral supernatants were collected, filtered through a 0.2-μm filter, and subsequently frozen.

CD34⁺ progenitors were transduced in 24-well culture dishes precoated with 10 μg/cm² recombinant human fibronectin fragment CH-296 (RetroNectin; Takara, Otsu, Japan) for 16 h. Transduction was performed by adding 0.5 ml viral supernatant to 0.5 ml medium containing 5 to 30 × 10⁴ cells. After 24 h, 0.7 ml medium was removed and 0.5 ml fresh viral supernatant was added together with 0.5 ml fresh medium. For in vitro experiments, cells were transduced in complete medium supplemented with GM-CSF, TFNα, SCF, and human serum in the concentrations described above. CD34⁺ progenitors used for in vivo experiments were transduced in IMDM medium (Invitrogen) supplemented with 8% heat inactivated FCS, penicillin/streptomycin, 50 μM β-mercaptoethanol, 2 mM l-glutamine, 50 ng/ml Flt3 ligand (Peprotech), and 50 ng/ml SCF.

For phenotypic analysis, cells were washed in PBS containing 1% BSA, 1% heat inactivated human serum, and 0.02% NaN₃. Labeling of cell surface markers was performed on ice, using fluorochrome-conjugated Abs against the following Ags: CD1a (HI149; BD Biosciences, Breda, The Netherlands), CD14 (MϕP9; BD Biosciences), Langerin/CD207 (DCGM4; Beckman Coulter, Woerden, The Netherlands), CD1c (AD5-8E7; Miltenyi Biotec), CD20 (L27; BD Biosciences), CD86 (Fun-1; BD Biosciences), CD83 (HB15e, BD Biosciences), HLA-DR (L243, BD Biosciences) and CCR7 (150503, R&D Systems). Apoptosis was detected by determination of phosphatidyl serine exposure and membrane permeability. Cells were harvested, washed in Annexin buffer (10 mM Hepes, 140 mM NaCl, 2.5 mM CaCl₂, pH 7.4), incubated with Annexin V-FITC (BD Biosciences) for 30 min on ice and subsequently taken up in 1 μg/ml propidium iodide (PI; Sigma-Aldrich). Transduced cells were incubated with Annexin V-PE (BD Biosciences) and 7-aminoactinomycin D (7-AAD; BD Biosciences) in Annexin buffer for 30 min. Assessment was performed using a FACS Calibur or FACS CantoII (BD Biosciences), and data were analyzed using Cell Quest Pro and FACS Diva software (BD Biosciences).

Cells were harvested, washed in PBS and taken up in PBS/5 mM EDTA. Ethanol was added and cells were fixed in 50% ethanol for 30 min on ice. Next, cells were washed and subsequently incubated with PBS/5 mM EDTA with 40 μg/ml RNAse A (Promega, Madison, WI) for 30 min at room temperature. PI (50 μg/ml) was added and fluorescence intensity was immediately measured by flow cytometry (FACS Calibur).

Subset-specific precursors or retrovirally transduced cells were sorted using a FACS Aria (BD Biosciences). Control cultures were harvested at day 6, incubated with fluorochrome-conjugated Abs against CD1a and CD14, life gated, and sorted into CD14⁺CD1a⁻ and CD14⁻CD1a⁺ fractions. eGFP⁺ cells were isolated 3–4 d after transduction.

Cells were washed in PBS, lysed in Laemni sample buffer (0.12 M Tris HCl [pH 6.8], 4% SDS, 20% glycerol, 0.05 μg/μl bromophenol blue, and 35 mM β-mercaptoethanol) and boiled for 5 min. Protein concentrations were determined by Lowry. Equal amounts of total lysate (10 μg/lane) were separated by 10% SDS-PAGE, transferred to a PVDF membrane (Millipore, Bedford, MA) and incubated with blocking buffer (Tris buffered saline/Tween20) containing 5% BSA or low-fat milk before probing with Abs against human phosphorylated S6 (rabbit polyclonal; Cell Signaling Technology, Danvers, MA) and tubulin (mouse monoclonal; Sigma-Aldrich). Subsequently, blots were incubated with HRP-conjugated secondary Abs (Dako, Glostrup, Denmark). Enhanced chemical luminescence was used as a detection method according to the manufacturer’s protocol (Amersham Pharmacia, Amersham, U.K.).

Total RNA was isolated using the RNAeasy Mini Kit (Qiagen, Valencia, CA) according to manufacturer’s instructions. OD260/280 ratios were measured to determine the quantity and purity of RNA preparations and RNA integrity was checked using the Agilent RNA 6000 nano assay kit according to manufacturer’s instructions (Agilent Technologies, Santa Clara, CA). Reverse transcription-multiplex ligation-dependent probe amplification (RT-MLPA) was performed as described before (29). RNA samples (40–60 ng of total RNA) were reverse transcribed using a gene-specific probe mix. The resulting cDNA was annealed to multiplex ligation-dependent amplification probes that had been prepared from two oligonucleotides, a short synthetic oligonucleotide (Biolegio, Malden, The Netherlands) and a phage M13-derived long probe oligonucleotide. The two oligonucleotides annealed to adjacent sites on a target sequence and were subsequently ligated by Ligase-65 (MRC-Holland, Amsterdam, The Netherlands). Ligation products were amplified by PCR, with each ligated probe giving rise to an amplification product of unique length. The resulting DNA mixture was analyzed by capillary sequencer and standard software for identification and relative quantification of the PCR products. Analyzed data were exported to Microsoft Excel (Microsoft, Redmond, WA) spreadsheet software for further analysis. To normalize for variations in total signal between samples, the sum of all peak data were set at 100% for each sample. Individual peaks were calculated relative to the 100% value. Signals below the detection limit in medium were assigned a value corresponding to the threshold value.

Protocols for mouse experiments were approved by the local animal experimental committee. The β2-microglobulin^−/− NOD/severe combined immune deficient (NOD/SCID) mice were bred and maintained under sterile conditions in microisolator cages and provided with autoclaved food and acidified water containing 111 mg/L ciprofloxacin (Ciproxin). Eight- to 10-wk-old mice, sublethally irradiated with 250 cGy x-rays, received transplants via tail vein injections with ∼0.5 × 10⁶ unsorted retrovirally transduced cord blood-derived CD34⁺ hematopoietic progenitors along with 10⁶ irradiated (1500 cGy) CD34⁺-depleted cord blood-derived accessory cells. Six weeks after transplantation, the mice were sacrificed and both tibiae and femora were flushed. Bone marrow cells were analyzed for the presence of human myeloid DCs using flow cytometry.

Ag uptake was analyzed as described before (30). Fluid phase and lectin-mediated endocytosis were measured as the cellular uptake of 100 μg/ml Lucifer Yellow dipotassium salt (Molecular Probes, Invitrogen) and 100 μg/ml dextran-FITC (dextran^FITC, 40,000 MW; Molecular Probes, Invitrogen), respectively. Approximately 5 × 10⁴ DCs were incubated for 2 h at 37°C in the presence of culture medium containing Lucifer Yellow or dextran^FITC. Negative controls were incubated with the respective Ag at 4°C. Staining was evaluated by flow cytometry, and Ag uptake was calculated by subtracting staining at 4°C from staining at 37°C.

Cytokine production was measured in supernatants of stimulated cells by multiplex particle-based flow cytometry or ELISA. Concentrations of IL-6, IL-10, and TNF-α were measured with the Bio-Plex system using Bio-Plex Manager software version 4.0 (Bio-Rad Laboratories, Hercules, CA), which uses Luminex xMAP technology as previously described (31, 32). The multiplex data (cytokine concentrations) were digitized to create a cytokine portrait enabling the complete spectrum of cytokines to be visualized. The detection limit of this procedure was 1.2 pg/ml. The commercially available ELISA kits for human IL-5 (eBioscience, San Diego, CA), IL-6 (eBioscience), IL-8 (Biosource, Invitrogen), IL-10 (eBioscience), IL-12p70 (eBioscience), IP-10 (Biosource, Invitrogen), and IFN-γ (eBioscience) were used according to manufacturer’s instructions. The detection limits of these assays were 2 pg/ml (IL-6, IL-10, and IP-10), 4 pg/ml (IL-5, IL-12p70, and IFN-γ), and 5 pg/ml (IL-8).

Responder T cells were isolated from a buffy coat. The mononuclear fraction was incubated with anti-CD15– and anti-CD235–coated microbeads (Miltenyi Biotec) and PE-labeled Abs against CD1c, CD14, CD19 (J4.119, Beckman Coulter, Woerden, The Netherlands), CD56 (MY31, BD Biosciences), and CD123 (SSDCLY107D2, Beckman Coulter), followed by incubation with anti-PE–coated microbeads (Miltenyi Biotec). T cells were isolated through negative selection according to manufacturer’s instructions (Miltenyi Biotec). Irradiated DCs (30 Gy) were added in graded doses to 2 × 10⁴ allogeneic T cells in 96-well round bottom plates in RPMI 1640 containing 8% heat inactivated FCS. Proliferation was quantified by incubation with 1 μCi (37 kBq) [methyl-³H]thymidine (NEN Life Science Products; Boston, MA) during the last 18 h of 6-d cultures.

The importance of PI3K signaling for the development of CD34-derived myeloid DCs was investigated by addition of a specific pharmacologic PI3K inhibitor, LY, during DC development in vitro. Continuous exposure to LY from day 0 did not affect differentiation, as shown by the presence of CD14⁺CD1a⁻ precursor intDC (pre-intDC) and CD14⁻CD1a⁺ precursor LC (pre-LC) at day 6 and 90% CD14⁻CD1a⁺ myeloid DCs present at day 13, regardless of the presence of PI3K inhibition (Fig. 1A). Also, the specific myeloid DC marker CD1c was expressed by the majority of the CD1a⁺ cells in the control as well as inhibitor cultures (data not shown). In addition, stimulation of these cells with LPS in the absence of LY induced comparable expression of CD83, CD86, and HLA-DR for all cells (Fig. 1B). Thus, blockade of PI3K during DC development did not significantly affect the phenotype of the DCs. However, a strong reduction (64.4 ± 7.9%) in the yield of CD1a⁺ DCs was observed at day 13 (Fig. 1C).

To determine the mechanism by which PI3K regulates DC development, the role of mTOR, one of the proteins whose activity is regulated by PI3K signaling, was examined. Both LY and Rapa, a specific pharmacologic inhibitor of mTOR, reduced the phosphorylation of S6 (Fig. 1D), a target downstream of mTOR (33). Similar to LY, addition of Rapa allowed development of pre-DCs and terminally differentiated DCs from CD34⁺ progenitors, but significantly reduced the DC yield (Fig. 1A, 1C). Also as observed for LY, LPS-induced costimulatory molecule expression was unaffected for these cells (Fig. 1B). Because LY and Rapa strongly reduced DC numbers without affecting their phenotype, we conclude that both PI3K and mTOR activity are required for CD34-derived myeloid DC development, but these proteins seem to regulate processes other than differentiation.

To investigate how PI3K and mTOR regulate DC development, we first focused on pre-DC development (day 0 to 6) and examined the consequences of PI3K and mTOR inhibition in more detail. Depending on the donor, control cultures resulted in a 3- to 50-fold increased cell yield after 6 d, compared with day 0. Addition of either LY or Rapa significantly reduced cell yields in a dose-dependent manner (Fig. 2A, 2B). At a concentration of 2 μM LY or 20 nM Rapa, which are both within the range of concentrations used in literature, cell growth was inhibited with 67 ± 9.7% and 83 ± 12.5% (Fig. 2A), respectively. Increasing the concentration of Rapa hardly induced additional effects (Fig. 2B). Increased concentrations of LY, however, appeared to be extremely toxic (Fig. 2B), which might be explained by mechanisms other than mTOR inhibition (34). These data show that the expansion of DC precursors requires both PI3K and mTOR activation.

To determine whether the reduced expansion in PI3K- or mTOR-inhibited cultures was due to decreased cell survival and/or proliferation, cell viability and cell cycle profiles were analyzed at different time points during the 6-d culture period. As demonstrated by increased V binding in both the PI⁻ and PI⁺ cell population upon culture in the presence of LY or Rapa, PI3K and mTOR inhibition induced apoptosis at all time points (Fig. 2C, 2D). In addition, a decreased cell cycle progression was observed, as shown by the increased proportion of cells in G0/1 (Fig. 2E, 2F). These data show that a combination of reduced survival and decreased proliferation was responsible for the reduction in pre-DC yields found upon inhibition of PI3K or mTOR activity.

Then, the importance of PI3K and mTOR activity during development of pre-DCs into intDCs and LCs was investigated. Day 6 pre-intDCs and pre-LCs were isolated by FACS sorting and subsequently cultured in the presence or in the absence of LY or Rapa. In both pre-intDC and pre-LC cultures, inhibition of PI3K or mTOR resulted in reduced cell survival by the induction of apoptosis within 1 d (Fig. 3A). Under control conditions, CD1a⁻CD14⁺ pre-intDCs differentiate to CD1a⁺CD14⁻ intDC via an intermediate CD1a⁺CD14⁺ state, while CD1a⁺CD14⁻ pre-LCs remain CD1a⁺ and acquire upon terminal differentiation (22). Interestingly, in intDC cultures both PI3K and mTOR inhibition induced a time-dependent relative increase in differentiated cells (Fig. 3B, 3C). After 1 d, the ratio of differentiated CD1a⁺ cells and CD1a^-CD14⁺ preintDC was similar for all intDC cultures irrespective of PI3K or mTOR inhibition. Two and 3 d of incubation with inhibitors resulted in a relative increase in CD1a⁺ cells compared with control cultures. Also, the percentage of langerin⁺ cells in end-stage LC cultures was increased upon addition of LY or Rapa from day 6 (Fig. 3D, 3E). However, this relative increase was not due to an increased differentiation, as demonstrated by the unchanged CD1a⁺ and langerin⁺ absolute cell numbers (Fig. 3F, 3G). Instead, it resulted from a specific loss of the less differentiated CD1a⁻ and langerin⁻ cells. Thus, PI3K and mTOR activity is required for the survival of pre-intDC and pre-LC, whereas further differentiated cells seem to be unaffected by inhibition of this pathway.

To further investigate this apparent reduction in dependency, the role of PI3K and mTOR in the survival of pre-DCs and terminally differentiated DCs was compared. LY or Rapa were added to pre-intDCs and pre-LCs, or to intDCs and LCs generated from these pre-DCs. In contrast to pre-intDCs (Fig. 3A), terminally differentiated intDCs survived in the presence of LY or Rapa as shown by an V/PI staining comparable to control cultures (Fig. 4A, 4B). Similarly, LY- and Rapa-induced apoptosis was reduced in LC cultures compared with pre-LC cultures (Fig. 4C), leading to a similar survival in inhibitor and control cultures (Fig. 4A). Thus, whereas pre-intDCs and pre-LCs highly depend on PI3K and mTOR activity, these proteins seem redundant for the survival of terminally differentiated DC.

To investigate whether these differences in survival regulation could be explained by changes in the expression of known regulators of apoptosis, RT-MLPA was used to compare the mRNA composition of pre-intDCs and intDCs as well as pre-LCs and LCs. Large differences in relative expression profiles were induced during development of both subsets (Table I). Some of these changes, including the increased relative expression of antiapoptotic IAP family members and the reduced expression of proapoptotic proteins such as Apaf, could reflect an overall reduced susceptibility to apoptosis. However, other differences, such as the reduced relative expression of the mTOR-regulated Mcl-1, suggest a specific reduction in PI3K/mTOR dependency for survival.

Although treatment with LY or Rapa induced overlapping results (Figs. 1–4), it remains possible that the observed effects of LY and Rapa treatment are due to unrelated consequences of PI3K and mTOR inhibition. Because PI3K-induced mTOR activation is mediated by the activation of PKB (11), the relation between LY- and Rapa-induced effects on DC development was investigated by transducing CD34⁺ HPCs with a bicistronic retroviral DNA construct coexpressing eGFP and a constitutively active form of PKBα (myrPKB). HPCs transduced with myrPKB showed increased expansion in relation to progenitors transduced with the control construct (Fig. 5A), and the viability of pre-DCs expressing myrPKB was increased compared with control cells (Fig. 5B). Furthermore, LY-induced apoptosis was completely abolished in myrPKB-transduced cells, whereas apoptosis induced by mTOR inhibition was unaffected (Fig. 5C), placing PKB downstream of PI3K and upstream of mTOR. Moreover, myrPKB was able to rescue cells from apoptosis induced by deprivation of GM-CSF (Fig. 5D), a cytokine known to trigger PI3K activation and involved in myeloid cell survival (34, 35). These data indicate that the roles of PI3K and mTOR in CD34-derived myeloid DC development can be ascribed to their function in the PI3K–PKB–mTOR signaling module and suggest that GM-CSF stimulation results in activation of this pathway in vitro.

To investigate human CD34-derived DC development in vivo, myrPKB- or control-transduced human CD34⁺ HPCs were transplanted into NOD/SCID β2-microglobulin^−/− mice. Six weeks after injection, human myeloid DCs originating from transduced HPCs could be recognized in the bone marrow by their expression of eGFP and CD1c⁺. The proportions of eGFP⁺CD1c⁺CD20⁻ human DCs in mice transplanted with progenitors expressing constitutively active PKB were enhanced 20-fold compared with mice that had received a transplant of control-transduced HPCs (Fig. 5E, 5F). Thus, in vivo CD34-derived human myeloid DC development is augmented by the ectopic expression of constitutively active PKB, which confirms the role of PI3K–PKB–mTOR signaling in the development of these cells.

While we could show that PI3K and mTOR activity was not critical for the phenotype or survival of terminally differentiated DC, inhibition of PI3K or mTOR activity during either DC differentiation or DC activation has been suggested to negatively affect DC function (13, 36–38). We therefore investigated the functionality of CD34-derived myeloid DC differentiated or activated in the presence of LY or Rapa. Both lectin-mediated and fluid phase endocytosis, as investigated by incubation with dextran^FITC and Lucifer Yellow, respectively, were reduced in cells generated in the presence of Rapa, while cells generated with LY showed only a slightly reduced Ag uptake capacity (Fig. 6). Ag uptake capacity of control CD34-derived myeloid DC, however, was not affected by the addition of LY or Rapa during incubation with dextran^FITC or Lucifer Yellow (data not shown).

Also the expression of costimulatory molecules was unaffected by PI3K or mTOR inhibition, and neither LY nor Rapa affected LPS-induced upregulation of CD83, CD86, or HLA-DR (Fig. 7A). However, LY and Rapa inhibited CCR7 expression (Fig. 7A), indicating that this signaling module may be involved in the regulation of lymph node migration. Although only low cytokine concentrations could be measured due to low cell densities, LPS stimulation induced the production of detectable levels of IL-6, IL-8, and IP-10, which were reduced by the addition of LY and Rapa during either differentiation or activation of the cells (Fig. 7B, 7C). Also the secretion of IL-10 and TNF-α depended on intact PI3K and mTOR activity (Fig. 7C). Secretion of IL-12p70, which could be detected only after stimulation with CD40L, was also reduced by the presence of LY or Rapa during either differentiation or activation (Fig. 7C and data not shown). In accordance with the stronger effects of Rapa compared with LY (Fig. 7), T cell proliferation induced by myeloid DCs pretreated with Rapa, but not LY, was reduced compared with control DCs (Fig. 8). Only when using DCs treated with Rapa during activation was this reduced proliferation clearly associated with reduced IFN-γ concentrations (Fig. 8). A reduction in IL-5 production was also observed in cocultures of T cells with Rapa-treated DCs, whereas IL-10 concentrations were similar in all cultures (Fig. 8). Like for the nonstimulated DCs, neither LY nor Rapa affected the survival of LPS-activated DCs (data not shown), indicating that the effects on DC function were not caused by reduced DC viability. These data show that whereas PI3K and mTOR activity are dispensable for the phenotype and survival of terminally differentiated CD34-derived myeloid DCs, these proteins play an important role in the functionality of these cells.

DCs play a crucial role in the induction and regulation of immunity. Despite the requirement of DC development and survival for both immunogenic and tolerogenic responses, the molecular mechanisms regulating these processes are still largely undefined. In this study, we have shown that PI3K–PKB–mTOR signaling is required for proliferation and survival, but not differentiation, of human CD34-derived myeloid DCs. Inhibition of PI3K–PKB–mTOR signaling reduced the in vitro development of CD34-derived myeloid DCs, whereas activation of this pathway increased DC development both in vitro and in vivo. Although DC phenotype was not affected by inhibition of this signaling module, its activity was required for the generation of fully functional DCs.

We and others have described reduced DC yields following Rapa treatment during murine and human myeloid DC development (19, 38, 39). In contrast to moDC development, the development of myeloid DCs from human CD34⁺ progenitors involves not only differentiation and survival, but also proliferation. The reduced intDC and LC yields found upon inhibition of PI3K or mTOR activity were partly caused by a block in cell cycle progression during the development of pre-DCs. An increase of cells in G0/1 has also been reported for other cell types in which inhibition of PI3K–PKB–mTOR signaling caused a G1 cell cycle arrest (8, 40–42). The LY- or Rapa-induced apoptosis could be either a direct consequence of cell cycle arrest or an independent process. A combined reduction in cell cycling and survival has also been reported in developing erythrocytes (8), but is not a prerequisite in all cell types (19, 41). Because only minimal proliferation was present from the stage of pre-DC while these cells underwent apoptosis upon PI3K or mTOR inhibition, it appears that the LY- and Rapa-induced effects on proliferation and survival are independent consequences of reduced signaling along this pathway.

PI3K signaling plays a critical role in the regulation of survival in various hematopoietic cell types, including murine bone marrow-derived DCs and human moDCs (8, 18, 43). As demonstrated here, also during CD34-derived myeloid DC development, survival was decreased upon inhibition of PI3K or mTOR activity and increased in pre-DCs with increased activity of this signaling module by transduction with myrPKB. Whereas CD34-derived myeloid pre-DC survival required PI3K–PKB–mTOR signaling, terminally differentiated CD34-derived DCs survived without active PI3K or mTOR. This lack of apoptosis induction by PI3K or mTOR inhibition in CD34-derived intDCs is in sharp contrast to moDCs, which share phenotypic and functional characteristics with CD34-derived intDCs, but die upon inhibition of this pathway (19, 44). Moreover, whereas CD34-derived pre-DCs require mTOR for their survival, monocytes, the direct precursors of moDC, are insensitive to Rapa (19, 44) (Fig. 9). These disparities in survival signaling challenge the hypothesis that DCs developing from monocytes and from direct DC precursors are completely identical (45).

Although the effects of PI3K and mTOR inhibition on survival and proliferation could theoretically be unrelated, this seems unlikely. The lack of effects induced by only mTOR or only PI3K inhibition as well as the complementary consequences of PI3K/mTOR inhibition and increased PKB activity suggest that the observed effects reflect a critical role of the PI3K–PKB–mTOR signaling module. In addition, the myrPKB-induced rescue from GM-CSF withdrawal- or PI3K inhibition-induced apoptosis, but not mTOR inhibition-induced apoptosis, supports the suggestion that PI3K regulates CD34-derived myeloid DC development via the activation of PKB and mTOR.

The apparently absent PI3K–PKB–mTOR dependency of terminally differentiated DCs could reflect a generally reduced apoptosis sensitivity of these cells compared with their precursors, for example owing to the decreased Apaf expression in these cells (46). Alternatively, signaling pathways other than those including mTOR might be involved in the survival regulation of terminally differentiated myeloid DC. This hypothesis is supported by the observation that the expression of the antiapoptotic Bcl2 like proteins Mcl-1 and Bcl-XL is reduced in differentiated CD34-derived DC compared with their precursors (Table I). Although the regulation of these proteins by mTOR is at the translational level (47, 48), the reduced mRNA expression suggests a diminished role in survival regulation and thus a reduced requirement of mTOR mediated translational control of these proteins. The increase in Bfl-1/A1 expression, a Bcl2 family member with similar functions as Mcl-1 but regulated by NF-κB (49, 50), may point toward a role for other signaling pathways in the survival regulation of differentiated CD34-derived myeloid DC.

Because the surviving cells demonstrated normal pre-DC and DC phenotype, PI3K–PKB–mTOR signaling appears to be redundant for differentiation of CD34-derived myeloid DC. However, tolerogenic effects of Rapa addition during DC differentiation have been reported (38, 51), and in particular activation-induced cytokine secretion has been shown to be regulated by PI3K and mTOR (13–16, 36, 52). In accordance, the function of CD34-derived myeloid DCs differentiated or activated in the presence of PI3K or mTOR inhibition was inhibited. As previously reported for mouse bone-marrow derived DCs and human moDCs (39, 53), human CD34-derived myeloid DCs differentiated in the presence of Rapa showed reduced macropinocytosis and lectin-mediated endocytosis. In contrast, LY or Rapa did not affect Ag uptake by control DCs when inhibitor and Ag were added simultaneously (data not shown), indicating that neither PI3K nor mTOR was directly involved in the regulation of endocytosis. Rather, mTOR inhibition during differentiation seems to induce development of altered DCs with relatively poor Ag uptake capacity. The reduced cytokine secretion by DCs generated in the presence of LY or Rapa also indicates the development of alternatively functioning DCs. However, because LY and Rapa also inhibited DC function when added during activation, we cannot fully exclude the possibility that these effects result from ongoing PI3K and/or mTOR inhibition during activation. Although both LY and Rapa inhibited LPS-induced DC function, LY did so to a lesser extent, which could explain the reduced T cell stimulatory capacity observed for Rapa- but not LY-treated DCs. These findings seem contradictory to some studies describing repression of cytokine production by PI3K and mTOR in moDC (16, 36), but are in accordance with others (38, 54–56). These data indicate that although PI3K–PKB–mTOR signaling is not required for the acquisition of a DC phenotype or the survival of terminally differentiated CD34-derived myeloid DCs, this pathway may still be involved in the regulation of other important processes in these cells.

The molecular mechanism underlying the impaired response to activation signals following mTOR inhibition is relatively unknown. Minor changes in TLR4 mRNA expression were observed (data not shown), which might provide an explanation for the reduced response to LPS. However, as LPS induced normal upregulation of costimulatory molecule expression, loss of TLR4 stimulation cannot fully account for the diminished function observed after mTOR inhibition. Whereas the role of mTOR in the regulation of proliferation and survival is mainly ascribed to translational control (11, 57), cytokine production has been suggested to be regulated on the transcriptional level (15, 58, 59). However, the exact mechanisms of the processes involved remain to be elucidated.

Whereas the decreased T cell proliferation induction by DCs treated with Rapa could result from the induction of regulatory T cell differentiation as suggested before (38, 51), Foxp3 expression in these T cell cultures was unchanged compared with control (data not shown). Whether the altered secretion of Th1 and Th2 cytokines after coculture with Rapa-treated DCs only reflects reduced T cell numbers or results from changed Th1 and Th2 differentiation as shown for murine Rapa-differentiated DCs (37) remains to be determined. Also, although the reduced IL-5 secretion but unaffected IFN-γ concentrations found in T cell cultures stimulated with Rapa-differentiated DCs might suggest skewing of T helper differentiation, this should be investigated further. In addition to reduced DC–T cell interaction as a consequence of mTOR inhibition in DCs, the reduced CCR7 expression following LY or Rapa treatment may also have consequences for the induction of immune responses. These data suggest that lymph node migration requires intact activity of the PI3K–mTOR signaling module, as has been suggested before (60).

In conclusion, PI3K–PKB–mTOR signaling is required for human CD34-derived myeloid DC development by regulating the proliferation and survival of DC precursors. Whereas this signaling module is redundant for differentiation and phenotypic maturation as well as survival of terminally differentiated CD34-derived intDCs and LCs, its activity is required for the generation of fully functional DCs. These findings could be used as a strategy to manipulate DC subset distribution and function to regulate immunity.

We thank Dr. W. de Jager (Department of Pediatric Immunology, University Medical Center, Utrecht, The Netherlands), E. van der Linden (Department of Immunology, University Medical Center, Utrecht, The Netherlands), and A. van den Bosch (Department of Gastroenterology and Hepatology, Erasmus MC-University Medical Center Rotterdam, Rotterdam, The Netherlands) for technical assistance; Dr. E. Eldering (Department of Experimental Immunology, Academic Medical Center, Amsterdam, The Netherlands) for providing the RT-MLPA facility and expertise; Dr. A. Boonstra (Department of Gastroenterology and Hepatology, Erasmus MC-University Medical Center Rotterdam, Rotterdam, The Netherlands) for critically reading the manuscript; Dr. G. Nolan (Stanford University School of Medicine, Stanford, CA) for providing the LZRS construct and the phoenix-ampho packaging cell line; and Dr. H. Spits (Academic Medical Center, Amsterdam, The Netherlands) for providing the LZRS construct expressing myrPKB.

Disclosures The authors have no financial conflicts of interest.