Osteopontin expression is essential for interferon-alpha production by plasmacytoid dendritic cells

Abstract

The observation that the T-bet transcription factor allows tissue-specific upregulation of intracellular osteopontin (Opn-i) in plasmacytoid dendritic cells (pDCs) suggests that Opn might contribute to the expression of interferon-alpha (IFN-alpha) in those cells. Here we show that Opn deficiency substantially reduced Toll-like receptor 9 (TLR9)-dependent IFN-alpha responses but spared expression of transcription factor NF-kappaB-dependent proinflammatory cytokines. Shortly after TLR9 engagement, colocalization of Opn-i and the adaptor molecule MyD88 was associated with induction of transcription factor IRF7-dependent IFN-alpha gene expression, whereas deficient expression of Opn-i was associated with defective nuclear translocation of IRF7 in pDCs. The importance of the Opn-IFN-alpha pathway was emphasized by its essential involvement in cross-presentation in vitro and in anti-herpes simplex virus 1 IFN-alpha response in vivo. The finding that Opn-i selectively coupled TLR9 signaling to expression of IFN-alpha but not to that of other proinflammatory cytokines provides new molecular insight into the biology of pDCs.

Figure 1

Opn expression in pDCs and cDCs 24 h after stimulation. (a) Opn mRNA in splenic pDCs (left) or cDCs (right) stimulated with 1 μg/ml of CpG-B (ODN-1668); 1 μg/ml of CpG-B plus 10 ng/ml of recombinant mouse IFN-γ protein; 1 μg/ml of lipopolysaccharide (LPS); or 5 μg/ml of anti-CD40. Relative Opn mRNA is based on Opn mRNA in pDCs without stimulation. (b) Opn mRNA and secreted protein in pDCs and culture supernatants, respectively. (c) Opn mRNA concentrations in T-bet wild-type (T-bet WT) and T-bet-deficient (T-bet KO) pDCs. Concentrations of CpG-B and recombinant mouse IFN-γ were 1 μg/ml and 10 ng/ml, respectively. Relative Opn mRNA is based on Opn mRNA in T-bet wild-type pDCs without stimulation. Opn mRNA and protein were measured by quantitative real-time PCR of cDNA and ELISA, respectively, with triplicate wells, and are presented with error bars. Data are representative of at least three independent experiments.

Figure 2

Cytokine production by pDCs. (a) IFN-α (left) and IL-6 (right) in 24-hour pDC culture supernatants of T-bet wild-type (T-bet WT) versus T-bet-deficient (T-bet KO) cells stimulated with various concentrations of CpG-B (ODN-1668). The pDCs were plated at a density of 1.0 × 10⁶ cells/ml. (b) Comparison of secreted IFN-α protein, cellular IFN-α mRNA, secreted IL-6 and TNF (left to right) for Opn wild-type (Opn WT) and Opn-deficient (Opn KO) pDCs plated at a density of 1.0 × 10⁶ cells/ml with CpG-B (1 μg/ml for mRNA samples; otherwise, concentrations along horizontal axes) and collected at 24 h. (c) Comparison of secreted IFN-α, IL-6 and TNF (left to right) for Opn wild-type (Opn WT) and Opn-deficient (Opn KO) pDCs after CpG-A stimulation. The pDCs (0.2 × 10⁶ cells/ml) were plated with CpG-A (ODN-D19; concentration, horizontal axis) and were collected at 24 h for analysis. (d) IFN-α from 24-hour pDC culture supernatants (1.0 × 10⁶ cells/ml) from Opn WT and Opn KO mice stimulated with imiquimod (R837). Protein and mRNA were measured by ELISA and real-time PCR in triplicate wells, except for c, for which the IFN-α ELISA was done in duplicate wells. Data are representative of two to three experiments.

Figure 3

Intracellular Opn is required for IFN-α production. (a–c) Lentivirus infection of pDCs stimulated with CpG-B (0.2 μg/ml). (a) Extracellular Opn protein measured by ELISA in 6-hour culture supernatants of sorted splenic Opn-deficient pDCs infected with lentivirus to express full-length Opn (fOpn), mutant Opn (ΔOpn) or GFP control (GFP). HIV capsid protein p24, determined by ELISA, indicates the concentration of lentivirus used for infection (horizontal axis). (b) IFN-α in 24-hour culture supernatants of splenic Opn-deficient pDCs transfected with lentiviral reagents (GFP, ΔOpn and fOpn). (c) IFN-α (left vertical axis) or intracellular Opn (right vertical axis) in culture supernatants and cells. ‘Titration’ of mutant Opn lentivirus (ΔOpn-lenti) into bone marrow–derived Opn-deficient pDCs (1 × 10⁶ cells/ml) was followed by CpG-B stimulation (ODN-1668; 0.2 μg/ml), then analysis 24 h later. (d) IFN-α production in 24-hour culture supernatants of Opn wild-type (WT) and Opn-deficient (KO) bone marrow–derived pDCs after a 20-minute pulse with DOTAP in complex with CpG-B (ODN-1668; left) or CpG-A (ODN-D19) with (+) or without (−) DOTAP (right). Data are representative of at least three independent experiments.

Figure 4

Biochemical analysis of Opn-i. (a) Nuclear translocation of IRF7 in total bone marrow–derived Opn wild-type (WT) and Opn-deficient (KO) DCs treated with CpG-B (0.4 μM) for 3 h before extraction of nuclear and total lysates. (b) Activation of NF-κB in total bone marrow–derived DCs stimulated for 1 h with 0.3 μM CpG-B. Control competitor and noncompetitor oligonucleotides were added to some samples (+, below graph) to confirm sequence-specific binding of NF-κB in lysate samples. (c) Immunoblot (IB) of protein complexes immunoprecipitated (IP) from HEK293 cells cotransfected for 48 h with Opn, Flag-MyD88 and TLR9 (lane 1) or Opn, Flag-Mink (serine kinase) and TLR2 (lane 2; negative control) followed by stimulation for 3 h with 0.3 μM CpG-B. (d) Opn expression and Ifna4 activation in HEK293 cells transiently transfected with mutant Opn–expressing lentivirus (concentrations, below graph) along with IRF7 and/or MyD88 cDNA. (e) MyD88-dependent Ifna4 activation after transfection of MyD88 expression vector (concentrations, below graph) into HEK293 cells along with mutant Opn and/or IRF7. Data are representative of at least three independent experiments.

Figure 5

Localization of Opn together with MyD88 and TLR9. (a, b) Immuno-fluorescence and confocal microscopy of Opn (green) and MyD88 (red; a) or TLR9 (red; b) in pDCs. After treatment of pDCs with 0.5 μM of CpG-B for 10 min, cells were fixed and Opn, MyD88 and TLR9 were detected with mouse anti-Opn (2A1), rabbit anti-MyD88 and rabbit anti-TLR9, followed by secondary antibodies. Original magnification, ×400 (first three columns) and ×800 (single-cell images, far right).

Figure 6

In vitro OVA cross-presentation by pDCs. (a) Purity of splenic CD11c⁺ cells (B220⁺CD11c⁺CD3ε⁻CD19⁻NK1.1⁻) after repeated sorting by flow cytometry. (b) OVA cross-presentation by Opn wild-type (WT) or Opn Opn-deficient (KO) pDCs. Purified bone marrow–derived pDCs were first cultured for 24 h with (+ CpG) or without (− CpG) 0.2 μg/ml of CpG-B in the presence of soluble OVA (50 μg/ml), then OT-I T cell proliferation was measured. (c) OT-I peptide presentation by Opn WT or Opn KO pDCs. Purified splenic pDCs were cultured for 18 h with (+) or without (−) 0.2 μg/ml of CpG-B in presence of OT-I peptide (1 μg/ml). (d) OVA cross-presentation (left and middle) and OT-I peptide direct presentation (right) by purified splenic pDCs from B6 wild-type mice (WT) and B6 mice deficient in transporter associated with antigen processing (TAP1-deficient). The pDCs were treated for 20 h with 50 μg/ml of OVA or 1 μg/ml of OT-I peptide with 0.2 μg/ml of CpG-B, then OT-I T cell proliferation and IFN-γ production were assessed. (e) IFN-α production by pDCs incubated with OVA and CpG (0.2 μg/ml of CpG-B) before culture together with OT-I T cells. Data are representative of at least four independent experiments.

Figure 7

IFN-α-dependent cross-presentation. (a, b) Purified bone marrow–derived Opn-deficient (Opn KO) and Opn wild-type (Opn WT) pDCs were treated with or without CpG-B (0.2 μg/ml) and/or recombinant IFN-α (rIFN-α; 1 × 10⁸ to 5 × 10⁸ units = 1 mg; key) in the presence of OVA (50 μg/ml). OT-1 T cell proliferation was measured. (a) Supplementary effect of recombinant IFN-α in antigen cross-presentation by CpG-treated pDCs. (b) Treatment with recombinant IFN-α without CpG is sufficient to ‘license’ pDCs for antigen cross-presentation. (c) ‘Licensing’ of cDCs from Opn KO and Opn WT mice after treatment of cells with (100 U/ml; key) or without (0 U/ml; key) recombinant IFN-α. Data are representative of at least three independent experiments.

Figure 8

Opn-dependent in vivo response to HSV-1 infection. (a) IFN-α concentrations from Opn wild-type (WT) and Opn-deficient (KO) pDC culture supernatants after incubation in vitro with ultraviolet irradiation–treated (UV-irradiated) HSV-1 for 24 h. (b) Serum IFN-α in B6, B6 TLR9-deficient (TLR9 KO) and B6 Opn-deficient (Opn KO) mice injected intraperitoneally with HSV-1 (1 × 10⁶ PFU/mouse). Data represent three mice per group. (c) HSV-1-specific delayed-type hypersensitivity responses in B6, TLR9 KO and Opn KO mice injected intraperitoneally with HSV-1 (5 × 10⁶ PFU/mouse) and challenged 6 d later in the left footpad with ultraviolet irradiation–treated HSV-1 (1 × 10⁵ PFU/mouse). Error bars indicate mean ± s.d. of three mice per group. (d) ⁵¹Cr-release assay of NK cells from B6 Opn wild-type (WT) and Opn-deficient (KO) mice, evaluated against YAC-1 cells. Draining lymph nodes of mice injected with CpG (ODN-1585; -CpG) or control GpC (–control ODN) were pooled. Data are representative of at least three independent experiments. (e) Survival of Opn WT and Opn KO mice injected intraperitoneally with 5 × 10³ B16 cells plus 100 μg CpG (ODN-1585) or control GpC (control ODN). There were five mice in each group.