Ikaros is required for plasmacytoid dendritic cell differentiation

Abstract

Plasmacytoid dendritic cells (pDCs) are specialized DCs that produce high levels of type I IFN upon viral infection. Despite their key immunoregulatory role, little is known about pDC ontogeny or how developmental events regulate their function. We show that mice expressing low levels of the transcription factor Ikaros (Ik(L/L)) lack peripheral pDCs, but not other DC subsets. Loss of pDCs is associated with an inability to produce type I IFN after challenge with Toll-like receptor-7 and -9 ligands, or murine cytomegalovirus (MCMV) infection. In contrast, conventional DCs are present in normal numbers and exhibit normal responses in vivo after challenge with MCMV or inactivated toxoplasma antigen. Interestingly, Ik(L/L) bone marrow (BM) cells contain a pDC population that appears blocked at the Ly-49Q- stage of differentiation and fails to terminally differentiate in response to Flt-3L, a cytokine required for pDC differentiation. This differentiation block is strictly dependent on a cell-intrinsic requirement for Ikaros in pDC-committed precursors. Global gene expression profiling of Ik(L/L) BM pDCs reveals an up-regulation of genes not normally expressed, or expressed at low levels, in WT pDCs. These studies suggest that Ikaros controls pDC differentiation by silencing a large array of genes.

Figure 1.

IkL/L mice show selective loss of pDCs. (A) Splenocytes from the indicated mice were stained with Alexa488 anti-120G8; PE anti-CD3, CD19, and NK1.1; APC anti-CD11c; PE-Cy7 anti-CD11b; APC-Cy5.5 anti-B220; and biotin anti-CD8α. Anti-CD8α antibodies were revealed with SA-PE-TexasRed. The resulting cells were analyzed on an LSR II flow cytometer by collecting 200 000 events for each file. Nonviable cells were eliminated from subsequent analyses by staining cells with DAPI. Numbers indicate percentage of events within the indicated gate as a function of the corresponding parent gate for each plot. The absolute number of events within each CD11c⁺ (CD3^– CD19^– NK1.1^–) gate are as follows: WT, 889; Ik^+/L, 795; and Ik^L/L, 1105. (B) Lymph node cells from the identical mice were stained and analyzed as in panel A. The absolute number of events within each CD11c⁺ (CD3^– CD19^– NK1.1^–) gate are as follows: WT, 237; Ik^+/L, 241; and Ik^L/L, 164. Absolute numbers of DCs within each subset in 3 mice per group are given in Table 1. These data are representative of 4 experiments.

Figure 2.

Ik^L/L mice do not produce IFNα upon stimulation with influenza virus or TLR ligands. (A) Purified CD11c⁺ spleen cells were stimulated with inactivated influenza virus (Flu) or CpG D19 for 24 hours, and supernatants were tested for the indicated cytokines. Results represent 1 of 2 experiments, on cells pooled from 5 mice/genotype per experiment. (B) Mice were injected with poly I:C, R848, or CpG 1668. Sera were collected at 2 hours (R848) and 4 hours (PBS, poly I:C, CpG) after injection and tested for the indicated cytokines. Results are representative of 2 experiments, with 3 mice/genotype per condition for each experiment. Means and SDs are shown for 1 representative experiment.

Figure 3.

Antiviral response to MCMV infection. (A) Animals were injected intraperitoneally with GFP-recombinant MCMV. At 1.5 days after challenge, the frequencies of pDCs and MCMV-infected DCs were measured in the spleen, and cytokines were titrated in the serum. *P = .005; **P = .001. (B) Distribution of pDCs (Ly-6C⁺ 120G8⁺) in WT and Ik^L/L spleens following infection. Contour plots are gated on CD11c⁺ cells. Numbers indicate percentage of gated cells in each plot. (C) Proportion of GFP⁺ WT and Ik^L/L spleen cells after infection. Results shown are representative of 5 experiments, each with 3 to 12 mice per group. Means and SDs are shown for 1 representative experiment.

Figure 4.

A pDC population in the BM of Ik^L/L mice. (A) CD11c⁺ BM cells were gated and analyzed for B220 and 120G8 expression. Numbers indicate percentage of cells in the gated contour plots. (B) 120G8⁺ spleen and BM cells were analyzed for their expression of the indicated surface antigens. The pDC population is indicated by the white histograms. Gray histograms indicate the isotype controls. Results are representative of 3 experiments. (C) Sorted CD11c⁺ 120G8⁺ BM cells were cultured with poly I:C, inactivated influenza virus, or CpG D19 for 24 hours, and supernatants were tested for the indicated cytokines. Results are representative of 2 experiments, on pools of cells from 5 mice/genotype per experiment. Means and SDs are shown for 1 representative experiment.

Figure 5.

Ik^L/L pDC precursors respond to Flt-3 signaling. (A) 120G8⁺ BM cells were analyzed for Flt-3 expression. The pDC population is indicated by the white histograms. Gray histograms indicate the isotype controls. Results are representative of 5 experiments. (B-C) Unfractionated BM cells were cultured with Flt-3L over a 9-day period. (B) Contour plots show B220 and 120G8 expression of CD11c⁺ cells after 7 days of culture. (C) Graph shows the percentage of 120G8⁺ cells among total cells for the indicated number of days. One representative experiment of 5 is shown.

Figure 6.

The defect in pDC differentiation is cell intrinsic. (A) BM from WT (Ly5^B6; 2 × 10⁶), Ik^L/L (Ly5^B6; 5 × 10⁶), or mixtures of WT:B6.Ly5^SJL (1 × 10⁶: 1 × 10⁶) or Ik^L/L:B6.Ly5^SJL (5 × 10⁶:1 × 10⁶) were used to reconstitute lethally-irradiated B6.Ly5^SJL recipients. Recipients were analyzed 6 to 10 weeks after BM transfer. All plots are gated on small Ly5^B6+CD11c-enriched cells. Numbers indicate the percentage of gated cells in each plot. One representative experiment of 4 is shown. (B) Ikaros expression in DC subsets. WT BM and splenic pDCs (CD11c⁺ B220⁺ 120G8⁺), and splenic CD11b⁺ cDCs (CD11c⁺ CD11b⁺120G8^–) and CD8α⁺ cDCs (CD11c⁺CD8α⁺CD11b^–120G8^–) were analyzed by RT-PCR for their expression of Ikaros, Aiolos, and Helios. Whole WT thymocytes were used as a control. One representative experiment of 3 is shown.

Figure 7.

Significant gene derepression in Ik^L/L BM pDCs. (A) Representative scatterplots comparing the gene expression profile between 2 independent WT samples (top), versus a WT and an Ik^L/L sample (bottom). The right panels are enlargements of boxed areas in the left panels. Red indicates increased genes; blue, decreased genes. (B) Genes that either “increased” or “decreased” in all of the 9 Ik^L/L/WT comparisons were selected and grouped according to the magnitude of change (average of the “fold change” [in log2] from all 9 comparisons). Note that the decreased genes exhibited mostly modest variations, while many increased genes exhibited large variations. correspond to the estimated number of nonspecific changes (“Materials and methods”). (C) Number of up-regulated genes in Ik^L/L samples whose expression was not detected in WT samples (▪) or were expressed in WT but not detected in Ik^L/L samples (□).