Chromatin structure undergoes global and local reorganization during murine dendritic cell development and activation

Abstract

Classical dendritic cells (cDCs) are essential for immune responses and differentiate from hematopoietic stem cells via intermediate progenitors, such as monocyte-DC progenitors (MDPs) and common DC progenitors (CDPs). Upon infection, cDCs are activated and rapidly express host defense-related genes, such as those encoding cytokines and chemokines. Chromatin structures, including nuclear compartments and topologically associating domains (TADs), have been implicated in gene regulation. However, the extent and dynamics of their reorganization during cDC development and activation remain unknown. In this study, we comprehensively determined higher-order chromatin structures by Hi-C in DC progenitors and cDC subpopulations. During cDC differentiation, chromatin activation was initially induced at the MDP stage. Subsequently, a shift from inactive to active nuclear compartments occurred at the cDC gene loci in CDPs, which was followed by increased intra-TAD interactions and loop formation. Mechanistically, the transcription factor IRF8, indispensable for cDC differentiation, mediated chromatin activation and changes into the active compartments in DC progenitors, thereby possibly leading to cDC-specific gene induction. Using an infection model, we found that the chromatin structures of host defense-related gene loci were preestablished in unstimulated cDCs, indicating that the formation of higher-order chromatin structures prior to infection may contribute to the rapid responses to pathogens. Overall, these results suggest that chromatin structure reorganization is closely related to the establishment of cDC-specific gene expression and immune functions. This study advances the fundamental understanding of chromatin reorganization in cDC differentiation and activation.

Keywords: chromatin structure; dendritic cell; hematopoiesis; infection; transcription factor.

Fig. 1.

Nuclear compartment dynamics during cDC differentiation. (A) A differentiation model of cDCs. The green rectangle indicates DC lineage populations subjected to Hi-C analysis. (B) Example of Hi-C contact matrix. The color represents normalized Hi-C interaction counts at indicated genomic regions in chromosome 8 using Homer software. Histograms of PC1 values of the corresponding genomic region are horizontally and vertically shown. Black and gray regions indicate regions in the A and B compartment, respectively. (C) Genome-wide assessment of changes in nuclear compartments during cDC differentiation. The mouse genome was subdivided into 25-kb bins, resulting in 103,531 regions, for PC1 value calculation. The genomic regions were clustered into five using the k-means method. Red and blue indicate the A and B compartments, respectively. CL, cluster. (D) University of California, Santa Cruz (UCSC) genome browser image of compartments in LMPPs, MDPs, CDPs, cDC1s, and cDC2s. In this view of chromosome 1, the number of regions in the A compartment increases from two (denoted as A₁ and A₂) in LMPPs to six (denoted as A₁, A₂, A₃, A₄, A₅, and A₆) in cDCs. (E) Kinetics of H3K27ac levels (z score), compartment PC1 values, and gene expression levels (z score) in CL1 or CL2 genomic regions shown by boxplots. Box and center line represent the 25th to 75th percentiles and the median, respectively. Whiskers indicate 1.5-fold of upper and lower interquartile range or the most extreme values. Two-tailed Student’s t test was performed to calculate statistical significance. In the statistical software we used, 10⁻³⁰⁸ was the lowest P value. P, P value.

Fig. 2.

Dynamics of intra-TAD interactions during cDC differentiation. (A) Kinetics of intra-TAD interaction frequency in cDC differentiation. TADs in each cell population were identified using Juicer software. By merging TADs across all cell types analyzed, we identified 7,311 unique TADs in total. Subsequently, the chromatin interaction frequency within the TADs and its z score were calculated and subdivided into seven clusters using the k-means method. (B) Examples of TADs with increased chromatin interactions during cDC differentiation. The Hi-C contact matrix of the MHC class II gene locus on chromosome 17 is shown. (C) Kinetics of H3K27ac levels, frequency of intra-TAD interactions, and RNA expression during cDC differentiation. Boxplots show the z scores. CL1, CL2, CL3, and CL7 TADs, shown in A, were analyzed. Two-tailed Student’s t test was performed to calculate statistical significance. (D) Pile-up images for TAD loops. The average interaction frequencies between TAD borders of each CL are shown as heatmaps. The signals represent loops formed between TAD borders.

Fig. 3.

Chromatin structure reorganization at cDC-specific gene loci. (A) Types of compartment dynamics of cDC-specific gene loci. Compartment dynamics types used here are the CLs shown in Fig. 1C. The cDC1- or cDC2-specific genes were identified from the RNA-seq data. (B and C) Temporal relationship between compartmentalization and TAD formation at genomic regions containing cDC-specific gene loci. CL1 (B) or CL3 (C) compartment regions containing cDC1- or cDC2-specific genes and TADs in these regions were analyzed. Two-tailed Student’s t test was performed to calculate statistical significance. (D) Schematic models for chromatin structure reorganization at cDC-specific gene loci during cDC differentiation.

Fig. 4.

Identification of transcription factors required for histone acetylation prior to the B-to-A compartment change. (A) De novo motif analysis of open chromatin regions in compartment CL1 defined in Fig. 1C. Open chromatin regions in each progenitor population were identified from assay for transposase-accessible chromatin sequencing (ATAC-seq) data. (B) Known motif analysis for PU.1-IRF binding sites. The “GGAANNGAAA” sequence was searched in the ATAC-seq peak regions in each progenitor population. (C) RNA expression of IRF family transcription factor genes in hematopoietic stem and progenitor populations. Values in the bar graph are the mean ± SD from two independent experiments. (D) Influence of IRF8 deficiency on H3K27ac accumulation in compartment CL1. ChIP-seq tag densities of H3K27ac in CL1 genomic regions with IRF8 binding (Left) or those with (Middle) or without (Right) known PU.1-IRF motifs (i.e., GGAANNGAAA) in open chromatin regions were analyzed. Two-tailed Student’s t test was performed to calculate statistical significance.

Fig. 5.

IRF8 induces switches to compartment A in DC progenitors. (A) Influence of IRF8 deficiency on nuclear compartment formation in MDPs. The mouse genome was subdivided into 25-kb bins, resulting in 103,531 regions as shown in Fig. 1C. PC1 values in each region were calculated in wild-type (WT) MDPs and Irf8^−/− MDPs. Red and blue dots indicate genomic regions where PC1 values were significantly down-regulated and up-regulated, respectively, in Irf8^−/− MDPs. We considered a q value less than 0.05 as significant. (B) Hierarchical clustering analysis of regions where PC1 values were significantly altered by the absence of IRF8 in MDPs. (C) Comparison of PC1 values at IRF8 binding sites between WT MDPs and Irf8^−/− MDPs. IRF8 binding sites were identified from IRF8 ChIP-seq data in WT MDPs. (D) Example of compartment change at IRF8 binding sites near the Itgb8 gene. (E) RNA expression of genes included in the 942 IRF8-bound genomic regions where PC1 values were significantly down-regulated. Two-tailed Student’s t test was performed to calculate statistical significance. (F) GSEA comparing WT CDPs with Irf8^−/− CDPs for genes included in the 942 genomic regions analyzed in E. (G) Schematic model of the role of IRF8 in compartment formation during cDC differentiation.

Fig. 6.

Chromatin structure reorganization at host defense-related genes in cDC1s. (A) Differentially expressed genes between cDC1s from uninfected mice (control [ctrl] cDC1) and those from T. gondii–infected mice (Tx cDC1). Red and blue dots indicate significantly up-regulated and down-regulated genes, respectively, in Tx cDC1s. (B) GO analysis of the genes up-regulated in Tx cDC1s. (C) Kinetics of chromatin structure reorganization at host defense-related gene loci. (Left) The mouse genome was subdivided into 25-kb bins, resulting in 103,531 regions as shown in Fig. 1C. Among these regions, those containing the genes up-regulated in Tx cDC1s were selected. TADs in these regions were then identified. Compartment PC1 values of the genomic regions and the z scores for H3K27ac signals, chromatin interactions within TADs, and RNA expression of up-regulated genes are shown as box plots. Two-tailed Student’s t test was performed to calculate statistical significance. (Right) The average frequencies of interactions between the TAD borders, that is, TAD loops, are shown as heatmaps. (D) Schematic model of chromatin structure reorganization at host defense-related gene loci during cDC differentiation and activation.