Landscape and Dynamics of the Transcriptional Regulatory Network During Natural Killer Cell Differentiation

Abstract

Natural killer (NK) cells are essential in controlling cancer and infection. However, little is known about the dynamics of the transcriptional regulatory machinery during NK cell differentiation. In this study, we applied the assay of transposase accessible chromatin with sequencing (ATAC-seq) technique in a home-developed in vitro NK cell differentiation system. Analysis of ATAC-seq data illustrated two distinct transcription factor (TF) clusters that dynamically regulate NK cell differentiation. Moreover, two TFs from the second cluster, FOS-like 2 (FOSL2) and early growth response 2 (EGR2), were identified as novel essential TFs that control NK cell maturation and function. Knocking down either of these two TFs significantly impacted NK cell differentiation. Finally, we constructed a genome-wide transcriptional regulatory network that provides a better understanding of the regulatory dynamics during NK cell differentiation.

Keywords: ATAC-seq; Dynamic regulatory network; EGR2; FOSL2; NK cell; Programmed differentiation.

Figure 1

DNA accessibility during NK cell differentiation A. Confocal microscopy images of membrane CD56 (red) in the cultured cells at day 7, day 14, day 21, day 28, and day 35. Scale bar, 30 µm. Nuclei are stained with DAPI. B. Schematic representation of the overall experimental design of this study. Chromosome opening at different time points were assessed using ATAC-seq data. The bioinformatics pipeline for data analysis is shown in the bottom. C. The gene expression profiles of SPI1 and TBX21 at different stages of NK cell differentiation. D. Flow cytometric measurement of PU.1/SPI1 and T-bet/TBX21 expression in cultured cells during a 35-day time course. The X-axis indicates the fluorescence intensity, and the Y-axis indicates the cell number density. E. Normalized ATAC-seq profiles of the SPI1 (top) and TBX21 (bottom) gene loci at different stages during NK cell differentiation. ATAC-seq signals were obtained from the UCSC Genome Browser. NK, natural killer; UCB, umbilical cord blood; HSC, hematopoietic stem cell; ATAC-seq, assay of transposase accessible chromatin with sequencing; SPI1, Spi-1 proto-oncogene; TBX21, T-box transcription factor 21; AU, arbitrary unit; ISO, isotype control.

Figure 2

Differential epigenetic regulation elements during NK cell differentiation A. Heatmap of 6401 differentially accessed regulatory elements during NK cell differentiation. Each column is a sample, and each row is a peak. Samples and peaks are organized by two-dimensional unsupervised hierarchical clustering. Color scale shows the relative ATAC-seq peak intensity centered by each peak summit. Bottom: samples at all time points are categorized into three groups: early stage (days 7–21; orange), interim stage (days 24–28; yellow), and late stage (day 35; green). Samples from the same stage are labeled with the same color. Left: differential peaks are categorized into three clusters. B. The top 10 most significant GO terms enriched in cluster I (upper panel) and cluster II (lower panel) peaks. C. Box plots showing expression levels of genes in different clusters during NK cell differentiation. Left: genes in cluster I show higher expression in the early stage. Right: genes in cluster II show higher expression in the interim and late stages. R1 and R2 represent biological replicates 1 and 2, respectively.P values are estimated from one-way ANOVA. D. The changes in ATAC-seq signal (red), gene expression signal (orange), and percentage of NK cell counts (green) at different time points during NK cell differentiation.

Figure 3

TFs enriched at different stagesof NK cell differentiation A. TF motifs enriched in cluster I (left) and cluster II (right) peaks, with enrichment P values estimated from HOMER. TFs known to regulate NK cell differentiation are colored in red. B. Enrichment of known TF motifs in differentially accessible elements in all samples. Each row is a TF motif and each column is a sample. Values in the matrix represent the significance levels expressed as –log₁₀P value of the enrichment estimated from Genomica. Red in the matrix indicates that the motif is enriched in the corresponding sample, whereas blue in the matrix indicates depletion. Red texts on the right indicate known key TFs that regulates NK cell development, while pink texts indicate new TFs whose functions will be experimentally tested later (see Figure 4). C. Prediction of TFs that may regulate NK cell differentiation. TFs known to regulate NK cell differentiation are shown at the top, TFs predicted to regulate the early stage of the differentiation process are shown in the bottom, and those predicted to regulate the interim and late stages are shown in the middle. The color of each circle represents the expression level of the gene encoding the corresponding TF, while the size of the circle represents the significance of the motifs estimated by P values. D. Visualization of the ATAC-seq footprints for STAT5, T-bet, FOSL2, and EGR2 motifs at five indicated time points during NK cell differentiation. ATAC-seq signals across all these motif binding sites in the genome were aligned on the motif and averaged. E. Normalized ATAC-seq profiles of the ETS1 (left) and GATA3 (right) gene loci at indicated time points during NK cell differentiation. ATAC-seq signals were obtained from the UCSC Genome Browser. The gray blocks indicate genomic loci that are gradually more accessible during NK cell differentiation. The locations of binding motifs for FOSL2 and EGR2 are depicted at the bottom. RUNX, Runt-related transcription factor; ETS1, ETS proto-oncogene 1; ETV, ETS variant transcription factor; ZNF416, zinc finger protein 416; SCL, TAL bHLH transcription factor 1; E2A, E2A immunoglobulin enhancer binding factors E12/E47; CEBP, CCAAT enhancer-binding protein; FOSL2, FOS-like 2; ERRA, estrogen related receptor alpha; EBF1, EBF transcription factor 1; STAT, signal transducer and activator of transcription; IRF, interferon regulatory factor; MEF2C, myocyte enhancer factor 2C; FOXO1, forkhead box O1; TFAP2C, transcription factor AP-2 gamma; EOMES, eomesodermin; EGR2, early growth response 2; GATA3, GATA binding protein 3; ATF3, activating transcription factor 3; BACH2, BTB domain and CNC homolog 2; ARNTL, aryl hydrocarbon receptor nuclear translocator like; BIN1, bridging integrator 1; NR4A1, nuclear receptor subfamily 4 group A member 1; ERG, ETS transcription factor ERG; CDC42EP3, CDC42 effector protein 3; KLF5, Kruppel like factor 5; CEBPB, CCAAT enhancer binding protein beta; ZNF467, zinc finger protein 467; TTK, TTK protein kinase; MAZ, MYC associated zinc finger protein; ZBTB18, zinc finger and BTB domain containing 18; ZNF711, zinc finger protein 711; NFE2, nuclear factor, erythroid 2; GFI1B, growth factor independent 1B transcriptional repressor; SMAD3, SMAD family member 3; TGIF1, TGFB induced factor homeobox 1; RXRA, retinoid X receptor alpha; MITF, melanocyte inducing transcription factor; TFBS, TF binding site.

Figure 4

Function of EGR2 and FOSL2 during NK cell differentiation A. The gating strategy for cultured cells transduced with lentiviruses expressing shFOSL2, shEGR2, or control shRNA via detection of GFP expression. B. Knockdown efficiency of FOSL2 and EGR2 with shFOSL2 and shEGR2. C. Flow cytometric analysis of CD56⁺ cells in the GFP⁺ cells at day 28 (top) and day 35 (bottom). GFP⁺ cells refer to cultured cells successfully transfected with lentiviruses expressing control shRNA (vector with a meaningless fragment), shFOSL2, or shEGR2. D. Quantification of CD56⁺ cells in the GFP⁺ cell population at day 28 (left) and day 35 (right). n = 5. E. Flow cytometric analysis of CD56⁺ in the GFP⁻ cells at day 28 (top) and day 35 (bottom). GFP⁻ cells refer to cultured cells not successfully transfected with lentiviruses expressing control shRNA, shFOSL2, or shEGR2. F. Quantification of CD56⁺ cells in the GFP⁻ cell population at day 28 (left) and day 35 (right). n = 5. G. Quantification of CD11b⁺ cells in the CD56⁺ cell population at day 35 (n = 5). Data are presented as the mean ± SEM. *, P < 0.05; **, P < 0.001; ***, P < 0.0005; ****, P < 0.0001. P values were estimated from Student’s t-test. SSC-A, side scatter area; FSC-A, forward scatter area; NIR, near infrared.

Figure 5

Transcriptional regulatory network during NK cell differentiation A.–E.Cis-regulatory networks between TFs (nodes) enriched in ≥1 gene set and specifically expressed (FC > 1.5) at day 7 (A), day 14 (B), day 21 (C), day 28 (D), and day 35 (E). Nodes represent TFs with gene expression levels and TF enrichment scores at day 7, day 14, day 21, day 28, and day 35 (from left to right). The arrow at the edge from node A to node B indicates that TF A regulates TF B by binding to the promoter region of TF B. The color of each node indicates the expression level of the gene encoding that TF, and the size of the circle represents the significance of the motif enrichment according to the P value. The types of edges indicate the Pearson correlation between the gene expression profiles of the connected TFs: positively correlated (PCC > 0.4); negatively correlated (PCC < −0.4); no correlation (−0.4 ≤ PCC ≤ 0.4, dashed line). FC, fold change; PCC, Pearson correlation coefficient.