Enhancer RNA Expression in Response to Glucocorticoid Treatment in Murine Macrophages

Abstract

Glucocorticoids are potent anti-inflammatory drugs; however, their molecular mode of action remains complex and elusive. They bind to the glucocorticoid receptor (GR), a nuclear receptor that controls gene expression in almost all tissues in a cell type-specific manner. While GR's transcriptional targets mediate beneficial reactions in immune cells, they also harbor the potential of adverse metabolic effects in other cell types such as hepatocytes. Here, we have profiled nascent transcription upon glucocorticoid stimulation in LPS-activated primary murine macrophages using 4sU-seq. We compared our results to publicly available nascent transcriptomics data from murine liver and bioinformatically identified non-coding RNAs transcribed from intergenic GR binding sites in a tissue-specific fashion. These tissue-specific enhancer RNAs (eRNAs) correlate with target gene expression, reflecting cell type-specific glucocorticoid responses. We further associate GR-mediated eRNA expression with changes in H3K27 acetylation and BRD4 recruitment in inflammatory macrophages upon glucocorticoid treatment. In summary, we propose a common mechanism by which GR-bound enhancers regulate target gene expression by changes in histone acetylation, BRD4 recruitment and eRNA expression. We argue that local eRNAs are potential therapeutic targets downstream of GR signaling which may modulate glucocorticoid response in a cell type-specific way.

Keywords: enhancer RNA; glucocorticoid receptor; inflammation; macrophages; tissue specificity; transcription.

Figure 1

Nascent transcription from GR binding sites in activated macrophages. (A) Strategy for eRNA identification based on transcript expression (left, 4sU labeling) and GR binding sites (GBS) (right, ChIP-seq). Brackets show the number of reads or GBSs at each step. Bar plots display the overlap of nascent transcripts (left) or GR binding sites (GBS) with nascent transcription (right) with indicated genomic features. eRNA—enhancer RNA, eGBS—eRNA expressing GBS, iGBS—intergenic GBS, iRNA—intergenic RNA, TPMs—transcripts per million, TSS—transcriptional start site, TTS—transcriptional termination site, UTR—untranslated region. (B) Venn diagram indicating the overlap of iRNAs identified by the transcript-based approach with intergenic GBSs (left). The proportion of iGBSs with and without eRNA expression characterized by the GBS-based approach is shown on the right. (C) Venn diagram of the eGBS overlap between both strategies. One-thousand six-hundred and eighty-four RNAs identified by the GBS-centric method were used for further analyses. (D) Motif occurrence among intergenic GBSs (FIMO). Motifs enriched at iGBS with or without eRNA expression (see Supplementary Figure S1D) as well as motifs associated with promoters were included. (E) Overlap of eGBSs with classical enhancer features such as openness (ATAC-seq), H3K27ac and H3K4me1 (ChIP-seq). Horizontal bars: intersection size, vertical bars: peak set size. (F) Mean genome browser tracks for Tsc22d3 (Gilz) and Ccr2/Ccr3 showing nascent transcription (4sU, n = 3), GR binding (GR ChIP-seq, n = 2), open chromatin (ATAC-seq, n = 4), H3K27ac (ChIP-seq, n = 2) and H3K4me1 (ChIP-seq, n = 3). Data are from primary murine macrophages treated with LPS and Dexamethasone.

Figure 2

GR-mediated eRNA expression patterns are tissue specific. (A) Venn diagram for overlapping eRNA-expressing GR binding sites (eGBS) and their nearest target genes in liver and macrophages (BMDMs). (B) Functional annotation for the genes associated with eGBSs specific to macrophages or livers (top 10 non-redundant biological processes q-value < 0.001, with examples). neg. = negative, pos. = positive, prod. = production, reg. = regulation, resp. = response (C) Heatmap of GR ChIP-seq, open chromatin (macrophage ATAC-seq, liver DHS) and nascent transcription (macrophage 4sU, liver GRO-seq). Heatmaps display the mean signal intensity across biological replicates. Coverage plots show the median signal per subset (blue=macrophage-specific, red=liver-specific, purple=shared). (D) Differential XSTREME motif analysis [68] of macrophage-specific, shared and liver-specific eGBSs. E = E-value; # = number of sites. (E) Example genome browser tracks for macrophage-specific (Ccr2/Ccr3), shared (Fos) and liver-specific (G6pc) eGBSs showing mean GR ChIP-seq signal, nascent transcription profiled by 4sU-seq in macrophages and GRO-seq in livers and open chromatin (macrophage ATAC-seq, liver DHS) (GR ChIPseq n = 2, ATAC-seq n = 4, DHS n = 1, 4sU n = 3, GRO-seq n = 1).

Figure 3

Glucocorticoids regulate differential eRNA expression. (A) Volcano plot displaying intergenic GBSs showing differential eRNA expression in LPS + Dex versus LPS-treated macrophages. Each dot represents one eRNA-producing GBS. Selected eRNAs are named by the closest TSS. (B) “Biological process” overrepresentation analysis of genes associated with Dex-induced or Dex-repressed eRNAs. Top 10 non-redundant processes with a q-value cutoff of 0.05 are shown and examples labelled. (C) Mean genome browser tracks of GR ChIP-seq (LPS + Dex; L + D, n = 2) and 4sU-seq coverage for LPS and LPS + Dex treated macrophages (n = 3). Shades indicate positions of qRT-PCR primers for D. (D) RT-qPCR for Tsc22d3 (Gilz) and Ccr3 and their eRNAs in macrophages treated with vehicle, LPS or LPS + Dex (n = 3). *** p < 0.001, ** p < 0.01, * p < 0.05 (Students t-test) (E) Differential motif enrichment analysis for GR-bound sites with activated eRNA expression over those with reduced expression. (24: number of sites with motif matches).

Figure 4

eRNA expression from GR binding sites correlates with H3K27ac and BRD4 recruitment. (A) Correlation of pairwise comparisons of histone marks and co-regulators between LPS and LPS + Dex stimulated macrophages at eGBSs. (B) Distribution of log2 fold-changes in chromatin features or co-regulators as Violin plot. eGBSs are grouped by significantly induced, repressed or non-significantly changed (adj. p-value > 0.05, Figure 3A). Each dot represents one enhancer. Red denotes significance (p < 0.1) (ChIP-seq, Wilcoxon-Mann-Whitney test, **** p < 0.00001, ** p < 0.001) (C) Mean heatmaps of ChIP-seq signals in LPS or LPS + Dex treated macrophages, grouped by eGBSs with induced, repressed or non-changing eRNA expression. eGBSs are sorted in descending order of the log2-fold change. Coverage plots above summarize median ChIP-seq densities. Colors as indicated in B. (D) Genome browser tracks of induced (Tsc22d3) and repressed (Ccr3) eRNAs, means from biological replicates. Red shades mark eGBSs with alterations of all displayed features. Red arrows point at significant signal changes at eGBSs (blue shades). (ATAC-seq n = 4; 4sU-seq n = 3; H3K27ac, RNAP2, BRD4, SETD1A, CXXC1 and GR n = 2 ChIP-seq).