Bromodomain-containing Protein 4 regulates innate inflammation via modulation of alternative splicing

Abstract

Introduction: Bromodomain-containing Protein 4 (BRD4) is a transcriptional regulator which coordinates gene expression programs controlling cancer biology, inflammation, and fibrosis. In the context of airway viral infection, BRD4-specific inhibitors (BRD4i) block the release of pro-inflammatory cytokines and prevent downstream epithelial plasticity. Although the chromatin modifying functions of BRD4 in inducible gene expression have been extensively investigated, its roles in post-transcriptional regulation are not well understood. Given BRD4's interaction with the transcriptional elongation complex and spliceosome, we hypothesize that BRD4 is a functional regulator of mRNA processing.

Methods: To address this question, we combine data-independent analysis - parallel accumulation-serial fragmentation (diaPASEF) with RNA-sequencing to achieve deep and integrated coverage of the proteomic and transcriptomic landscapes of human small airway epithelial cells exposed to viral challenge and treated with BRD4i.

Results: We discover that BRD4 regulates alternative splicing of key genes, including Interferon-related Developmental Regulator 1 (IFRD1) and X-Box Binding Protein 1 (XBP1), related to the innate immune response and the unfolded protein response (UPR). We identify requirement of BRD4 for expression of serine-arginine splicing factors, splicosome components and the Inositol-Requiring Enzyme 1 IREα affecting immediate early innate response and the UPR.

Discussion: These findings extend the transcriptional elongation-facilitating actions of BRD4 in control of post-transcriptional RNA processing via modulating splicing factor expression in virus-induced innate signaling.

Keywords: BRD4; IFRD1; RSV (respiratory syncytial virus); alternative splicing; innate inflammatory response; unfolded protein response.

Figure 1

BRD4 inhibition alters alternative splicing in the context of airway viral infection. (A) Volcano plot indicating differential isoform fraction (dIF) and isoform switch q-value of isoform switching events with predicted gene-level consequences. Non-significant changes ( $Q_{Val}>$ .05 or |dIF| $<0.1$ are plotted in grey. Significant changed transcripts also found to be significantly changed due to RSV-infection are plotted in red. Significant switch events only detected due to BRD4 inhibition are plotted in blue. (B) Gene Ontology Biological Process Over-representation analysis of BRD4i-mediated splicing events. Count refers to the number of alternatively spliced genes detected in the significantly over-represented pathway. (C) Transcript-level consequences of alternatively spliced genes as induced by RSV or BRD4 inhibition.

Figure 2

BRD4 inhibition induces isoform switching in key innate inflammation co-regulators. Transcript maps, RNA-seq quantitation in Transcripts per-Kilobase Million (TPM), and isoform usage of detected transcripts belonging to (A) XBP1, (B) IFRD1, and (C) ATP11A. Mapped transcripts are plotted in compressed genomic coordinates and color coded according to gene body and coding status. Barplots represents the mean of n=4 replicates per group. Error bars indicate the 95% confidence interval. Transcripts with less than 5% of total detected transcript abundance were omitted from this figure. * $P_{Adj}<0.05$ ; *** $P_{Adj}<0.001$ .

Figure 3

Validation of BRD4i-induced alternative splicing. (A) RT-PCR Amplification of IFRD1-204 and IFRD1-Total. (B) Gel Densitometry quantitation of IFRD1-204/IFRD1-Total Ratio. (C) q-RT-PCR quantitation of XBP1s/XBP1-Total Ratio. (D, E) Protein-level (diaPASEF) quantitation of ATP11A and IFRD1 abundance. Boxplots represent median and interquartile ranges of n=3 biological replicates per experimental group. Open circles represent imputed data points. *P_Adj< 0.05; **P_Adj< 0.01; ***P_Adj< 0.001.

Figure 4

IFRD1 knockdown reduces NF- B-mediated expression in hSAECs. Timecourse q-RT-PCR experiment measuring (A) IL6, (B) (NFKBIA), and (C) (TNFAIP3) gene expression in hSAECs depleted of IFRD1. Data represents the geometric mean fold change ± standard error of n=3 biological replicates per experimental group. (D) Western blot and quantitation of RelA Lysine-310 Acetylation (K310Ac) as a function of IFRD1 knockdown by siRNA in A549 cells. Barplot represents the mean ± standard error of n=3 biological replicates. *P_Adj< 0.05; ***P_Adj< 0.001.

Figure 5

BRD4 regulates the transcript- and protein-level abundance of core splicing factors. (A) diaPASEF quantitation of core splicing factor protein abundance. Open circles indicate imputed data points. (B) q-RT-PCR quantitation of transcript-level splicing factor abundances. Boxplots represent median and interquartile range of n=3 biological replicates per experimental group. *P_Adj < 0.05; **P_Adj < 0.01; ***P_Adj < 0.001.

Figure 6

BRD4 controls the transcript- and protein-level abundance of IRE1α. (A) RNA sequencing read counts of summarized IRE1α transcripts. (B) Protein-level (diaPASEF) quantitation of IRE1α protein. Open circles represent imputed data points. (C) Immunoblot and Densitometric Quantitation of IRE1α from non-infected hSAECs treated with the BRD4 inhibitor ZL0454. (D) ChIP-PCR analysis of BRD4 binding to the IRE1α promoter. (E) ChIP-PCR analysis of BRD4 binding to the XBP1 promoter. Boxplots represent median and interquartile ranges of n=4 (RNA-seq) or n=3 (diaPASEF, Immunoblot, ChIP) biological replicates per experimental group. *P_Adj < 0.05; ***P_Adj < 0.001.