IRF1 mediates airway epithelial innate memory through priming BRD4- EP300 activity

Sentinel cells of the airway epithelium are repetitively exposed to RNA virus infections. Here, the production of pathogen-associated molecular patterns (PAMPs) activates the innate immune response (IIR), inducing antiviral defenses, activating adaptive immunity and stimulating epithelial repair. The question whether repeated exposures produce innate "memory", where heightened inflammatory and injury/repair responses are produced upon subsequent exposure, is unknown.

We establish a reproducible model of innate memory in tumor protein (TP63)+/keratin (KRT)5+ human small airway basal epithelial cells by repetitive activation of the Toll like receptor (TLR3) pathway (a.k.a. "training"). Mechanisms of IRF1-dependent target gene remodeling are investigated by CRISPR/Cas9 knockdown, chromatin immunoprecipitation, co-immunoprecipitation and through use of selective inhibitors of the BRD4 chromatin remodeling complex (CRC).

We observe "training" increases basal expression of a network of intrinsic IIR genes, reducing Respiratory Syncytial Virus (RSV) replication upon subsequent exposure. In addition, trained hSAECs enter an epithelial mesenchymal plasticity (EMP) program, characterized by cell shape changes, increased expression of cytoplasmic vimentin and core mesenchymal regulators. Mechanistically, we find that Interferon Regulatory Factor 1 (IRF1) is required for innate memory because IRF1-deficient cells fail to activate basal intrinsic IIR. In TLR3 activation, IRF1 is incorporated into the bromodomain containing protein 4 (BRD4) chromatin remodeling complex (CRC), a complex containing the histone H3 acetyltransferase, EP300/p300. In acute TLR3 activation, the binding of BRD4 is increased on intrinsic IIR promoters, yet this binding is unaffected by training. Instead, we find training induces EP300 binding and increases H3K27 acetylation (H3K27ac) at these regions. Small molecule inhibitors show that BRD4 is required for establishing innate memory and H3K27ac accumulation. Finally, we find that training enhances BRD4-EP300 interactions and that memory-induced H3K27ac formation is EP300-dependent.

We conclude that hSAECs acquire innate memory through IRF1 coupling with the BRD4-EP300 CRC to activate intrinsic IIR and cell-state transition genes. Our data suggest that innate training is the consequence of dynamic BRD4 CRC interactions through enhanced EP300 recruitment that primes transcriptional responsiveness.