Stromal Interferon Regulatory Factor 3 Can Antagonize Human Papillomavirus Replication by Supporting Epithelial-to-Mesenchymal Transition

Abstract

Epithelia contribute to the innate immune system through barrier formation and through signaling to immune cells. When the barrier is breached, epithelial cells undergo epithelial-to-mesenchymal transition (EMT) as part of the wound healing process. EMT is largely directed by signals from the stromal microenvironment, including transforming growth factor beta (TGFβ1), and antagonizes normal epithelial differentiation. How EMT and innate immunity may be connected molecularly has not been explored, although both processes are likely to occur simultaneously. Keratinocytes are the host cell type for human papillomaviruses (HPV), which can induce EMT in certain conditions but also depend on differentiation for their replication. We previously found that the innate immune factor interferon regulatory factor 3 (IRF3) inhibits epithelial differentiation and reduces the expression of HPV16 late genes. Here we report that IRF3 in the stroma compartment promotes an EMT-like pattern of gene expression in an HPV16-containing epithelium. The depletion of stromal IRF3 resulted in the downregulation of TGFβ1-related signaling in both the stroma and epithelium. IRF3 binds to the TGFB1 promoter in human foreskin fibroblasts and is necessary for TGFB1 mRNA production. Because an EMT-like state is unfavorable for differentiation-dependent HPV16, we observed that all EMT markers examined were reduced in the presence of episomal HPV16. Together, we show that stromal IRF3 can disrupt epithelial differentiation and act as an anti-HPV factor through the regulation of EMT, linking wound healing and immunity.

Keywords: epithelial-to-mesenchymal transition (EMT); human papillomavirus type 16 (HPV16); interferon regulatory factor 3 (IRF3); transforming growth factor beta 1 (TGFβ1).

Figure 1

Stromal IRF3 suppresses HPV16 L1 transcript levels. (a) Preparation of organotypic (raft) cultures. HFFs are imbedded in a collagen matrix and keratinocytes are seeded on the surface. The construct is lifted onto a wire grid, exposed to the air on top, and fed by tissue culture medium from the bottom. During incubation, the cells proliferate, stratify, and differentiate. RNAs were isolated from either the stromal (fibroblast-containing) layer or the epithelial (keratinocyte-containing) layer of raft cultures made from HFFs in which IRF3 was knocked down or HFFs containing non-target control shRNA (NTC). (b) Representative H&E images of NTC and IRF3 KD raft cultures. RNAs were subjected to RT-qPCR analysis to measure levels of (c) stromal IRF3 or (d) epithelial L1 RNAs. N = 3. *** = p < 0.001.

Figure 2

Stromal IRF3 promotes an EMT-like gene expression pattern in the epithelium. (a) Hypothesis: IRF3 supports expression of a paracrine factor (X) whose downstream signaling suppresses HPV late gene expression. Knockdown of IRF3 would reduce levels of X, resulting in increased HPV late gene expression. Total RNA-seq was performed on RNAs from both the epithelial and stromal compartments of rafts with or without IRF3 knockdown in the stromal fibroblasts (N = 3 each). Keratinocytes contained episomal HPV16. (b) Pathway analysis from epithelial RNAs showing a subset of pathways regulated by stromal IRF3. (c) RT-qPCR analysis of selected RNAs from the epithelial layer. N = 6. (d) Raft cultures were processed, imbedded, sectioned, and stained using antibodies against E-cadherin (top, green) or β-catenin (bottom, green). Intensity of staining was quantified (right). N = 4. NS = not significant; * = p < 0.05; ** = p < 0.01; *** = p < 0.001.

Figure 3

Stromal IRF3 promotes expression of cytokines and growth factors in the stroma. RNAs were isolated from the stromal layer of rafts containing HFFs with IRF3 knocked down (or controls) and subjected to RT-qPCR analysis. N = 4. * = p < 0.05.

Figure 4

IRF3 is necessary for TGFβ1 expression in isolated fibroblasts. HFFs were transfected with either IRF3-specific siRNAs or scrambled controls. (a) Total protein was harvested from transfected cells and immunoblotted for IRF3 (top) and TGFβ1 (bottom) with GAPDH as a loading control. Mean band intensities are indicated below the blots (N = 5). (b) Total RNAs from transfected cells were subjected to RT-qPCR analysis for TGFB1 and FGF7 transcripts (N = 6). (c) (Top) putative IRF3 binding sites in the TGFB1 promoter. (Bottom) Chromatin was prepared from HFFs and subjected to chromatin immunoprecipitation, followed by PCR for each putative IRF3 binding site. (d) Chromatin was prepared from untreated HFFs grown at either high or low density and subjected to chromatin immunoprecipitation using antibodies targeting IRF3. PCR was performed on DNA isolated from the immunoprecipitates using primers targeting Site 3 (top). N = 3. (e) HFFs were grown at either high or low density, fixed, and stained with antibodies against IRF3 (green), with DAPI counterstain (blue). N = 3. NS = not significant; * = p < 0.05; ** = p < 0.01.

Figure 5

TGFB1 gene expression does not respond to the canonical IRF3 pathway. HFFs were treated with (a) poly I:C (PIC), (b) IFNβ (25 U/mL), or ruxolitinib (10 µM) for 24 h. Total RNAs were subjected to RT-qPCR analysis using the indicated primers. N = 3. NS = not significant; * = p < 0.05; *** = p < 0.001.

Figure 6

TGFβ1 contributes to a network of growth factors in the stroma. RNAs were harvested from raft cultures containing HFFs with TGFB1 knocked down (or controls). (a) RNAs from the epithelium were subjected to RT-qPCR analysis to measure the levels of L1. N = 6. (b) Levels of the indicated transcripts in the stroma of TGFB1 KD or NTC rafts were measured by RT-qPCR analysis. N = 8. NS = not significant; * = p < 0.05; ** = p < 0.01; *** = p < 0.001.

Figure 7

HPV16 suppresses EMT-related factors in the epithelium and IRF3/TGFB1 in the stroma. Raft cultures containing either uninfected HFKs or HPV16-containing keratinocytes were harvested for RNA isolation or sectioning. (a) RNAs from the epithelial compartment were subjected to RT-qPCR analysis using the indicted primers. N = 4. (b) Fixed sections were stained using antibodies specific for filaggrin (green). The dotted line represents the basement membrane. (c) RNAs from the stromal fibroblasts were subjected to RT-qPCR using the indicated primers. N = 9. NS = not significant; * = p < 0.05; ** = p < 0.01; *** = p < 0.001.