Pulmonary fibrosis distal airway epithelia are dynamically and structurally dysfunctional

Abstract

The airway epithelium serves as the interface between the host and external environment. In many chronic lung diseases, the airway is the site of substantial remodeling after injury. While, idiopathic pulmonary fibrosis (IPF) has traditionally been considered a disease of the alveolus and lung matrix, the dominant environmental (cigarette smoking) and genetic (gain of function MUC5B promoter variant) risk factor primarily affect the distal airway epithelium. Moreover, airway-specific pathogenic features of IPF include bronchiolization of the distal airspace with abnormal airway cell-types and honeycomb cystic terminal airway-like structures with concurrent loss of terminal bronchioles in regions of minimal fibrosis. However, the pathogenic role of the airway epithelium in IPF is unknown. Combining biophysical, genetic, and signaling analyses of primary airway epithelial cells, we demonstrate that healthy and IPF airway epithelia are biophysically distinct, identifying pathologic activation of the ERBB-YAP axis as a specific and modifiable driver of prolongation of the unjammed-to-jammed transition in IPF epithelia. Furthermore, we demonstrate that this biophysical state and signaling axis correlates with epithelial-driven activation of the underlying mesenchyme. Our data illustrate the active mechanisms regulating airway epithelial-driven fibrosis and identify targets to modulate disease progression.

Fig. 1. The IPF distal airway is recapitulated in vitro and persists in the unjammed phase.

Representative distal human airways have increased KRT5 and MUC5B protein (a) when compared to their proximal airway or control counterpart. Primary human bronchiolar epithelial cells were isolated via (b) bronchial brushing (distal airways from 6th to 7th generation, generally <2 mm diameter) or CT-guided isolations (proximal > 10 mm diameter and distal < 2 mm diameter) and expanded from passage 0 to 2 prior to experimentation. Distal epithelia from control and IPF bronchial brushes were seeded onto porous Transwells and stained for (c) KRT5 or MUC5B, 4 or 14 days after establishing ALI demonstrating a similar phenotype to the in vivo patient airway. Control and IPF distal epithelia were imaged for 3 h during epithelial maturation. Utilizing particle image velocimetry (PIV – see methods), cellular speeds (d), mean-squared displacement (MSD) (e), and the overlap parameter (Q) (f), illustrate a biophysical divergence between control and IPF epithelia between days 6 and 8 of ALI. The perimeter-to-area ratio (q) of segmented epithelia (see methods) from control distal epithelia fell below the predicted jamming transition threshold (p₀ = 3.81) around day 8 of ALI, whereas IPF epithelial maintain an elongated cell body through day 14 (g) correlating with the persistence of a migratory, or unjammed state. Representative F-actin staining and speed maps (h) of control and IPF epithelia illustrate the biophysical differences in speed and cell shape temporally (scale bar = 100 $\mathrm{\mu }$m). Utilizing cells from CT-guided dissections, we seeded IPF epithelia from proximal or distal airways from within the same patient. Time-lapse imaging for 3 h demonstrated that this unjammed state, assessed via speed (i), MSD (j), and Q (k), and q (l) was specific to the distal airways of IPF patient samples with proximal airways entering the jammed phase ~day 9 of ALI. This phenotype was appreciated by representative F-actin-stained epithelia and speed maps (m) illustrating the biophysical differences of intra-airway epithelia. Shown: mean $\pm$ 95% confidence interval for n = 4 donors (control epithelia) and n = 3 (IPF epithelia) with $\ge$3 replicates per donor and 100 μm scale bar.

Fig. 2. EGFR and YAP activation induce unjamming in airway epithelia.

Representative distal airway epithelial cells were stained for KRT5 and MUC5B protein, and AREG mRNA in control or IPF (a) patient samples with increased AREG present in the IPF epithelium. Representative images of control or IPF distal epithelia stained for ERBB2, or YAP (b) demonstrate increases in ERBB2 protein and YAP nuclear localization in IPF distal epithelia. This difference in ERBB protein (c) and YAP nuclear localization (d) was most pronounced at later ALI days with protein/localization differences becoming most apparent as control distal epithelia entered the jammed state. These protein level changes were also accompanied by significant gene level changes in ERBB family receptors EGFR (e) and ERBB2 (f) as well as YAP target genes AREG (g) and CTGF (h) with control epithelial differences becoming pronounced on day 8, across the jamming transition. Control epithelia from distal airways on day 14 of ALI were treated with XMU (a small molecule YAP activator) or AREG (an EGFR ligand) to elicit unjamming of the static, quiescent epithelium. Upon treatment, XMU and AREG elicited a collective migratory phenotype as measured by root-mean-squared velocity V_RMS (i). This induced unjammed state was also quantified via cell shape changes with the perimeter-to-area (q) (j) and aspect ratios (k), cells were analyzed every 12 h after treatment was initiated. Representative speed maps taken every 16 h after treatment (l) and cell boundary segmentation at 48 h after treatment (m) demonstrated concordant migration and cell body elongation consistent with the induction of an unjammed state. Shown: mean $\pm$ 95% confidence interval for n = 4 donors (control epithelia) and n = 3 donors (IPF epithelia) with $\ge$2 replicates per donors and scale bars representing 100 μm.

Fig. 3. Inhibition of EGFR and YAP induce jamming in IPF distal epithelia.

Distal airway IPF cells were treated at day 14 for ALI with Verteporfin (YAP inhibitor) or AG1478 (EGFR inhibitor) (a) and filmed continuously for 48 h demonstrating a cessation of collective migration as assessed by V_RMS. Cell shape ratio measurements, perimeter-to-area (b) and aspect (c) were taken every 12 h after treatment initiation and illustrated the acquisition of a cobblestone-like phenotype which persisted for the entire duration of the measurement (up to 48 h). Representative speed maps taken every 16 h after treatment (d) and cell boundary segmentation at 48 h after treatment (e) demonstrate a concordant cessation in migration and acquisition of a cobblestone-like apical surface, color coded by aspect ratio (AR) of the epithelium consistent with a jammed state. IPF epithelia treated with Verteporfin or AG1478 had reduced YAP target gene expression AREG (f) and CTGF (g) as well as ERBB family receptor ERBB2 (h) and EGFR (i) 48 h after treatment, nearly back to control epithelial levels. In addition, gene expression of MUC5B (j) was also decreased after treatment. Inhibition of YAP or EGFR did not significantly impact barrier function as assessed via TEER (k) with marginal increased barrier function occurring after treatment. Shown: mean $\pm$ 95% confidence interval for n = 3 donors (control epithelia) and n = 4 donors (IPF epithelia) with $\ge$2 replicates and scale bars representing 100 μm.

Fig. 4. The signaling and biophysical state of epithelia direct fibroblast activation.

Primary human lung fibroblasts (HLF) were seeded on to a 3 kPa hydrogel and treated for 72 h with conditioned media from control distal epithelia (see methods) treated with either XMU or AREG (a) to induce unjamming. Representative immunofluorescence images of HLFs stained for F-actin (blue), $\mathrm{\alpha }$-SMA (yellow), or vimentin (red) showed minimal staining in the HLFs treated with control, untreated conditioned media (b). However, fibroblast activation became apparent in control epithelial conditioned media treatment of HLFs from XMU or AREG epithelia. These differences nearly reached the level of HLF activation from untreated IPF distal epithelial conditioned media in terms of cell number (c), vimentin (d), and $\mathrm{\alpha }$-SMA (e) quantity. Schematic of direct fibroblast inhibition (f) and representative immunofluorescence images of HLFs (g) demonstrating that treating HLFs with SB431542, AREG neutralizing antibody, or a combination of the two during treatment with IPF distal epithelial conditioned media showed an attenuation of the ability for IPF epithelial media to induce an intense pro-fibrotic phenotype indicated by cell number (h), vimentin (i), and $\mathrm{\alpha }$-SMA (j) approached the control-like phenotype. Utilizing media from IPF distal epithelia treated with Verteporfin, AG1478, or AREG neutralizing antibody (k) demonstrated the ability to prevent fibroblast activation (l). Cell number (m), vimentin (n), and $\mathrm{\alpha }$-SMA (o) were all restored to baseline, control-like levels when using conditioned media from inhibited IPF distal epithelia. Shown: mean $\pm$ 95% confidence interval for n = 4 donors with three replicates and scale bars representing 100 μm.