Chronic airway epithelial hypoxia exacerbates injury in muco-obstructive lung disease through mucus hyperconcentration

Abstract

Unlike solid organs, human airway epithelia derive their oxygen from inspired air rather than the vasculature. Many pulmonary diseases are associated with intraluminal airway obstruction caused by aspirated foreign bodies, virus infection, tumors, or mucus plugs intrinsic to airway disease, including cystic fibrosis (CF). Consistent with requirements for luminal O₂, airway epithelia surrounding mucus plugs in chronic obstructive pulmonary disease (COPD) lungs are hypoxic. Despite these observations, the effects of chronic hypoxia (CH) on airway epithelial host defense functions relevant to pulmonary disease have not been investigated. Molecular characterization of resected human lungs from individuals with a spectrum of muco-obstructive lung diseases (MOLDs) or COVID-19 identified molecular features of chronic hypoxia, including increased EGLN3 expression, in epithelia lining mucus-obstructed airways. In vitro experiments using cultured chronically hypoxic airway epithelia revealed conversion to a glycolytic metabolic state with maintenance of cellular architecture. Chronically hypoxic airway epithelia unexpectedly exhibited increased MUC5B mucin production and increased transepithelial Na⁺ and fluid absorption mediated by HIF1α/HIF2α-dependent up-regulation of β and γENaC (epithelial Na⁺ channel) subunit expression. The combination of increased Na⁺ absorption and MUC5B production generated hyperconcentrated mucus predicted to perpetuate obstruction. Single-cell and bulk RNA sequencing analyses of chronically hypoxic cultured airway epithelia revealed transcriptional changes involved in airway wall remodeling, destruction, and angiogenesis. These results were confirmed by RNA-in situ hybridization studies of lungs from individuals with MOLD. Our data suggest that chronic airway epithelial hypoxia may be central to the pathogenesis of persistent mucus accumulation in MOLDs and associated airway wall damage.

Figure 1:. Evidence of airway epithelial hypoxia in mucus-plugged airways from patients and HBE cells exposed to chronic hypoxia.

(A) Representative RNA-ISH images demonstrating EGLN3 mRNA localization in the airways from control (CTRL) and mucus-plugged airways from cystic fibrosis (CF), non-CF bronchiectasis (NCFB), primary ciliary dyskinesia (PCD), and COVID-19 excised or autopsy lungs. Red dots indicate positive signals, and hematoxylin staining was used for counterstaining. Scale bar = 100μm. The arrows point to the area that is enlarged in the inset. (B) Quantification of EGLN3 mRNA expression defined by the fraction of epithelial surface area containing staining for EGLN3 RNA-ISH signals in control (n=4), muco-obstructive lung disease airways (MOLDs; n=11, square: CF, triangle: NCFB, and diamond: PCD), and COVID-19 (n=5) lung. Each value shown is the mean of 5 randomly selected airways or lungs. Dotted line = group mean. (C) H&E and Alcian Blue PAS (AB-PAS) stained images of normoxic (N) and chronically hypoxic (CH; 1% O₂ for 5 days) HBE cultures. Scale bar = 20μm. (D) LDH cytotoxicity assays in normoxic (N) and chronic hypoxic (CH) HBE cultures. Each value represents duplicates from n=3 different HBE cultures grown at separate times. (E) Transepithelial electrical resistance (R_t) measurements in normoxic (N), acute hypoxic (AH), and chronically hypoxic (CH) HBE cultures. Each value represents the mean value of quadruplicates from n=15 (N), n=3 (AH), and n=12 (CH) different HBE cultures grown at separate times. (F) Pie charts depicting the cell-type composition (19) determined by scRNA-seq of HBE cells cultured under normoxia, hypoxia for 6h, or hypoxia for 5 days (5d). (G) Representative whole-mount images (left panels) showing staining of a secretory cell marker (SCGB1A1 (CCSP); magenta) and a ciliated cell marker (α-tubulin; green) for normoxic vs chronic hypoxic HBE cultures, and the quantification results of %SCGB1A1 positive and %α-tubulin positive areas (right panels). Each value represents the mean value from n=5 different HBE cultures grown at separate times. Nuclei are stained with DAPI (blue), and cell borders are stained with phalloidin (white). Scale bar = 30μm. Statistical analysis was performed using Kruskal-Wallis test followed by Dunn’s test (B), Wilcoxon test (D) and (G), and one-way ANOVA followed by Dunnett multiple comparison test (E). *: p<0.05, **: p<0.01, ns: not significant. Data are means ±SEM.

Figure 2:. HBE characteristics under normoxic, acute, and chronic hypoxia.

(A) Correlation analysis of differentially expressed pathways (MSigDB C2 curated genes) derived from scRNA-seq dataset for 6h vs 5d hypoxia HBE cultures. E-score=-log₁₀(p.adj) x Normalized enrichment score. Key pathways cited in the text are highlighted (full results available in Data File S1). Red = significantly (p.adj < 0.05) up-regulated pathways; Blue = significantly (p.adj<0.05) down-regulated pathways. (B) Violin plots depicting expression value within the six major cell-type clusters of epithelial cells for mitochondrially encoded phosphoglycerate kinase 1 (PGK1), vascular endothelial growth factor A (VEGFA), Solute Carrier Family 1 Member 1 (SLC1A1; also known as GLUT1), IGF-binding protein-3 (IGFBP3), NADH-ubiquinone oxidoreductase chain 1 (MT-ND1), proteasome activator subunit 1 (PSME1), and EGLN3 expression in normoxic and hypoxic (6h and 5d) HBE cultures. Basal cell includes Basal.1, 2, 3, and 4 clusters, Secretory cell includes Secretory.1, 2, 3, 4, 5, 6, and 7 clusters, Ciliated cell includes FOXN4⁺ ciliated, Ciliated.1, 2, and 3 clusters, and Cycling cell includes Cycling.1 and 2 clusters (see Fig. S5). (C) Representative images of p16^INK4A, p21, and Ki67 immunostaining for HBE cultures under normoxia (N) and chronic hypoxia (CH) and (D) quantification (n=4–5/group). Red arrows indicate positively stained cells. Each value represents the mean value of quantitation from four randomly selected regions per membrane. Statistical analyses were performed using two-tailed paired t-tests. *: p<0.05, **: p<0.01, ns: not significant. Data are means ±SEM.

Figure 3:. Effects of chronic hypoxia on HBE mucus hydration and ion transport.

(A) Percent mucus solids content, an index of hydration of apical mucus, as measured from HBE cells after 5d culture under normoxia (N) vs chronic hypoxia (CH). Each value represents the mean of triplicates from n=4 different HBE cultures grown at separate times. (B) Effect of chronic hypoxia on MUC5B expression and protein secretion. MUC5B (left graph), MUC5AC (right graph) mRNA expression normalized with B2M (n=5). (C) Representative image and densitometry analysis of secreted MUC5B from HBE cells under 5d normoxic vs hypoxic conditions from n=8 different HBE cultures grown at separate time by western blotting. (D and E) Measurement of ion transport activities in normoxic vs hypoxic HBE cells in Ussing chambers. Experiments were conducted in HBE cultures utilizing protocols for both acute (~2h) (AH) and chronic (5d) hypoxia (CH). Indices of basal Na⁺ transport rates (basal Isc (D, left graph), amiloride-sensitive Isc (D, middle graph)) and basal (post-amiloride Isc (D, right graph)) and stimulated Cl⁻ transport (forskolin-stimulated (E, left graph), CFTR 172-inhibited (E, right graaph)) are depicted. Each value represents the mean value of quadruplicates from n=13 (N), n=5 (AH), and n=9 (CH) different HBE cultures grown at separate times. (F) Expressions of ENaC gene subunits in normoxic vs chronically hypoxic HBE cells (n=13). (G) Representative images of Western blotting of surface γ-ENaC by biotinylation for normoxic and chronically hypoxic HBE cells and quantification of results (n=4). RPL13a is as an internal control of whole cell lysate. γ-ENaC signal intensity was normalized with RPL13a signal intensity. (H) Basal and Amiloride-sensitive short-circuit current (Isc) in negative control (NC) guide RNA and SCNN1G knockout HBE cells measured in Ussing chambers. Each value represents the mean value of quadruplicates from n=3. The relative gene expression was normalized with B2M using the 2^−ΔΔCT method. Statistical analysis was performed using two-tailed paired t-test (A), (B), (C), (F), and (G), one-way ANOVA followed by Dunnett multiple-comparison analysis (D and E), and two-way ANOVA followed by Tukey multiple-comparison analysis (H). *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001, ns: not significant. Data are means ±SEM.

Figure 4:. HIF pathways contribute to the upregulation of sodium absorption in HBE during chronic hypoxia.

(A) Representative image of HIF1α Western blot of HBE cells cultured under hypoxia (1% O₂) for 24–240h compared to normoxic controls (N). TATA-Box Binding Protein (TBP) is used as an internal loading control. (B) Representative images of HIF1α immunostaining for HBE cells cultured under normoxic (N) vs chronically hypoxic (CH) conditions (left panels). Positively stained cells were stained with dark brown and Nuclear Fast Red was used for counterstaining. The fraction of HIF1α positive nuclei was quantified by counting 300 nuclei per independent HBE cultures (right panel). (C) Basal and amiloride-sensitive short-circuit current (Isc) in negative control (NC) guide and HIF1A knockout HBE cultures measured in Ussing chambers. Each value represents the mean value of quadruplicates from n=3 individual donor-derived cultures. (D) SCNN1G mRNA expression in negative control (NC) guide RNA and HIF1A knockout (KO) HBE cultures. (n=6 donors) (E) Basal and amiloride-sensitive short-circuit current (Isc) in the negative control (NC) guide and EPAS1 (HIF2α) knockout HBE cultures measured in Ussing chambers. Each value represents the mean value of quadruplicates from n=3 individual donor-derived cultures. (F) SCNN1G mRNA expression in negative control (NC) guide RNA and EPAS1 knockout (KO) HBE cultures (n=4). (G) SCNN1G expression in response to FG-4592 compared to chronic hypoxia (CH) and trichostatin A (TSA) under basal conditions or chronic hypoxia. N = normoxia, N+TSA = normoxia + trichostatin A (TSA, 400nM), FG-4592 = normoxia + FG-4592 (PHD inhibitor, 20μM), CH = hypoxia 5d, CH+TSA = hypoxia 5d + TSA (400nM), n=8. Statistical analysis was performed using two-tailed paired t-test (B), and two-way ANOVA followed by Tukey multiple-comparison analysis (C), (D), and (G). *: p<0.05, **: p<0.01, ***: p<0.001, ****: p<0.0001, ns: not significant. Data are means ±SEM.

Figure 5:. EGLN3 and SCNN1G expression in healthy and diseased human lung tissue.

(A) Representative RNA-ISH images demonstrating EGLN3 (red) and SCNN1G (green) mRNA localization in the airways from non-disease (CTRL), cystic fibrosis (CF), non-CF bronchiectasis (NCFB), primary ciliary dyskinesia (PCD), and COVID-19 excised lungs. Nuclei are stained with DAPI (blue). Differential interference contrast (DIC) scan was used to enhance the tissue shape. Scale bar = 20μm. (B) Representative RNA-ISH images demonstrating SCNN1G mRNA localization in the airways from CTRL, CF, NCFB, PCD and COVID-19 lungs. Red dots indicate positive signals and hematoxylin was used for counterstaining. (C) Quantification of SCNN1G positive stained area per airway epithelial areas in excised lungs. Each value shown is the mean of 5 randomly selected airways per lung. (Note log₁₀ scale, Dotted line = group mean, n=5 for control, n=8 for MOLDs [square: CF, triangle: NCFB, and diamond: PCD], n=5 for COVID-19). Statistical analysis was performed using Kruskal-Wallis test. *: p<0.05.

Figure 6:. Pathobiologic responses of human airway epithelia to chronic hypoxia.

(A) Up-regulation of selected pathways derived from 2^nd batch of bulk RNA-seq analysis of normoxic vs 5d hypoxic HBE cultures (n=8; full pathway data presented in Data File 2). (B) Log2 fold-change of representative genes from bulk RNA-seq analyses of normoxic vs 5d hypoxic HBE cultures categorized by inflammation (IL1B, IL1A, CXCL8, and IL6), host defense (PIGR, SCGB1A1, SLPI, and DEFB4A), and airway remodeling or angiogenesis (COL1A1, TGFB1, MMP1, MMP7, MMP9, PGF, and VEGFA). (C) Gene expression patterns per cell type from scRNA-seq data of normoxic vs chronically hypoxic HBE cells for selected genes. N=normoxic, AH=acute hypoxia (6h); CH = chronic hypoxia (5d). (D) Representative images of RNA-ISH detection of VEGFA in control, CF, NCFB, and PCD. Scale bar = 20μm. Red arrowheads indicate basal cells that locate around the basal membrane.