CFTR Correctors and Antioxidants Partially Normalize Lipid Imbalance but not Abnormal Basal Inflammatory Cytokine Profile in CF Bronchial Epithelial Cells

Abstract

A deficiency in cystic fibrosis transmembrane conductance regulator (CFTR) function in CF leads to chronic lung disease. CF is associated with abnormalities in fatty acids, ceramides, and cholesterol, their relationship with CF lung pathology is not completely understood. Therefore, we examined the impact of CFTR deficiency on lipid metabolism and pro-inflammatory signaling in airway epithelium using mass spectrometric, protein array. We observed a striking imbalance in fatty acid and ceramide metabolism, associated with chronic oxidative stress under basal conditions in CF mouse lung and well-differentiated bronchial epithelial cell cultures of CFTR knock out pig and CF patients. Cell-autonomous features of all three CF models included high ratios of ω-6- to ω-3-polyunsaturated fatty acids and of long- to very long-chain ceramide species (LCC/VLCC), reduced levels of total ceramides and ceramide precursors. In addition to the retinoic acid analog fenretinide, the anti-oxidants glutathione (GSH) and deferoxamine partially corrected the lipid profile indicating that oxidative stress may promote the lipid abnormalities. CFTR-targeted modulators reduced the lipid imbalance and oxidative stress, confirming the CFTR dependence of lipid ratios. However, despite functional correction of CF cells up to 60% of non-CF in Ussing chamber experiments, a 72-h triple compound treatment (elexacaftor/tezacaftor/ivacaftor surrogate) did not completely normalize lipid imbalance or oxidative stress. Protein array analysis revealed differential expression and shedding of cytokines and growth factors from CF epithelial cells compared to non-CF cells, consistent with sterile inflammation and tissue remodeling under basal conditions, including enhanced secretion of the neutrophil activator CXCL5, and the T-cell activator CCL17. However, treatment with antioxidants or CFTR modulators that mimic the approved combination therapies, ivacaftor/lumacaftor and ivacaftor/tezacaftor/elexacaftor, did not effectively suppress the inflammatory phenotype. We propose that CFTR deficiency causes oxidative stress in CF airway epithelium, affecting multiple bioactive lipid metabolic pathways, which likely play a role in CF lung disease progression. A combination of anti-oxidant, anti-inflammatory and CFTR targeted therapeutics may be required for full correction of the CF phenotype.

Keywords: bronchial epithelial cell; ceramide species; cystic fibrosis; cystic fibrosis transmembrane conductance regulator corrector therapy; cytokine array; lipidomics; oxidative stress; polyunsaturated (essential) fatty acids.

Figure 1

Oxidative stress, ceramides, and abnormal lipid metabolism in CF human bronchial epithelial cells in ALI culture. CF HBEC-ALI from five different donors (Table 1) compared to non-CF (WT, five different donors, see methods), were analyzed as described in methods. Data were merged from three separate experiments, comparing each donor in triplicate or quadruplicate in every experiment, data points are shown representing a single membrane (CF N = 40 filters, non-CF N = 25 filters). Horizontal bars represent average, error bars ± SD. (A) malondialdehyde (MDA) pmol/nmol fatty acid. (B) Total ceramides in CF HBEC-ALI. (C) Docosohexaenoic acid (DHA) levels (% nmol total fatty acids) are strongly reduced in CF HBEC-ALI compared to WT. (D) arachidonic acid (AA) levels (% nmol total fatty acids). (E) Long-chain ceramide (CER14:0). (F) Very long-chain ceramide (CER24:0) expressed as pmol/nmol phosphate. Statistical analysis (unpaired T-test) was performed with grouped averages (CF vs. Non-CF) for each different donor.

Figure 2

Enhanced long-chain to very long-chain ceramide species ratio in CF airway cells and lung tissue. Lipids were extracted and analyzed by mass spectrometry, as described in methods merged and averaged data. The bars represent the average concentration of individual ceramide species expressed as pmol/nmol total lipid phosphate (pmol/nmol P, error bars: SEM). (A) Human bronchial epithelial cells in ALI culture (HBEC-ALI) CF (BCF000174; N = 12 inserts) compared to WT (N = 12, three donors, four filters each), data merged from three independent experiments. Four other CF donors (Table 1) give similar results to BCF000174 as shown in (Figures 1E,F). (B) Pig bronchial epithelial cells in ALI culture (PIG ALI) on membrane inserts cystic fibrosis transmembrane conductance regulator (CFTR) KO (N = 6) and WT littermate (N = 9). (C) Total lung of age- and sex-matched adult homozygous mutant F508del CFTR (DD; N = 19) or control (WT) littermate mouse (N = 15), as described in methods.

Figure 3

Oxidative stress and abnormal lipid balance in CFTR KO PIG bronchial cells in air-liquid interface (ALI) culture is partially corrected by Fenretinide. (A) malondialdehyde (MDA) pmol/nmol fatty acid. (B) Total ceramides in CF HBEC-ALI. (C) DHA levels and (D) AA levels (% nmol total fatty acids). (E) Long-chain ceramide (CER14:0) and (F) Very long-chain ceramide (CER24:0) expressed as pmol/nmol phosphate. All parameters show significant differences when comparing CFTR KO (N = 9) to WT (N = 12), and CFTR KO filters treated with fenretinide (N = 6) in parallel experiments (****Padj < 0.001, unpaired ANOVA, Dunnet, using KO as reference control). Results are from two CFTR KO and two wildtype newborn pigs cultured in parallel in two separate experiments. Symbols in the figures represent data from single filters, average ± SD.

Figure 4

Oxidative stress and abnormal lipid levels in F508del CFTR variant mouse lung. Total lungs from adult male and female mice, homozygous for the F508del CFTR allele (DD, N = 19) or age- and sex-matched normal littermates (WT, N = 14), were extracted and lipids were analyzed as described, each data point represents a single individual. (A) malondialdehyde (pmol/nmol fatty acid). (B) Nitro-tyrosine. (C) DHA. (D) AA. (E) Total ceramide species after TLC purification. (F) LCC(Cer16:0)/VLCC(Cer26:0) ratio (compare Figure 2C). CF and WT lungs differed significantly in all parameters (unpaired T-test p < 0.001).

Figure 5

Partial correction of oxidative stress and lipid imbalance in CF HBEC-ALI by ivacaftor/lumacaftor. Treatment of the high responder (Table 1) homozygous F508del CFTR BCF174 HBEC ALI with VX-809 (2 μM) and VX-770 10 nM for 24 h (Ork) causes partial correction of lipid imbalance, compared to carrier control (Ctr; N = 12). Non-CF (WT) cells are represented by three donors in quadruplicate each (N = 12). (A) malondialdehyde (MDA) pmol/nmol fatty acid. (B) Total ceramides in CF HBEC-ALI. (C) DHA levels (% nmol total fatty acids). (D) AA levels (% nmol total fatty acids). (E) Long-chain ceramide (CER14:0). (F) Very long-chain ceramide (CER24:0) both expressed as pmol/nmol phosphate. Most relevant comparisons are indicated (unpaired ANOVA, Dunnet, *** p < 0.001, untreated CF as control).

Figure 6

Partial correction of oxidative stress and lipid imbalance in CF HBEC-ALI by elexacaftor/tezacaftor/ivacaftor triple therapy. CF HBEC ALI were treated for 24 or 72 h with the triple combination elexacaftor/tezacaftor/ivacaftor (VX-445 + VX-661 + VX-770; Triple) or vehicle alone (Ctr). CF HBEC ALI inserts of three donors, defined by their rescue by ivacaftor/lumacaftor and in Ussing chamber experiments Medium responder (BCF000191) low responder (BCF000554, squares) and high responder (BCF000584, triangles; Supplementary Table S1). Filters were analyzed in triplicate at every time point, data points (N = 9) represent single inserts. Avarage Levels ± SEM are shown after treatment with elexacaftor/tezacaftor/ivacaftor or control vehicle (Ctr) for 24 or 72 h. Treatment of Non-CF HBEC in parallel with elexacaftor/tezacaftor/ivacaftor had no detectable effect at either timepoint (WP945, N = 6, not shown). For comparison, gray bars represent average ± SEM of parallel Non-CF data obtained in parallel (three donors, N = 12). (A) malondialdehyde (MDA) pmol/nmol fatty acid. (B) Total ceramides. (C) DHA levels (% nmol total fatty acids). (D) AA levels (% nmol total fatty acids). (E) Long chain ceramide (CER14:0). (F) Very long chain ceramide (CER24:0) both expressed as pmol/nmol phosphate. Most relevant comparisons are indicated (unpaired ANOVA, Dunnet, ^*Padj < 0.05, **p < 0.01, ***p < 0.001, using parallel carrier treated control at 24 and 72 h, respectively, as control).

Figure 7

Partial correction of oxidative stress and lipid imbalance by anti-oxidants in CF HBEC-ALI. Treatment of CF HBEC ALI with extracellular glutathione (10 mM GSH, three different donors (BCF000174, N = 15; BCF000554, N = 3; BCF000889. N = 3; N = 21), or deferoxamine (Def, BCF000191 N = 3, BCF000584 N = 3) for 24 h, compared to carrier treated parallel controls of the same donors (Ctr), causes partial correction towards non-CF untreated values indicated as shaded bars representing untreated non-CF average ± SD (five donors analyzed in parallel, Figure 1). (A) malondialdehyde (MDA) pmol/nmol fatty acid. (B) Total ceramides. (C) DHA levels (% nmol total fatty acids). (D) AA levels (% nmol total fatty acids). (E) Long-chain ceramide (CER14:0). (F) Very long-chain ceramide (CER24:0) both expressed as pmol/nmol phosphate. Most relevant comparisons are indicated (unpaired one-way ANOVA, Dunnet, *p < 0.05, **p < 0.01, ***p < 0.001).

Figure 8

Basal shedding of cytokines and growth factors by primary CF and non-CF airway epithelial cells. Markers of inflammation and tissue remodeling were measured by a Fluidigm-based protein array (Olink 96×96 Oncology/inflammation) in basal medium (BEGM minus EGF), 24 h after the medium change. CF HBEC-ALI (five donors, Table 1, 31 samples) compared to non-CF (WT) cultured in parallel (four donors, 17 samples). Complete array data are summarized in Supplementary Table S1. Data are expressed as NPX (−Log2 expression relative to the internal control). (A) CXCL5 (C-X-C chemokine ligand 5) involved in neutrophil activation and angiogenesis is on average four-fold (two logs) higher in CF media compared to non-CF. (B) CCL17 (C-C motif chemokine ligand 17) involved in trafficking and maturation of T-cells is four-fold lower in CF HBEC [statistical analysis on total merged array data by multiple T-test p < 0.001, q < 0.001 (Supplementary Table S2), and nested T-test of donor-specific data (p < 0.02) shown here].