Pharmacological Improvement of Cystic Fibrosis Transmembrane Conductance Regulator Function Rescues Airway Epithelial Homeostasis and Host Defense in Children with Cystic Fibrosis

Abstract

Rationale

Pharmacological improvement of cystic fibrosis transmembrane conductance regulator (CFTR) function with elexacaftor/tezacaftor/ivacaftor (ETI) provides unprecedented improvements in lung function and other clinical outcomes in patients with cystic fibrosis (CF). However, ETI effects on impaired mucosal homeostasis and host defense at the molecular and cellular levels in the airways of patients with CF remain unknown.

Objectives

To investigate effects of ETI on the transcriptome of nasal epithelial and immune cells from children with CF at the single-cell level.

Methods

Nasal swabs from 13 children with CF and at least one F508del allele aged 6 to 11 years were collected at baseline and 3 months after initiation of ETI, subjected to single-cell RNA sequencing, and compared with swabs from 12 age-matched healthy children.

Measurements and Main Results

Proportions of CFTR-positive cells were decreased in epithelial basal, club, and goblet cells, but not in ionocytes, from children with CF at baseline and were restored by ETI therapy to nearly healthy levels. Single-cell transcriptomics revealed an impaired IFN signaling and reduced expression of major histocompatibility complex classes I and II encoding genes in epithelial cells of children with CF at baseline, which was partially restored by ETI. In addition, ETI therapy markedly reduced the inflammatory phenotype of immune cells, particularly of neutrophils and macrophages.

Conclusions

Pharmacological improvement of CFTR function improves innate mucosal immunity and reduces immune cell inflammatory responses in the upper airways of children with CF at the single-cell level, highlighting the potential to restore epithelial homeostasis and host defense in CF airways by early initiation of ETI therapy.

At a Glance Commentary

Scientific Knowledge on the Subject

Previous studies using single-cell RNA sequencing to identify changes in the transcriptome of airway epithelial and immune cells of adult patients with cystic fibrosis (CF) compared with healthy control subjects at single-cell resolution reported a disruption of epithelial homeostasis and an altered immune cell repertoire with mostly immature proinflammatory neutrophils in CF airways. However, whether pharmacological improvement of cystic fibrosis transmembrane conductance regulator (CFTR) function with the highly effective CFTR modulator therapy elexacaftor/tezacaftor/ivacaftor (ETI) restores airway homeostasis and altered immune cell states in patients with CF has not been studied.

What This Study Adds to the Field

We performed single-cell RNA sequencing of epithelial and immune cells from the nasal mucosa of children with CF aged 6 to 11 years at baseline and after initiation of ETI therapy and compared the results with those in age-matched healthy children. We found that pharmacological improvement of CFTR function improves innate mucosal immunity and reduces immune cell inflammatory responses in the upper airways of children with CF at the single-cell level, highlighting the potential to restore epithelial homeostasis and host defense in CF airways by early initiation of ETI therapy.

Cystic fibrosis (CF) is the most common fatal genetic disorder in White populations and is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene (1). In the airways, the CFTR anion channel plays an important role in epithelial homeostasis and mucosal host defense, and CFTR dysfunction leads to impaired homeostasis of ion, fluid, and pH regulation on airway surfaces (2–5), resulting in highly viscoelastic airway mucus, compromised mucociliary clearance, and impaired bacterial killing, which ultimately cause chronic inflammation and progressive structural lung damage (1). The triple combination CFTR modulator therapy elexacaftor/tezacaftor/ivacaftor (ETI) showed unprecedented improvement in lung function and other clinical outcomes in adults and children with CF and at least one F508del mutation in pivotal clinical trials, providing highly effective treatment for up to 90% of genetically eligible patients (6–15). Recent real-world observational studies in adolescent and adult patients with CF and at least one F508del allele showed an increase of CFTR function to ∼40–50% of normal CFTR activity, leading to improvement in airway mucus obstruction as well as sputum viscoelastic properties, airway infection, and inflammation (16–20). However, the extent to which treatment of the underlying defect can restore epithelial homeostasis and host defense in the airways of patients with CF remains unknown.

We previously demonstrated the potential of single-cell RNA sequencing (scRNA-seq) to study the molecular pathogenesis of SARS-CoV-2 on the basis of epithelial and immune cells obtained from noninvasively collected nasal swabs (21–23). Previous studies comparing CF and healthy airway epithelial and immune cells using the scRNA-seq technique in bronchial epithelial cells in adult patients showed an overabundance of epithelial cells transitioning to specialized ciliated and secretory cell subsets accompanied by a decrease in cycling basal cells, as well as an altered immune cell repertoire with mostly immature proinflammatory neutrophils (24, 25). However, the effect of ETI on the single-cell transcriptome of epithelial and immune cells of the airway mucosa has not been studied so far. Therefore, the aim of this study was to investigate the effects of ETI on upper airway epithelial and immune cells at the single-cell level to test the hypothesis that ETI therapy corrects epithelial and immune cell abnormalities observed in children with CF and restores airway homeostasis to healthy levels. To achieve this goal, we performed scRNA-seq of nasal cells from children with CF aged 6 to 11 years at baseline and after initiation of ETI and compared the results with those in age-matched healthy children. Examining nasal cells of children with CF at baseline and while receiving ETI treatment allowed us not only to analyze the intricate interaction of epithelial and immune cells underlying the complex in vivo CF pathophysiology but also to study the molecular and cellular effects of ETI therapy on the airway mucosa of patients with CF.

Methods

Detailed methods are provided in the online supplement.

Study Design and Participants

This prospective, observational postapproval study was approved by the ethics committee of the Charité – Universitätsmedizin Berlin (EA2/220/18; EA2/066/20). Written informed consent was obtained from all patients, their parents, or their legal guardians. Patients were eligible to participate if they were 6 to 11 years old, homozygous for F508del or compound heterozygous for F508del and a minimal function mutation, had no prior exposure to ETI, and were willing to remain on a stable medication regimen including ETI for the duration of study participation. Age- and sex-matched healthy children without any medical history of respiratory disease were included as control subjects. Exclusion criteria were an acute respiratory infection or pulmonary exacerbation at baseline. Nasal swabs for scRNA-seq were collected at baseline and 3 months after initiation of ETI in children with CF and at a single time point in healthy children. In children with CF, sweat chloride concentration, spirometry, anthropometry, and multiple-breath washout were assessed at both time points.

Single-Cell Isolation from Nasal Swabs and Preparation for Single-Cell RNA Libraries

Sample handling, library preparation, and sequencing were conducted as described previously (21–23). Nasal swabs were placed in Dulbecco’s modified Eagle medium/F12 medium (Gibco) and incubated with dithiothreitol (AppliChem) and subsequently Accutase (Thermo Fisher Scientific) for cell dissociation. A total of 17,500 cells per sample were resuspended in the master mix and loaded onto the chip and transferred to the 10× Chromium Controller (both from 10× Genomics). All subsequent steps, including reverse transcription, cDNA cleanup, and amplification, were performed following the manufacturer’s instructions. Final 3′ RNA libraries were pooled for sequencing using the NovaSeq 6000 Sequencing System (Illumina) with paired-end, single-indexing configuration.

Data Analyses and Statistics

The preprocessing of fastq files was performed using Cell Ranger version 3.0.1, with subsequent alignment of data to the human hg19 reference genome (10× Genomics, version 3.1.0). After data quality assessments and integration, all subsequent analyses were performed in R version 4.1 and Seurat version 3.2.2 (26). Differential gene expression analysis were conducted using the MAST (Model-based Analysis of Single-cell Transcriptomics) package (27). All P values were corrected for multiple comparisons using the Benjamini-Hochberg method (28).

Results

Characteristics of the Study Population

This prospective observational study enrolled 13 children with CF aged 6 to 11 years to perform anthropometry, sweat chloride concentration, spirometry, multiple-breath washout, and nasal swabs for scRNA-seq at baseline and 3 months after initiation of ETI. Furthermore, anthropometry and nasal swabs for scRNA-seq were collected from 12 age-matched healthy children. The clinical characteristics of the children with CF at both time points and basic characteristics of the healthy children are summarized in Table 1. Consistent with previous clinical trials (11, 12), ETI treatment resulted in an improvement in sweat chloride concentration and the lung clearance index and a trend in mean percent predicted FEV₁ (Figures 1B–1D and the Supplemental Results section in the online supplement).

Table 1.

Clinical Characteristics of Healthy Children and Children with CF, at BL and after Initiation of ETI

Clinical Characteristics	Healthy	CF BL	CF ETI
Number	12	13	13
Age, yr, mean (±SD)	8.7 (±2.1)	9.3 (±2.4)	9.7 (±2.4)
Time on ETI therapy, d, mean (±SD)	—	—	82.0 (±23.0)
Sex (female), n (%)	6 (50.0%)	5 (38.5%)
Genotype, n (%)
F508del/F508del	—	10 (76.9%)
F508del/MF		3 (23.1%)
No prior CFTR modulator, n (%)	—	3 (23.1%)
LUM/IVA therapy		10 (76.9%)
Pancreatic insufficiency (%), n (%)	0 (0%)	13 (100%)	13 (100%)
Sweat chloride concentration, mmol/L, mean (±SD)	—	86.4 (±18.7)	33.9 (±12.5)
Change (±SD)	—	—	−52.5 (±14.3)
P value	—	—	<0.001
FEV1% predicted, mean (±SD)	—	84.5 (±21.9)	96.4 (±16.1)
Change (±SD)	—	—	11.9 (±24.1)
P value	—	—	0.150
MEF25% predicted, mean (±SD)	—	100.3 (±61.4)	137.4 (±61.1)
Change (±SD)	—	—	37.1 (±49.3)
P value	—	—	< 0.05
LCI2.5, mean (±SD)	—	8.2 (±2.8)	6.9 (±1.2)
Change (±SD)	—	—	−1.3 (±1.8)
P value	—	—	<0.05
BMI, kg/m2, mean (±SD)	15.5 (±1.8)	15.9 (±1.2)	16.5 (±2.0)
Change (±SD)	—	—	0.6 (±1.5)
P value	—	—	0.183
BMI-for-age Z-score, mean (±SD)	−0.62 (±0.97)	−0.41 (±0.75)	−0.45 (±0.78)
Change (±SD)	—	—	−0.04 (±0.34)
P value	—	—	0.702

Definition of abbreviations: BL = baseline; BMI = body mass index; CF = cystic fibrosis; CFTR = cystic fibrosis transmembrane conductance regulator; ETI = elexacaftor/tezacaftor/ivacaftor; IVA = ivacaftor; LCI_2.5 = lung clearance index; LUM = lumacaftor; MEF = mean expiratory flow; MF = minimal function mutation.

P values describe the changes between paired measurements from BL and ETI therapy with a paired Student’s t test with the Benjamini-Hochberg correction.

Figure 1.

Effects of elexacaftor/tezacaftor/ivacaftor (ETI) on clinical outcomes and cystic fibrosis transmembrane conductance regulator (CFTR) expression in upper airway epithelial cells from children with cystic fibrosis (CF). (A) Nasal swabs of 13 children with CF at baseline (BL) and 3 months after initiation of ETI therapy as well as of 12 sex- and age-matched healthy children (Healthy) were collected for single-cell RNA sequencing. (B–D) Paired measurements of sweat chloride concentrations (B) and lung function parameters, including FEV₁ percent predicted (ppFEV₁) (C) and lung clearance index (LCI_2.5) (D), for all children with CF. Red lines represent mean values, dashed line indicates diagnostic threshold for cystic fibrosis (>60 mmol/L) or for healthy people (<30 mmol/L). *P < 0.05, ***P < 0.001, Student’s t test with Benjamini-Hochberg correction. (E) UMAP plots showing all identified immune and epithelial (highlighted within the dashed circle) cell types and stages. (F) Feature plot depicting CFTR expression in all epithelial cells from E. For detailed cell type annotation, please refer to E. (G) Bar plot showing the proportion of CFTR⁺ cells of the total count in each epithelial cell type per group. CFTR⁺ cells: cells with n ⩾ one CFTR read; *P < 0.05, **P < 0.01, ***P < 0.001, Fisher’s exact test with Benjamini-Hochberg correction. BMI = body mass index; CD4 bulk = CD4⁺ T cells; CD8 bulk = CD8⁺ T cells; Cil diff = differentiating ciliated cells; DC bulk = dendritic cells; DNT = double-negative T cells; lowRNA = cells with low RNA; MC = mast cells; mo/nrMa = monocyte-derived and nonresident macrophages; NK = natural killer cells; rMa = resident macrophages; Sec diff = differentiating secretory cells; Sec-Cil = secretory ciliated cells; UMAP = Uniform Manifold Approximation and Projection.

Upper Airway Cell Types and States and Their CFTR Expression

The transcriptomes of 163,432 cells from a total of 38 samples were analyzed. We could identify all major airway epithelial and immune cell types. The epithelial compartment included ciliated, secretory, and basal cells, as well as the rare FOXN4⁺ cells and ionocytes; the immune cell compartment included neutrophils, macrophages, and dendritic, B, and T cells (Figures 1E and E1A and E1B in the online supplement). No significant differences in cell type composition were observed between the study groups. Consistent with previous studies in airway epithelial cells in adults (24, 29, 30), overall, CFTR was primarily expressed in ionocytes, with over 90% of all ionocytes showing CFTR expression, and, to a lesser extent, in secretory, basal, and ciliated cells (Figures 1F and 1G and Tables E1A and E1B). Analysis of the proportions of CFTR⁺ cells within a particular nasal epithelial cell type revealed reduced CFTR⁺ basal cells (odds ratio [OR], 0.53; 95% confidence interval [CI], 0.32 to 0.86; P < 0.05) and club and goblet cells (club cells, OR, 0.78; 95% CI, 0.68 to 0.90; P < 0.01; goblet cells, OR, 0.65; 95% CI, 0.58 to 0.72; P < 0.001) in children with CF at baseline compared with healthy children (Tables E1A and E1B). ETI treatment increased the proportion of CFTR⁺ cells in club and ciliated cells to the level of healthy control subjects, whereas the proportion of CFTR⁺ cells in goblet cells remained decreased compared with healthy control subjects (OR, 0.81; 95% CI, 0.74 to 0.89; P < 0.001) (Figure 1G and Tables E1A and E1B). In ionocytes, the proportions of CFTR⁺ cells did not change after initiation of ETI therapy. In immune cells, we observed a low number of CFTR⁺ cells, representing a frequency of 0–1.2% of the corresponding immune cell type (Tables E1A and E1B), which was unaffected by ETI.

IFN Signaling Is Blunted in CF Nasal Epithelial Cells and Restored by ETI

Genes significantly repressed in nasal epithelial cells of children with CF at baseline, compared with those in healthy children, showed enrichment for various gene ontology (GO) terms related to cytokine signaling, IFN response, and antigen processing and presentation. Conversely, upregulated genes were enriched for terms associated with the translation and sorting of proteins (Figure 2A and Tables E2 and E3). Given the prominence of IFN-related GO terms among the top 10 enriched pathways, we calculated a gene set score per cell based on all differentially expressed genes in these sets. Representing this score as a two-dimensional Uniform Manifold Approximation and Projection showed that nasal epithelial cells of children with CF had a lower IFN signaling score at baseline than that in healthy children (Figure 2B). This impairment was largely restored after initiation of ETI without reaching healthy levels (Figure E2 and Table E4).

Figure 2.

IFN signaling is impaired in upper airway epithelial cells from children with cystic fibrosis (CF) and partially restored by ETI. (A) Top 10 gene ontology (GO) terms (Biological Process, 2021) derived from EnrichR analysis, including all significant genes that were downregulated (coefficient, less than −0.1; false discovery rate [FDR], <0.05) or upregulated (coefficient, >0.1; FDR, <0.05) in children with CF at BL compared with healthy children (Healthy). Color intensities indicate log₁₀ (P value). (B) Scaled feature plots colored by module scores based on the overlap of all significant, differentially expressed genes of the four different IFN terms shown in A (GO:0060333, GO:0071346, GO:0071357, GO:0060337, |coefficient| > 0.05; FDR, <0.05; scaled for 15,000 cells in each group). Scores are shown for all epithelial cells; please refer to Figure 1E for detailed annotations. BL = baseline; ETI = elexacaftor/tezacaftor/ivacaftor.

Genes induced in nasal epithelial cells by ETI therapy compared with CF at baseline formed an intricate network comprising seven distinct modules, including IFN response and antigen processing and presentation (major histocompatibility complex [MHC]) as well as modules linked to the activator protein 1 (AP-1) transcription factor complex, xenobiotics/glutathione metabolism, hypoxia, and translation (Figure E3 and E4 and Tables E5 and E6). Within the IFN response module, we observed distinct expression of genes involved in pathogen sensing and type I and type II IFN-stimulated genes (ISGs) (Figures 3A and E3). Expression of ISGs was repressed in club, goblet, and ciliated cells in children with CF at baseline compared with healthy children and induced by ETI therapy. In secretory cells, ETI induced genes related to pathogen sensing to levels even higher than in healthy children, whereas ISGs were induced without reaching the expression level in healthy children (Figure 3A and Tables E7 and E8). The PROGENy gene set score confirmed a generally lower activation of the IFN response–mediating Janus kinase/signal transducer and activator of transcription pathway across the different epithelial cell types in children with CF at baseline, which was increased by ETI therapy to the level of healthy control subjects in ciliated cells and ionocytes, whereas it remained impaired in basal and secretory cells (Figure 3B and Table E9).

Figure 3.

Elexacaftor/tezacaftor/ivacaftor (ETI) elicits the activation of genes associated with pathogen sensing, the IFN response, and major histocompatibility complex (MHC) encoding across various nasal epithelial cell types. (A) Dot plot depicting average gene expression of selected genes involved in pathogen sensing and IFN signaling identified in Figure 2B for different epithelial cell types. Significance is marked as a full red circle for cystic fibrosis (CF) baseline (BL) compared with healthy children (Healthy), dashed circles for CF ETI compared with CF BL, and asterisks between CF ETI and Healthy, calculated with MAST (Model-based Analysis of Single-cell Transcriptomics) (Benjamini-Hochberg–corrected false discovery rate [FDR], <0.05). *CF ETI vs. Healthy. (B) Box plots showing PROGENy scores for Janus kinase/signal transducer and activator of transcription (JAK-STAT) signaling in different epithelial cell types. Boxes represent 25–75% quantile; lines indicate the median. Significance values are shown for CF BL (red) compared with Healthy (blue), CF ETI (green) compared with CF BL, and CF ETI compared with Healthy. ***P < 0.001, Benjamini-Hochberg–corrected two-sided Wilcoxon test (FDR, <0.05). (C) Average expression of different MHC classes I and II genes are shown as dot plots for different epithelial cells. Full red circles indicate significance between CF BL and Healthy, dashed circles represent significance between CF ETI and CF BL, and asterisks denote significance between CF ETI and Healthy. Significance was derived from Benjamini-Hochberg–corrected MAST P values (FDR, <0.05); Ave. exp. = average gene expression; Pct. exp. = percentage of cells expressing the gene. *CF ETI vs. Healthy.

ETI Restores MHC Gene Expression in CF Nasal Epithelial Cells

The MHC or HLA complex plays an important role in antigen processing and presentation by epithelial cells in the airways (31, 32). At baseline, the expression of genes comprising the MHC/phagosome module was notably reduced in basal, club, goblet, and ciliated cells of children with CF, whereas in ionocytes, all MHC I but only a few MHC II genes were repressed (Figure 3C). With ETI treatment, there was an overall increase in the expression of MHC class I/II genes in club cells and, for the majority of these genes, in goblet and ciliated cells. However, only some MHC I gene expression levels were restored to healthy levels in ionocytes. MHC I/II genes were only partially restored in basal cells. In comparison with healthy children, most of these genes remained repressed across all epithelial cells despite ETI treatment (Figure 3C and Table E10).

The Proinflammatory Phenotype of CF Immune Cells Is Reduced by ETI

In immune cells, the enrichment analysis found cytokine-mediated signaling (GO:0019221) as the most enriched biological process among genes induced in children with CF and for genes repressed by ETI (Figures 4A and 4B and Tables E11A and E11B). The gene set score for this term suggested a pronounced decrease in the proinflammatory phenotype by ETI, particularly in neutrophils and macrophages (Figure 4C). Statistical comparison confirmed a decrease in the gene set score of neutrophils and monocyte-derived/nonresident macrophages (mo/nrMa) after initiation of ETI therapy (mo/nrMa, Z-score, −11.70; 95% CI, −0.087 to −0.063; neutrophils, Z-score, −26.00; 95% CI, −0.042 to −0.036; all P < 0.001). In mo/nrMa, cytokine-mediated signaling was restored to levels of healthy children, whereas neutrophils still showed elevated scores (Z-score, 31.50; 95% CI, 0.044 to 0.050; P < 0.001) (Figure 4D and Table E12).

Figure 4.

Elexacaftor/tezacaftor/ivacaftor (ETI) decreases activation of macrophages and neutrophils in upper airways of children with cystic fibrosis (CF). (A and B) Top 10 gene ontology (GO) terms (Biological Process, 2021) found to be enriched on the basis of differentially expressed genes (DEGs) calculated with MAST (Model-based Analysis of Single-cell Transcriptomics) (false discovery rate [FDR], <0.05). (A) On the left for genes upregulated (coefficient, >0.1) in children with CF at baseline (BL) compared with healthy children (Healthy) and (B) on the right for genes downregulated (coefficient, less than −0.1) in children with CF after initiation of ETI therapy (CF ETI) compared with CF BL. Color intensities indicate log₁₀ (P value). (C and D) Pathway activity score calculated per cell, based on all significant DEG overlapping with the most significant GO term cytokine-mediated signaling pathway (GO:0019221) (259 genes; FDR, <0.05; |coefficient| > 0.05). (C) Feature plots (scaled to 38,000 cells per group) showing the distribution of the module scores in Healthy, CF BL, and CF ETI. Red indicates higher scores, blue lower scores (please refer to Figure 1E for detailed annotations). (D) Box plots depicting the score for monocyte-derived and nonresident macrophages (mo/nrMa) and neutrophils. Boxes represent 25–75% quantile; lines indicate the median. Significance values are shown for the comparisons of CF BL (red) with Healthy (blue), CF ETI (green) with CF BL, and CF ETI with Healthy. ***P < 0.001 derived from Benjamini-Hochberg–corrected two-sided Wilcoxon signed-rank test (FDR, <0.05).

ETI Partially Restores Altered IL-1 and TNF Signaling Genes in Neutrophils and Macrophages of Children with CF

The expression of genes associated with cytokine-mediated signaling showed characteristic patterns in immune cells (Figures 5A and E5). There were two distinct groups of genes: those that were downregulated in CF compared with healthy children and upregulated by ETI treatment and those that were induced in children with CF at baseline and downregulated by ETI toward the level of healthy children. Furthermore, there is a subgroup that remains upregulated despite ETI (Table E2). On one hand, similar to what was observed in nasal epithelial cells, genes upregulated by ETI were involved in antigen presentation and processing. On the other hand, genes associated with IL-1 and tumor necrosis factor (TNF)-α signaling pathways were among the genes repressed after initiation of ETI compared with baseline (Figure 5A). Expression of IL1A/IL1B and their respective receptor subunits IL1R1 and IL1R2 was elevated in mo/nrMa and in resident macrophages at baseline. In particular, mo/nrMa showed activation of IL-1 downstream mediators and of proinflammatory chemokines such as CCL3 and CCL4 in children with CF at baseline, which were reduced by ETI to expression levels below those of healthy children. Although also in neutrophils the initiation of ETI reduced the activation of genes related to IL-1 signaling in children with CF compared with baseline, the expression levels in the majority of these genes were still elevated and did not reach the levels observed in healthy children (Figures 5B and 5C and Tables E13 and E14). Furthermore, ETI reduced the elevated transcription of AQP9 (encoding aquaporin 9), which is involved in IL-1B secretion, in neutrophils and mo/nrMa of children with CF (Figure 5D and Table E15).

Figure 5.

Elexacaftor/tezacaftor/ivacaftor (ETI) partially alleviates the activated IL-1 and TNF-α signaling in macrophages and neutrophils of children with cystic fibrosis (CF). (A) Heat map showing scaled expression in all immune cells of those 259 significant genes of the pathway activity score in Figure 4C and Figure 4D. Only selected genes are annotated; please refer to Figure E5 for complete annotation; red indicates higher expression, blue lower expression. (B and C) Dot plots depict expression of genes related to IL-1 signaling including IL1A/B, IL-1 receptors (IL1R1/2), and receptor-associated coding genes (B) and representative downstream genes (C). Significance is marked as a full red circle for CF BL compared with healthy children (Healthy), dashed circle for CF ETI compared with CF BL, and asterisks between CF ETI compared with Healthy calculated with MAST (Model-based Analysis of Single-cell Transcriptomics) and Benjamini-Hochberg adjustment (false discovery rate [FDR], <0.05). *CF ETI vs. Healthy. (D) Violin plots showing the expression of aquaporin 9 (AQP9) in mo/nrMa and neutrophils. Lines indicate the median. Significance values are marked for the comparisons of CF BL (red) with Healthy (blue), CF ETI (green) with CF BL, and CF ETI with Healthy. ***P < 0.001 from Benjamini-Hochberg–corrected two-sided Wilcoxon test (FDR, <0.05). (E) Box plots depicting TNF signaling score calculated with PROGENy for mo/nrMa and neutrophils. Boxes represent 25–75% quantile; lines indicate the median. Significance values are shown for CF BL (red) compared with Healthy (blue), CF ETI (green) compared with CF BL, and CF ETI compared with Healthy. ***P < 0.001 from Benjamini-Hochberg–corrected two-sided Wilcoxon test (FDR, <0.05). (F) Circle plots representing significant ligand–receptor (TNF-TNFRSF1A/B) interactions of TNF signaling between different cell types and stages indicated by different colors. Plots are scaled by the weight, and arrows indicate the direction of interactions. Interactions derived from CellChat. Ave. exp. = average gene expression; CD4 bulk = CD4⁺ T cells; CD8 bulk = CD8⁺ T cells; Cil diff = differentiating ciliated cells; DC bulk = dendritic cells; DNT = double-negative T cells; MC = mast cell; mo/nrMa = monocyte-derived and non-resident macrophages; NK = natural killer cells; Pct. exp. = percentage of cells expressing the gene; rMa = resident macrophages; Sec diff = differentiating secretory cells; Sec-Cil = secretory-ciliated cells.

Similar to IL-1 signaling, a general ETI-dependent decrease in expression of TNF (mo/nrMa, coefficient, −0.15; 95% CI, −0.21 to 0.09; P < 0.001) (Figure E6) to levels comparable with those of healthy children and a reduced transcriptional activation of the TNF signaling pathway were observed in mo/nrMa and neutrophils (mo/nrMa, Z-score, −16.90; 95% CI, −343.90 to −275.96; neutrophils, Z-score, −40.20; 95% CI, −105.90 to −96.01; all P < 0.001) (Figure 5E and Table E16). Notably, although neutrophils still had increased TNF signaling after ETI treatment, mo/nrMa showed a reduction to levels lower than in healthy children. In addition, children with CF showed stronger TNF-associated interactions among the different cell types. Especially at baseline, mo/nrMa exhibited a robust intrinsic TNF-mediated activation and interacted with nearly all other cell types, in particular neutrophils, a pattern not observed in the other study groups. In addition, CD4⁺ T cells and B cells strongly interacted with mo/nrMa and neutrophils. ETI led to a general decrease in TNF ligand–receptor interaction strength between the different cell types, comparable to that in healthy children (Figure 5F).

Discussion

This is the first study, to our knowledge, that provides molecular data on the effect of pharmacological improvement of CFTR function by ETI therapy on epithelial homeostasis and host defense in the upper airways of patients with CF at a single-cell resolution. By scRNA-seq of epithelial and immune cells obtained by nasal swabs from children with CF and healthy children, we found that in CF, the proportion of CFTR⁺ cells is reduced in most epithelial cell types but partially restored after initiation of ETI. Furthermore, our analysis revealed an impaired IFN signaling and MHC gene expression, which was partially restored by ETI therapy. Finally, ETI therapy decreased the activation of macrophages and neutrophils in the nasal mucosa of children with CF. However, despite this improvement by ETI, a proinflammatory phenotype of neutrophils persisted in CF compared with healthy children. Collectively, our data provide novel insights into the role of the upper airway epithelium in impaired host defense in CF and in vivo effects of ETI on epithelial and immune cells, suggesting a partial restoration of airway mucosal homeostasis and immunity in children with CF.

Similar to previous scRNA-seq studies in bronchial epithelial cells from adult patients with CF (24, 29, 30, 33), in our present study, we observed the highest expression of CFTR in ionocytes, followed by secretory (club/goblet), basal, and ciliated cells, in upper airway epithelial cells of children with CF. Interestingly, the proportion of CFTR⁺ cells was reduced in children with CF at baseline, especially in secretory cells, and increased toward normal levels with ETI treatment. We hypothesize that a reduced proportion of CFTR⁺ nasal epithelial cells may indicate an early stage of epithelial remodeling in CF and that its restoration with ETI therapy may reflect airway epithelial regeneration, given the role of these cells in important homeostatic functions, including airway surface liquid and pH regulation, and in maintaining airway integrity. However, in line with previous results (24), the proportion of CFTR⁺ cells was not reduced in ionocytes of patients with CF. The reason for these differential changes in CFTR expression in different epithelial cell types in children with CF needs to be further investigated in future studies. CFTR expression is also present in immune cells (34–37). However, in contrast to the substantial CFTR expression in epithelial cells, its expression in immune cells was comparatively low, with no observable variations between the study groups. This finding aligns with that in another single-cell study that similarly showed limited CFTR expression in immune cells derived from sputum samples (25). Furthermore, in vitro studies have shown a role of CFTR in immune cell activation and phagocytosis, particularly in macrophages and neutrophils (35, 37, 38). This is consistent with our findings, and, although we did not observe any differences in CFTR expression within the immune cell compartment, the overall functional improvement of CFTR by ETI (17) may have contributed as a secondary effect to the observed ameliorations in the immune cell transcriptional profiles. Given that hyperconcentrated mucus per se was shown to cause abnormal polarization and impair efferocytosis and phagocytosis capacities of airway macrophages (39), these reduced proinflammatory transcriptional profiles of immune cells may also be explained by improvement of mucus properties and mucociliary clearance after initiation of ETI therapy (20, 40). Taken together, our results in the upper airway cells of children with CF confirm the cell type distribution and CFTR expression pattern previously observed in bronchial airway epithelial and sputum cells in adult patients with CF and provide novel insight into epithelial regeneration after initiation of ETI therapy.

The airway epithelium covered by a mucus layer whose properties are critically regulated by CFTR is the first line of defense against inhaled pathogens, thereby protecting the underlying tissue from infection and damage. Essential for an efficient innate immune response of the airway epithelium is an appropriate IFN response. IFNs stimulate other neighboring cells to produce antiviral molecules, facilitate the recruitment of immune cells, and can activate and regulate the immune system. Furthermore, IFNs can contribute to the enhancement and renewal of the airway barrier through enhanced cellular integrity and promoting cell proliferation (41). We have previously shown that the airway epithelium of children is primed for host defense by the activation of genes encoding pathogen-sensing molecules and those involved in IFN response, associated with milder disease progression than in adults after SARS-CoV-2 infection (22). Our present data show that children with CF appear to have reduced expression of IFN signaling genes compared with healthy children, suggesting ineffective priming of epithelial cells. This observation is supported by in vitro studies in airway epithelial cell lines expressing endogenous F508del-CFTR showing that CFTR dysfunction can indeed abolish the epithelial IFN response to Pseudomonas aeruginosa (42). Of note, therapy with ETI improved the repression of genes encoding MHC I/II and peptide-cleaving subunits of the proteasome at baseline in all epithelial cell types, although consistently showing expression levels below those of healthy children. This indicates an improvement of their ability to adequately recognize and present pathogens and initiate innate immune responses. Recent clinical studies corroborate this hypothesis, describing improvements in the airway microbiome composition, including a reduced burden of P. aeruginosa and improved phagocytosis and bacterial killing of macrophages in patients with CF treated with ETI (19, 20, 35, 37). However, P. aeruginosa is not cleared in most patients, suggesting that the partial restoration of the MHC I/II complex might not be sufficient to eradicate chronic infection. Although epithelial cells are considered nonprofessional antigen-presenting cells, recent evidence suggests that antigen presentation by epithelial cells is a critical regulator of T cell recruitment and maintenance in the airways and in the gut (31, 32). Furthermore, it has recently been shown that efficient antigen presentation by epithelial cells limits macrophage and neutrophil activation in a murine colitis model (43). Although this is in line with our findings in the nasal epithelium, future studies are needed to determine if this mechanism is relevant in CF and plays a role in the improvement of the proinflammatory phenotype of neutrophils and macrophages by ETI therapy. Moreover, in children with CF, we observed a decrease in the expression of genes that encode members of the AP-1 transcription factor complex, which was restored by ETI. In the lung, AP-1 has been linked with alveolar homeostasis, and its deletion in epithelial cells has been shown to induce lung inflammation (44). Thus, the reduced expression of AP-1 members in children with CF could contribute to the impaired epithelial function that can be improved by ETI. Collectively, our results support the hypothesis that altered innate immune functions of the airway epithelium may play an important role in impaired host defense in CF airways and that this epithelial defect can be improved by pharmacological rescue of CFTR function.

Chronic airway inflammation is a critical element of CF lung disease, playing an important role in its pathophysiology and disease progression (14, 45, 46). Airway inflammation is already found in infants and young children, even when asymptomatic and in the absence of detectable lower airway infection (47). Inflammatory processes, even in the absence of detectable pathogens, are characterized by the recruitment of innate immune cells and the production of proinflammatory cytokines and chemokines, especially IL-1 and TNF. In this line, we observed an exacerbated expression of genes involved in IL-1 signaling in innate immune cells, specifically in mo/nrMa, of clinically stable children with CF with normal spirometry, which was ameliorated by ETI. The release of IL-1A and IL-1B by macrophages and epithelial cells plays a crucial role in eliciting a neutrophilic response and thus contributes to inflammation (48–51). Accordingly, expression of AQP9, which encodes a water channel involved in the inflammasome-dependent secretion of IL-1B (52, 53), showed a similar response, suggesting not only reduced IL-1 signaling activity but also reduced IL-1B release after initiation of ETI. This may limit further recruitment and activation of neutrophils and ameliorate neutrophilic inflammation. In line with observations in sputum samples (20), we observed reduced TNF expression with ETI therapy in conjunction with reduced TNF pathway activation that was more pronounced in mo/nrMa than in neutrophils. However, also in line with our previous results from sputum samples (20), TNF expression in neutrophils was still elevated after initiation of ETI compared with healthy control subjects. Notably, our analysis revealed strong interactions of TNF-expressing mo/nrMa and CD4⁺ T and B cells especially with neutrophils in CF, which may contribute to further activation and increased overall inflammation in the CF lung. However, this interaction was no longer observed after initiation of ETI. Thus, by reducing the proinflammatory state of neutrophils and macrophages, ETI may contribute to breaking the self-perpetuating proinflammatory loop in the airways.

Limitations

This study also has limitations. First, the results are based on samples from a limited number of patients with CF with a small range of age and lung disease severity. Nevertheless, this is the first single-cell study to investigate the effects of ETI therapy in strictly paired patients with CF, as well as an equal number of age- and sex-matched healthy children, to minimize heterogeneity and allow comparisons between CF and healthy airway mucosa, in addition to studying the mucosal response to ETI therapy. Most patients were treated with a dual CFTR modulator therapy at baseline, potentially superimposing some effects. All patients were clinically stable at both visits; however, we did not perform PCRs for respiratory viruses to exclude potential effects of an acute infection on our results. Differences in the expression patterns of epithelial cells have been described in relation to their location in the airways, so the differential expression we observed in the nasal environment may not directly reflect the changes in the lung. However, these samples exhibit major hallmarks observed in previous scRNA studies in sputum and lower airway tissue of adult patients with CF, such as the primary epithelial cell populations expressing CFTR and accumulation of proinflammatory neutrophils (24, 25). Recent reports showed residual airway infection and inflammation based on sputum studies of adolescent and adult patients with CF (18–20) and identified proinflammatory stem cells with neutrophilic properties in adult patients with CF with end-stage lung disease resistant to ETI therapy (46). In light of these findings and our observation that some CF-related transcriptional changes remain unaffected by ETI therapy, future long-term studies including patients with CF of different age groups and different stages of lung disease are needed. These studies will be crucial to determine to what extent the restoration of molecular signatures of innate epithelial host defense and resolution of neutrophil- and macrophage-driven airway inflammation observed in our study can also be achieved in older patients with more severe lung disease. Finally, it will be important to determine if the changes in innate epithelial immunity and the inflammatory response we observed in school-age children can be prevented by even earlier restoration of CFTR function in preschool children or infants with CF.

Conclusions

In summary, by using scRNA-seq of samples from the upper airways of children with CF and healthy children, this study shows for the first time, that pharmacological restoration of CFTR function by ETI restores the impaired innate immune response in epithelial cells and reduces the hyperinflammatory status of immune cells, including neutrophils and macrophages, in CF airways. These results highlight the potential of CFTR-directed therapies to halt the vicious circle of infection, inflammation, and progressive structural lung damage in CF and to restore epithelial homeostasis and host defense, especially if treatment is started in children with preserved lung function.