Where Is the Cystic Fibrosis Transmembrane Conductance Regulator?

Cystic fibrosis (CF) is a mucoobstructive pathology associated with chronic inflammation and chronic bacterial infection of the lungs. Mutations in the CF gene lead to dysfunction of the CFTR (CF transmembrane conductance regulator) and development of clinical symptoms. Two central characteristics of CF lung disease are an inadequate hydration and a defective transport of the mucous layer that covers and protects the airway surface. Despite the huge progress made in CF care, the exact relationships between primary defects and different manifestations of the disease are lacking. In that context, it is important to establish a precise map of CFTR expression along the airways. Such studies, begun immediately after the discovery of the CF gene, gradually concluded that CFTR was detected in airway surface epithelial cells, including in airway multiciliated cells, as well as in rare “CFTR hot” cells near or within airway submucosal gland acini or gland duct cells (1, 2).

In 2018, Montoro and colleagues (3) and Plasschaert and colleagues (4) applied single-cell RNA sequencing (scRNAseq), a technology that allows unbiased transcriptional profiling of tens of thousands of individual cells. They generated a catalog of the cells expressed in the lung, and, more particularly, they revealed a population of rare cells that they entitled pulmonary ionocytes. This name reflected a population of cells found in fish gills and frog skin, which contribute to ion homeostasis and hydration. Interestingly, pulmonary ionocytes were not only established by Montoro and colleagues and Plasschaert and colleagues as the sites of highest CFTR expression in airway cells but were also characterized by their high expression of other ion-transport genes, including subunits of the amiloride sensitive Na⁺ channel, and components of H⁺-ATPases. This peculiar gene expression program, also showing similarities with renal intercalated cells, suggests that ionocytes could be directly involved in active absorption of fluids (5) and/or regulation of acid-base homeostasis (6). In the murine tracheal epithelium, the majority of CFTR was present in pulmonary ionocytes but basal and secretory cells also expressed CFTR (3). Remarkably, little to no CFTR expression was detected in multiciliated cells, which were thought to be the main harbor of CFTR expression.

In this issue of the Journal, Okuda and colleagues (pp. 1275–1289) are now providing a comprehensive description of CFTR-expressing cell types in normal human conducting airways (7). To do so, they have combined scRNAseq technologies, single-cell quantitative RT-PCR, and RNA in situ hybridization methods, validating some of their results by electrophysiological approaches. Their measurements also provide information about variations of gene expression between large and small airway epithelia. This work confirms that CFTR is strongly expressed in ionocytes but also underlines the rarity of these cells in human small airway epithelium. The authors’ conclusion is that ionocytes represent a fraction of the total CFTR signal. Instead, more abundant cell types that express lower individual levels of CFTR represent a much larger fraction of the total signal. Secretory cells are thus the dominant cell type that expresses CFTR in the surface epithelium of large and small airways. CFTR is also significantly expressed in basal cells, suprabasal cells, and, to a lesser extent, multiciliated cells. Finally, the authors directly measured CFTR-mediated Cl⁻ secretory function, demonstrating a better correlation between this signal and the presence of secretory cell types than with ionocytes. Secretory cells from CF airway epithelia, but not multiciliated cells, were capable of CFTR-mediated Cl⁻ secretion after transduction with wild-type CFTR.

The results of Okuda and colleagues fit well with independent data sets that were recently published on human lung and airway (Table 1). Deprez and colleagues provided an scRNAseq atlas of 77,969 cells from 35 healthy human airway samples derived from 10 subjects, in which they defined 28 distinct cell types/states (8). They confirmed the high expression of CFTR in ionocytes (2.5% of total CFTR signal in 0.15% of total cells), but they also detected it in secretory (65.5% of signal), suprabasal (23.3% of signal), basal (5.1% of signal), and multiciliated cells (1.5% of signal) at four different levels of the airways. Habermann and colleagues analyzed 114,000 cells from 20 pulmonary fibrosis and 10 control lungs, including cells from the parenchyma (9). After defining 31 distinct cell types, they detected CFTR in alveolar type 2 cells (48.1% of total signal), secretory cells (34.6% of total signal), multiciliated cells (8.4% of total signal), and basal cells (4.9% of total signal). This data set includes a few ionocytes, which were not formally typed but which express bona fide ionocyte markers, including CFTR. Miller and colleagues analyzed 8,443 human fetal lung cells (11.5–21 wk of development), defining 12 distinct epithelial cell types. In this data set, the major sites of expression for CFTR correspond to secretory progenitors (totaling 39.1% of the CFTR signal) and bud tip adjacent cells (17.8% of the CFTR signal) (10). Finally, Goldfarbmuren and colleagues defined 10 epithelial cell clusters in 36,248 epithelial cells isolated from 15 donors (either smokers or never-smokers), in which they detected 46.3% in KRT8-high intermediate cells, 19% of total CFTR signal in basal cells, 17.9% in mucus secretory cells, and 11.2% in ionocytes (11). Collectively, these results draw a very consistent picture with secretory cells as major sites of expression for CFTR in human airways. It illustrates the interest of integrating well-standardized data sets from different origins in powerful atlases (12). Sharing of common ontologies, proper definitions of gene markers, and continuous improvement of the existing resources are key to ensure their overall quality (13). As nicely illustrated by Okuda and colleagues, maintaining a constant dialogue between production of single-cell data and independent biological results is also important.

Table 1.

Characteristics of the Five Different Data Sets Discussed Herein

Authors	Data Set	CFTR Expression	Reference
Okuda et al.	16,643 cells, 7 donors	9 clusters (10x): secretory cells (43.6%) > suprabasal cells (23.5%) > cycling/deuterosomal cells (15.9%) > basal cells (9.2%) > ionocyte + neuroendocrine cells (4.8%) > multiciliated cells (3%)	7
	26,319 cells, 7 donors	11 clusters (Drop-Seq scRNAseq): secretory cells (43.6%) > suprabasal cells (20.7%) > multiciliated cells (5.6%) > ionocytes + neuroendocrine cells (3.6%) > cycling/deuterosomal cells (3.2%) > basal cells (2.9%) https://cells.ucsc.edu/?bp=lung&ds=lung-airway
Deprez et al.	77,969 cells, 35 human airway samples, 10 healthy volunteers	28 distinct cell types, including ionocytes (10x): secretory cells (65.5%) > suprabasal cells (23.3%) > basal cells (5.1%) > ionocytes (2.5%) > multiciliated (1.5%) https://www.genomique.eu/cellbrowser/HCA	8
Habermann et al.	114,000 cells, 20 pulmonary fibrosis and 10 control lungs	31 distinct cell types: alveolar type 2 cells (48.1%) > secretory cells (34.6%) > multiciliated cells (8.4%) > basal cells (4.9%) https://cells.ucsc.edu/?ds=lung-pf-control	9
Miller et al.	8,443 EPCAM+ cells, 8 human fetal lung samples (11.5–21 wk of development)	12 distinct cell types: secretory progenitors (39.1%) > bud tip adjacent cells (17.8%) https://cells.ucsc.edu/?ds=covid19atlas+c19a-miller20	10
Goldfarbmuren et al.	36,248 epithelial cells, 15 donors (including 6 never-smokers and 6 heavy smokers, i.e., 15 pack-years)	10 epithelial cell clusters: KRT8-high intermediate cells (46.3%) > basal cells (19%) > mucus secretory cells (17.9%) > ionocytes (11.2%) > multiciliated cells, submucosal gland cells https://cells.ucsc.edu/?ds=lung-smoking-effect	11

Definition of abbreviations: CFTR = cystic fibrosis transmembrane conductance regulator; Drop-Seq = droplet-based scRNAseq; EPCAM = epithelial cell adhesion molecule; scRNAseq = single-cell RNA sequencing.

By establishing a cellular context that link CFTR with MUC5AC and MUC5B, that is, the two main synthetic and secreted respiratory mucins, the work by Okuda and colleagues defines possible scenarios to finely control mucus hydration. This may also fit well with a recent observation of some very early posttranslational modifications that can lead to hypersialylation of mucin O-sugars in a CF pig model (14).

The airway surface liquid is made of two distinct components: a mucus layer principally made of MUC5AC and MUC5B and a periciliary layer (gel) formed by tethered macromolecules, including MUC1, MUC4, and MUC16 (15). Small changes in mucus concentrations alter the osmotic pressures in the two phases, resulting in a severe impairment of the mucociliary transport rate. A direct control of ion fluxes by CFTR in cells that do secrete MUC5B and MUC5AC makes a lot of sense and should now be tested quantitatively.

A final point is that the importance of secretory cells does not preclude additional roles of CFTR in other cells. The function of ionocytes and ionocyte-localized CFTR remains unknown. It is also likely that CFTR located in different cell types subtends different functions in different parts of the respiratory tree (5).