Single-cell RNA sequencing reveals novel cell differentiation dynamics during human airway epithelium regeneration

Sandra Ruiz Garcia, Marie Deprez, Kevin Lebrigand, Agnes Paquet, Amelie Cavard, Marie-Jeanne Arguel, Virginie Magnone, Ignacio Caballero, Sylvie Leroy, Charles Hugo Marquette, Brice Marcet, Pascal Barbry, Laure-Emmanuelle Zaragosi

Preprint posted on October 24, 2018

By single cell RNA sequencing airway epithelial cells during their differentiation from basal progenitor cells to mature epithelial cell types, the authors of this preprint throw light on the step-wise differentiation of the respiratory epithelium.

Selected by Rob Hynds


The airways conduct air to the alveoli in order for gas exchange to occur. The epithelium that lines the airways is an important barrier to the outside world: it is challenged by inhaled particulates and pathogens and mediates mucociliary clearance by producing mucous and actively beating trapped matter out of the lungs to be swallowed. Recently, a number of single cell RNA sequencing (scRNAseq) studies have begun to reveal new insights into the composition of this epithelium in humans. Most notably, two papers published in Nature by the Regev and Rajogopal laboratories and the Klein and Jaffe laboratories revealed the presence of cystic fibrosis transmembrane conductance regulator (CFTR)-rich population of cells that were dubbed ‘ionocytes’ because of their similarity to ion transporter-rich ionocyte cells found in Xenopus larval skin and populations of mammalian kidney and inner ear cells. A novel epithelial population, which expresses transcripts associated with both airway and alveolar lineages, has also been discovered in idiopathic pulmonary fibrosis.

Key findings

In their pre-print, Sandra Ruiz Garcia, Marie Deprez and colleagues in Pascal Barbry’s team at Université Côte d’Azur in France study the regeneration of the airway epithelium from cultured basal cells in the widely used air-liquid interface cell culture model which allows differentiation of basal cells to mature airway cell types.

In their first analysis, the authors find a picture which fits neatly with previous studies; two populations of basal cells [cycling (Ki67+) and non-cycling (Ki67-)], a suprabasal population, which represent an intermediate ‘mid-differentiation’ cell type, secretory cells (Scgb1A1+), goblet cells (MUC5AC+) and ciliated cells (FOXJ1+). Using sub-clustering, in which less discrimination between clusters is accepted, the authors are able to distinguish important intermediate states within each of these baseline populations. Non-cycling basal cells can be divided into two subsets, one of which expresses genes associated with cell motility. Pseudotime analysis suggests that these cells are the most differentiated basal cells, or pre-suprabasal cells. Our knowledge around basal cell motility in humans is limited but very motile cells can be selected in airway cell cultures and clonal relationships between spatially separated pre-malignant airway lesions make this an interesting area for future study. Three populations of suprabasal cells and three populations of secretory cell can also be distinguished from one another. Intriguingly, one of the secretory populations differentially expresses transcripts which make it a candidate immune regulatory population, including TNF, IFNG, IL1 and IL6. As the authors acknowledge, future work might investigate the spatial distribution of these cells in human airways and also establish how dynamic the secretory subpopulation lineage hierarchy is. Finally, the authors are able to distinguish mature multiciliated cells from those that are actively undergoing multiciliogenesis (labelled ‘deuterosomal cells’). These cells selectively express genes such as DEUP1, which is involved in the massive centriole amplification necessary for multiciliation, and reactivate some cell cycle associated genes.

Figure 5E from the pre-print summarises the lineage hierarchy of the airway epithelium.

In further analyses, the authors investigate the dynamic changes in keratin gene expression during airway epithelial differentiation and investigate the expression of key airway cell signalling pathway components in a cell type specific manner. Notch signalling has been widely described as a key mediator of airway epithelial cell fate choice with no or little Notch activation in basal cells, Notch activation during secretory differentiation and then repression of Notch to generate ciliated cells. The new analysis suggests several potential Notch repressors that become expressed at the deuterosomal stage and finds, consistent with a previous report, that suprabasal cells are the first to activate Notch signalling by NOTCH3 expression. Interestingly, the putative immune-regulatory secretory population selectively express Wnt5a and Wnt4 and basal cells express both Wnt10a and Wnt target genes, suggesting their behaviour might be regulated in an autocrine fashion. Future experimental work is required to confirm these effects and to investigate how the epithelial gene expression landscape is altered by non-epithelial cell types in vivo.


  • Analysis of airway epithelial regeneration from basal progenitor cells in vitro adds to our understanding of differentiation intermediates.
  • Airway basal cells differentiate via suprabasal cells to secretory lineages which can further differentiate towards multiciliated cells given the correct cues.
  • A subpopulation of relatively differentiated basal cells is enriched in migration-associated transcripts.
  • A subpopulation of secretory cells is enriched in immune-modulatory genes and might orchestrate mucosal immunity.

Further reading

Travaglini, K.J. and Krasnow, M.A.(2018). Profile of an unknown airway cell. Nature, 560, 313-314.

Questions for the authors

Q1. How much difference is there in practice between a secretory cell and a goblet cell? Is it an oversimplification to describe them by Scgb1a1 and MUC5AC in human proximal airway samples where club cells are rarely seen? Put another way, is there a meaningful difference between a BEGM ALI culture (which has no Scgb1a1+ or MUC5AC+ cells but ‘secretory-like’ cells that don’t express these specific markers) and a Pneumacult ALI culture (which has both populations)?

Q2. In the pre-print, cultured cells are airway epithelial cells but your fresh samples are nasal. Do you think that scRNAseq studies of different airway locations will reveal meaningful differences in different parts of the respiratory tree as suggested by cell culture studies?

Q3. The deuterosomal population is found consistently across samples – do the authors know how transient a state it is? (i.e. how long does multiciliogenesis take in human airways?)

Q4. The authors see significant differences between late-stage ALI cultures and fresh biopsy tissue. Since ALI cultures are imperfectly differentiated with incomplete ciliation and some squamous differentiation, do the authors think that the intermediate cells found in ALI cultures are physiologically relevant differentiation intermediates?

Tags: differentiation, epithelial cells, lungs, rnaseq

Posted on: 10th November 2018 , updated on: 17th November 2018

Read preprint (No Ratings Yet)

  • Author's response

    Pascal Barbry shared

    Q1. See our “Supplementary figure 4: Comparison of fully differentiated epithelia cell population between Pneumacult and BEGM media”, in which we have performed such comparison which shows that secretory-like cells from BEGM are very well correlated with secretory/goblet cells from Pneumacult. Gene signatures of the two cell populations are highly similar and differ by secretoglobins and muc5ac expression. A precise comparison would probably require the use of cells derived from the same individual (which was not the case here), grown with the 2 different cell culture medium.

    Q2. Yes, we believe so. We have indeed recently published a study in which we compared the bulk transcriptomes of nasal and bronchial brushings, findings interesting differences between those two sites (Giovannini-Chami et al. 2018. The “one airway, one disease” concept in light of Th2 inflammation. European Respiratory Journal 52 (4), 1800437, Figure 2), and we have an ongoing work to extend this description at different levels of the airways.

    Q3. We have no exact quantification of the process, but the small size of the population suggests that this is a transient state for a given cell. That said, the population is usually renewed in our cultures, except in very old PneumaCult cells.  In homeostatic tissues, we also find a small subset of deuterosomal cells, indicating that what we observe in our cell cultures is consistent to what is seen in this slow-renewing tissue.

    Q4. Different situations occur, with significant differences between PneumaCult and BEGM cultures. In our hands, PneumaCult led to a better differentiation. Alternative models are certainly worth being investigated, such as organoids, and look very promising. This approach can perhaps better provide a better picture of what’s going on in fresh tissue. It remains that the principal steps of differentiation are conserved in all models that we have tested and that main cell populations are found both in Pneumacult cultures and fresh tissues. We are also investigating more closely subpopulations of fresh tissues on datasets containing high cell numbers.

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here