SOX21 modulates SOX2-initiated differentiation of epithelial cells in the extrapulmonary airways

Essential revisions:

In particular, the following major points require additional experimental data (details are outlined below):

1) The functional Sox21 role in the adult lung is not clear.

– Mouse studies would benefit from inclusion of overexpression of Sox21 in the airways.

We added the analysis of the effects of overexpressing SOX21 (and SOX2) on the epithelial differentiation of the airways late in gestation at E18.5 in figure 3E. Moreover, we studied the effect of SOX21 (and SOX2) overexpression during adult homeostasis using additional mouse experiments, and added figure 4E and figure 4—figure supplement 1E to the manuscript to discuss the results.

– In human airways and the models used (organoids and ALI cultures) gain and loss of function studies are not performed and would also behelpful to further determine the phenotype of the intermediate para-basal cells in the human airway.

We have performed SOX2 and SOX21 knock down experiments in our human Air Liquid Interface (ALI) cultures and added the data to figure 7 and figure7-supplement 1.

2) Evidence for genetic interaction between Sox21 and Sox2 in the adult lung needs to be expanded.

For the genetic interaction, we have tried to obtainSox2^+/-Sox21^-/- mice, but the main difficulty resides in the fact that it is nearly impossible to breed these mice, and we did not obtain (sufficient) mice to perform reasonable experiment. We did analyze Sox2^+/-, Sox21^+/-, and Sox21^-/- E18.5 lungs, which we added to the manuscript (figure 3E, supplemental figure 1E). We obtained only one Sox2^+/-Sox21^+/- mouse pup at E18.5, of which the lungs were analyzed. Since this is only one sample, we did not include it in the manuscript, but we like to share these preliminary data with the reviewer (see Author response image 1). It appears that the Sox2^+/-Sox21^+/- lung is comparable to the individual Sox2 and Sox21 heterozygous lungs.

Author response image 1

Download asset Open asset

Previously, we observed a direct interaction between SOX2 and SOX21, which was already published by others (Mallanna et al., Stem Cells 28, 1715-1727, 2010). Thus, there is a direct link between the two proteins. Moreover, we performed SOX2 and SOX21 ChIP-seq experiments using human ALI cultures. However, we will not report on the full data set (yet), but have included four genome tracks to show single SOX2 or SOX21, and double SOX2-SOX21 occupancy at different genomic loci (Figure 7—figure supplement 1H). We included the supplemental figure, and refer to the data in the discussion line (452-457), to underly the importance of studying the interaction between these proteins further. Of note, the human ALI system is a culture method with a diverse range of cell types and all cell types are in different status of differentiation. It is challenging to study the interaction and function(s) of SOX2 and SOX21 at the cellular level. As described previously and in the discussion (line 446-448), the interaction of both SOX2 and SOX21 are highly context dependent, emphasizing the importance of obtaining a suitable system to study this interaction in the human and mouse airway epithelium.

3) Presence of Sox9 positive basal cells as additional atypical basal cells needs to be further investigated.

We have further investigated the putative presence of SOX9+ basal cells in our experiments, and found that there are some rare SOX9+ basal-like cells. To acknowledge these findings, we have adapted the text (line 178-184) and added supplemental figure2-supplement 2E. As SOX9+ basal cells lack expression of the KRT5 basal marker (Utisyan 2016), we added a KRT5 staining of Sox2^+/-, Sox21^+/- and Sox21^-/- E18.5 lungs to analyze the number of basal cells (Figure 3E and figure 3 supplement 1E).

4) Quantification the % of club cells in the corn oil and naphthalene treated tracheas of the different mutants as these are the main cells that should be replaced upon naphthalene injury.

We have quantified the percentage of Club cells before and after the treatment with naphthalene, and we did not observe significant differences between the different genotypes, or between treated and non-treated animals. We have included these data in figure 5D

Reviewer #2:

This is an interesting but also a bit confusing manuscript by Evelien Eenjes and colleagues.

I was wondering whether the seemingly conflicting effect of Sox21 on basal cell differentiation could be due to an increase or decrease in Sox9 positive basal cells. These atypical basal cells have been described in a number of manuscripts. For example https://www.ncbi.nlm.nih.gov/pubmed/26869074.

It appears from the figures that Sox21 may increase Sox9 expression. Can the authors look at the presence of Sox9 positive basal cells in the trachea of the different mutants?

The reviewer points out an interesting observation. In Figure 1C we indeed find that when Sox21 is induced in the lung bud, SOX9 remains expressed. Furthermore, in the airway we find some iSOX21 cells that are still low in expression of SOX9, but these are scarce and not all iSOX21 expressing cells do.

We did find the presence of TRP63 expressing SOX9+ basal cells in the SPC induced animals. However, we did not elaborate! As described by Ustiyan et al., the observed SOX9+ TRP63+ atypical basal cells are KRT5 negative. We included therefore KRT5 in our staining panel to analyze the number of basal cells in the proximal airways upon induction of Sox2 or Sox21 using the CC10 rTta (Figure 2B-E). All basal cells analyzed were also positive for KRT5. KRT5 was also added to analyze the number of basal cells in Sox2^+/-, Sox21^+/- and Sox21^-/- mice at E18.5 (Figure 3E).

We have included these findings in the text and figures:

“Previously, conditional deletion of βcatenin in SPC+ cells during lung development showed the appearance of cystic structures, similar to the induction of SOX2 and SOX21. […] Similarly we also found SOX9+ basal cells in the iSox2^SPC-rtTA and iSox21^SPC-rtTA mouse lungs, and sporadic SOX9+ basal cells were also detected in the control (Figure 2—figure supplement 1E).”

It would be important to quantify the % of club cells in the corn oil and naphthalene treated tracheas of the different mutants as these are the main cells that should be replaced upon naphthalene injury. I am not sure that people normally look at the number of ciliated cells reappearing after naphthalene injury as ciliated cells are not thought to be killed by naphthalene.

The reviewer has indeed a good point, and we have quantified the number of Club cells before and after treatment with naphthalene. We have adapted the figure and the text accordingly:

“To determine whether SOX2 or SOX21 deficiency affects regeneration after naphthalene injury, we quantified the percentage of secretory, ciliated, non-dividing and dividing basal cells at 20 DPI. Fewer FOXJ1+ ciliated cells were observed in Sox2^+/- TE compared to WT, Sox21^+/- , or to Sox2^+/- CO exposed mice (Figure. 5D, Figure 5-supplement 1D).”

The authors should also look at Sox9 positive basal cells in the tracheas of the corn oil and naphthalene treated ctrl and mutant mice.

The SOX9+ basal cells is missing in the analysis and unfortunately were not able to add it to the naphthalene experiments, due to lack of material to quantify. However, we did stain TRP63 and SOX9 at 5 days post injury and SOX9+ cells are present at the surface epithelium together with TRP63+ basal cells , we cannot exclude that these basal cells are SOX9+ and whether there is a difference. However, at day 5 post-injury we did not see a difference in the number of basal cells between genotypes but a small difference in SOX9+ cells.

Are the intermediate para-basal cells in the human airway differentiation experiment club cells? Why do the authors think these cells are on their way to become ciliated cells? Doesn't Sox21 inhibit ciliated cell differentiation?

The SOX21 high expressing cells in the ALI culture did not correlate with SCGB1A1 expression (Author response image 2, n=1). We did not continue these experiments, as the single cell RNA seq of human ALI cultures shows a high expression of SOX21 in para-basal cells (Figure 7- figure supplement 1E) and the localization of high expressing SOX21 on human sections are found in the intermediate layer (Figure 6D). As there are some FOXJ1+ nuclei high in expression of SOX21 (Figure 7—figure supplement 1D) we interpreted the data as cells transition to ciliated cells. We observed with the added knock-down experiments that SOX21, similar as in mouse, inhibits ciliated cell differentiation in human ALI cultures. We excluded this possibly premature statement from the manuscript and now state the following:

“We hypothesize that these cells are intermediate cells transitioning from basal cells to a luminal cell fate, the para-basal cell. […] This analysis confirmed that high levels of SOX21 mRNA denote an intermediate cell type.”

Author response image 2

Download asset Open asset

Reviewer #3:

[…] The development data is clear based on the studies presented. The immunostaining and figures are elegant. It does not appear the Sox21 is necessary for airway development. A key question raised by this work is how important and what precisely is the role of Sox21? For example, overexpression of Sox21 in the developing and adult airways might have been more instructive in answering these key questions rather than under the control of the SPC-promoter.

The reviewer raises a good point, and indeed, the SOX21 knockout mice suggests that SOX21 is not necessary for lung development. However, SOX21 has a clear influence on the distribution of the different cell types, as we report.

Initially, we choose to use the SPC-driven rtTA, as this transgene is expressed early in development in most epithelial cells (e.g. Gontan et al., 2008). These data were included in the original manuscript. We have now expanded this data set using a mouse model to induce SOX2 or SOX21 in differentiating epithelial cells, using the rtTA under the control of the CC10 promoter. This resulted in additional embryonic data, as well as new adult data, which we have included. For the embryonic data, we included the following text and accompanying figures:

“To further investigate the role of SOX2 and SOX21 in differentiation of airway epithelium during development, we induced expression of the MYC-tagged Sox2 or Sox21 transgenes in secretory cells using a CC10/SCGB1A1 driven rtTA inducible model (iSox2^CC10-rtTA and iSox21^CC10-rtTA) (Tichelaar, Lu et al. 2000). […] So, ectopic expression of SOX2 and SOX21 during development stimulates the appearance of basal cells and decrease the number of ciliated cells.”

And for the adult data set:

“To further understand the role for SOX2 and SOX21 in adult airway epithelium, we used the iSox2^CC10-rtTA and iSox21^CC10-rtTA mice to induce expression of Sox2 or Sox21 in secretory cells for 6 weeks (Figure 4E). […] Similar as during development, the MYC+ transgene expressing cells were mostly negative for other cell specific markers in both mouse models (Figure 4—figure supplement 1E).”

The adult mouse data and human data seem to overall suggest that SOX2 and SOX21 are upregulated in differentiating cells in vitro and are expressed at lower levels in basal cells but real mechanism or phenotype related to SOX21 function is observed in the SOX21 het mouse cells.

We are not sure what the exact question is of the reviewer. In light of this, we have now expanded the human data with knock down experiments using shRNA constructs in human ALI cultures (figure 7 and figure 7—figure supplement 1). These data show comparable effects of reduced SOX2 or SOX21 as observed in the mouse cells.

The attempt to assess human fetal airways is admirable. However, sections of fetal lungs would be more relevant. Can the authors address how similar fetal lung cultures are to in vivo airways and at what developmental time point ?

We have performed immunofluorescence staining on human sections throughout gestation and included these in figure 6A. We have restructured the figures, so figure 6 now displays human fetal lung sections, organoids (fetal tip organoids and differentiated organoids), and adult bronchi. The ALI cultures we have used for the mechanistic experiments mostly resemble the adult bronchus.

Overall the relevance of SOX21 expression in the adult airways is unclear.

We have performed additional experiments to investigate the role of SOX2 and SOX21 in the differentiating human epithelium by shRNA knockdown experiments, described in the text, figure 7 and supplemental figure 6. We have added the following text:

“Next, we assessed the effects of reduced levels of SOX2 or SOX21 on human basal cell differentiation by transducing ALI cultures with lentiviruses expressing short hairpin RNAs (shRNA) to knockdown SOX2 or SOX21. […] Overall, our data suggest that both SOX2 and SOX21 suppress basal to ciliated cell differentiation, while both seem to have a positive role in the differentiation of goblet cells in the human airways.”

Combined with the mouse data (Sox2^+/-, Sox21^+/- and Sox21^-/-; Naphthalene; CC10rTTA adult), we hope to better indicate the relevance of SOX21 in the adult airways.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Essential revisions:

The manuscript improved significantly several additional experiments included that have been ask for in the first revision. These experiments further clarified and support the dynamic and context-dependent role for Sox21 in airway cell differentiation. No further experimental data are required, however, some textual changes to clarify the results and enhance readability.

1) The manuscript is still very hard to read, especially given the variety in model systems and different results based on these. Please carefully revise and modify the text accordingly to clarify the results and enhance readability.

We agree with the reviewers that the inclusion of new experimental data in the revised version resulted in loss of clarification of the some of our data, which regrettably led to reduced readability. We have carefully revised the new manuscript, and made several textual changes to enhance readability. To do so, we have consulted three independent researchers, including a native speaker, to better foreword our data, conclusions and discussion. We hope the reviewers agree that the manuscript has indeed been improved. Since we have made several changes, we have included a marked-up version to better highlight the changes to the manuscript.

2) Adding 1-2 small figures to explain / visualize the experimental setups/ findings would also be helpful.

We have included two new panels (figure 2F and figure 4F) to summarize and highlight our findings. We believe that these two extra cartoons indeed clarify the data in an appropriate manner, and thank the reviewers for this suggestion. Moreover, throughout the manuscript’s figures and figure supplements, we have several cartoons to explain our experimental setup (e.g. 2B, 4A, 4E, 5B, 7A), as well as three data-summarizing cartoons (Figures 3G, Figure 2-supplement 1D, Figure 4-supplement 1D).

In addition, we have made several (minor) changes to the figures to comply with the changes in the text:

Figure 1

Panel A: Add “Trachea” and “Carina”

Panel C: Add the duration of Dox treatment Figure 1—figure supplement 1:

Panel B: changed text color of DOX to red

Panel C: Line made red to show that the mice received doxycycline from E6 to E18

Figure 2

Panel B: Removed “1st branch” above cartoon, added “Carina”

Panel F: New scheme to illustrate the expression of MYC-SOX2 or MYC-SOX21 in secretory cells during development.

Figure 4

Panel F.: Added a scheme to illustrate the expression of MYC-SOX2 or MYC-SOX21 in SCGB1A1+ secretory cells in adult tracheal epithelium.

Figure 7

Panel B: Changed “luminal” to “apical” Figure 7—figure supplement 1:

Panel C: Changed graph y-ax to apical.